Weed and Seed Performance Measures: Analyzing and Improving Data Resources Final Report
Size 218.3 kB - File type application/pdf
Full screen
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Final Report
Submitted to the
Community Capacity Development Office
Office of Justice Programs
U.S. Department of Justice
By the
May 31, 2006
Acknowledgements
Principal JRSA authors: Jack O’Connell, Matthew Perkins, and Jim Zepp
JRSA research assistance: Jessica Blair, Jesse Handsher, Marina Karzag, Eileen
McDermott, Mary Poulin, and Jason Trask
Justice Research and Statistics Association
777 North Capitol St., NE, Suite 801
Washington, DC 20002-4239
(202) 842-9330 - phone
(202) 842-9329 - fax
cjinfo@jrsa.org - email
jrsa.org - Web site
This research described in this report was supported by Grant No. 2005-WS-Q5-K007 awarded
by the Community Capacity Development Office, Office of Justice Programs, U.S. Department
of Justice. Points of view in this document are those of the authors and do not necessarily
represent the official position or policies of the U.S. Department of Justice.
Table of Contents
EXECUTIVE SUMMARY .......................................................................................................... 1
OVERVIEW OF THE NATIONAL WEED AND SEED INITIATIVE ................................. 6
PREVIOUS WEED AND SEED EVALUATION EFFORTS .................................................. 9
METHODOLOGY ..................................................................................................................... 13
D
ATA
C
OLLECTION
.................................................................................................................... 13
D
ATA
A
NALYSIS
........................................................................................................................ 14
COMMUNITY PROBLEM IDENTIFICATION AND SITE CLASSIFICATION ............ 16
D
ESCRIPTION OF
C
OMMUNITY
P
ROBLEM
C
ATEGORIES
.............................................................. 17
F
REQUENCY OF
P
ROBLEMS
I
DENTIFIED BY
S
ITES
....................................................................... 20
IDENTIFICATION OF STRATEGIC SOLUTIONS AND SITE CLASSIFICATION...... 26
L
AW
E
NFORCEMENT
A
CTIVITIES
(LE) ....................................................................................... 26
C
OMMUNITY
P
OLICING
A
CTIVITIES
(CP)................................................................................... 28
P
REVENTION
/I
NTERVENTION
/T
REATMENT
(PIT) ....................................................................... 29
N
EIGHBORHOOD
R
ESTORATION
A
CTIVITIES
(NR) ..................................................................... 32
PERFORMANCE MEASURE RECOMMENDATIONS ...................................................... 34
REVISED HOMICIDE DATA ANALYSIS USING EXPANDED CCDO
ADMINISTRATIVE DATA SET.............................................................................................. 41
O
FFICIAL
R
ECOGNITION
V
ERSUS
T
HE
S
ITE
R
EPORTED
“S
TART
” D
ATE
: W
HICH IS THE
R
EAL
B
EGINNING FOR
S
TRATEGY
I
MPLEMENTATION
? ........................................................................ 43
P
RE
/P
OST
S
TRATEGY
I
MPLEMENTATION
A
NALYSIS
U
SING
2005 W
EED AND
S
EED
H
OMICIDE
D
ATA
S
ET
.................................................................................................................................. 44
C
OMPARING THE
P
RE
/P
OST
S
TRATEGY
I
MPLEMENTATION
A
NALYSIS
R
ESULTS TO THE
D
RAFT
2005 W
EED AND
S
EED
H
OMICIDE
D
ATA
A
NALYSIS
.................................................................. 45
APPENDIX A: STRATEGY CATEGORIES......................................................................... 49
APPENDIX B: PROGRAM/SERVICE CLASSIFICATION GUIDE .................................. 50
APPENDIX C: SUMMARY OF EXPANDED CCDO ADMINISTRATIVE RECORDS
DATA SET FINDINGS .............................................................................................................. 51
BIBLIOGRAPHY ....................................................................................................................... 54
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 1
Executive Summary
This project had three major objectives that would help to expand the available information for
assessing the performance of individual local sites as well as provide indicators for the overall
National Weed and Seed Initiative.
1. A systematic analysis was conducted of the problems and solutions identified by local
sites as the focus of their Weed and Seed Strategies. The purpose of this effort was to be
able to categorize the sites according to particular community issues. This would allow
future evaluation and performance measurement efforts to concentrate on sites that are
clearly involved with specific concerns and would reasonably be expected to have results
related to that topic.
2. Additional research work was done with 100 local sites to explore data resources cited in
their strategy documents as potential information sources for future evaluation studies
and performance measures. A listing of possible measures and considerations was
developed from this effort.
3. An updated analysis using expanded information from CCDO’s administrative records on
key implementation dates and the GPRA homicide data set was done that provides
additional documentation of the National Weed and Seed Initiative’s impact on crime as
indicated by homicide statistics.
Objective 1
Although the site strategies are reviewed individually through the Weed and Seed Official
Recognition application process, an aggregate analysis of the site strategies had not been done to
identify the commonalities and differences across the sites regarding the specific community
problems being addressed and the range of solutions being pursued. This study was able to
determine the range of issues and the relative complexity of the overall strategies developed by
the sites.
To accomplish this task, the strategy documents were retrieved from CCDO’s administrative
records for the 309 local sites that received Official Recognition (OR) status from FY2000 to
FY2005. A rigorous review process was used to develop a classification scheme for the
community problems and solutions cited in the strategies. Following this step, each site strategy
was then categorized using this approach.
The report presents the community problem/strategy solutions classification schemes, the
frequency of specific problem/solution references by the sites, and an analysis of the distribution
of sites by the categories.
The conclusion of this effort is that while, as a whole, the sites share many commonalities in the
problems being addressed and the solutions adopted, individually there can be substantial
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 2
differences in the focus of a site’s strategy. The design of any future evaluation studies and
performance measures should take this diversity across the sites into account.
Objective 2
The follow-up contacts with the local sites found that many have information resources that
could contribute to research on Weed and Seed’s effects on communities. However, there is
great variability in both content and reporting capabilities across jurisdictions that has to be
considered when designing any national information collection effort. For answering specific
questions about Weed and Seed efforts, these resources could provide additional data that would
respond to various inquiries.
Objective 3
Previously studies have examined the homicide statistics submitted in the local sites’ annual
Government Performance and Results Act (GPRA) reports as a common indicator of Weed and
Seed’s impact on crime. A limitation of these analyses has been incomplete information on key
strategy implementation dates for many sites. One of this project’s tasks was a search of
CCDO’s administrative records to locate these missing dates. Additional date information was
located for 321 sites.
Because the analysis of crime levels before and after a site’s strategy implementation requires a
complete set of implementation dates and multi-year homicide statistics, the number of sites
available for study can be affected by several factors. Due to the initial step of obtaining OR
status before a local site can receive federal funding, there are actually three possible dates that
could be used for analyzing pre/post strategy implementation. Not all dates were available for all
sites due to various recordkeeping issues and the age of some sites. Changes in site boundaries
and other developments prevent some sites from having valid data for a time series analysis.
Some sites have not provided complete, multi-year data sets, or inconsistencies were found in
some records that could not be readily resolved. Because some sites have not existed for a
sufficient time to have data for all of the years needed for the analysis, their records are currently
incomplete.
As was previously mentioned, three possible dates could be used as the start of a site’s strategy
implementation
⎯the OR award date, first grant award date, and the date that sites report for the
beginning of activities. The elapsed times between each of the three dates were calculated to see
if there was a significant difference for studying the implementation of strategies. The vast
majority of sites were found to have times between OR and the start of activities that were
insignificant for studying pre/post implementation effects. Of the five sites with at least a year
between OR and their activity starting date, only one site had any substantial difference in
pre/post results.
This project was able to perform a pre/post strategy implementation analysis for 100 sites. This
was nearly double the number of sites available for a previous 2004 study. These sites had an
average decline of 1.4 homicides following implementation of their strategies compared to a
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 3
decrease of only 0.5 homicides for the remainder of their jurisdictions. A further analysis was
done using a Relative Change Scale that characterizes the sites’ performance as a function of the
changing conditions in their respective host jurisdictions. This analysis also found that the sites
have a very positive effect on crime in their target areas when viewed from the perspective of the
trends occurring in the surrounding jurisdictions.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 4
Introduction
This report expands the information available for the study of the National Weed and Seed
Initiative. Specifically, it identifies and categorizes the diverse community problems and
solutions cited across the local Weed and Seed sites’ strategies, explores possible performance
measures and data sources for future site assessments, and reanalyzes the local sites’
Government Performance and Results Act (GPRA) homicide data using different key dates in
the process of site strategy implementation to determine possible effects on crime trends. The
intent of this research effort was to provide a more comprehensive understanding of the common
features and differences among Weed and Seed participants and to explore the potential for
additional data sources for measuring the sites’ performance and impact.
Unlike other federal programs which are generally structured around a particular client group,
need, or service such as providing subsidized meals to poor school children or the rehabilitation
of substandard housing, Weed and Seed is a comprehensive, strategic approach using locally
developed plans to solve persistent crime problems and other contributing conditions in targeted
community or neighborhood areas. The belief is that improved coordination and collaboration
across agencies and the community spanning a broad range of responsibilities and resources will
result in a synergistic effect that would be more effective than the existing networks of isolated
decision-making and service delivery.
Consequently, the over 300 local Weed and Seed sites have activities and impacts related to a
wide range of crime and other community concerns such as social, educational, and health needs;
poverty and economic development problems; and housing and environmental quality issues.
Performing accurate data collection and analysis tasks across this extensive universe of different
efforts and outcomes would consume a tremendous amount of evaluation resources. Conversely,
not all local sites are severely affected by every possible crime or community problem or are
directly involved in their solutions, which will minimize their relevance for the study of specific
concerns. A goal of this project was to provide a basis for focusing a national evaluation study
on the issues of greatest interest to the Community Capacity Development Office (CCDO) and
the subsets of local sites most appropriate for inclusion in various research efforts.
Recognizing the diversity of the community problems that Weed and Seed sites are attempting to
address through their strategies as well as the variety of solutions being employed, this project
reviewed the strategy statements submitted to CCDO from 2000 through 2005. The purpose was
to classify the problems and activities that the sites present in the applications for Official
Recognition (OR), the status conferred by CCDO upon Weed and Seed sites that allows a site to
apply for federal funding. These findings are intended to assist a future national evaluation effort
by providing information on which issues and local sites are most likely to have meaningful
impacts and potential performance measures a study of the National Weed and Seed Initiative’s
effects on specific concerns such as drugs, gangs, or distressed living conditions.
In addition to the classification analysis activities discussed above, sites were contacted to
determine data sources that may be commonly available to Weed and Seed sites.
Representatives from 100 active or formerly active sites were interviewed to obtain a broad-
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 5
based understanding of potential sources and considerations for consistent and reasonable data
collection. By collecting information on common problems and strategies, as well as building a
database of available data sources, JRSA will provide valuable information for any subsequent
evaluation efforts. As a result of this research, JRSA has developed a series of recommendations
for performance measures related to common strategies for Weed and Seed sites.
Additional elements of this project include an effort to supplement the previous analyses of
Weed and Seed efforts that rely exclusively on OR dates or the first date of federal grant awards
as single arbitrary starting points for Weed and Seed efforts. An intensive search of
administrative records was done to complete a database of local sites’ OR dates as well as first
Weed and Seed grant award dates. A comparison of homicide trends using different key dates
would resolve questions regarding the 1) the amount of time between when sites receive OR
status and their first federal funding and 2) the relative effects of receiving OR status versus
receiving Federal grant funds in terms of crime reduction.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 6
Overview of the National Weed and Seed Initiative
The National Weed and Seed Initiative is currently consists of 331 local sites, which are located
in 46 states, the District of Columbia, Puerto Rico, and the U.S. Virgin Islands. Sponsored by
the U.S. Department of Justice (DOJ) and overseen by CCDO, the Weed and Seed effort is a
cooperative strategy involving United States Attorneys’ Offices, community and social service
organizations, neighborhood residents and businesses, and local law enforcement and other
government agencies. The goal of Weed and Seed is to prevent, control, and reduce crime in
targeted high crime areas through interagency and community coordination and collaboration.
Based on research that indicated effective crime control efforts must reach beyond the scope of
traditional criminal justice agencies and resources to address the underlying factors that make
some communities susceptible to persistent and high levels of crime, Weed and Seed takes a
comprehensive and strategic approach to systemic change in selected neighborhoods. As a
geographically based program, Weed and Seed brings a different perspective to the delivery of
services and allocation of resources by providing a community-wide focus and insuring ongoing
communications across efforts that might otherwise operate in an isolated and uncoordinated
manner.
Site characteristics are widely variable; for example, site boundaries may range from a few
neighborhood blocks to an entire county. Weed and Seed sites focus their goals and objectives on
site-specific problems while simultaneously securing human services and economic resources in
and around the target area. Although their problems and solutions are locally chosen and
prioritized, sites may utilize a wide variety of federal, state, and local resources in their Weeding
and Seeding programs.
One of the key tenets of the Weed and Seed strategy is the recognition that even in the most
crime-plagued neighborhoods, it is likely that a variety of anti-crime efforts are already being
done. In addition to regular policing services being provided, there may be some form of
community policing activity in the area. There are probably a number of prevention efforts
sponsored by various government, community-based and faith-based organizations, as well as
grass roots activities by residents and other interested stakeholders (e.g., local business owners).
It is through organizing and coordinating these ongoing efforts, and securing resources to fill
gaps in existing activities, that Weed and Seed is designed to affect crime and community
livability. In order to enter the federal Weed and Seed program, local sites must show the
capacity and resolve needed to bring about long-term change affecting individuals, groups, and
institutions.
A Weed and Seed area's U.S. Attorney’s Office (USAO), community leaders, local law
enforcement, and other key stakeholders form a steering committee that sets the site’s general
direction and acts as a focal point for planning and coordination of the local site’s goals and
activities. Once the structure and leadership has been established for the proposed site, the
steering committee must develop a strategy based on local resources and needs. The structure of
the Weed and Seed strategy is based on four major components: law enforcement, community
policing, prevention/intervention/treatment, and neighborhood restoration. There are also four
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 7
fundamental principles that guide the development and implementation of a Weed and Seed
strategy: collaboration, coordination, community participation, and leveraging of resources.
Individual sites strive to weed out crime, seed the community against future crime and crime-
related risk factors, and sustain gains made through the initial Weed and Seed efforts. Local law
enforcement agencies and prosecutors, with support from federal law enforcement resources,
including the USAOs, take the lead in weeding efforts while governmental and community-
based public and private organizations pursue seeding strategies with various human and
neighborhood resources to restore the community’s infrastructure. The community-policing
component provides a bridge between the weeding and seeding aspects of a site strategy. The
long-term goal is to go beyond any immediate reductions in crime by making the community
resistant to future threats that may result in the return of high levels of criminal activity.
A significant challenge for the local sites’ leadership is insuring that their strategies remain
appropriate, as community conditions may change over time and new concerns emerge. These
may be otherwise positive developments such as improved housing and increasing property
values that may result in resident displacement or, negative influences such as the arrival of new
gangs or the sudden proliferation of new street drugs.
A national priority for CCDO is building the capacity of the local sites to sustain Weed and Seed
efforts beyond the expiration of the initial federal funding. By espousing long-term
sustainability for all Weed and Seed sites, the strategy addresses the problems and concerns of
local residents and promotes the financial stability and community resources necessary for
continued growth and development of the designated area. Sustainability depends on a site’s
continued evaluation of neighborhood problems and unmet needs and its ability to maintain the
steering committee as a principal mechanism in securing existing and new resources and funding
sources.
Successful weeding strategies require a joint effort of law enforcement agencies and prosecutors,
as well as the cooperation and leadership of neighborhood residents. Community engagement is
a key component of community policing and is considered an essential element in fostering long-
term community change. Community policing is combined with intensive law enforcement,
including problem-oriented policing strategies, to present a comprehensive crime control
approach in the Weed and Seed sites. These two mechanisms compel law enforcement
personnel to improve their interaction and relationship with the community while allowing
residents to participate in crime prevention and law enforcement activities within their own
neighborhood.
Weed and Seed prevention, intervention, and treatment goals focus on access to needed human
services in the designated area. Through their strategies and activities, sites identify, advocate
for, and, in some instances, provide essential services to at-risk and criminal populations in order
to reduce and control criminogenic influences and improve the overall quality of life. This often
requires collaboration from multilevel social service providers. As part of the neighborhood
restoration strategy, improved human services contribute to the development of the economy,
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 8
employment opportunities, the physical environment, and the community’s general
infrastructure.
To join the federal Weed and Seed effort, an interested community must apply for Official
Recognition (OR) by submitting a strategic plan to CCDO for review and approval in order to be
eligible for Weed and Seed funding. A potential Weed and Seed site must first notify the U.S.
Attorney’s Office in its district of its intent to apply for OR. The U.S. Attorney’s Office may also
assist the potential site with its administrative and strategic development, according to CCDO
guidelines. Once a site receives OR status, it is then eligible to apply for federal funding on a
yearly basis during a 5-year funding window. During this process of preparing their OR
applications, potential sites will first organize their community, begin the collaboration process,
and state the needs and strategies upon which this report is based.
A site’s Weed and Seed OR status lasts for a five-year period. In special circumstances, the OR
status can be extended up to a maximum of ten years. During this time, a site is eligible for
grants from CCDO that require an annual application. The cumulative total received during the
entire OR period cannot exceed $1 million.
Following the expiration of its OR status, a site may continue its participation in the National
Weed and Seed Initiative by applying for Graduated Site status. While it is no longer eligible for
direct CCDO funding, this allows a local site to be eligible for other Weed and Seed activities
such as conferences and technical assistance and for any special considerations available to
Weed and Seed sites when applying for funding assistance from other programs.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 9
Previous Weed and Seed Evaluation Efforts
Since its inception in the early 1990s, the National Weed and Seed Strategy has considered
evaluation research an important part of its efforts. National and locally initiated studies, yearly
data collection, and academic efforts are some of the varied methods by which the program has
been examined. Despite these efforts, there are inherent barriers to conducting a uniform and
consistent national evaluation across the local sites. These include differences in criminal code
definitions, procedures, and reporting capabilities across jurisdictions; wide ranges in site size,
local organizational structures and responsibilities, and agency participation; site strategies that
include a broad universe of community problems and solutions; modest grant award amounts
that must pay for program operations before local evaluations can be pursued; overlapping
funding sources for activities affecting the sites; limited authority by the sites’ management over
partner agencies; and other logistical impediments.
The need to improve current evaluation components and build upon previous work is fully
recognized by CCDO. For example, the Government Accountability Office (GAO) stated its
concern that Weed and Seed “generally collected activity data rather that measuring outcomes”
and warned against CCDO relying on homicide as an outcome measure (GAO, 2004). Since
1999, homicide statistics have been a primary crime control performance measure used by
CCDO for budgetary reporting on the effectiveness of Weed and Seed. The GAO report was
published before the final release of CCDO's first Crime Pattern Data Analysis, which provided a
multi-crime, multi-year, view of offending patterns in participating Weed and Seed sites.
Another criticism is that the federal Weed and Seed program has not undergone a thorough
evaluation encompassing all local sites. Given the number of sites across the nation and funding
available to CCDO for awards and administration of the program, any type of comprehensive all-
site evaluation would prove to be prohibitively expensive, or it would be so cursory as to be
effectively useless.
The very nature of the National Weed and Seed Initiative poses a number of challenges in terms
of conducting a methodologically rigorous evaluation. Even in the collection and analysis of
simple crime data, there are substantial problems due to the variation in legal definitions and
procedures for jurisdictions across the nation. Other difficulties are the inconsistencies in or lack
of ready availability of data at the neighborhood level that prevent the collection of uniform and
reliable national measures for many concerns. CCDO encourages sites to conduct local
evaluations and provides resources and technical assistance to those doing so. Many locally
initiated studies have been conducted that examined individual sites as well as there have been
several studies of multiple sites within a state. There also have been national efforts that have
attempted to measure effectiveness through the study of multiple sites.
The largest of the national efforts is a cross-site analysis funded by The National Institute of
Justice (NIJ, 1999). This study examined both the implementation and outcomes of the Weed
and Seed Initiative across sites. Case studies were performed for each of 12 target areas in eight
jurisdictions to determine how well the strategy lived up to expectations, what elements were
strongest, what elements failed, and lessons to be learned over the life of the project.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 10
In order to answer these questions, funding applications and other documents were reviewed;
program administrators, senior law enforcement staff, service providers, managers of seeding
activities, and community leaders were interviewed; crime incident and arrest data were provided
by police departments and analyzed; group interviews were conducted with seeding participants;
and residents were surveyed at two different time periods. Although findings varied across sites,
significantly positive results were found for Weed and Seed’s impact on several outcome
measures.
In six of the study sites Part I crimes decreased more than in the host jurisdiction during the same
period. There also appeared to be an association between a decrease in Part I crimes and drug
crime. These same six sites also experienced an initial increase in drug arrests (probably due to
concentrated law enforcement), and then a decline. Findings also examined the relationship
between crime trends and the concentration of resources in sites. Factors that appeared to be
related to success in implementation included the community setting (i.e., strength of social and
institutional infrastructure, economic development possibility, and transience of population) and
program design (i.e., mix of weeding and seeding activities and the sequence of implementation).
Another national effort was the Crime Pattern Data Analysis conducted by the Justice Research
and Statistics Association (JRSA, 2003). For this study, sites were asked to submit data for three
crimes that were considered the greatest problem in their site. Data were requested for a six-year
period, beginning two years prior to, the first year of, and three years following Weed and Seed
project implementation. Crime data for the entire jurisdiction of which the target area was a part
were also reported for the same time period. A pre/post analysis of the change in crime over
time was conducted for each site as well as a comparison of the target area and host jurisdiction
over time.
A total of 98 sites submitted data, with 20 providing complete six-year data sets, 19 with five-
year data sets, and the remainder submitting some data but for less than five years. Sites most
frequently reported drugs and violent crime, and approximately 62% named violent crime in
their top three most problematic crimes. In a analysis of the sites that submitted usable data sets,
a pattern of an initial increase in reported crime during the year of implementation and the first
year following was identified, followed by decreased levels of reported crime in later years of the
effort. These decreased levels of crime were well below the starting values experienced at the
onset of Weed and Seed activities.
Based on research into crime control strategies, including intensive law enforcement and/or
community policing components, it is to be expected that any program employing these
strategies should initially experience increases in crime rates as a consequence of higher levels of
police activity and citizen reporting of incidents. The data then show an apparent decrease in the
study sites in years three and four - a decrease not mirrored by the remainder of the jurisdiction.
Similar trends were found for both sites that submitted partial multi-year data sets as well as the
sites that provided a complete six-year data set.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 11
This analysis of yearly patterns of offending was repeated for data broken down by specific
crime types. Five categories were used
⎯homicide, illicit drugs, violence and weapons, property
crimes, and other (prostitution, public order crimes, etc.). A pattern of initial increase in reported
crimes, followed later in the Weed and Seed cycle by dramatically lowered reported crime is
seen throughout. Crimes that are likely to be dealt with in a more urgent manner (e.g., violent
crimes) show decreases earlier than other crimes. This is especially true in the case of drug-
related crimes, where reported offenses are heavily influenced by police activity. Therefore,
because intensive law enforcement is an integral and ongoing component of the Weed and Seed
strategy, it is understandable that drug crimes take much longer before showing a decrease
through police counts of incidents. Overall findings indicate that Weed and Seed sites present a
general pattern of relatively positive change or stability in crime levels as compared to their host
jurisdictions.
Another major source for Weed and Seed performance data is the annual Grantee Site
Characteristics and Activity Data Report, which is part of CCDO’s responsibilities under the
Government Performance and Results Act (GPRA). Each site is required to submit a report for
every year it operates as an Officially Recognized Weed and Seed site or for any year in which
the site spends federal monies. The report includes questions about law enforcement operations,
community policing activities, site-specific as well as surrounding-jurisdiction crime data,
community/social services provided by the sites, neighborhood restoration projects, and funding
other than CCDO money leveraged for use in the site.
The homicide data collected through the GPRA has been used to analyze trends in the sites in
comparison with their host jurisdictions. This analysis, conducted by JRSA (2003), focused on
220 sites in 174 jurisdictions. Over half of the study sites experienced a decrease in homicides
during the study period, 14% stayed the same, 8% increased but at a lower rate than the
jurisdiction, and 23% rose at a faster rate than the jurisdiction. The authors classified 77% of the
Weed and Seed sites as having successful outcomes, i.e., they either had a decline in homicides
or displayed greater stability or control in homicide trends than those found in the host
jurisdiction, exclusive of the Weed and Seed site area. The latter finding was considered
reasonable due to the belief that the Weed and Seed target areas were specifically chosen
because of their persistent crime problems. Consequently, any negative crime trends affecting
the jurisdiction at large should be reflected in the Weed and Seed sites in an equal or more
intense level.
The greatest source of information on Weed and Seed effectiveness however, has been locally
initiated studies of single sites. These studies have generally been overlooked in efforts to assess
the Weed and Seed program but are far from rare and are generally consistent in their findings.
Through 2005, JRSA had collected copies of at least 62 local site evaluations. In 2004, JRSA
conducted a review of locally initiated site studies and examined 34 that met acceptable
methodological standards for consideration as part of the 2004 Weed and Seed Local Evaluation
Meta-Analysis. In examining law enforcement/crime control and community safety effects, the
authors found 21 of 34 sites reporting positive outcomes in one of these categories, with the
other 13 evaluations reporting little or no change or no findings in the component examined.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 12
None of the local evaluations reported an increase in crime greater than that experienced in the
surrounding area.
The above-mentioned studies do not include all of the Officially Recognized Weed and Seed
sites and there continues to be gaps in assessment on the local and national level. However,
CCDO has taken an aggressive approach in response to criticism that the Weed and Seed
strategy has not been properly assessed, including the commitment of substantial financial
resources. Overall the studies presented here provide very positive results of Weed and Seed’s
impact on crime and stabilizing neighborhood conditions. Additional research work can still be
done, and this study is another step in showing the positive influence of Weed and Seed in some
of the nation's most dangerous neighborhoods.
In reviewing the past national evaluation effort, a major drawback has been that the
individualized aspects of each site’s strategy have not been recognized. Because the
fundamental concept underlying the National Weed and Seed Initiative is the establishment of a
locally developed strategy to address the specific crime and related community concerns for each
local site, an appropriate evaluation strategy should be based on a similar approach.
Consequently, a new national evaluation effort for Weed and Seed should begin by identifying
the problems being addressed by the local sites and which sites are involved with each issue.
This would enable the evaluators to develop measures that are specific to the community
problems being studied for Weed and Seed’s impact and to select local sites for inclusion in the
evaluation that should reasonably have results for the chosen issues.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 13
Methodology
This project required the collection and analysis of previously unexamined data in CCDO’s
administrative records and the merging of some of this information with JRSA’s GPRA data.
These collection activities included extracting information from the U.S. Department of Justice’s
(DOJ) Grants Management System (GMS) and CCDO’s paper files to determine key dates for
the implementation of the local sites’ strategies and to obtain the OR applications that contain the
strategies for recent Weed and Seed sites. The key dates for receipt of OR status and first grant
awards allow an analysis of the time between these events and changes in homicide crime
patterns associated with these dates. The major task of this project was the development of a
typology for the crime and community problems being addressed by the local sites and the
solutions identified in the local Weed and Seed strategies. Based on this classification scheme,
the existing sites could then be categorized according to the main focus of their strategies.
Data Collection
The analysis of the site OR applications was designed to identify both the site-specific problems
and the strategies employed to address neighborhood problems. The process involved an
intensive review of multi-year OR data and marked the first attempt by CCDO to systematically
analyze multi-year OR application data. The parameters of this project were the examination of
available OR applications from sites that applied for and were granted Official Recognition from
fiscal year 2000 through fiscal year 2005. JRSA identified 309 sites that had been granted
Official Recognition during this time period. Once these sites had been identified, the next step
was to locate and collect their relevant grant documents. In addition to the OR applications,
other pertinent documentation, such as grant status and funding information, was retrieved via
the DOJ GMS files.
The GMS database records grant-specific data, including date of first award, grant amounts,
fiscal agent and other data. In recent years GMS’s capabilities were expanded to include the
ability to store electronic copies of grant applications and supporting documents. JRSA staff
worked on site at CCDO offices to retrieve and enter grant information, including; Weed and
Seed site name, city and state, OR date, date of first award information, start dates and tracking
information related to the retrieval of these electronic records. While much of this information
was available via GMS, there were some instances of missing data elements. In cases for which
information was unavailable through the GMS system, JRSA staff went through CCDO's paper
files for relevant information. Since GMS lacks the capability to electronically extract data for
analysis purposes, needed information was retrieved record by record within GMS. These data
then had to be manually entered into a project database, slowing the process and increasing the
potential for human error.
The collection of OR application information was affected by changes in electronic storage of
application data. Beginning in 2004, GMS included the ability to store electronic versions of OR
applications. Prior to the 2004 application cycle, CCDO relied almost exclusively upon
hardcopies stored on-site at the CCDO office. To reduce the burden of converting paper copies
to electronic files, sites granted OR prior to 2004 were contacted in hopes of obtaining electronic
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 14
copies of these earlier applications. A number of sites were able to provide electronic versions
of their original applications. All paper copies were electronically scanned, converted to PDF
format and saved for future reference. The PDF format was chosen to ensure the integrity of the
electronic documents. For this reason, all existing electronic copies were also converted into a
PDF format before being saved. Along with this report, JRSA will provide CCDO with
electronic copies of the reviewed OR materials.
This report is based on a review of 273 of the 309 OR applications that were granted Official
Recognition status between fiscal years 2000 and 2005. Thirty-six of the 309 sites could not be
included in this analysis because the OR applications for 26 sites could not be located in
CCDO’s files, while 10 others failed to clearly state the problems and their proposed strategies
and therefore classification of these OR applications was not possible.
Data Analysis
The OR application process allows the applicants to indicate any problems confronting their
communities and any strategies they plan to implement to address their communities' problems.
This latitude in selection coupled with the differences in local communities' challenges
introduces a wide degree of variation in the stated problems and strategies. The sites describe
these issues in a narrative text with optional charts and other supporting data. Given the
unstructured nature of the strategy documents, a process was developed for the consistent
classification of sites by discrete categories of problems and solutions that could then be used for
future performance and evaluation studies.
To effectively analyze the reported problems and strategies, a categorization scheme was
developed that enabled the grouping of these items based on observed commonalities. The
categories used in this scheme were developed utilizing a random sample review process that
included 40 applications. Based on this review, 73 strategy types were identified (See Appendix
A for full list of strategy types). Each of these strategies fall within one of the four broader
Weed and Seed elements
⎯Intensive Law Enforcement, Community Policing, Prevention
Intervention and Treatment, and Neighborhood Restoration. Following the development of these
strategy types, the 40 applications were reviewed again and coded to reflect problems and
strategies among the categories. The applications were reviewed by at least two and up to four
different reviewers to ensure consistency in the review process. Adjustments to the categories
were made whenever reviewer interpretations were inconsistent, and a key was developed for the
classification of common activities that could reasonably be assigned to more than one of the
strategy types (See Appendix B).
Due to the local nature of the Weed and Seed program, sites applying for Official Recognition
have a great deal of latitude in selecting the problems that they intend to address. The initial OR
analysis revealed that the 273 applications contained 2,083 problems, which included some
repetitions of issues across sites. Also, some strategies may refer to crime problems in general,
while others identify specific concerns, such as drugs or weapons violations. Attempting to
analyze 2,083 problems without categorization would be extremely difficult; therefore, for the
purposes of this analysis, these problems were grouped into 15 problem area categories.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 15
The decision to place a problem into a specific category was based on commonalities within the
descriptions of the problem. In a substantial number of cases, the applications only provided
general descriptions of their problems. To capture these responses, a series of general categories
were developed. More specific categories were created for instances in which sufficient detail
was provided. For example, if a site listed educational concerns, it was grouped into the
Educational Concerns–General category. If the site reported that they had school dropout
problem, it was grouped into the Educational Concerns-Dropout category. Therefore, the
number of problems reported in the general categories may not equal the combined number of
problems reported in the related subcategories.
The following analysis is based on the review and classification process described above, and is
the first step in the development of a performance measures guide for Weed and Seed sites. It
details the common problems and strategies for reducing crime and disorder cited by Weed and
Seed sites in their OR applications. By classifying these efforts, it should be possible to develop
performance assessments that more accurately reflect the intent of Weed and Seed grantees.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 16
Community Problem Identification and Site Classification
Weed and Seed strategies are developed for at-risk areas that face a multitude of challenges and a
disproportionate level of offending. This is evident from the large number of problems
referenced in the OR applications. In the 273 applications, 2,083 problem citations were
identified through the review process. These were grouped into 16 broad categories. For eight of
the general problem categories, the sites provided sufficiently detailed descriptions and
supporting data that these could be further broken down into subcategories. The following
section details the community problems reported in the applications and the categorization of
these problems.
Table P1
Problem Category
Count
% of All
Problems
Problem Category
Count
% of All
Problems
Blight
139
6.67% Family/Life Skills
26
1.25%
Code Violation / Enforcement
13
0.62% Fear of Crime
34
1.63%
Community Cohesion
104
4.99% Health
51
2.45%
General Community Problems
48
2.30%
General Health Problems
28
1.34%
Lack of Community Involvement
33
1.58%
Teen Pregnancy
23
1.10%
Traffic
23
1.10%
Housing
115
5.52%
Crime
833
39.99%
Drug Crime
225
10.80%
Offender-Related
14
0.67%
Violent Crime
150
7.20%
Reentry
10
0.48%
Gang-Related Crime
98
4.70%
Probation
4
0.19%
Juvenile Crime
95
4.56%
General Crime Problem
65
3.12%
Police-Related
47
2.26%
Domestic / Family Violence
57
2.74%
Property Crime
55 2.64%
Service / Program Gap
132
6.34%
Weapons Offenses
48
2.30% General Services / Programs
82
3.94%
Prostitution
40
1.92% Youth Services / Programs
33
1.58%
Adult Services / Programs
7
0.34%
Economic
220
10.56% Elderly Services / Programs
6
0.29%
Unemployment / Underemployment
101
4.85%
Reentry Services / Programs
4
0.19%
Poverty
64
3.07%
General Economic Problems
55
2.64%
Substance Abuse
70
3.36%
Education
149
7.15% Youth –Related
85
4.08%
Truancy
57
2.74%
Alternatives for Youth
48
2.30%
General Education Problems
40
1.92%
General Youth Problems
23
1.10%
Education Level
26
1.25%
At-risk Youth
14
0.67%
Dropouts
22
1.06%
Literacy
4
0.19%
Other
51
2.45%
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 17
Table P1 displays the overall frequency of problems being cited in the site strategies. Due to the
pattern of multiple problem topics across most sites and the potential for multiple subcategories
of a general problem category (e.g., both drugs and gangs being identified as concerns for the
same local site), the problem counts may exceed the actual number of sites working on a specific
issue. However, the problem counts indicate the relative incidence of each concern across the
site strategies.
Perhaps appropriate to a U.S. DOJ-sponsored initiative, crime-related problems are the most
frequently cited concerns. Reflecting the comprehensive approach of the Weed and Seed
strategy to crime control, the majority of the problems being cited are non-justice related topics.
An examination of the distribution of reported problems revealed that the fewest number of
problems reported by any site was two and that the largest was 20. The greatest number of sites,
41, identified seven problems, while the smallest number of sites, one, identified 20 different
problems. The mean number of problems was eight, while the median and modal numbers of
problems were both seven. A distribution of the reported problems by the number of sites can be
seen below in Chart P1.
Chart P1
Distribution of Reported Problems
0
5
10
15
20
25
30
35
40
45
0
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16 17 18 19 20
Number of Problems
Number of Sites
Description of Community Problem Categories
Blight refers to a host of issues that were described as having a deleterious effect on the physical
environment of the designated Weed and Seed target areas. Some of the issues grouped into this
category included, but were not limited to, graffiti, dilapidated buildings, abandoned cars,
neglected properties and litter. More than half of the applications, 140, referred to some form of
blight as being a problem.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 18
Code Violations/Enforcement represents the opportunity to use city or jurisdiction codes to
improve conditions in the Weed and Seed area by pressuring property owners to adhere to
community standards and to take a more proactive approach to combating neighborhood
disorder. While most of the identified issues involved violations, some did involve effective
enforcement of existing codes. Thirteen sites indicated that these issues were a problem in their
designated Weed and Seed areas.
The Community Cohesion Problems category includes a variety of problems that negatively
impact the community as a whole, but were not appropriate for assigning to the other categories.
Because of the detailed descriptions provided by the local sites, many of the references to
problems in this category could be characterized by the three subcategories
⎯General
Community Problems, Lack of Community Involvement and Traffic Problems. Traffic
Problems subcategory involves issues related to traffic safety concerns, such as speeding. The
General category consists of a variety of problems, including, but not limited to, immigration or
minority population issues, transient populations, poor lighting, and community pride.
Combined, the sites reported a total of 157 Community Cohesion Problems. The majority of
these, 100, did not warrant a separate subcategory and were therefore grouped into the General
category. Lack of Community Involvement was the most commonly identified community
problem, followed by Traffic Problems.
Crime Problems were not only most frequently reported, but they were also described with
greater specificity. For this reason, Crime Problems were grouped into nine subcategories.
While most of these subcategories are self-explanatory, some deserve further explanation. The
Weapons Offenses category consists of gun and other weapons-related crimes, many of which
were simply identified as 'weapons' related, without reference to a specific type of weapon.
Domestic violence and child abuse/neglect were each reported frequently enough to warrant
being separate from the general crime, but not frequently enough to distinguish each crime from
one another. As a result, these crimes were collapsed into one subcategory, Domestic/Family
Violence.
The Economic Problems category contains General Economic Issues, Poverty, and
Unemployment/Underemployment. While most sites specifically identified unemployment,
there were a substantial number that identified problems closely related to unemployment. These
problems mainly dealt with the overall employability of residents and included problems such as
lack of requisite job skill. The General category included problems such as lack of investment in
the community, lack of local businesses, and poor tax base. Since most applications did not
detail poverty and how it was measured, the Poverty subcategory included references to poverty,
including, but not limited to, people not earning a livable wage, children living in poverty, large
numbers of residents on public assistance, and lack of financial resources.
Education Problems were grouped into the following five subcategories; Truancy, Education
Level, Dropouts, Literacy, and General. As was the case with other general categories, these
references were assigned to a specific subcategory when sufficient detail was provided; any
nonspecific problem statements were put into the General subcategory. Some examples of
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 19
General Education Problems are education, lack of education, school in dire need of repair, and
strained educational assistance.
Program/Service Gap Problems were commonly reported, in part because every Weed and
Seed Strategy must identify both existing community programs and unmet service-related needs.
These concerns form the Program/Service Gap Problem category. Whenever possible, these
unmet needs were assigned to subcategories by the intended target population
⎯general, youth,
adult, elderly or returning offenders.
Family/Life Skills Problems included an array of issues such as lack of parental involvement,
poor parenting skill, marginal life skills, and poor personal financial management. No
subcategory of the Family/Life Skills group was reported frequently enough to warrant it own
category, therefore, all of these problems were included under the main category name.
Fear of Crime are expressions of concerns about public safety concerns that are widespread and
pervasive and, if left unaddressed, will contribute to further community deterioration as residents
and businesses as flee an area. While some applications did contain other terms such as “feeling
unsafe” and “fear of retribution from criminals,” most specifically listed “fear of crime.” Since
no more specific problems related to fear of crime emerged, all of the problems dealing with this
issue were grouped into this one category.
Housing Problems were frequently identified in the sites’ strategies. However, while a large
number of applications indicated Housing Problems as a concern, these were so closely related
that further distinctions were not reasonable. Some examples of entries in this category are lack
of affordable housing, poor quality housing, and lack of home ownership.
Health Problems included a host of problems like community health, lack of affordable
healthcare, lack of healthcare providers and problems with healthcare. Only one health-related
problem, Teen Pregnancy, was mentioned frequently enough to warrant its own category.
The Police – Related Problems category was mostly composed of problems related to the
relationship between the police and the community. However, this category was kept general in
nature to capture other problems such as low visibility of officers, lack of officer training, poor
inter-agency cooperation and insufficient police capacity.
Substance Abuse Problems contains entries dealing with the full range of alcohol and other
drugs
⎯some dealing with one of these elements, others with multiple issues. It includes
responses such as substance abuse, alcohol abuse, drug abuse, and alcohol and drug abuse.
Youth Related Problems reported in the applications could be classified into one of three
subcategories General, Alternatives for Youth and At-Risk Youth. In many cases, the
applications simply listed that there was a need for Alternatives for Youth. When more specific
information was provided, the more commonly mentioned alternatives were lack of after-school
activities and recreational activities. The program shortages discussed in this section differ from
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 20
those discussed in Service/Program Gap section, because these involve needs for recreation and
entertainment resources.
At-Risk Youth was a difficult problem to classify. There are varying definitions for at-risk, and
any youth living in a Weed and Seed area could be considered at-risk. The youth could be at-
risk of becoming delinquent, succumbing to possible health threats, or being subjected to a
variety of other risk factors. For this reason, an application was only classified as having a
problem with at-risk youth if it contained that specific phrasing. The General subcategory was
used to capture the other Youth-Related Problems that were either too few in number or too
vague to further classify. This subcategory included problems such as unsupervised youth, youth
harassing other residents, and court-involved youth.
Frequency of Problems Identified by Sites
Crime
As previously mentioned, crime-related problems were most frequently identified. The
applications contained a combined total of 833 references to various crime problems, which
represents 40% of the total 2,083 problems cited in the 273 applications. In an effort to better
describe the specific nature of these crimes, they were classified into nine subcategories. The
most frequently identified crime problems were drug crime, violent crime, gang-related crime,
and juvenile crime. Combined, these crimes accounted for over two-thirds of the identified
crime problems.
In terms of local sites, 99% of the sites identified one or more crime-related issues as concerns
being addressed through their strategies (see Table P2). However, for the specific crime
subcategories, only drugs and violent crime were cited by a majority of sites. The other crime
subcategories were mentioned by a third or less of the sites.
In this report the term “drug crime” refers to the illegal distribution, purchase, and/or possession
of narcotics. Not only was drug crime the most frequently identified crime problem, but it was
also the most frequently identified problem overall. Drug crime was reported as a problem in
226 of the applications accounting for over one-quarter of all the crime problems and over 10%
of all problems identified.
The next most common crime problem was violent crime. While most sites listed specific
violent crimes, many simply identified violent crime in general as being a problem in their
targeted area. Violent crime was a listed as a problem in 150 of the 273 applications, accounting
for almost one in five identified crime problems and about 7% of all problems.
Problems classified as gang-related crime were also common. Ninety-eight applications detailed
some form of gang-related crime as being a problem in their designated area. Crimes classified
as juvenile crime were those instances for which the application made specific reference to the
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 21
offenders being juveniles
1
. Combined, the applications contained a total of 95 instances of
juvenile crime problems. Both gang-related and juvenile crimes represent a little over 10% of all
crime problems and less than 5% of all problems.
The general crime subcategory represents all crimes that did not fit into any of the other crime-
type subcategories. The crimes in this subcategory were either too few in number or were
described in such a vague manner that the no further categorization was possible. For these
reasons, these crimes were placed into a single subcategory of miscellaneous concerns.
Table P2
Crime Problem
Number of
Sites
Percent of Sites
Crime problem specified
269
98.53%
No identified crime problem
4
1.47%
Type of Crime Specified
Drug crime
225
82.42%
Violent crime
150
54.95%
Gang crime
98
35.90%
Juvenile crime
95
34.80%
Crime (general)
65
23.81%
Domestic violence/child abuse
57
20.88%
Property crime
55
20.15%
Weapons violations
48
17.58%
Prostitution 40
14.65%
Total 833
Economic Problems
Economic problems were the second most frequently identified problem. There were a total of
220 references to economic problems in the 273 applications, accounting for 10.5 % of the total
number of problems. This category was broken down into three subcategories
⎯general, poverty
and unemployment/underemployment. As was the case with crime, problems that were
described too generally to be classified further were included in the general subcategory.
Of all the reported economic problems, unemployment/underemployment was the most common.
Combined, the applications listed this as a problem 101 times, accounting for just over 45% of
economic problems and almost 5% of all problems. Poverty accounted for almost one-third of
economic problems and 3% of all problems. There were also a total of 55 general economic
problems identified, accounting for another quarter of the identified economic problems.
Of all local sites, 153 or 56% have some economic-related issues in their strategies (see Table
P3). Depending on the nature of the economic problems, this may affect roughly a fifth to a third
1
Truancy, however, was grouped into the Education category rather than a crime category.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 22
of the sites. Also, economic conditions, such as poverty and unemployment that could be viewed
as contributing factors to a community’s overall distress, may be more successfully addressed as
the area’s persistent crime problems are brought under control and may require longer term
solutions that would be more likely components of subsequent iterations of a site’s strategy.
Table P3
Economic Problem
Number of
Sites
Percent of Sites
Economic problem specified
153
56.04%
No identified Economic problem
120
43.96%
Type of Economic Problem Specified
Unemployment/underemployment 101
37.00%
Poverty 64
23.44%
General economic problems
55
20.15%
Total 220
Education Problems
Problems related to education were the next most commonly identified problem type in the 273
applications. A total of 114 site strategies indicated that they had an education-related problem
that impacted the residents of their designated area (see table P4). Truancy accounted for the
largest number of education problems, identified in 57 or 21% of all sites and accounting for
almost 40% of the education problems indicated. An additional 40 general problems were
described in the applications accounting for over one-quarter of the education problems. The
remaining education problems related to Education Level, Dropouts, and Literacy accounted for
another third of these problems, though Literacy was far less often mentioned than the preceding
subcategories.
Table P4
Education Problem
Number of
Sites
Percent of Sites
Education problem specified
114
41.76%
No identified Education problem
159
58.24%
Type of Education Problem Specified
Truancy 57
20.88%
General education problems
40
14.65%
Education level
26
9.52%
Dropouts 22
8.06%
Literacy 4
1.47%
Total 149
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 23
Blight
Blight was the fourth most common problem and, according to the “Broken Windows Theory”
of crime control, can be a significant factor contributing to existing community crime conditions.
As was explained above, blight is used to describe a wide variety of problems that have a
deleterious effect on the physical environment of the designated communities. While many
applications contained references specifically to blight, others listed the problems individually.
Due to the diversity of blight-related problems reported, it was not possible to further classify
these problems. Therefore, these problems were combined into one larger Blight category. Of
the 273 strategies, 139 sites, or 51%, included blight as a problem affecting their communities.
This represents 6.7% of total number of identified problems.
Service/Program Gaps
The next most frequently identified problem type after blight concerned the provision, delivery,
or availability of services and programs. There were 132 references to these problems for 102,
or 37 % of all sites, accounting for over 5% of the total number of problems (see Table P5). The
greatest number of these applications, 83, contained references to this problem that did not
identify a specific target population. These general references accounted for three of five service
or program gap problems and about 4% of all identified problems. Thirty-three applications
outlined youth-related service or program gaps, while the remaining three populations mentioned
in the applications were adults, elderly and returning offenders. These populations accounted for
13% of the service/program gap problems as a group.
Table P5
Service/Program Gap Problems
Number of
Sites
Percent of Sites
Service/program gap problem specified
102
37.36%
No identified service/program gap problem
171
62.64%
Type of Service/Program Gap Specified
General services/program
82
30.04%
Youth services/program
33
12.09%
Adult services/program
7
2.56%
Elderly services/program
6
2.20%
Reentry services/program
4
1.47%
Total 132
Housing Problems
Housing Problems contains a variety of housing-related conditions in the Weed and Seed areas.
Some of the problems contained in the applications were lack of affordable housing, sub-
standard housing, and lack of home ownership. There were not sufficient numbers of any
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 24
specific problem to warrant the creation of subcategories. Combined, the 273 applications
contained 116 references to housing problems making it the sixth most frequently mentioned
problem type representing just over 5% of the total number of recognized problems.
Community Cohesion Problems
The Community Cohesion Problems category was the next most frequently identified problem
type after housing problems. Eighty-six strategies, or 32% of all sites, mentioned negative
conditions in this category. In all, community problems accounted for 105 problems, or 5% of
the total number of identified problems. This category includes three subcategories
⎯general
community problems, lack of community involvement and traffic. The greatest number of
applications, though, outlined problems that were described in such a vague manner or so unique
to their community that they were not classifiable beyond general community problems. The
single most common identifiable problem listed in the applications was a lack of community
involvement. Thirty-four applications, or 32% of those listing Community Problems, indicated
that community involvement was an issue that was negatively affecting their designated area. Of
the applications that identified a community problem, 23 identified some type of traffic concern.
See Table P6 for a breakdown of the community cohesion problems.
Table P6
Community Cohesion Problem
Number of
Sites
Percent of Sites
Community problem specified
86
31.50%
No identified community problem
187
68.50%
Type of Community Cohesion Problem Specified
General community problems
48
17.58%
Lack of community involvement
33
12.09%
Traffic
23
8.42%
Total 104
Remaining Community Problems
The remaining nine categories accounted for a combined total of 391 problems or almost one-
fifth of the total problems. These categories were broken down as Youth-Related, Substance
Abuse, Other, Health, Police-Related, Fear of Crime, Family/Life Skills, Offender-Related, and
Code Violations/Enforcement. There were a total of 85 references to youth-related problems, of
which 48 identified a lack of alternatives for youth, 23 identified general youth-related problems,
and 14 identified at-risk youth.
There were 70 instances in which Substance Abuse was identified as a problem accounting for
barley 3% of all problems. The problems that could not be classified or that did not warrant their
own category were grouped together into the Other Problems category. A similar number of a
Health-related issues were found, half of which were identified as related to teen pregnancy. The
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 25
Other Problems and Health-related categories were noted in 52 and 51 instances, respectively,
numbering about 2.5% of the total number of problems each.
While most Police-related problems dealt with the relationship between the police and the
community, the Police-related problems category also included other problems, such as
equipment and staffing issues. Police-related problems were reported in 47 instances making up
just over 2% of the total problems. Fear of Crime was identified as a problem in 34 instances
while there were 26 instances of Family or Life Skills as a problem.
One of the least often mentioned problems include Offender-related problems accounting for
only 14 of the 2,083 problems. Most of these identified returning offenders as the main issue,
while the remaining Offender-related problems identified probationers as the issue. The final
problem category is Code Violations/Enforcement. The applications contained only thirteen
instances where code violations or enforcement was cited as a problem. See Table P7 below for
a detailed breakout of these less numerous problem categories and subcategories.
Table P7
Misc. Problem Category
Total
% of All
Sites
Youth - Related
85
31.14%
Alternatives for youth
48
17.58%
General youth problems
23
8.42%
At-risk youth
14
5.13%
Substance Abuse
70
25.64%
Other Problems
51
18.68%
Health Problems
51
18.68%
General health problems
28
10.26%
Teen pregnancy
23
8.42%
Police - Related
47
17.22%
Fear of Crime
34
12.45%
Family/Life Skills Issues
26
9.52%
Offender - Related
14
5.13%
Reentry
10
3.66%
Probation
4
1.47%
Code Enforcement
13
4.76%
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 26
Identification of Strategic Solutions and Site Classification
In order to group the OR applications by commonalities among the strategic solutions being
pursued, a coding scheme was developed. This coding scheme consisted of the four overall
strategy elements and a subset of 69 activity types. The four strategy elements reflect the core
elements of the Weed and Seed Strategy, which are Law Enforcement (LE), Community
Policing (CP), Prevention/ Intervention/ Treatment (PIT), and Neighborhood Restoration (NR).
The 69 activity types each belong to the most appropriate strategy element and can be viewed in
Appendix A.
The review of the OR applications revealed that most sites were implementing an array of
activities to address the problems affecting their targeted neighborhoods. Combined, the 273
applications contained 7,884 activities. The smallest number of activities reported in any
application was four, while the largest was 57. The mean number of activities was 28.88, the
median was 29, and the mode was 28. See Chart S1 below.
Chart S1
Law Enforcement Activities (LE)
The Law Enforcement strategy element encompasses a variety of enforcement-oriented activities
that are designed to address the crime problems in the designated areas. These activities do not
include police activities that are intended to bolster relationships between the police and the
communities that they serve. Upon examination of the Law Enforcement (LE) activities, 15
activity types were developed.
Site Strategy Frequencies
0
2
4
6
8
10
12
14
16
18
4 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 47 49 51 57
Number of Strategies
Number of Sites
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 27
Combined, the 273 OR applications contained 1,757 references to LE activities, which accounted
for almost one-quarter of all of the identified activities. All but one strategy contained at least
one LE activity. See Table LE1.
Table LE1
Law Enforcement Activity
Frequency Percent of Sites
Law Enforcement activity specified
272
99.63%
No Law Enforcement activity specified
1
0.37%
Type of Law Enforcement Activity Specified
Intensive drug enforcement
226
82.78%
Intensive supply reduction
91
33.33%
Intensive demand reduction
53
19.41%
Prosecution/sentencing coordination
184
67.40%
Increased police patrol/visibility
153
56.04%
Task forces
148
54.21%
Targeted crime enforcement
142
52.01%
Known/repeat offender strategies
128
46.89%
Juvenile crime enforcement
117
42.86%
Parole/probation strategies
108
39.56%
Public order crime enforcement
106
38.83%
Gang reduction strategies
104
38.10%
Weapons reduction strategies
96
35.16%
Use of technology
91
33.33%
Hotspot policing
67
24.54%
Other law enforcement strategies
52
19.05%
Traffic violation enforcement 35
12.82%
Total Law Enforcement Activities
1,757
A closer examination of the LE activity types revealed that intensive drug enforcement was the
most common LE activity type employed by the sites. In all, 226 of the 273 (83%) of the
applications indicated that intensive drug enforcement was part of their overall strategy.
Whenever possible, the focus of these efforts was also determined. Of the 226 applications that
detailed intensive drug enforcement strategies, 91 indicated that these efforts were focused on
reducing the supply of drugs, while 53 indicated that these efforts were focused on reducing the
demand for drugs. This represents almost one-third and one-fifth, respectively, of all
applications. No focus could be determined in the remaining 66 cases.
The next most frequently identified LE activities fell under the prosecution/sentencing
coordination category. These activities involved inter-agency cooperation designed to target
criminal offenders in the designated Weed and Seed neighborhoods. The analysis showed that
184 of the 273 (67%) applications identified this type of activity as part of their overall LE
strategy.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 28
The next three most common LE activities were increased patrol/visibility, task forces, and
targeted crime enforcement. A total of 153 applications (56%) identified the use of increased
patrol/visibility, 148 applications (54%) identified the use of task forces, and 142 applications
(52%) identified some form of targeted crime enforcement
Community Policing Activities (CP)
Community Policing activities are designed to foster the relationships between these groups in
communities
⎯residents, stakeholders and local law enforcement officials. The community
policing (CP) activities were divided into 13 activity-type categories, which are listed in Table
CP2.
The 273 applications contained a total of 1,491 CP activities. The maximum number of
identified activities was 12. As was the case with LE activities, only one site failed to identify a
single CP activity. See Table CP1, below.
Table CP1
Total Number of Community Policing
Activities 1,491
Minimum 0
Maximum
12
Mean
5.46
Median
5
Mode 6
The most common types of CP activities that the sites were conducting were those aimed at
building or strengthening police-community relations. Of the 273 applications, 211 (77%)
indicated that this was a focus of their strategy. Another commonly identified activity was the
use of crime watch /resident patrol programs. A total of 195 applications (73%) specified that
these types of programs were part of their overall strategy.
The next two most commonly identified community policing activities were community
awareness and education. While these activities seem very similar, for this analysis the
following distinction was made. Community awareness refers to activities such as officer-
initiated referrals or efforts to make community residents aware of the department’s activities or
the availability of community or social services. Community education refers to activities such
as classes or the distribution of information pertaining to topics such as crime prevention or
individual/human rights.
A nearly equal number of applications identified the use of each of these activities. A total of 179
applications (66%) reported that increasing community awareness was part of their overall
strategy, while 178 (65%) reported activities designed to educate the community. The next most
commonly identified CP activity type were those activities focused on youth. Youth-focused
activities were identified in 147 of the 273 applications (54%). A complete listing of the 13 CP
activities can be seen below in Table CP2.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 29
Table CP2
Community Policing Activity
Frequency Percent of Sites
Community Policing activity specified
272
99.63%
No Community Policing activity specified
1
0.37%
Type of Community Policing Activity Specified
Police-community relationship building
211
77.29%
Crime watch/ resident patrol
195
71.43%
Community awareness
179
65.57%
Community education
178
65.20%
Youth-focused efforts
147
53.85%
Training for law enforcement personnel
96
35.16%
Bike/horse patrol
92
33.70%
Foot patrol
92
33.70%
Hotlines/anonymous reporting
74
27.11%
Substations 69
25.27%
Improve public perceptions of safety
65
23.81%
Senior citizen focus
50
18.32%
Other 43
15.75%
Total Community Policing Activities
1491
Prevention/Intervention/Treatment (PIT)
Prevention/Intervention/Treatment (PIT) includes a wide variety of social programs and services
that are designed to address various problems or needs in the communities. Governmental
agencies or non-profit community groups commonly provide these services. From the OR
applications, a total of 28 different PIT activities were identified, which are listed in Table PIT2.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 30
Table PIT 1
Total Number of Prevention, Intervention,
and Treatment Activities
3,077
Minimum 2
Maximum
25
Mean
11.27
Median
11
Mode 10
Combined, the 273 sites listed a total of 3,077 PIT activities. The maximum number of
identified PIT activities was 25 and the minimum reported number was two. Both the mean and
median number of identified activities was 11 (see Table PIT1 and Chart PIT1).
An examination of the reported PIT information revealed that the most commonly identified PIT
activities were those directly related to Safe Havens. A total of 228 applications (84%) indicated
that Safe Havens were a major focus of their PIT activities. The second most common activity
types were those related to academic/enrichment programs. For the purpose of this analysis,
these programs were defined as programs or activities involving arts, culture, homework
assistance, and/or tutoring. Of the 273 applications, 202 (74%) indicated that they had some
form of academic/enrichment program.
Life skills/family programs were also clearly a priority for the sites. This category was defined
as programs or activities involving counseling, budgeting, leadership training and parenting
training. Combined, 195 applications (73%) indicated that their effort included these types of
programs. A similar number of applications (194) indicated that they had or planned to
implement programs or activities intended to provide residents with employability/job training.
Other common PIT program or activity types were Community/Social Service Access,
Mentoring, and Drug and Alcohol Education/Prevention. Community/Social Service Access
Number of Identified PIT Activities
0
5
10
15
20
25
30
1
3
5
7
9
11
13
15
17
19
21
23
25
27
Number of PIT Activities
Number of Applications
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 31
strategies were defined as activities directly related to accessibility of programs and involve such
things as centralized service locations, community service centers, and transportation to and from
service providers. A review of the applications revealed that 178 of the 273 sites (65%)
referenced these types of activities. Mentoring and Drug and Alcohol Education Prevention
Strategies appeared with a similar frequency, being noted 168 and 166 times, respectively. A
complete listing of the 28 PIT activities can be seen below in Table PIT2. Additionally, an
explanation of the various activities and or programs included in each activity type can be found
in Appendix B.
Table PIT2
Prevention, Intervention, and Treatment Activity
Frequency
Percent
of Sites
Prevention, Intervention, and Treatment activity specified
273 100%
No prevention, Intervention, and Treatment activity specified
0
0%
Type of Prevention, Intervention, and Treatment Activity Specified
Safe Havens
228 83.52%
Academic enrichment
202 73.99%
Life skills/family programs
195 71.43%
Employability/job training
194 71.06%
Recreation
184 67.40%
Community/social service access
178 65.20%
Mentoring
168 61.54%
Drug and alcohol education/prevention
166 60.81%
Juvenile justice programs
142 52.01%
Drug and alcohol treatment
129 47.25%
Community service volunteer programs
121 44.32%
Health-related
107 39.19%
Job placement
96 35.16%
Other
91 33.33%
Domestic violence prevention/intervention
82 30.04%
Resource directory
79 28.94%
Truancy programs
79 28.94%
Gang prevention
77 28.21%
Conflict resolution
75 27.47%
Literacy
72 26.37%
Reentry initiatives
72 26.37%
GED/school completion
64 23.44%
Victim services
60 21.98%
Boys and Girls Clubs
51 18.68%
English as a Second Language (ESL)
44 16.12%
Housing assistance
44 16.12%
Internship programs
40 14.65%
Teen pregnancy programs
37 13.55%
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 32
Neighborhood Restoration Activities (NR)
Neighborhood Restoration (NR) refers to an assortment of programs and activities that share the
overall goals of stimulating and revitalizing both the economy and physical appearance of the
neighborhoods. In all, 13 neighborhood restoration categories were developed. These 13
activity types are listed below.
Combined, the 273 sites identified 1,425 NR activities. The maximum number of NR-related
activities for a site was 12 and three sites did not indicate the use of any NR activities (see Table
NR1).
Table NR1
Total Number of Neighborhood Restoration
Activities 1,425
Maximum Reported Number
12
Minimum Reported Number
0
Mean Reported Number
5.8
Median Reported Number
6.0
The most common types of NR activities outlined in the strategies were those dealing with
community space restoration. These activities involved restoring previously unusable public
spaces and included activities such as neighborhood, park and vacant lot cleanup projects. In all,
226 strategies (83%) contained descriptions of various community space restoration activities.
The second most commonly identified type of NR was code enforcement. Based on the
information contained in the applications, it was clear that code enforcement was a tool utilized
to address a wide array of issues ranging from problem establishments to absentee property
owners. A total of 189 applications (69%) indicated that code enforcement was a characteristic
of their NR activities and their overall strategy.
The next most frequently identified NR activity was housing improvement/restoration. These
activities generally included assisting residents with making necessary repairs to their homes or
properties. The elderly were a major focus of many of these initiatives. In all, 173 applications
(63%) indicated these types of activities as being part of their strategy. Other commonly
indicated strategies include Community Involvement Initiatives, Home Ownership, and Public
Infrastructure Improvement activities. Of the 273 applications, 167 (61%) indicated that
Community Involvement Initiatives were a focus of their strategy. An equal number of
applications (127) indicated that Home Ownership programs and Public Infrastructure
Improvement were aspects of their overall strategy
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 33
Table NR2
Neighborhood Restoration Activity
Frequency
Percent of
Sites
Neighborhood Restoration activity specified
270
98.90%
No Neighborhood Restoration activity specified
3
1.10%
Type of Neighborhood Restoration Activity Specified
Community space restoration
226
82.78%
Code enforcement
189
69.23%
Housing improvement/ rehab
173
63.37%
Community involvement initiatives
167
61.17%
Home ownership
127
46.52%
Public infrastructure improvement
127
46.52%
Small business development
118
43.22%
New home development 92
33.70%
Landlord/property owner training
87
31.87%
Loan programs
83
30.40%
Building demolition
71
26.01%
Entrepreneurial training
54
19.78%
Other 45
16.48%
Total Neighborhood Restoration Activities
1559
This analysis has demonstrated that while there are many commonalities across sites in the
problems and solutions identified in their strategies, individual sites may have substantial
variations in the number and specific nature of the issues being addressed and the services and
activities being employed. This has major implications for future performance measurement and
evaluation efforts for framing specific research questions and site selection.
Because useful performance measures should reflect those changes that result from Weed and
Seed efforts, it is these strategies that should serve as the basis for developing indicators of local
sites’ successes. Considering that this analysis found a substantial number of site strategies,
which do not clearly present a linkage between the identified community problems and the
solutions proposed, planning for future measures should recognize this difficulty. As expected,
the comprehensive nature of the Weed and Seed philosophy, combined with the large number of
applicants, results in a great variety of strategies and potential sources for data and measures of
efficacy. Later in this report, some performance measures are recommended based in large part
on the strategies and problems laid out in this section. Following are suggestions for future
problem and strategy definition by applicants seeking to achieve OR status and the possibility of
federal funding for their Weed and Seed effort.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 34
Performance Measure Recommendations
As with the wide range of problems and strategies found among Weed and Seed sites, there is a
wide range of strategies used by program evaluators in conducting reviews of Weed and Seed
efforts on the local level. Many locally initiated site studies have been, and continue to be,
conducted. These studies have been funded in part by Special Emphasis grants from CCDO,
other supplemental grants, and local sources.
With the local origin of these efforts, the methodology used is often a product of individual site
strategies, availability of preferred data, and the strategies of individual evaluators. The strength
of this variability is that local evaluators can examine aspects that are most germane to the
specific strategies in place in their community. A weakness is that the evaluations are less
comparable than if there had been consistent standards followed for the types of performance
measures to be used. For example, several site evaluations are based on the effect of Weed and
Seed efforts on reported crime in the Weed and Seed area, while others base conclusions of
efficacy on changes in the number of risk factors to which youth are subject in target areas. Both
types of reports allow for judgments on whether an effect was seen in Weed and Seed areas, but
comparisons of relative success (or failure) become problematic.
In order to facilitate cross-site comparability, this report includes suggested performance
measures for Weed and Seed sites. These suggested performance measures should not be
viewed as a limit to variables that can be examined by Weed and Seed sites sponsoring or
conducting an assessment of their efforts, but a guide of variables to be examined in addition to
other data that will then allow for a more unified look across site evaluations. See below for
specific recommended performance measures for selected common strategies from the review of
OR materials.
Activity
Performance Measure
Notes
Intensive Drug
Enforcement
Reported drug arrest rates for sales
and possession in the intensive
drug enforcement area
For this measure, as with all measures in
this table, it is critical that there be
consistency in the reporting period.
Crime data are currently reported to
CCDO most often on an annual basis,
with a given calendar year compared to
prior and later years. Snapshots may
also be used (for example, comparing
January 2005 to January 2006), but
limiting the pool of data increases the
probability that unrelated factors (such
as a warm week in the month of January)
may influence your measure. At a
minimum, measures should be taken
annually.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 35
Activity
Performance Measure
Notes
Prosecution/Sentencing
Coordination
Survey of criminal justice
participants (e.g. police,
prosecutors) to determine
perceptions of coordination
Example may be annual survey
questions such as "How would you rate
cooperation between police and local
prosecutors for the sentencing of repeat
offenders?" with an ordinal scale for
reply.
Increased Police Patrol/
Visibility
-and-
Foot/Bike/Horse Patrol
Community survey question to
determine level of awareness of
increased police presence in the
community
Example may be annual survey
questions such as "Are the police around
the neighborhood more than before
(baseline date)?" with an ordinal scale
for reply.
Task Forces
Changes in reported crime rates in
area covered by task force
Data should apply to crimes for which
the task force was created to address.
Presumably, for most that will be drug
crimes, but a violent crime task force
would, of course, be measured by
changes in violent crime rates.
Target Crime
Enforcement
Reported crime rates for target
crimes
As with task force data, the crimes
examined should be those for which the
strategy is developed. Investigators may
want to also look at related crimes,
where appropriate.
Juvenile Crime
Enforcement
Changes in rates of arrests of
juveniles
Parole/Probation
Strategies
Changes in rates of parole
revocations/actions for Weed and
Seed area residents
Changes in the rate of Weed and
Seed area residents on probation
A critical element of this is the use of
rates, rather than raw numbers, as the
number of Weed and Seed residents on
parole or probation may vary greatly
from year to year.
Public Order Crime
Enforcement
-and-
Improved Public
Perceptions of Safety
Community survey question to
gauge resident perceptions of
disorder in the Weed and Seed area
Example may be annual survey
questions such as "How do you rate your
neighborhood as a place that is safe and
orderly?" with an ordinal scale for reply.
Gang Reduction
Strategies
-and-
Gang Prevention
Change in rates of reported gang-
related crimes
Investigator may encounter problems of
classification, such as the definition of a
gang-related crime (e.g., is it any crime
committed by a known gang member or
is it any crime with a suspected tie to
gang activities?).
If crime data for gang-related crimes are
not available, a second recommended
measure would be a question in a regular
and reliably conducted youth survey
asking respondents about their
involvement with gangs.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 36
Activity
Performance Measure
Notes
Weapons Reduction
Strategies
Changes in rates of reported
weapons crimes
There will likely be wide variation in the
manner in which local law enforcement
agencies track crimes involving a
weapon. Some agencies may track
weapon involvement for every incident.
In some cases, it may be necessary to
rely upon reports of crime that
specifically deal with the use/misuse of
weapons (such as carrying a concealed
weapon, or improper discharge of a
firearm).
Police-Community
Relationship Building
Community survey question
measuring the satisfaction with
police-community relations
Example may be annual survey
questions such as "How do you rate your
satisfaction with the performance of your
local police department?" with an ordinal
scale for reply.
Crime Watch/Resident
Patrol
Number of incidents reported by
crime watch/resident patrols and
number of patrols by groups
These strategies almost demand a two-
part measure that can reflect both the
growth (or lack thereof) in patrols and
their ability to recognize and report
criminal activity.
Academic Enrichment
Increased grades among program
participants
Due to privacy concerns, local schools
will be hesitant, at best, to provide grade
information for program participants.
The most effective method of getting
grade information is usually by having
participants' parents agree to provide
information from report cards.
Local schools should be able to provide
cumulative grade information as a point
of comparison.
Employability/Job
Training
-and-
Job Placement Strategies
Change in unemployment rate
within the Weed and Seed area or
the number of program participants
employed at given points or for
given periods of time after
program completion.
The measure chosen will depend on the
number of residents served and whether
the intervention is general or specific.
These programs are normally assessed
with process measures (such as
graduation rates or participation
numbers).
Mentoring
Change in the rate of arrests of
program-involved youth
Investigator will want to have an
appropriate point of comparison and will
need to find a group that matches as
closely as possible the group entering the
mentoring program. For example, if the
mentoring program limits itself to at-risk
youth, the comparison group should be
similarly at risk youth.
Also of interest will be the degree of
involvement in the program. Does a
greater degree of participation forecast
probabilities of success?
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 37
Activity
Performance Measure
Notes
Juvenile Justice
Programs
Change in rate of arrests of
program-involved youth
Primary prevention programs should use
the overall rate of arrest among youths in
the Weed and Seed area.
Health-Related Strategies
Changes in emergency room
admissions and changes in use of
preventative care facilities
Specific programs may have varying
degrees of emphasis on preventative
care. Youth-focused efforts may also
look to obtain school absenteeism data
for participants or for the entire school as
appropriate for program focus.
Resource Directory
Change in the number of inquires
and service requests received by
providers listed in the resource
directory
While all local providers are likely to be
listed in the directory, limiting the
potential for a comparison group of
providers not listed, it would be possible
to request the same information from
providers outside the Weed and Seed
area not included in the contact
directory. These providers could then
serve as a comparison group.
Truancy Programs
Changes in local school
absenteeism rates for participating
youth
As with earlier measures, the scope of
youth for whom absenteeism rates
should be examined will depend on the
scope of the truancy programs. Any
community-wide general programs
should examine truancy for an entire
affected school or schools. Programs
that focus on specific populations could
be confined to participating youth.
In the first case, comparisons with trends
in schools not in the target area would be
appropriate. In the latter case,
comparison groups based on factors
related to admission into the program
would be appropriate.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 38
Activity
Performance Measure
Notes
Reentry Initiatives
Changes in recidivism rates over a
specified time period
Any examination of recidivism requires
a number of decisions to be made before
data collection. Among these are the
time period to be considered. Given the
limited timeframe of the federal Weed
and Seed funding cycle, a period of two
years following reentry would be
reasonable (assuming this allows for
completion of the program and a
reasonable post-treatment period).
Other issues are whether to consider
arrest or conviction as a marker of
recidivism (arrest may be a high
standard, especially if police look for 'the
usual suspects,' conviction may require a
longer time frame for comparison to
allow for resolution of cases). Also, a
decision about the seriousness of contact
with criminal justice authorities should
be made; is any arrest or conviction a
marker for recidivism? Or only if a
felony is involved? This will depend in
part on the pool of those being studied.
GED/School Completion
Changes in the proportion of
community residents that are not
high school graduates or GED
recipients
As with earlier measures, the scope of
population for whom GED/school
completion should be examined will
depend on the scope of the program.
Any general programs should examine
rates for an entire affected area while
those focusing on specific populations
could be confined to participants.
As with job training programs, resource
intensive GED/school completion
programs may want to examine cost
versus benefit
Community Space
Restoration
-and-
Building Demolition
Changes in the numbers of
reported crimes at specific
locations
As the purpose of these efforts is often to
remove magnets for offending, changes
under this program should make specific
places less prone to crime. Investigators
with access to GIS analytical systems
may also want to examine changes in the
number of crimes in a surrounding buffer
zone as well.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 39
Activity
Performance Measure
Notes
Housing Improvement/
Rehabilitation
-and-
New Home Development
Changes in the proportion of
owner occupied dwellings within
the Weed and Seed target area
By building new homes and improving
homes in the area, these strategies
encourage long-term residency that will
increase the stake of residents in the
long-term health of the community.
Improved housing stock that is used for
rentals is less likely to have the
ameliorative effect on crime sought
under such programs.
Home Ownership
Increase in the proportion of
owner-occupied housing units in
Weed and Seed area.
Because Weed and Seed is
geographically based, the emphasis will
be on seeing a change within the Weed
and Seed site. While other programs
may show success in program
participants buying homes in any part of
a jurisdiction, Weed and Seed-related
efforts should examine changes within
the Weed and Seed area boundaries.
Small Business
Development
Increase in the number of small
businesses in the target area
A Weed and Seed site of limited size
may be able to collect this and other
information (such as the number of
empty commercial buildings) by
completing a physical survey of the
Weed and Seed area; actually traveling
the streets and noting visible
facilities/lots, etc. An alternate source of
data is the number of businesses paying
taxes within the area or changes in
business tax revenue.
Loan Programs
Change in number of business
financed through loan programs
In addition to the number of businesses
funded, there are other measures that
may shed light on the efficacy of such
programs. Included among these are the
rates of loans paid back and the number
of jobs created by businesses funded
through the program.
Several of the recommended performance measures involve community or resident surveys.
Such surveys pose challenges to Weed and Seed sites that are generally pursuing evaluative
efforts under tight financial constraints. The quality of data collection, however, is critical in any
survey. Some reminders for data collection that will yield useful and reliable data are:
2
• Insure the reliability of any surveys by doing a random sample of residents that represent
all sectors of the community. Do not just survey people who come to Weed and Seed
meetings, for example. If there is a substantial non-English speaking population, the
surveys should be translated into the relevant languages or dialects, reviewed for cultural
appropriateness for the intended subjects, and the surveyors should be prepared to
2
Additional pointers from the American Association for Public Opinion Research can be found here:
http://www.aapor.org/pdfs/best_pra.pdf
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 40
communicate with these groups. In these instances, the cooperation of some individuals
may be improved by also obtaining the support of key leaders in the community. If you
only conduct surveys in the day, you are likely to miss a lot of residents who are at work;
likewise, if you do it at night, you will miss residents who work at night.
• The survey data collection method chosen (i.e., phone, mail, or in-person interviews) may
skew the results by affecting likely respondents, but may have cost tradeoffs. Phone
surveys tend to undercount the young (who use cell phones instead) and the poor (who do
not have phones at as high a rate) and the mobile (for whom phone records are more
likely to be out of date) and so may be ill suited for Weed and Seed areas. In-person
interviews may avoid this but will be much more expensive to conduct and may have
other limitations for accuracy. Sometimes a combination of methods provides a balance
between reliability and cost concerns. Surveyors should be trained so they are consistent
in the delivery of the instrument and do not 'hint' at specific answers (a survey of
opinions on law enforcement, for example, will be much different if collected by
uniformed law enforcement officers.
• Given standards for numbers of surveys needed, methods for assuring random sampling,
and proper instrument design, expert assistance in designing such a process is highly
desirable.
• Because there can be substantial time lags between when changes in public services such
as local policing practices or community conditions like affordable housing availability
are implemented and a widespread awareness of these efforts among residents is
achieved, a survey should only be adopted as a performance measure when there is
sufficient time for this information to diffuse throughout the area. Residents’ perceptions
may be longstanding opinions that will require substantial and sustained counterevidence
before significant change is reflected in survey results.
• Also relative levels of resident satisfaction as expressed in surveys may be filtered
through different expectations for services or community change, e.g., is a neighborhood
considered safer when crime is reduced or only when crime is totally eliminated?
Though surveys can be time consuming and expensive, there are few sources that can provide as
useful and accurate information for some questions as a properly conducted survey.
Other sources and types of data have their own limitations on how they should be used or
viewed. In analyzing crime data, for instance, one should remember that the data generally
reflect incidents reported to police, and that the majority of all crimes are never reported. Certain
tactics (e.g., community policing, intensive law enforcement) have been found to change the rate
of crime reporting, so even if crime is going down, in the short term the number of reported
crimes may rise. Given differences in local laws, the same event may be reported as different
offenses (or not be an offense in extreme cases) in different jurisdictions. Try to understand the
limitations of your data and how they can affect your analysis.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 41
Revised Homicide Data Analysis Using Expanded CCDO
Administrative Data Set
As part of this project’s tasks, CCDO and JRSA conducted a detailed review of the
administrative records for local sites including Official Recognition (OR) notifications, grant
awards and progress reports with the goal of obtaining detailed and accurate information
regarding the key dates for the implementation of local Weed and Seed sites’ activities. Because
these dates are not part of the GPRA information processed by JRSA for CCDO, it has been
sporadically obtained at various times in the past. Consequently, this has meant that for various
reasons these dates were missing for some sites, which prevented their inclusion in analyses
using site implementation dates.
With more complete and accurate Weed and Seed site start date information, it would be possible
to determine the time periods between when sites receive OR status, obtain their first Weed and
Seed grant awards, and commence activities. The Weed and Seed homicide and crime trend
reports could be substantially improved because it would expand the number of site records
available for any analysis efforts, which would result in more definitive pre/post periods for
comparison of crime and programmatic statistics. While limited pre/post Weed and Seed site
analyses has been conducted in the past, it was hoped that hundreds rather than a score of sites
could be more rigorously studied.
To date, the most extensive Weed and Seed impact assessment is a draft 2005 GPRA Weed and
Seed homicide data study that has information regarding 435 individual Weed and Seed sites.
The draft 2005 GPRA Weed and Seed homicide study is an update of a report entitled, A
Comparison of Homicide Trends in Local Weed and Seed Sites Relative to their Host
Jurisdictions, which was published in 2003. The homicide data file that was developed for the
draft 2005 GPRA Weed and Seed homicide study was used for these analyses because it includes
information for about 100 additional sites and has more recent homicide data than the published
2003 report.
Of these 435 sites, 274 sites have sufficiently complete and accurate historical data for producing
time series comparisons of homicides within Weed and Seed sites and their host jurisdictions.
There are two main reasons for the drop-off in the number of sites available for analysis. First,
many of the sites were so new (110) that they were either in their initial years of operation or
they have only a single year (sometimes only a part of a year) of operations. Second, some sites
(51) had seriously flawed GPRA homicide reports that could have been due to a wide range of
reasons that rendered them invalid for inclusion in this study.
The methodology for the draft 2005 GPRA Weed and Seed homicide data study was simply an
examination of the following questions:
• Have homicides decreased in the Weed and Seed sites over time?
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 42
• And, how does the change in homicides in the Weed and Seed sites compare to the
homicide statistics for the remainder of their host jurisdictions?
That analysis found the number of homicides had decreased in 56% of the Weed and Seed sites,
remained stable in 19%, and increased at a lower rate than the host jurisdiction in 5% of the sites.
In only 21% of the Weed and Seed sites had homicides increased.
The draft 2005 GPRA Weed and Seed homicide data study examined the trend in homicides
without knowing precisely when the Weed and Seed operations actually began; therefore it is
impossible to conduct a pre/post comparison examining the timing associated with Weed and
Seed implementation. The following assessment combines the homicide time series data from
the 274 sites included in the draft 2005 GPRA Weed and Seed homicide data study with the
expanded Weed and Seed start date information found in the CCDO administrative records. It
was anticipated that there would be a strong likelihood of matches between these two data sets
and that it would provide a more effective test of the hypothesis that following the
implementation of a Weed and Seed strategy, homicides should decrease.
In 2004, JRSA did perform a four-year homicide data comparison using the OR award and first
grant award date information that was available to it at that time. In this first effort to capture
Weed and Seed “start dates”, significant data collection issues were encountered. Consequently,
the 2004 pre/post implementation homicide crime analysis was based on data for only 54 Weed
and Seed sites. For these sites, a 30% reduction in homicides occurred three years after
implementation of their Weed and Seed strategies. One of this project’s objectives was to further
expand the number of useable site records for studying Weed and Seed’s effects on community
crime levels.
Local Site Strategy Implementation and Other Key Event Dates
Key strategy implementation dates collected for the 2005 expanded CCDO administrative
records data set include the Official Recognition date, date of the first Weed and Seed grant
award, and the site “start date”.
Official Recognition status refers to the extensive planning proposal that is submitted for
review before a jurisdiction can be invited to submit a Weed and Seed grant proposal.
The first award date is the date CCDO gives to a site’s first approved grant proposal.
The “start date” is the day that a Weed and Seed site reports actually beginning Weed
and Seed activity.
Of these three initial Weed and Seed strategy implementation dates, the OR date and the actual
activity start date reported by the local sites were considered the two most important dates for
this analysis.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 43
In addition to these three key strategy implementation dates, other important administrative
information regarding the Weed and Seed sites was collected. For instance, whether or not the
grant application was initially accepted or denied is recorded. Often when the initial grant is
denied, the site will have a delayed start. Whether or not a denied site eventually does begin
operations can be verified by confirming if the site has ever provided first and subsequent years’
progress reports. A site with a denied first grant and no subsequent progress reports was judged
as not ever starting. The lack of GPRA reporting for these sites provides additional verification
that these sites never really began operations.
The availability of the expanded CCDO administrative records for 321 sites raised the possibility
that there could be significant overlap with the 274 sites that have valid homicide data sets and
were used in the draft 2005 GPRA Weed and Seed homicide data study. This would allow a
more authoritative analysis using the various strategy implementation dates. In the end, it was
possible to link the expanded CCDO administrative records with 100 sites in the draft 2005
GPRA Weed and Seed homicide data study, which was nearly a doubling of the number of sites
for the previous pre/post analysis. The details regarding the record linking process and reasons
that information for many local sites could not be linked can be found later in this report in
Appendix C - Summary of CCDO Administrative Findings.
Official Recognition Versus The Site Reported “Start” Date: Which is the Real Beginning for
Strategy Implementation?
As previously stated, for a community to receive the federal Weed and Seed designation or OR
status, it has to pass an extensive pre-screening process. This requires the submission and
approval of an OR plan that is reviewed by CCDO. The review process has multiple criteria,
which require the interested jurisdictions to describe their plans and organizational structure for
implementing a Weed and Seed strategy that complies with the national model. A failing score
on any of the major criteria can result in not being invited to submit an application for funding.
The process of creating an OR plan requires substantial interagency cooperation and
organizational development which lead some to believe that achieving this by itself should be
viewed as the date that the Weed and Seed strategy actually begins. Others argue that while the
OR process is an important initial stage in creating a Weed and Seed site, most sites’ strategies
are not meaningfully implemented until federal funding and CCDO support are guaranteed and
the site staff are hired and working.
At first glance, there appears to be some validity to the question as to whether the OR date or the
“start date” for Weed and Seed sites is the most appropriate measure indicating the beginning of
Weed and Seed. For instance, there is an average period of 270 days between the OR date and
the actual program start date. However, of the 100 Weed and Seed sites where the CCDO
administrative records and the 2005 GPRA Weed and Seed homicide study data sets overlap, the
records for 40 (while having an OR date) do not have a start date.
Moreover, in 52 of the 60 other sites, the OR and start dates actually occur within the same
calendar year. Since homicide statistics are reported annually, this means that in the vast
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 44
majority of cases, the OR and start dates for Weed and Seed homicide analysis are not
substantially different for the purpose of measuring the effects of strategy implementation.
Of the eight cases, in which the OR and start date are not in the same calendar year, the OR
versus start date analysis becomes moot for three, because when the OR date is used as the
initiation point for a Weed and Seed site, it coincides with the first year of homicide reporting
⎯thus there is no baseline history⎯and therefore no “pre” time period for the pre/post analysis.
In the remaining five cases, in which the OR precedes the start date by at least one calendar year,
the comparative results are only substantially different in one case.
Finally, while there is generally about a nine-month lag between the OR and the Weed and Seed
start dates, this amount of time is not long enough to create a significant difference for the annual
Weed and Seed homicide statistics. Therefore, in the remainder of this report, where the Weed
and Seed start date is available, it is used as the pre/post demarcation. Where the actual start date
is missing from the CCDO administrative records, which includes 40% of cases in the final
analysis, the OR date is used as a surrogate start date with the assumption that if the actual start
date was known, it would not be significantly different from the OR date.
Pre/Post Strategy Implementation Analysis Using 2005 Weed and Seed Homicide Data Set
In an effort to maximize the number of Weed and Seed sites in the pre/post analysis, both the
official recognition and start dates were considered as possible “start dates.” This method of
measurement is used because, as shown above, the analytical difference between the two is
minimal.
Using this type of measurement for Weed and Seed start dates, there were 100 sites with a match
between the expanded CCDO administrative date data set and the draft 2005 GPRA homicide
data set. This means that the pre/post analysis includes only 37% (100 out of 271) of the GPRA
homicide database’s sites. The majority of these matched sites (60 percent) have both a start date
and an OR date. For the remaining 40 percent of the sites only the OR dates were used. This
degree of matching with CCDO’s administrative records, while less than ideal, still results in an
improved data set for “exploratory” pre/post analysis.
Further attrition in the number of matched Weed and Seed sites occurred because some of the
CCDO start dates are not the original site start dates – but appear to be start dates for an interim
award or grant revision that took place after the actual site start date. When the full draft 2005
GPRA homicide database was compared to the CCDO “known” start or official recognition date,
it was clear that GPRA forms were far more complete and included more historical reporting
than the CCDO start date database. Sites where CCDO start dates seem to be an interim date
instead of the actual start date were removed from the analysis. Increasing the number of “valid”
sites in for future pre/post analyses would involve further searches of CCDO’s files to identify
the actual site start dates and may require reviewing archived records for some of the pre-1999
local sites.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 45
The first step taken in the CCDO start date analysis was to pick the start year for each site’s
strategy implementation. This same start year was also used for homicides reported in the
remainder of the host jurisdiction.
Because of the variability of homicide statistics for a single year, the average number of
homicides was calculated for the pre/post periods for each Weed and Seed site and each host
jurisdiction. Because of the high percentage of mismatches between the CCDO administrative
records and the GPRA homicide data, most of the jurisdictions with multiple active Weed and
Seed Sites were reduced to only one site where a start date was known, thus minimizing the
validity of the “jurisdictional” analysis used in the draft 2005 GPRA homicide report.
Furthermore, it is not uncommon for jurisdictions with multiple Weed and Seed sites to have
different start dates for each of the sites. Therefore, in this analysis the multiple sites are treated
separately, each with its own unique start date without consideration of the impact of other active
Weed and Seed sites within the host jurisdiction.
The first analysis involved a comparison of the average number of homicides per site prior to the
start date compared to the average number of homicides following the start date.
Findings: Analysis of Pre/Post Weed and Seed Strategy Implementation Changes in
Homicides
For the 100 sites where the draft 2005 GPRA homicide database could be matched to the
expanded CCDO administrative records “start date” database, the average number of homicides
prior to Weed and Seed was 4.93. Following the implementation of Weed and Seed in these
sites, the average number of homicides decreased to 3.57 – an average decrease of 1.4 homicides
per site – representing a 28% decrease in homicides.
This may appear to be very small until the change in homicides in the remainder of the host
jurisdictions is considered. In the rest of the host jurisdictions, which have much larger
populations and volume of homicides than the Weed and Seed sites, there was also a decrease in
the number of homicides after the implementation of Weed and Seed. The average decrease in
the remainder of the host jurisdictions was, however, only .5 homicides. The decrease in
homicides within Weed and Seed sites is three times that which occurred in the remainder of the
host jurisdictions. In other words, the pre/post analysis shows that there was a significant
decrease in homicides following the implementation of Weed and Seed in the sites (even when
combining the very successful and the unsuccessful sites). During this same time period, there
was very little change in the host jurisdictions’ homicide rates.
Comparing the Pre/Post Strategy Implementation Analysis Results to the Draft 2005 Weed
and Seed Homicide Data Analysis
The pre/post analysis for the 100 sites, for which the draft 2005 GPRA homicide database and
expanded CCDO administrative records database can be matched, provides a more valid analysis
of Operation Weed and Seed. While this method provides promising results, it is still limited by
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 46
the number of cases (about one-third of sites with data available). The pre/post analysis for the
100 sites also does not consider the sites’ performance relative to comparable trends for their
host jurisdictions. These limitations raise the question as to whether the simpler analysis used
for the GPRA homicides studies, which cover 274 sites, might still be of value.
To help address this question, a more detailed review of the similarities and differences of the
pre/post strategy implementation analysis and the draft 2005 GPRA homicide data study was
undertaken. The bridge between the two analysis methods is the Weed and Seed Homicide
Relative Change Scale. The results from both the pre/post strategy implementation analysis and
the draft 2005 Weed and Seed homicide data analysis can be translated into scores using the
Homicide Relative Change Scale. This analytic approach and the differences in the finding are
discussed below.
The Homicide Relative Change Scale
Because the sites’ target areas were selected due to their persistent and severe crime problems, it
is reasonable to assess their performance as a function of change relative to the surrounding
jurisdictions. This alternative approach adopts the viewpoint that, given this history, these
communities would be expected to have crime trends equal or worse than their jurisdiction
without the implementation of a Weed and Seed strategy. This was the method used for the draft
2005 GPRA Weed and Seed homicide report. The pre/post average number of homicides for
sites and the remainder of the host jurisdiction can easily be translated to the Homicide Relative
Change Scale used in the draft 2005 GPRA homicide report. It is just a matter of switching from
a linear measure (i.e., employing an arithmetic average) to a ranked order score as was used for
the draft 2005 GPRA homicide data study.
The Weed and Seed Homicide Relative Change Scale assesses homicide trends for local sites in
relation to similar changes occurring in their host jurisdiction. The Relative Change Scale ranks
all of the possible combinations of changes in both the sites and the jurisdictions according to the
degree of success for a site. The most successful category or ranking are those local sites that
have experienced substantial decreases in homicides while their host jurisdictions have had
increasing homicide trends. The least successful category or ranking is when homicides
increased in the Weed and Seed site but decreased in the remainder of the host jurisdiction. The
different combinations of homicides decreasing, increasing or remaining stable results in a 13-
step scale from the most successful to the least and is represented numerically from a +7 to a -5.
For example, in the draft 2005 GPRA Weed and Seed homicide data study, if the number of
homicides in a Weed and Seed site data series decreased by five homicides and the remainder of
the jurisdiction only declined by three homicides during the same time period, the site would be
represented as a site where its decrease in homicides was greater than the decrease in homicides
in the remainder of the host jurisdiction – and would be scored as a + 6.
Translating the pre/post average number of homicides for the same site and host jurisdiction was
done in the following manner. The average number of homicides in the Weed and Seed site
preceding strategy implementation was 12 and for the post-implementation period, the average
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 47
was 5. Therefore, the average number of homicides decreased by 7 after the implementation of
Weed and Seed. At the same time, the number of homicides in the remainder of the jurisdiction
decreased from an average of 75 prior to Weed and Seed implementation to 70 for the time
period afterwards. This means that homicides only decreased by 5 in the remainder of the host
jurisdiction and so this would be classified as a site where homicides decreased more in the
Weed and Seed site than in the remainder of the jurisdiction – and would also be scored as a +6.
Weed and Seed Homicide Relative Change Scale
7 Site Decreases & Jurisdiction Increases
6 Site Decreases > Jurisdiction Decreases
5 Site Decreases = Jurisdiction Decreases
4 Site Decreases < Jurisdiction Decreases
3 Site Decreases & Jurisdiction Stable
2 Site Stable & Jurisdiction Increases
1 Site Stable & Jurisdiction Stable
0 Site Stable & Jurisdiction Decrease
-1 Site Increases < Jurisdiction Increase
-2 Site Increases & Jurisdiction Stable
-3 Site Increases = Jurisdiction Increases
-4 Site Increases > Jurisdiction Increases
-5 Site Increases & Jurisdiction Decreases
Findings: Comparison Between the Pre/Post Analysis Results and the Draft 2005 Weed and
Seed Homicide Analysis Using the Relative Change Scale
How do the average relative change scores for the pre/post analysis data set compare to the draft
2005 GPRA homicide study results? Applying the Homicide Relative Change Scale to the 100
sites used in the pre/post strategy implementation analysis, the average score was +2.0. This was
just a little bit less than the result of +2.15 for the draft 2005 homicide data analysis.
While the difference between the two techniques does not change the direction or the magnitude
of success, it is very important to note that the more accurate results from the pre/post strategy
implementation data set differ significantly from the Relative Change cores from the 2005
GPRA homicide report in 28 cases (28%). Interestingly, 14 of the 100 sites switch from a
positive to a negative Relative Change Scale score – when compared to the original GPRA
homicide study. On the other hand, the number of CCDO sites that changed from a negative to
positive score also number 14. Therefore, it is to some degree serendipity that the pre/post
average relative change scores are so similar to the draft 2005 homicide study results. However,
the most important thing is that despite the methodological differences and issues, the Weed and
Seed homicide studies both indicate strong positive outcomes for the Weed and Seed strategy.
The following summary of findings indicates how the pre/post strategy implementation data set
differs from the larger draft 2005 homicide analysis data set. As noted above, while the results
for the two data sets are very similar, the distribution across the possible range of scores varies.
The results from the pre/post strategy implementation data set tend to push sites to the extremes,
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 48
when compared to the draft 2005 homicide study data set. In other words, for the pre/post
strategy implementation data set, there are more very successful sites (23% versus 16%) and
very unsuccessful site (24% versus 13%: using -4 and -5).
Distribution of Sites by Relative Change Scores:
Comparing the Draft 2005 Homicide Data Study Data Set
To the Pre/Post Strategy Implementation Analysis Data Set
2005 Homicide Data Study Pre/Post Analysis
16%
7 Site Decreases & Jurisdiction Increases
23%
9%
6 Site Decreases > Jurisdiction Decreases
7%
3%
5 Site Decreases = Jurisdiction Decreases
3%
22%
4 Site Decreases < Jurisdiction Decreases
27%
4%
3 Site Decreases & Jurisdiction Stable
2%
5%
2 Site Stable & Jurisdiction Increases
3%
3%
1 Site Stable & Jurisdiction Stable
0%
8%
0 Site Stable & Jurisdiction Decrease
3%
11%
-1 Site Increases < Jurisdiction Increase
3%
1%
-2 Site Increases & Jurisdiction Stable
1%
4%
-3 Site Increases = Jurisdiction Increases
4%
5%
-4 Site Increases > Jurisdiction Increases
14%
8%
-5 Site Increases & Jurisdiction Decreases
10%
The pre/post strategy implementation data set provides a more precise measure for assessing the
success of the Weed and Seed strategy across sites. This approach’s main shortcoming is that
the results are limited to about only a third of the sites for which valid homicide data are
available because of incomplete site start date records. While the site/host jurisdiction homicide
trend comparison results may not be as precise as the pre/post strategy implementation analysis,
they provide useful information for almost three times as many sites. Until more complete
administrative information is available for site start dates, a combination of both methods to
track and assess the performance of local sites can provide a more comprehensive picture of the
National Weed and Seed Initiative.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 49
Appendix A: Strategy Categories
Law Enforcement
___ Intensive Drug Crime Enforcement
___ Supply
___ Demand
___ Firearm Crime Reduction Strategies
___ Gang Crime Reduction Strategies
___ Hotspot Policing
___ Increased Police Patrol/Visibility
___ Juvenile Crime Enforcement
___ Known/Repeat Offender Strategies
___ Parole/Probation Involvement Strategies (youth &
adult)
___ Prosecution/Sentencing Coordination Efforts
___ Public Order Crime Enforcement
___ Task Forces
___ Target Crime Enforcement________________
(Crime type)
___ Traffic Violation Enforcement
___ Use of Technology
___ Other (Describe): ______________________
Community Policing
___ Bike/Horse Patrol
___ Community Awareness
___ Community Education (rights, LE efforts, etc.)
___ Crime Watch/Resident Patrol Programs
___ Foot Patrol
___ Hotlines/Anonymous Reporting
___ Improving Public Perceptions of Safety
___ Police-Community Relationship Building
___ Senior Citizen-Focused Efforts
___ Substations/Storefronts/Mobile Command Centers
___ Training for Law Enforcement Personnel
___ Youth-Focused Efforts
___ Other (Describe): _________________________
Prevention/Intervention/Treatment
___ Academic/Enrichment Programs
___ Boys & Girls Club
___ Conflict Resolution
___ Community/Social Service Access
___ Community Service/Volunteer Programs
___ Domestic Violence Prevention/Intervention
___ Drug/Alcohol Education/Prevention
___ Drug/Alcohol Treatment
___ Employability/Job Training
___ ESL Programs
___ Gang Prevention Programs
___ GED/School Completion Programs
___ Health-Related Programs
___ Housing Assistance
___ Internship Programs
___ Job Placement
___ Juvenile Justice Programs
___ Life Skills/Family Programs
___ Literacy Programs
___ Mentoring Programs
___ Recreational Programs
___ Re-entry Initiatives
___ Resource Directory
___ Safe Havens
___ Teen Pregnancy Programs
___ Truancy Programs
___ Victim Services
___ Other (Describe): __________________________
Neighborhood Restoration
___ Building Demolition
___ Code Enforcement
___ Community Involvement Initiatives
___ Community Space Restoration/Improvement
___ Entrepreneurial Training
___ Home Ownership Programs
___ Housing Improvement/Rehab
___ Landlord/Property Owner Training
___ Loan Programs
___ New Home Development
___ Public Infrastructure Restoration/Improvement
___ Small Business Development/Expansion
___ Other (Describe): __________________________
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 50
Appendix B: Program/Service Classification Guide
Program/Service Name/Type
Category
Law Enforcement
SRO (School Resource Officer)
Juvenile Crime Enforcement
GREAT
Gang Crime Reduction Strategies
Littering
Loitering
Noise
Nuisance abatement/crime enforcement
Trespassing
Public Order Crime Enforcement
Community Policing
Crime prevention techniques
Human/individual rights
Community Education
Available community/social services
What the department is doing in the community
Community Awareness
Prevention/Intervention/Treatment
Cultural programs
Homework help
Tutoring
Arts/Enrichment
Centralized service location
Community service centers
Transportation to/from services
Community/Social Service Access
Domestic violence counseling
Domestic violence prevention
Domestic Violence Prevention/Intervention
DEFY – Drug Education for Youth
Drug/Alcohol Education/Prevention
Health education/information
Sex education
Health Programs
Aftercare
At-risk youth programs
Youth council/court
Youth offender programs
Juvenile Justice Programs
Budgeting
Counseling
Leadership
Parenting
Life Skills/Family Programs
Distribute list of services/where to get
Database of available services
Resource directory
Neighborhood Restoration
Beautification
Clean-ups
Creating community gardens
Murals
Cleaning up vacant lots/public places
Community Space Restoration
Building improvement/restoration
Improving public transportation system
Improving/creating sidewalks & roads
Lighting
Public Infrastructure Restoration/Improvement
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 51
A
PPENDIX
C: Summary of Expanded CCDO Administrative Records
Data Set Findings
This CCDO/JRSA effort resulted in a more complete documentation of 321 separate Weed and
Seed sites. One site was duplicated (site 435 Athens GA) which reduced the count from 320. In
contrast, the GPRA homicide database has 274 sites that have sufficiently complete and accurate
historical data that are used for a “crime change” analysis. These data are used to assess the
change in homicide patterns in Weed and Seed sites, which are then compared to the change in
homicides for the sites’ host jurisdictions. In addition, the GPRA homicide database has partial
information for an additional 160 sites for a total of 434 sites. Of these, 51 sites have extensive
historical site data, but that information is unreliable with inexplicable differences across the
annual GPRA reports. The other 109 Weed and Seed sites are so new that they have little actual
Weed and Seed activity (none to one year) with which to justify a change analysis of their
homicide data.
It was hoped that there would be a significant overlap between the 320 sites with information
regarding implementation start dates for Weed and Seed sites and the 274 sites with complete,
multi-year homicide data. The first review of the CCDO start date database showed that:
317 out of 320 sites (99%) had an Official Recognition date, and
221 out of 320 sites (69%) had a first grant award and site activity start dates.
3 out of 320 (1%) did not have an Official Recognition date. These were Site ID No. 471 -
Minneapolis MN, Site ID No. 057 - Des Moines IA, and Site ID No. 473 - St. Paul MN.
First Award Dates and Start Dates
The relationship between first grant award dates and the site activity start dates is complex. On
average, the reported site activity start date for Weed and Seed sites occurs 88 days before the
first award date. But this average is somewhat misleading because some sites activity “starts”
before the first award date and some start after.
Of the 221 sites with both first award and start dates before the first award, 167 (76%) start
before the first award day, on average 125 days before the first award. This means that about
three-fourths of the sites actually begin their local Weed and Seed activity prior to, in
anticipation of the receipt of their first grant award.
Of the 221 sites with both first award and start dates, 54 (24%) start after the first award day, the
average delay between the first award date and the site start date is 30 days.
Official Recognition versus the Start Dates
There are 219 sites out of 320 (69%) with Official Recognition dates and start dates and first
award dates.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 52
For these 219 sites, the Official Recognition dates precede the start dates by an average of 270
days.
Also for these 219 sites, the Official Recognition dates precede the first award dates by an
average of 359 days.
Actually as described above, the relationships between Official Recognition dates and either the
start date and the first award date can go to another level of definition where the Official
Recognition dates are compared to those sites where the first award dates precede the site start
date and visa versa.
Chronology of Sites with Official Recognition (OR) But Missing Start Dates
There are 98 sites with Official Recognition dates but are missing both the first grant award and
the site activity start dates. It would be natural to expect that the missing dates would be heavily
weighted to the older sites. While there is a significant number of missing start dates for older
sites, the table below shows that the missing information is by no means limited to these old sites
OR Year Number of Missing Start Dates
1998 1
1999 17
2000 25
2001 10
2002 8
2003 7
2004 9
2005 21
M
ATCHING
CCDO OR
AND
S
TART
D
ATES AND
GPRA 2005 H
OMICIDE
D
ATA
Out of the 320 CCDO sites, 284 (89%) can be matched to the 2005 homicide analysis database
which has 434 sites going back to 1996. However, the 2005 homicide database has data quality
limitations that result in only 274 sites being analyzed as to their level of success at reducing
homicides within the Weed and Seed sites. These 274 sites have GPRA homicide reports that
have reliable reporting over the years.
There also 51 sites in the 2005 homicide database, while having multi-year information, are not
included in the homicide analysis because are large differences across the GPRAs submitted for
these sites that cannot be easily reconciled.
There are another 63 sites in the GPRA 2005 homicide database that have 3 years of history and
either a single year or a partial year of Weed and Seed site activity and are not yet included in the
homicide change analysis.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 53
Finally, there 43 new sites that have only historical homicide data being reported as part of the
2005 homicide database and cannot be used as part of the homicide change analysis.
The following table indicates how well the GPRA homicide database and the CCDO official
recognition and site start data match for each of the GPRA homicide database groups.
Matching GPRA Homicide Weed and Seed Sites
And CCDO OR and Start Dates
Category of GPRA
2005 Homicide Data GPRA Sites CCDO OR Sites CCDO Start Date Sites
Valid Data
274
147 (54%)
97 (35%)
Questionable Data
51
30 (59%)
25 (49%)
New Sites: Only 1
st
Year 63 56 (89%) 48 (76%)
New Sites: History Only 46 43 (93%) 37 (80%)
As discussed earlier in the CCDO start date database, there are many more records with only
Official Recognition dates than those with both Official Recognition and start dates. This
situation limits the ability to match these records to the GPRA homicide database. While 54
percent of the sites with OR dates match to the GPRA homicide database’s records with
complete and valid data; only 35 percent with both dates overlap. While there may be some
expectation that older sites may be more difficult to match, it is probably even more
disconcerting that the overlap between the newest Weed and Seed sites (i.e., those initiated
within the past two years) and the GPRA homicide database is lower than expected. These new
sites are still missing 20 to 24 percent of the start dates. This may be due to the carryover of
funding for these sites to the next fiscal year.
Weed and Seed Performance Measures:
Analyzing and Improving Data Resources
Justice Research and Statistics Association
Page 54
Bibliography
Dunworth, T., Mills, G., Cordner, G., & Greene, J. (1999). National Evaluation of Weed and
Seed Cross-Site Analysis [NCJ 176358]. Washington, DC: National Institute of Justice.
Retrieved February 22, 2006 from
http://www.jrsa.org/weedandseedinfo/studies-national.htm
General Accounting Office. (2004). Grants Management: Despite Efforts to Improve Weed and
Seed Program Management, Challenges Remain (GAO 04-245). Washington, DC: Author.
Retrieved February 22, 2006 from
http://www.gao.gov/new.items/d04245.pdf
O’Connell, J., Perkins, M. & Zepp, J. (2004). Weed and Seed Crime Pattern Data Analysis.
Washington, DC: Justice Research and Statistics Association. Retrieved February 22, 2006 from
http://www.jrsa.org/weedandseedinfo/studies-other.htm
O’Connell, J., Perkins, M. & Zepp, J. (2003). A Comparison of Homicide Trends in Local Weed
and Seed Sites Relative to Their Host Jurisdictions, 1996 to 2001. Washington, DC: Justice
Research and Statistics Association. Retrieved February 22, 2006 from
http://www.jrsa.org/weedandseedinfo/studies-other.htm
Perkins, M. & Zepp, J. (2004). Weed and Seed local Evaluations Meta Analysis. Washington,
DC: Justice Research and Statistics Association. Retrieved February 22, 2006 from
http://www.jrsa.org/weedandseedinfo/studies-other.htm
File contents
Acknowledgements
JRSA research assistance: Jessica Blair, Jesse Handsher, Marina Karzag, Eileen
McDermott, Mary Poulin, and Jason Trask
777 North Capitol St., NE, Suite 801
Washington, DC 20002-4239
(202) 842-9330 - phone
(202) 842-9329 - fax
cjinfo@jrsa.org - email
jrsa.org - Web site
This research described in this report was supported by Grant No. 2005-WS-Q5-K007 awarded
by the Community Capacity Development Office, Office of Justice Programs, U.S. Department
of Justice. Points of view in this document are those of the authors and do not necessarily
represent the official position or policies of the U.S. Department of Justice.
EXECUTIVE SUMMARY .......................................................................................................... 1
OVERVIEW OF THE NATIONAL WEED AND SEED INITIATIVE ................................. 6
PREVIOUS WEED AND SEED EVALUATION EFFORTS .................................................. 9
METHODOLOGY ..................................................................................................................... 13
REVISED HOMICIDE DATA ANALYSIS USING EXPANDED CCDO
ADMINISTRATIVE DATA SET.............................................................................................. 41
APPENDIX B: PROGRAM/SERVICE CLASSIFICATION GUIDE .................................. 50
APPENDIX C: SUMMARY OF EXPANDED CCDO ADMINISTRATIVE RECORDS
DATA SET FINDINGS .............................................................................................................. 51
BIBLIOGRAPHY ....................................................................................................................... 54
Analyzing and Improving Data Resources
This project had three major objectives that would help to expand the available information for
assessing the performance of individual local sites as well as provide indicators for the overall
National Weed and Seed Initiative.
able to categorize the sites according to particular community issues. This would allow
future evaluation and performance measurement efforts to concentrate on sites that are
clearly involved with specific concerns and would reasonably be expected to have results
related to that topic.
2. Additional research work was done with 100 local sites to explore data resources cited in
and performance measures. A listing of possible measures and considerations was
developed from this effort.
additional documentation of the National Weed and Seed Initiative’s impact on crime as
indicated by homicide statistics.
Objective 1
Although the site strategies are reviewed individually through the Weed and Seed Official
Recognition application process, an aggregate analysis of the site strategies had not been done to
identify the commonalities and differences across the sites regarding the specific community
problems being addressed and the range of solutions being pursued. This study was able to
determine the range of issues and the relative complexity of the overall strategies developed by
the sites.
To accomplish this task, the strategy documents were retrieved from CCDO’s administrative
records for the 309 local sites that received Official Recognition (OR) status from FY2000 to
FY2005. A rigorous review process was used to develop a classification scheme for the
community problems and solutions cited in the strategies. Following this step, each site strategy
was then categorized using this approach.
The report presents the community problem/strategy solutions classification schemes, the
frequency of specific problem/solution references by the sites, and an analysis of the distribution
of sites by the categories.
The conclusion of this effort is that while, as a whole, the sites share many commonalities in the
problems being addressed and the solutions adopted, individually there can be substantial
Analyzing and Improving Data Resources
performance measures should take this diversity across the sites into account.
Objective 2
The follow-up contacts with the local sites found that many have information resources that
could contribute to research on Weed and Seed’s effects on communities. However, there is
great variability in both content and reporting capabilities across jurisdictions that has to be
considered when designing any national information collection effort. For answering specific
questions about Weed and Seed efforts, these resources could provide additional data that would
respond to various inquiries.
Objective 3
Previously studies have examined the homicide statistics submitted in the local sites’ annual
Government Performance and Results Act (GPRA) reports as a common indicator of Weed and
Seed’s impact on crime. A limitation of these analyses has been incomplete information on key
strategy implementation dates for many sites. One of this project’s tasks was a search of
CCDO’s administrative records to locate these missing dates. Additional date information was
located for 321 sites.
Because the analysis of crime levels before and after a site’s strategy implementation requires a
complete set of implementation dates and multi-year homicide statistics, the number of sites
available for study can be affected by several factors. Due to the initial step of obtaining OR
status before a local site can receive federal funding, there are actually three possible dates that
could be used for analyzing pre/post strategy implementation. Not all dates were available for all
sites due to various recordkeeping issues and the age of some sites. Changes in site boundaries
and other developments prevent some sites from having valid data for a time series analysis.
Some sites have not provided complete, multi-year data sets, or inconsistencies were found in
some records that could not be readily resolved. Because some sites have not existed for a
sufficient time to have data for all of the years needed for the analysis, their records are currently
incomplete.
As was previously mentioned, three possible dates could be used as the start of a site’s strategy
implementation
if there was a significant difference for studying the implementation of strategies. The vast
majority of sites were found to have times between OR and the start of activities that were
insignificant for studying pre/post implementation effects. Of the five sites with at least a year
between OR and their activity starting date, only one site had any substantial difference in
pre/post results.
This project was able to perform a pre/post strategy implementation analysis for 100 sites. This
was nearly double the number of sites available for a previous 2004 study. These sites had an
average decline of 1.4 homicides following implementation of their strategies compared to a
Analyzing and Improving Data Resources
done using a Relative Change Scale that characterizes the sites’ performance as a function of the
changing conditions in their respective host jurisdictions. This analysis also found that the sites
have a very positive effect on crime in their target areas when viewed from the perspective of the
trends occurring in the surrounding jurisdictions.
Analyzing and Improving Data Resources
This report expands the information available for the study of the National Weed and Seed
Initiative. Specifically, it identifies and categorizes the diverse community problems and
solutions cited across the local Weed and Seed sites’ strategies, explores possible performance
measures and data sources for future site assessments, and reanalyzes the local sites’
Government Performance and Results Act (GPRA) homicide data using different key dates in
the process of site strategy implementation to determine possible effects on crime trends. The
intent of this research effort was to provide a more comprehensive understanding of the common
features and differences among Weed and Seed participants and to explore the potential for
additional data sources for measuring the sites’ performance and impact.
Unlike other federal programs which are generally structured around a particular client group,
need, or service such as providing subsidized meals to poor school children or the rehabilitation
of substandard housing, Weed and Seed is a comprehensive, strategic approach using locally
developed plans to solve persistent crime problems and other contributing conditions in targeted
community or neighborhood areas. The belief is that improved coordination and collaboration
across agencies and the community spanning a broad range of responsibilities and resources will
result in a synergistic effect that would be more effective than the existing networks of isolated
decision-making and service delivery.
Consequently, the over 300 local Weed and Seed sites have activities and impacts related to a
wide range of crime and other community concerns such as social, educational, and health needs;
poverty and economic development problems; and housing and environmental quality issues.
Performing accurate data collection and analysis tasks across this extensive universe of different
efforts and outcomes would consume a tremendous amount of evaluation resources. Conversely,
not all local sites are severely affected by every possible crime or community problem or are
directly involved in their solutions, which will minimize their relevance for the study of specific
concerns. A goal of this project was to provide a basis for focusing a national evaluation study
on the issues of greatest interest to the Community Capacity Development Office (CCDO) and
the subsets of local sites most appropriate for inclusion in various research efforts.
Recognizing the diversity of the community problems that Weed and Seed sites are attempting to
address through their strategies as well as the variety of solutions being employed, this project
reviewed the strategy statements submitted to CCDO from 2000 through 2005. The purpose was
to classify the problems and activities that the sites present in the applications for Official
Recognition (OR), the status conferred by CCDO upon Weed and Seed sites that allows a site to
apply for federal funding. These findings are intended to assist a future national evaluation effort
by providing information on which issues and local sites are most likely to have meaningful
impacts and potential performance measures a study of the National Weed and Seed Initiative’s
effects on specific concerns such as drugs, gangs, or distressed living conditions.
In addition to the classification analysis activities discussed above, sites were contacted to
determine data sources that may be commonly available to Weed and Seed sites.
Representatives from 100 active or formerly active sites were interviewed to obtain a broad-
Analyzing and Improving Data Resources
collection. By collecting information on common problems and strategies, as well as building a
database of available data sources, JRSA will provide valuable information for any subsequent
evaluation efforts. As a result of this research, JRSA has developed a series of recommendations
for performance measures related to common strategies for Weed and Seed sites.
Additional elements of this project include an effort to supplement the previous analyses of
Weed and Seed efforts that rely exclusively on OR dates or the first date of federal grant awards
as single arbitrary starting points for Weed and Seed efforts. An intensive search of
administrative records was done to complete a database of local sites’ OR dates as well as first
Weed and Seed grant award dates. A comparison of homicide trends using different key dates
would resolve questions regarding the 1) the amount of time between when sites receive OR
status and their first federal funding and 2) the relative effects of receiving OR status versus
receiving Federal grant funds in terms of crime reduction.
Analyzing and Improving Data Resources
The National Weed and Seed Initiative is currently consists of 331 local sites, which are located
in 46 states, the District of Columbia, Puerto Rico, and the U.S. Virgin Islands. Sponsored by
the U.S. Department of Justice (DOJ) and overseen by CCDO, the Weed and Seed effort is a
cooperative strategy involving United States Attorneys’ Offices, community and social service
organizations, neighborhood residents and businesses, and local law enforcement and other
government agencies. The goal of Weed and Seed is to prevent, control, and reduce crime in
targeted high crime areas through interagency and community coordination and collaboration.
Based on research that indicated effective crime control efforts must reach beyond the scope of
traditional criminal justice agencies and resources to address the underlying factors that make
some communities susceptible to persistent and high levels of crime, Weed and Seed takes a
comprehensive and strategic approach to systemic change in selected neighborhoods. As a
geographically based program, Weed and Seed brings a different perspective to the delivery of
services and allocation of resources by providing a community-wide focus and insuring ongoing
communications across efforts that might otherwise operate in an isolated and uncoordinated
manner.
Site characteristics are widely variable; for example, site boundaries may range from a few
neighborhood blocks to an entire county. Weed and Seed sites focus their goals and objectives on
site-specific problems while simultaneously securing human services and economic resources in
and around the target area. Although their problems and solutions are locally chosen and
prioritized, sites may utilize a wide variety of federal, state, and local resources in their Weeding
and Seeding programs.
One of the key tenets of the Weed and Seed strategy is the recognition that even in the most
crime-plagued neighborhoods, it is likely that a variety of anti-crime efforts are already being
done. In addition to regular policing services being provided, there may be some form of
community policing activity in the area. There are probably a number of prevention efforts
sponsored by various government, community-based and faith-based organizations, as well as
grass roots activities by residents and other interested stakeholders (e.g., local business owners).
It is through organizing and coordinating these ongoing efforts, and securing resources to fill
gaps in existing activities, that Weed and Seed is designed to affect crime and community
livability. In order to enter the federal Weed and Seed program, local sites must show the
capacity and resolve needed to bring about long-term change affecting individuals, groups, and
institutions.
A Weed and Seed area's U.S. Attorney’s Office (USAO), community leaders, local law
enforcement, and other key stakeholders form a steering committee that sets the site’s general
direction and acts as a focal point for planning and coordination of the local site’s goals and
activities. Once the structure and leadership has been established for the proposed site, the
steering committee must develop a strategy based on local resources and needs. The structure of
the Weed and Seed strategy is based on four major components: law enforcement, community
policing, prevention/intervention/treatment, and neighborhood restoration. There are also four
Analyzing and Improving Data Resources
strategy: collaboration, coordination, community participation, and leveraging of resources.
Individual sites strive to weed out crime, seed the community against future crime and crime-
related risk factors, and sustain gains made through the initial Weed and Seed efforts. Local law
enforcement agencies and prosecutors, with support from federal law enforcement resources,
including the USAOs, take the lead in weeding efforts while governmental and community-
based public and private organizations pursue seeding strategies with various human and
neighborhood resources to restore the community’s infrastructure. The community-policing
component provides a bridge between the weeding and seeding aspects of a site strategy. The
long-term goal is to go beyond any immediate reductions in crime by making the community
resistant to future threats that may result in the return of high levels of criminal activity.
A significant challenge for the local sites’ leadership is insuring that their strategies remain
appropriate, as community conditions may change over time and new concerns emerge. These
may be otherwise positive developments such as improved housing and increasing property
values that may result in resident displacement or, negative influences such as the arrival of new
gangs or the sudden proliferation of new street drugs.
A national priority for CCDO is building the capacity of the local sites to sustain Weed and Seed
efforts beyond the expiration of the initial federal funding. By espousing long-term
sustainability for all Weed and Seed sites, the strategy addresses the problems and concerns of
local residents and promotes the financial stability and community resources necessary for
continued growth and development of the designated area. Sustainability depends on a site’s
continued evaluation of neighborhood problems and unmet needs and its ability to maintain the
steering committee as a principal mechanism in securing existing and new resources and funding
sources.
Successful weeding strategies require a joint effort of law enforcement agencies and prosecutors,
as well as the cooperation and leadership of neighborhood residents. Community engagement is
a key component of community policing and is considered an essential element in fostering long-
term community change. Community policing is combined with intensive law enforcement,
including problem-oriented policing strategies, to present a comprehensive crime control
approach in the Weed and Seed sites. These two mechanisms compel law enforcement
personnel to improve their interaction and relationship with the community while allowing
residents to participate in crime prevention and law enforcement activities within their own
neighborhood.
Weed and Seed prevention, intervention, and treatment goals focus on access to needed human
services in the designated area. Through their strategies and activities, sites identify, advocate
for, and, in some instances, provide essential services to at-risk and criminal populations in order
to reduce and control criminogenic influences and improve the overall quality of life. This often
requires collaboration from multilevel social service providers. As part of the neighborhood
restoration strategy, improved human services contribute to the development of the economy,
Analyzing and Improving Data Resources
infrastructure.
To join the federal Weed and Seed effort, an interested community must apply for Official
Recognition (OR) by submitting a strategic plan to CCDO for review and approval in order to be
eligible for Weed and Seed funding. A potential Weed and Seed site must first notify the U.S.
Attorney’s Office in its district of its intent to apply for OR. The U.S. Attorney’s Office may also
assist the potential site with its administrative and strategic development, according to CCDO
guidelines. Once a site receives OR status, it is then eligible to apply for federal funding on a
yearly basis during a 5-year funding window. During this process of preparing their OR
applications, potential sites will first organize their community, begin the collaboration process,
and state the needs and strategies upon which this report is based.
A site’s Weed and Seed OR status lasts for a five-year period. In special circumstances, the OR
status can be extended up to a maximum of ten years. During this time, a site is eligible for
grants from CCDO that require an annual application. The cumulative total received during the
entire OR period cannot exceed $1 million.
Following the expiration of its OR status, a site may continue its participation in the National
Weed and Seed Initiative by applying for Graduated Site status. While it is no longer eligible for
direct CCDO funding, this allows a local site to be eligible for other Weed and Seed activities
such as conferences and technical assistance and for any special considerations available to
Weed and Seed sites when applying for funding assistance from other programs.
Analyzing and Improving Data Resources
Since its inception in the early 1990s, the National Weed and Seed Strategy has considered
evaluation research an important part of its efforts. National and locally initiated studies, yearly
data collection, and academic efforts are some of the varied methods by which the program has
been examined. Despite these efforts, there are inherent barriers to conducting a uniform and
consistent national evaluation across the local sites. These include differences in criminal code
definitions, procedures, and reporting capabilities across jurisdictions; wide ranges in site size,
local organizational structures and responsibilities, and agency participation; site strategies that
include a broad universe of community problems and solutions; modest grant award amounts
that must pay for program operations before local evaluations can be pursued; overlapping
funding sources for activities affecting the sites; limited authority by the sites’ management over
partner agencies; and other logistical impediments.
The need to improve current evaluation components and build upon previous work is fully
recognized by CCDO. For example, the Government Accountability Office (GAO) stated its
concern that Weed and Seed “generally collected activity data rather that measuring outcomes”
and warned against CCDO relying on homicide as an outcome measure (GAO, 2004). Since
1999, homicide statistics have been a primary crime control performance measure used by
CCDO for budgetary reporting on the effectiveness of Weed and Seed. The GAO report was
published before the final release of CCDO's first Crime Pattern Data Analysis, which provided a
multi-crime, multi-year, view of offending patterns in participating Weed and Seed sites.
Another criticism is that the federal Weed and Seed program has not undergone a thorough
evaluation encompassing all local sites. Given the number of sites across the nation and funding
available to CCDO for awards and administration of the program, any type of comprehensive all-
site evaluation would prove to be prohibitively expensive, or it would be so cursory as to be
effectively useless.
The very nature of the National Weed and Seed Initiative poses a number of challenges in terms
of conducting a methodologically rigorous evaluation. Even in the collection and analysis of
simple crime data, there are substantial problems due to the variation in legal definitions and
procedures for jurisdictions across the nation. Other difficulties are the inconsistencies in or lack
of ready availability of data at the neighborhood level that prevent the collection of uniform and
reliable national measures for many concerns. CCDO encourages sites to conduct local
evaluations and provides resources and technical assistance to those doing so. Many locally
initiated studies have been conducted that examined individual sites as well as there have been
several studies of multiple sites within a state. There also have been national efforts that have
attempted to measure effectiveness through the study of multiple sites.
The largest of the national efforts is a cross-site analysis funded by The National Institute of
Justice (NIJ, 1999). This study examined both the implementation and outcomes of the Weed
and Seed Initiative across sites. Case studies were performed for each of 12 target areas in eight
jurisdictions to determine how well the strategy lived up to expectations, what elements were
strongest, what elements failed, and lessons to be learned over the life of the project.
Analyzing and Improving Data Resources
In order to answer these questions, funding applications and other documents were reviewed;
program administrators, senior law enforcement staff, service providers, managers of seeding
activities, and community leaders were interviewed; crime incident and arrest data were provided
by police departments and analyzed; group interviews were conducted with seeding participants;
and residents were surveyed at two different time periods. Although findings varied across sites,
significantly positive results were found for Weed and Seed’s impact on several outcome
measures.
In six of the study sites Part I crimes decreased more than in the host jurisdiction during the same
period. There also appeared to be an association between a decrease in Part I crimes and drug
crime. These same six sites also experienced an initial increase in drug arrests (probably due to
concentrated law enforcement), and then a decline. Findings also examined the relationship
between crime trends and the concentration of resources in sites. Factors that appeared to be
related to success in implementation included the community setting (i.e., strength of social and
institutional infrastructure, economic development possibility, and transience of population) and
program design (i.e., mix of weeding and seeding activities and the sequence of implementation).
Another national effort was the Crime Pattern Data Analysis conducted by the Justice Research
and Statistics Association (JRSA, 2003). For this study, sites were asked to submit data for three
crimes that were considered the greatest problem in their site. Data were requested for a six-year
period, beginning two years prior to, the first year of, and three years following Weed and Seed
project implementation. Crime data for the entire jurisdiction of which the target area was a part
were also reported for the same time period. A pre/post analysis of the change in crime over
time was conducted for each site as well as a comparison of the target area and host jurisdiction
over time.
A total of 98 sites submitted data, with 20 providing complete six-year data sets, 19 with five-
year data sets, and the remainder submitting some data but for less than five years. Sites most
frequently reported drugs and violent crime, and approximately 62% named violent crime in
their top three most problematic crimes. In a analysis of the sites that submitted usable data sets,
a pattern of an initial increase in reported crime during the year of implementation and the first
year following was identified, followed by decreased levels of reported crime in later years of the
effort. These decreased levels of crime were well below the starting values experienced at the
onset of Weed and Seed activities.
Based on research into crime control strategies, including intensive law enforcement and/or
community policing components, it is to be expected that any program employing these
strategies should initially experience increases in crime rates as a consequence of higher levels of
police activity and citizen reporting of incidents. The data then show an apparent decrease in the
study sites in years three and four - a decrease not mirrored by the remainder of the jurisdiction.
Similar trends were found for both sites that submitted partial multi-year data sets as well as the
sites that provided a complete six-year data set.
Analyzing and Improving Data Resources
crime types. Five categories were used
crimes, followed later in the Weed and Seed cycle by dramatically lowered reported crime is
seen throughout. Crimes that are likely to be dealt with in a more urgent manner (e.g., violent
crimes) show decreases earlier than other crimes. This is especially true in the case of drug-
related crimes, where reported offenses are heavily influenced by police activity. Therefore,
because intensive law enforcement is an integral and ongoing component of the Weed and Seed
strategy, it is understandable that drug crimes take much longer before showing a decrease
through police counts of incidents. Overall findings indicate that Weed and Seed sites present a
general pattern of relatively positive change or stability in crime levels as compared to their host
jurisdictions.
Another major source for Weed and Seed performance data is the annual Grantee Site
Characteristics and Activity Data Report, which is part of CCDO’s responsibilities under the
Government Performance and Results Act (GPRA). Each site is required to submit a report for
every year it operates as an Officially Recognized Weed and Seed site or for any year in which
the site spends federal monies. The report includes questions about law enforcement operations,
community policing activities, site-specific as well as surrounding-jurisdiction crime data,
community/social services provided by the sites, neighborhood restoration projects, and funding
other than CCDO money leveraged for use in the site.
The homicide data collected through the GPRA has been used to analyze trends in the sites in
comparison with their host jurisdictions. This analysis, conducted by JRSA (2003), focused on
220 sites in 174 jurisdictions. Over half of the study sites experienced a decrease in homicides
during the study period, 14% stayed the same, 8% increased but at a lower rate than the
jurisdiction, and 23% rose at a faster rate than the jurisdiction. The authors classified 77% of the
Weed and Seed sites as having successful outcomes, i.e., they either had a decline in homicides
or displayed greater stability or control in homicide trends than those found in the host
jurisdiction, exclusive of the Weed and Seed site area. The latter finding was considered
reasonable due to the belief that the Weed and Seed target areas were specifically chosen
because of their persistent crime problems. Consequently, any negative crime trends affecting
the jurisdiction at large should be reflected in the Weed and Seed sites in an equal or more
intense level.
The greatest source of information on Weed and Seed effectiveness however, has been locally
initiated studies of single sites. These studies have generally been overlooked in efforts to assess
the Weed and Seed program but are far from rare and are generally consistent in their findings.
Through 2005, JRSA had collected copies of at least 62 local site evaluations. In 2004, JRSA
conducted a review of locally initiated site studies and examined 34 that met acceptable
methodological standards for consideration as part of the 2004 Weed and Seed Local Evaluation
Meta-Analysis. In examining law enforcement/crime control and community safety effects, the
authors found 21 of 34 sites reporting positive outcomes in one of these categories, with the
other 13 evaluations reporting little or no change or no findings in the component examined.
Analyzing and Improving Data Resources
surrounding area.
The above-mentioned studies do not include all of the Officially Recognized Weed and Seed
sites and there continues to be gaps in assessment on the local and national level. However,
CCDO has taken an aggressive approach in response to criticism that the Weed and Seed
strategy has not been properly assessed, including the commitment of substantial financial
resources. Overall the studies presented here provide very positive results of Weed and Seed’s
impact on crime and stabilizing neighborhood conditions. Additional research work can still be
done, and this study is another step in showing the positive influence of Weed and Seed in some
of the nation's most dangerous neighborhoods.
In reviewing the past national evaluation effort, a major drawback has been that the
individualized aspects of each site’s strategy have not been recognized. Because the
fundamental concept underlying the National Weed and Seed Initiative is the establishment of a
locally developed strategy to address the specific crime and related community concerns for each
local site, an appropriate evaluation strategy should be based on a similar approach.
Consequently, a new national evaluation effort for Weed and Seed should begin by identifying
the problems being addressed by the local sites and which sites are involved with each issue.
This would enable the evaluators to develop measures that are specific to the community
problems being studied for Weed and Seed’s impact and to select local sites for inclusion in the
evaluation that should reasonably have results for the chosen issues.
Analyzing and Improving Data Resources
This project required the collection and analysis of previously unexamined data in CCDO’s
administrative records and the merging of some of this information with JRSA’s GPRA data.
These collection activities included extracting information from the U.S. Department of Justice’s
(DOJ) Grants Management System (GMS) and CCDO’s paper files to determine key dates for
the implementation of the local sites’ strategies and to obtain the OR applications that contain the
strategies for recent Weed and Seed sites. The key dates for receipt of OR status and first grant
awards allow an analysis of the time between these events and changes in homicide crime
patterns associated with these dates. The major task of this project was the development of a
typology for the crime and community problems being addressed by the local sites and the
solutions identified in the local Weed and Seed strategies. Based on this classification scheme,
the existing sites could then be categorized according to the main focus of their strategies.
Data Collection
The analysis of the site OR applications was designed to identify both the site-specific problems
and the strategies employed to address neighborhood problems. The process involved an
intensive review of multi-year OR data and marked the first attempt by CCDO to systematically
analyze multi-year OR application data. The parameters of this project were the examination of
available OR applications from sites that applied for and were granted Official Recognition from
fiscal year 2000 through fiscal year 2005. JRSA identified 309 sites that had been granted
Official Recognition during this time period. Once these sites had been identified, the next step
was to locate and collect their relevant grant documents. In addition to the OR applications,
other pertinent documentation, such as grant status and funding information, was retrieved via
the DOJ GMS files.
fiscal agent and other data. In recent years GMS’s capabilities were expanded to include the
ability to store electronic copies of grant applications and supporting documents. JRSA staff
worked on site at CCDO offices to retrieve and enter grant information, including; Weed and
Seed site name, city and state, OR date, date of first award information, start dates and tracking
information related to the retrieval of these electronic records. While much of this information
was available via GMS, there were some instances of missing data elements. In cases for which
information was unavailable through the GMS system, JRSA staff went through CCDO's paper
files for relevant information. Since GMS lacks the capability to electronically extract data for
analysis purposes, needed information was retrieved record by record within GMS. These data
then had to be manually entered into a project database, slowing the process and increasing the
potential for human error.
The collection of OR application information was affected by changes in electronic storage of
application data. Beginning in 2004, GMS included the ability to store electronic versions of OR
applications. Prior to the 2004 application cycle, CCDO relied almost exclusively upon
hardcopies stored on-site at the CCDO office. To reduce the burden of converting paper copies
to electronic files, sites granted OR prior to 2004 were contacted in hopes of obtaining electronic
Analyzing and Improving Data Resources
of their original applications. All paper copies were electronically scanned, converted to PDF
format and saved for future reference. The PDF format was chosen to ensure the integrity of the
electronic documents. For this reason, all existing electronic copies were also converted into a
PDF format before being saved. Along with this report, JRSA will provide CCDO with
electronic copies of the reviewed OR materials.
This report is based on a review of 273 of the 309 OR applications that were granted Official
Recognition status between fiscal years 2000 and 2005. Thirty-six of the 309 sites could not be
included in this analysis because the OR applications for 26 sites could not be located in
CCDO’s files, while 10 others failed to clearly state the problems and their proposed strategies
and therefore classification of these OR applications was not possible.
Data Analysis
The OR application process allows the applicants to indicate any problems confronting their
communities and any strategies they plan to implement to address their communities' problems.
This latitude in selection coupled with the differences in local communities' challenges
introduces a wide degree of variation in the stated problems and strategies. The sites describe
these issues in a narrative text with optional charts and other supporting data. Given the
unstructured nature of the strategy documents, a process was developed for the consistent
classification of sites by discrete categories of problems and solutions that could then be used for
future performance and evaluation studies.
To effectively analyze the reported problems and strategies, a categorization scheme was
developed that enabled the grouping of these items based on observed commonalities. The
categories used in this scheme were developed utilizing a random sample review process that
included 40 applications. Based on this review, 73 strategy types were identified (See Appendix
A for full list of strategy types). Each of these strategies fall within one of the four broader
Weed and Seed elements
strategy types, the 40 applications were reviewed again and coded to reflect problems and
strategies among the categories. The applications were reviewed by at least two and up to four
different reviewers to ensure consistency in the review process. Adjustments to the categories
were made whenever reviewer interpretations were inconsistent, and a key was developed for the
classification of common activities that could reasonably be assigned to more than one of the
strategy types (See Appendix B).
Due to the local nature of the Weed and Seed program, sites applying for Official Recognition
have a great deal of latitude in selecting the problems that they intend to address. The initial OR
analysis revealed that the 273 applications contained 2,083 problems, which included some
repetitions of issues across sites. Also, some strategies may refer to crime problems in general,
while others identify specific concerns, such as drugs or weapons violations. Attempting to
analyze 2,083 problems without categorization would be extremely difficult; therefore, for the
purposes of this analysis, these problems were grouped into 15 problem area categories.
Analyzing and Improving Data Resources
The decision to place a problem into a specific category was based on commonalities within the
descriptions of the problem. In a substantial number of cases, the applications only provided
general descriptions of their problems. To capture these responses, a series of general categories
were developed. More specific categories were created for instances in which sufficient detail
was provided. For example, if a site listed educational concerns, it was grouped into the
Educational Concerns–General category. If the site reported that they had school dropout
problem, it was grouped into the Educational Concerns-Dropout category. Therefore, the
number of problems reported in the general categories may not equal the combined number of
problems reported in the related subcategories.
The following analysis is based on the review and classification process described above, and is
the first step in the development of a performance measures guide for Weed and Seed sites. It
details the common problems and strategies for reducing crime and disorder cited by Weed and
Seed sites in their OR applications. By classifying these efforts, it should be possible to develop
performance assessments that more accurately reflect the intent of Weed and Seed grantees.
Analyzing and Improving Data Resources
disproportionate level of offending. This is evident from the large number of problems
referenced in the OR applications. In the 273 applications, 2,083 problem citations were
identified through the review process. These were grouped into 16 broad categories. For eight of
the general problem categories, the sites provided sufficiently detailed descriptions and
supporting data that these could be further broken down into subcategories. The following
section details the community problems reported in the applications and the categorization of
these problems.
Table P1
Analyzing and Improving Data Resources
pattern of multiple problem topics across most sites and the potential for multiple subcategories
of a general problem category (e.g., both drugs and gangs being identified as concerns for the
same local site), the problem counts may exceed the actual number of sites working on a specific
issue. However, the problem counts indicate the relative incidence of each concern across the
site strategies.
Perhaps appropriate to a U.S. DOJ-sponsored initiative, crime-related problems are the most
frequently cited concerns. Reflecting the comprehensive approach of the Weed and Seed
strategy to crime control, the majority of the problems being cited are non-justice related topics.
An examination of the distribution of reported problems revealed that the fewest number of
problems reported by any site was two and that the largest was 20. The greatest number of sites,
41, identified seven problems, while the smallest number of sites, one, identified 20 different
problems. The mean number of problems was eight, while the median and modal numbers of
problems were both seven. A distribution of the reported problems by the number of sites can be
seen below in Chart P1.
Description of Community Problem Categories
Blight refers to a host of issues that were described as having a deleterious effect on the physical
environment of the designated Weed and Seed target areas. Some of the issues grouped into this
category included, but were not limited to, graffiti, dilapidated buildings, abandoned cars,
neglected properties and litter. More than half of the applications, 140, referred to some form of
blight as being a problem.
Analyzing and Improving Data Resources
improve conditions in the Weed and Seed area by pressuring property owners to adhere to
community standards and to take a more proactive approach to combating neighborhood
disorder. While most of the identified issues involved violations, some did involve effective
enforcement of existing codes. Thirteen sites indicated that these issues were a problem in their
designated Weed and Seed areas.
The Community Cohesion Problems category includes a variety of problems that negatively
impact the community as a whole, but were not appropriate for assigning to the other categories.
Because of the detailed descriptions provided by the local sites, many of the references to
problems in this category could be characterized by the three subcategories
Problems subcategory involves issues related to traffic safety concerns, such as speeding. The
General category consists of a variety of problems, including, but not limited to, immigration or
minority population issues, transient populations, poor lighting, and community pride.
Combined, the sites reported a total of 157 Community Cohesion Problems. The majority of
these, 100, did not warrant a separate subcategory and were therefore grouped into the General
category. Lack of Community Involvement was the most commonly identified community
problem, followed by Traffic Problems.
Crime Problems were not only most frequently reported, but they were also described with
greater specificity. For this reason, Crime Problems were grouped into nine subcategories.
While most of these subcategories are self-explanatory, some deserve further explanation. The
Weapons Offenses category consists of gun and other weapons-related crimes, many of which
were simply identified as 'weapons' related, without reference to a specific type of weapon.
Domestic violence and child abuse/neglect were each reported frequently enough to warrant
being separate from the general crime, but not frequently enough to distinguish each crime from
one another. As a result, these crimes were collapsed into one subcategory, Domestic/Family
Violence.
The Economic Problems category contains General Economic Issues, Poverty, and
Unemployment/Underemployment. While most sites specifically identified unemployment,
there were a substantial number that identified problems closely related to unemployment. These
problems mainly dealt with the overall employability of residents and included problems such as
lack of requisite job skill. The General category included problems such as lack of investment in
the community, lack of local businesses, and poor tax base. Since most applications did not
detail poverty and how it was measured, the Poverty subcategory included references to poverty,
including, but not limited to, people not earning a livable wage, children living in poverty, large
numbers of residents on public assistance, and lack of financial resources.
Education Problems were grouped into the following five subcategories; Truancy, Education
Level, Dropouts, Literacy, and General. As was the case with other general categories, these
references were assigned to a specific subcategory when sufficient detail was provided; any
nonspecific problem statements were put into the General subcategory. Some examples of
Analyzing and Improving Data Resources
strained educational assistance.
Program/Service Gap Problems were commonly reported, in part because every Weed and
Seed Strategy must identify both existing community programs and unmet service-related needs.
These concerns form the Program/Service Gap Problem category. Whenever possible, these
unmet needs were assigned to subcategories by the intended target population
Family/Life Skills Problems included an array of issues such as lack of parental involvement,
poor parenting skill, marginal life skills, and poor personal financial management. No
subcategory of the Family/Life Skills group was reported frequently enough to warrant it own
category, therefore, all of these problems were included under the main category name.
Fear of Crime are expressions of concerns about public safety concerns that are widespread and
pervasive and, if left unaddressed, will contribute to further community deterioration as residents
and businesses as flee an area. While some applications did contain other terms such as “feeling
unsafe” and “fear of retribution from criminals,” most specifically listed “fear of crime.” Since
no more specific problems related to fear of crime emerged, all of the problems dealing with this
issue were grouped into this one category.
Housing Problems were frequently identified in the sites’ strategies. However, while a large
number of applications indicated Housing Problems as a concern, these were so closely related
that further distinctions were not reasonable. Some examples of entries in this category are lack
of affordable housing, poor quality housing, and lack of home ownership.
Health Problems included a host of problems like community health, lack of affordable
healthcare, lack of healthcare providers and problems with healthcare. Only one health-related
problem, Teen Pregnancy, was mentioned frequently enough to warrant its own category.
The Police – Related Problems category was mostly composed of problems related to the
relationship between the police and the community. However, this category was kept general in
nature to capture other problems such as low visibility of officers, lack of officer training, poor
inter-agency cooperation and insufficient police capacity.
Substance Abuse Problems contains entries dealing with the full range of alcohol and other
drugs
Youth Related Problems reported in the applications could be classified into one of three
subcategories General, Alternatives for Youth and At-Risk Youth. In many cases, the
applications simply listed that there was a need for Alternatives for Youth. When more specific
information was provided, the more commonly mentioned alternatives were lack of after-school
activities and recreational activities. The program shortages discussed in this section differ from
Analyzing and Improving Data Resources
entertainment resources.
At-Risk Youth was a difficult problem to classify. There are varying definitions for at-risk, and
any youth living in a Weed and Seed area could be considered at-risk. The youth could be at-
risk of becoming delinquent, succumbing to possible health threats, or being subjected to a
variety of other risk factors. For this reason, an application was only classified as having a
problem with at-risk youth if it contained that specific phrasing. The General subcategory was
used to capture the other Youth-Related Problems that were either too few in number or too
vague to further classify. This subcategory included problems such as unsupervised youth, youth
harassing other residents, and court-involved youth.
Crime
As previously mentioned, crime-related problems were most frequently identified. The
applications contained a combined total of 833 references to various crime problems, which
represents 40% of the total 2,083 problems cited in the 273 applications. In an effort to better
describe the specific nature of these crimes, they were classified into nine subcategories. The
most frequently identified crime problems were drug crime, violent crime, gang-related crime,
and juvenile crime. Combined, these crimes accounted for over two-thirds of the identified
crime problems.
In terms of local sites, 99% of the sites identified one or more crime-related issues as concerns
being addressed through their strategies (see Table P2). However, for the specific crime
subcategories, only drugs and violent crime were cited by a majority of sites. The other crime
subcategories were mentioned by a third or less of the sites.
In this report the term “drug crime” refers to the illegal distribution, purchase, and/or possession
of narcotics. Not only was drug crime the most frequently identified crime problem, but it was
also the most frequently identified problem overall. Drug crime was reported as a problem in
226 of the applications accounting for over one-quarter of all the crime problems and over 10%
of all problems identified.
The next most common crime problem was violent crime. While most sites listed specific
violent crimes, many simply identified violent crime in general as being a problem in their
targeted area. Violent crime was a listed as a problem in 150 of the 273 applications, accounting
for almost one in five identified crime problems and about 7% of all problems.
Problems classified as gang-related crime were also common. Ninety-eight applications detailed
some form of gang-related crime as being a problem in their designated area. Crimes classified
as juvenile crime were those instances for which the application made specific reference to the
Analyzing and Improving Data Resources
crime problems and less than 5% of all problems.
The general crime subcategory represents all crimes that did not fit into any of the other crime-
type subcategories. The crimes in this subcategory were either too few in number or were
described in such a vague manner that the no further categorization was possible. For these
reasons, these crimes were placed into a single subcategory of miscellaneous concerns.
Economic Problems
Economic problems were the second most frequently identified problem. There were a total of
220 references to economic problems in the 273 applications, accounting for 10.5 % of the total
number of problems. This category was broken down into three subcategories
described too generally to be classified further were included in the general subcategory.
Of all the reported economic problems, unemployment/underemployment was the most common.
Combined, the applications listed this as a problem 101 times, accounting for just over 45% of
economic problems and almost 5% of all problems. Poverty accounted for almost one-third of
economic problems and 3% of all problems. There were also a total of 55 general economic
problems identified, accounting for another quarter of the identified economic problems.
Of all local sites, 153 or 56% have some economic-related issues in their strategies (see Table
P3). Depending on the nature of the economic problems, this may affect roughly a fifth to a third
Analyzing and Improving Data Resources
as contributing factors to a community’s overall distress, may be more successfully addressed as
the area’s persistent crime problems are brought under control and may require longer term
solutions that would be more likely components of subsequent iterations of a site’s strategy.
Education Problems
Problems related to education were the next most commonly identified problem type in the 273
applications. A total of 114 site strategies indicated that they had an education-related problem
that impacted the residents of their designated area (see table P4). Truancy accounted for the
largest number of education problems, identified in 57 or 21% of all sites and accounting for
almost 40% of the education problems indicated. An additional 40 general problems were
described in the applications accounting for over one-quarter of the education problems. The
remaining education problems related to Education Level, Dropouts, and Literacy accounted for
another third of these problems, though Literacy was far less often mentioned than the preceding
subcategories.
Analyzing and Improving Data Resources
Blight
Blight was the fourth most common problem and, according to the “Broken Windows Theory”
of crime control, can be a significant factor contributing to existing community crime conditions.
As was explained above, blight is used to describe a wide variety of problems that have a
deleterious effect on the physical environment of the designated communities. While many
applications contained references specifically to blight, others listed the problems individually.
Due to the diversity of blight-related problems reported, it was not possible to further classify
these problems. Therefore, these problems were combined into one larger Blight category. Of
the 273 strategies, 139 sites, or 51%, included blight as a problem affecting their communities.
This represents 6.7% of total number of identified problems.
Service/Program Gaps
The next most frequently identified problem type after blight concerned the provision, delivery,
or availability of services and programs. There were 132 references to these problems for 102,
or 37 % of all sites, accounting for over 5% of the total number of problems (see Table P5). The
greatest number of these applications, 83, contained references to this problem that did not
identify a specific target population. These general references accounted for three of five service
or program gap problems and about 4% of all identified problems. Thirty-three applications
outlined youth-related service or program gaps, while the remaining three populations mentioned
in the applications were adults, elderly and returning offenders. These populations accounted for
13% of the service/program gap problems as a group.
Housing Problems
Housing Problems contains a variety of housing-related conditions in the Weed and Seed areas.
Some of the problems contained in the applications were lack of affordable housing, sub-
standard housing, and lack of home ownership. There were not sufficient numbers of any
Analyzing and Improving Data Resources
contained 116 references to housing problems making it the sixth most frequently mentioned
problem type representing just over 5% of the total number of recognized problems.
Community Cohesion Problems
The Community Cohesion Problems category was the next most frequently identified problem
type after housing problems. Eighty-six strategies, or 32% of all sites, mentioned negative
conditions in this category. In all, community problems accounted for 105 problems, or 5% of
the total number of identified problems. This category includes three subcategories
applications, though, outlined problems that were described in such a vague manner or so unique
to their community that they were not classifiable beyond general community problems. The
single most common identifiable problem listed in the applications was a lack of community
involvement. Thirty-four applications, or 32% of those listing Community Problems, indicated
that community involvement was an issue that was negatively affecting their designated area. Of
the applications that identified a community problem, 23 identified some type of traffic concern.
See Table P6 for a breakdown of the community cohesion problems.
Remaining Community Problems
The remaining nine categories accounted for a combined total of 391 problems or almost one-
fifth of the total problems. These categories were broken down as Youth-Related, Substance
Abuse, Other, Health, Police-Related, Fear of Crime, Family/Life Skills, Offender-Related, and
Code Violations/Enforcement. There were a total of 85 references to youth-related problems, of
which 48 identified a lack of alternatives for youth, 23 identified general youth-related problems,
and 14 identified at-risk youth.
There were 70 instances in which Substance Abuse was identified as a problem accounting for
barley 3% of all problems. The problems that could not be classified or that did not warrant their
own category were grouped together into the Other Problems category. A similar number of a
Health-related issues were found, half of which were identified as related to teen pregnancy. The
Analyzing and Improving Data Resources
numbering about 2.5% of the total number of problems each.
While most Police-related problems dealt with the relationship between the police and the
community, the Police-related problems category also included other problems, such as
equipment and staffing issues. Police-related problems were reported in 47 instances making up
just over 2% of the total problems. Fear of Crime was identified as a problem in 34 instances
while there were 26 instances of Family or Life Skills as a problem.
One of the least often mentioned problems include Offender-related problems accounting for
only 14 of the 2,083 problems. Most of these identified returning offenders as the main issue,
while the remaining Offender-related problems identified probationers as the issue. The final
problem category is Code Violations/Enforcement. The applications contained only thirteen
instances where code violations or enforcement was cited as a problem. See Table P7 below for
a detailed breakout of these less numerous problem categories and subcategories.
Analyzing and Improving Data Resources
In order to group the OR applications by commonalities among the strategic solutions being
pursued, a coding scheme was developed. This coding scheme consisted of the four overall
strategy elements and a subset of 69 activity types. The four strategy elements reflect the core
elements of the Weed and Seed Strategy, which are Law Enforcement (LE), Community
Policing (CP), Prevention/ Intervention/ Treatment (PIT), and Neighborhood Restoration (NR).
The 69 activity types each belong to the most appropriate strategy element and can be viewed in
Appendix A.
The review of the OR applications revealed that most sites were implementing an array of
activities to address the problems affecting their targeted neighborhoods. Combined, the 273
applications contained 7,884 activities. The smallest number of activities reported in any
application was four, while the largest was 57. The mean number of activities was 28.88, the
median was 29, and the mode was 28. See Chart S1 below.
Law Enforcement Activities (LE)
The Law Enforcement strategy element encompasses a variety of enforcement-oriented activities
that are designed to address the crime problems in the designated areas. These activities do not
include police activities that are intended to bolster relationships between the police and the
communities that they serve. Upon examination of the Law Enforcement (LE) activities, 15
activity types were developed.
Analyzing and Improving Data Resources
for almost one-quarter of all of the identified activities. All but one strategy contained at least
one LE activity. See Table LE1.
most common LE activity type employed by the sites. In all, 226 of the 273 (83%) of the
applications indicated that intensive drug enforcement was part of their overall strategy.
Whenever possible, the focus of these efforts was also determined. Of the 226 applications that
detailed intensive drug enforcement strategies, 91 indicated that these efforts were focused on
reducing the supply of drugs, while 53 indicated that these efforts were focused on reducing the
demand for drugs. This represents almost one-third and one-fifth, respectively, of all
applications. No focus could be determined in the remaining 66 cases.
The next most frequently identified LE activities fell under the prosecution/sentencing
coordination category. These activities involved inter-agency cooperation designed to target
criminal offenders in the designated Weed and Seed neighborhoods. The analysis showed that
184 of the 273 (67%) applications identified this type of activity as part of their overall LE
strategy.
Analyzing and Improving Data Resources
targeted crime enforcement. A total of 153 applications (56%) identified the use of increased
patrol/visibility, 148 applications (54%) identified the use of task forces, and 142 applications
(52%) identified some form of targeted crime enforcement
Community Policing activities are designed to foster the relationships between these groups in
communities
CP2.
identified activities was 12. As was the case with LE activities, only one site failed to identify a
single CP activity. See Table CP1, below.
Activities 1,491
Minimum 0
Maximum
building or strengthening police-community relations. Of the 273 applications, 211 (77%)
indicated that this was a focus of their strategy. Another commonly identified activity was the
use of crime watch /resident patrol programs. A total of 195 applications (73%) specified that
these types of programs were part of their overall strategy.
The next two most commonly identified community policing activities were community
awareness and education. While these activities seem very similar, for this analysis the
following distinction was made. Community awareness refers to activities such as officer-
initiated referrals or efforts to make community residents aware of the department’s activities or
the availability of community or social services. Community education refers to activities such
as classes or the distribution of information pertaining to topics such as crime prevention or
individual/human rights.
A nearly equal number of applications identified the use of each of these activities. A total of 179
applications (66%) reported that increasing community awareness was part of their overall
strategy, while 178 (65%) reported activities designed to educate the community. The next most
commonly identified CP activity type were those activities focused on youth. Youth-focused
activities were identified in 147 of the 273 applications (54%). A complete listing of the 13 CP
activities can be seen below in Table CP2.
Analyzing and Improving Data Resources
Prevention/Intervention/Treatment (PIT) includes a wide variety of social programs and services
that are designed to address various problems or needs in the communities. Governmental
agencies or non-profit community groups commonly provide these services. From the OR
applications, a total of 28 different PIT activities were identified, which are listed in Table PIT2.
Analyzing and Improving Data Resources
and Treatment Activities
Maximum
Combined, the 273 sites listed a total of 3,077 PIT activities. The maximum number of
identified PIT activities was 25 and the minimum reported number was two. Both the mean and
median number of identified activities was 11 (see Table PIT1 and Chart PIT1).
An examination of the reported PIT information revealed that the most commonly identified PIT
activities were those directly related to Safe Havens. A total of 228 applications (84%) indicated
that Safe Havens were a major focus of their PIT activities. The second most common activity
types were those related to academic/enrichment programs. For the purpose of this analysis,
these programs were defined as programs or activities involving arts, culture, homework
assistance, and/or tutoring. Of the 273 applications, 202 (74%) indicated that they had some
form of academic/enrichment program.
Life skills/family programs were also clearly a priority for the sites. This category was defined
as programs or activities involving counseling, budgeting, leadership training and parenting
training. Combined, 195 applications (73%) indicated that their effort included these types of
programs. A similar number of applications (194) indicated that they had or planned to
implement programs or activities intended to provide residents with employability/job training.
Other common PIT program or activity types were Community/Social Service Access,
Mentoring, and Drug and Alcohol Education/Prevention. Community/Social Service Access
Analyzing and Improving Data Resources
things as centralized service locations, community service centers, and transportation to and from
service providers. A review of the applications revealed that 178 of the 273 sites (65%)
referenced these types of activities. Mentoring and Drug and Alcohol Education Prevention
Strategies appeared with a similar frequency, being noted 168 and 166 times, respectively. A
complete listing of the 28 PIT activities can be seen below in Table PIT2. Additionally, an
explanation of the various activities and or programs included in each activity type can be found
in Appendix B.
Analyzing and Improving Data Resources
Neighborhood Restoration (NR) refers to an assortment of programs and activities that share the
overall goals of stimulating and revitalizing both the economy and physical appearance of the
neighborhoods. In all, 13 neighborhood restoration categories were developed. These 13
activity types are listed below.
activities for a site was 12 and three sites did not indicate the use of any NR activities (see Table
NR1).
Activities 1,425
Maximum Reported Number
spaces and included activities such as neighborhood, park and vacant lot cleanup projects. In all,
226 strategies (83%) contained descriptions of various community space restoration activities.
The second most commonly identified type of NR was code enforcement. Based on the
information contained in the applications, it was clear that code enforcement was a tool utilized
to address a wide array of issues ranging from problem establishments to absentee property
owners. A total of 189 applications (69%) indicated that code enforcement was a characteristic
of their NR activities and their overall strategy.
The next most frequently identified NR activity was housing improvement/restoration. These
activities generally included assisting residents with making necessary repairs to their homes or
properties. The elderly were a major focus of many of these initiatives. In all, 173 applications
(63%) indicated these types of activities as being part of their strategy. Other commonly
indicated strategies include Community Involvement Initiatives, Home Ownership, and Public
Infrastructure Improvement activities. Of the 273 applications, 167 (61%) indicated that
Community Involvement Initiatives were a focus of their strategy. An equal number of
applications (127) indicated that Home Ownership programs and Public Infrastructure
Improvement were aspects of their overall strategy
Analyzing and Improving Data Resources
This analysis has demonstrated that while there are many commonalities across sites in the
problems and solutions identified in their strategies, individual sites may have substantial
variations in the number and specific nature of the issues being addressed and the services and
activities being employed. This has major implications for future performance measurement and
evaluation efforts for framing specific research questions and site selection.
Because useful performance measures should reflect those changes that result from Weed and
Seed efforts, it is these strategies that should serve as the basis for developing indicators of local
sites’ successes. Considering that this analysis found a substantial number of site strategies,
which do not clearly present a linkage between the identified community problems and the
solutions proposed, planning for future measures should recognize this difficulty. As expected,
the comprehensive nature of the Weed and Seed philosophy, combined with the large number of
applicants, results in a great variety of strategies and potential sources for data and measures of
efficacy. Later in this report, some performance measures are recommended based in large part
on the strategies and problems laid out in this section. Following are suggestions for future
problem and strategy definition by applicants seeking to achieve OR status and the possibility of
federal funding for their Weed and Seed effort.
Analyzing and Improving Data Resources
As with the wide range of problems and strategies found among Weed and Seed sites, there is a
wide range of strategies used by program evaluators in conducting reviews of Weed and Seed
efforts on the local level. Many locally initiated site studies have been, and continue to be,
conducted. These studies have been funded in part by Special Emphasis grants from CCDO,
other supplemental grants, and local sources.
With the local origin of these efforts, the methodology used is often a product of individual site
strategies, availability of preferred data, and the strategies of individual evaluators. The strength
of this variability is that local evaluators can examine aspects that are most germane to the
specific strategies in place in their community. A weakness is that the evaluations are less
comparable than if there had been consistent standards followed for the types of performance
measures to be used. For example, several site evaluations are based on the effect of Weed and
Seed efforts on reported crime in the Weed and Seed area, while others base conclusions of
efficacy on changes in the number of risk factors to which youth are subject in target areas. Both
types of reports allow for judgments on whether an effect was seen in Weed and Seed areas, but
comparisons of relative success (or failure) become problematic.
In order to facilitate cross-site comparability, this report includes suggested performance
measures for Weed and Seed sites. These suggested performance measures should not be
viewed as a limit to variables that can be examined by Weed and Seed sites sponsoring or
conducting an assessment of their efforts, but a guide of variables to be examined in addition to
other data that will then allow for a more unified look across site evaluations. See below for
specific recommended performance measures for selected common strategies from the review of
OR materials.
Enforcement
and possession in the intensive
drug enforcement area
this table, it is critical that there be
consistency in the reporting period.
Crime data are currently reported to
CCDO most often on an annual basis,
with a given calendar year compared to
prior and later years. Snapshots may
also be used (for example, comparing
January 2005 to January 2006), but
limiting the pool of data increases the
probability that unrelated factors (such
as a warm week in the month of January)
may influence your measure. At a
minimum, measures should be taken
annually.
Analyzing and Improving Data Resources
Coordination
participants (e.g. police,
prosecutors) to determine
perceptions of coordination
questions such as "How would you rate
cooperation between police and local
prosecutors for the sentencing of repeat
offenders?" with an ordinal scale for
reply.
Visibility
-and-
Foot/Bike/Horse Patrol
determine level of awareness of
increased police presence in the
community
questions such as "Are the police around
the neighborhood more than before
(baseline date)?" with an ordinal scale
for reply.
area covered by task force
the task force was created to address.
Presumably, for most that will be drug
crimes, but a violent crime task force
would, of course, be measured by
changes in violent crime rates.
Enforcement
crimes
examined should be those for which the
strategy is developed. Investigators may
want to also look at related crimes,
where appropriate.
Enforcement
juveniles
Strategies
revocations/actions for Weed and
Seed area residents
Changes in the rate of Weed and
Seed area residents on probation
rates, rather than raw numbers, as the
number of Weed and Seed residents on
parole or probation may vary greatly
from year to year.
Enforcement
-and-
Improved Public
Perceptions of Safety
gauge resident perceptions of
disorder in the Weed and Seed area
questions such as "How do you rate your
neighborhood as a place that is safe and
orderly?" with an ordinal scale for reply.
Strategies
-and-
Gang Prevention
related crimes
classification, such as the definition of a
gang-related crime (e.g., is it any crime
committed by a known gang member or
is it any crime with a suspected tie to
gang activities?).
If crime data for gang-related crimes are
not available, a second recommended
measure would be a question in a regular
and reliably conducted youth survey
asking respondents about their
involvement with gangs.
Analyzing and Improving Data Resources
Strategies
weapons crimes
manner in which local law enforcement
agencies track crimes involving a
weapon. Some agencies may track
weapon involvement for every incident.
In some cases, it may be necessary to
rely upon reports of crime that
specifically deal with the use/misuse of
weapons (such as carrying a concealed
weapon, or improper discharge of a
firearm).
Relationship Building
measuring the satisfaction with
police-community relations
questions such as "How do you rate your
satisfaction with the performance of your
local police department?" with an ordinal
scale for reply.
Patrol
crime watch/resident patrols and
number of patrols by groups
part measure that can reflect both the
growth (or lack thereof) in patrols and
their ability to recognize and report
criminal activity.
participants
will be hesitant, at best, to provide grade
information for program participants.
The most effective method of getting
grade information is usually by having
participants' parents agree to provide
information from report cards.
Local schools should be able to provide
cumulative grade information as a point
of comparison.
Training
-and-
Job Placement Strategies
within the Weed and Seed area or
the number of program participants
employed at given points or for
given periods of time after
program completion.
number of residents served and whether
the intervention is general or specific.
These programs are normally assessed
with process measures (such as
graduation rates or participation
numbers).
program-involved youth
appropriate point of comparison and will
need to find a group that matches as
closely as possible the group entering the
mentoring program. For example, if the
mentoring program limits itself to at-risk
youth, the comparison group should be
similarly at risk youth.
Also of interest will be the degree of
involvement in the program. Does a
greater degree of participation forecast
probabilities of success?
Analyzing and Improving Data Resources
Programs
program-involved youth
the overall rate of arrest among youths in
the Weed and Seed area.
admissions and changes in use of
preventative care facilities
degrees of emphasis on preventative
care. Youth-focused efforts may also
look to obtain school absenteeism data
for participants or for the entire school as
appropriate for program focus.
and service requests received by
providers listed in the resource
directory
listed in the directory, limiting the
potential for a comparison group of
providers not listed, it would be possible
to request the same information from
providers outside the Weed and Seed
area not included in the contact
directory. These providers could then
serve as a comparison group.
absenteeism rates for participating
youth
youth for whom absenteeism rates
should be examined will depend on the
scope of the truancy programs. Any
community-wide general programs
should examine truancy for an entire
affected school or schools. Programs
that focus on specific populations could
be confined to participating youth.
In the first case, comparisons with trends
in schools not in the target area would be
appropriate. In the latter case,
comparison groups based on factors
related to admission into the program
would be appropriate.
Analyzing and Improving Data Resources
specified time period
a number of decisions to be made before
data collection. Among these are the
time period to be considered. Given the
limited timeframe of the federal Weed
and Seed funding cycle, a period of two
years following reentry would be
reasonable (assuming this allows for
completion of the program and a
reasonable post-treatment period).
Other issues are whether to consider
arrest or conviction as a marker of
recidivism (arrest may be a high
standard, especially if police look for 'the
usual suspects,' conviction may require a
longer time frame for comparison to
allow for resolution of cases). Also, a
decision about the seriousness of contact
with criminal justice authorities should
be made; is any arrest or conviction a
marker for recidivism? Or only if a
felony is involved? This will depend in
part on the pool of those being studied.
community residents that are not
high school graduates or GED
recipients
population for whom GED/school
completion should be examined will
depend on the scope of the program.
Any general programs should examine
rates for an entire affected area while
those focusing on specific populations
could be confined to participants.
As with job training programs, resource
intensive GED/school completion
programs may want to examine cost
versus benefit
Restoration
-and-
Building Demolition
reported crimes at specific
locations
remove magnets for offending, changes
under this program should make specific
places less prone to crime. Investigators
with access to GIS analytical systems
may also want to examine changes in the
number of crimes in a surrounding buffer
zone as well.
Analyzing and Improving Data Resources
Rehabilitation
-and-
New Home Development
owner occupied dwellings within
the Weed and Seed target area
homes in the area, these strategies
encourage long-term residency that will
increase the stake of residents in the
long-term health of the community.
Improved housing stock that is used for
rentals is less likely to have the
ameliorative effect on crime sought
under such programs.
owner-occupied housing units in
Weed and Seed area.
geographically based, the emphasis will
be on seeing a change within the Weed
and Seed site. While other programs
may show success in program
participants buying homes in any part of
a jurisdiction, Weed and Seed-related
efforts should examine changes within
the Weed and Seed area boundaries.
Development
businesses in the target area
may be able to collect this and other
information (such as the number of
empty commercial buildings) by
completing a physical survey of the
Weed and Seed area; actually traveling
the streets and noting visible
facilities/lots, etc. An alternate source of
data is the number of businesses paying
taxes within the area or changes in
business tax revenue.
financed through loan programs
funded, there are other measures that
may shed light on the efficacy of such
programs. Included among these are the
rates of loans paid back and the number
of jobs created by businesses funded
through the program.
Several of the recommended performance measures involve community or resident surveys.
Such surveys pose challenges to Weed and Seed sites that are generally pursuing evaluative
efforts under tight financial constraints. The quality of data collection, however, is critical in any
survey. Some reminders for data collection that will yield useful and reliable data are:
meetings, for example. If there is a substantial non-English speaking population, the
surveys should be translated into the relevant languages or dialects, reviewed for cultural
appropriateness for the intended subjects, and the surveyors should be prepared to
Analyzing and Improving Data Resources
may be improved by also obtaining the support of key leaders in the community. If you
only conduct surveys in the day, you are likely to miss a lot of residents who are at work;
likewise, if you do it at night, you will miss residents who work at night.
surveys tend to undercount the young (who use cell phones instead) and the poor (who do
not have phones at as high a rate) and the mobile (for whom phone records are more
likely to be out of date) and so may be ill suited for Weed and Seed areas. In-person
interviews may avoid this but will be much more expensive to conduct and may have
other limitations for accuracy. Sometimes a combination of methods provides a balance
between reliability and cost concerns. Surveyors should be trained so they are consistent
in the delivery of the instrument and do not 'hint' at specific answers (a survey of
opinions on law enforcement, for example, will be much different if collected by
uniformed law enforcement officers.
desirable.
are implemented and a widespread awareness of these efforts among residents is
achieved, a survey should only be adopted as a performance measure when there is
sufficient time for this information to diffuse throughout the area. Residents’ perceptions
may be longstanding opinions that will require substantial and sustained counterevidence
before significant change is reflected in survey results.
considered safer when crime is reduced or only when crime is totally eliminated?
Though surveys can be time consuming and expensive, there are few sources that can provide as
useful and accurate information for some questions as a properly conducted survey.
Other sources and types of data have their own limitations on how they should be used or
viewed. In analyzing crime data, for instance, one should remember that the data generally
reflect incidents reported to police, and that the majority of all crimes are never reported. Certain
tactics (e.g., community policing, intensive law enforcement) have been found to change the rate
of crime reporting, so even if crime is going down, in the short term the number of reported
crimes may rise. Given differences in local laws, the same event may be reported as different
offenses (or not be an offense in extreme cases) in different jurisdictions. Try to understand the
limitations of your data and how they can affect your analysis.
Analyzing and Improving Data Resources
Administrative Data Set
As part of this project’s tasks, CCDO and JRSA conducted a detailed review of the
administrative records for local sites including Official Recognition (OR) notifications, grant
awards and progress reports with the goal of obtaining detailed and accurate information
regarding the key dates for the implementation of local Weed and Seed sites’ activities. Because
these dates are not part of the GPRA information processed by JRSA for CCDO, it has been
sporadically obtained at various times in the past. Consequently, this has meant that for various
reasons these dates were missing for some sites, which prevented their inclusion in analyses
using site implementation dates.
With more complete and accurate Weed and Seed site start date information, it would be possible
to determine the time periods between when sites receive OR status, obtain their first Weed and
Seed grant awards, and commence activities. The Weed and Seed homicide and crime trend
reports could be substantially improved because it would expand the number of site records
available for any analysis efforts, which would result in more definitive pre/post periods for
comparison of crime and programmatic statistics. While limited pre/post Weed and Seed site
analyses has been conducted in the past, it was hoped that hundreds rather than a score of sites
could be more rigorously studied.
To date, the most extensive Weed and Seed impact assessment is a draft 2005 GPRA Weed and
Seed homicide data study that has information regarding 435 individual Weed and Seed sites.
The draft 2005 GPRA Weed and Seed homicide study is an update of a report entitled, A
Comparison of Homicide Trends in Local Weed and Seed Sites Relative to their Host
Jurisdictions, which was published in 2003. The homicide data file that was developed for the
draft 2005 GPRA Weed and Seed homicide study was used for these analyses because it includes
information for about 100 additional sites and has more recent homicide data than the published
2003 report.
time series comparisons of homicides within Weed and Seed sites and their host jurisdictions.
There are two main reasons for the drop-off in the number of sites available for analysis. First,
many of the sites were so new (110) that they were either in their initial years of operation or
they have only a single year (sometimes only a part of a year) of operations. Second, some sites
(51) had seriously flawed GPRA homicide reports that could have been due to a wide range of
reasons that rendered them invalid for inclusion in this study.
The methodology for the draft 2005 GPRA Weed and Seed homicide data study was simply an
examination of the following questions:
Analyzing and Improving Data Resources
That analysis found the number of homicides had decreased in 56% of the Weed and Seed sites,
remained stable in 19%, and increased at a lower rate than the host jurisdiction in 5% of the sites.
In only 21% of the Weed and Seed sites had homicides increased.
The draft 2005 GPRA Weed and Seed homicide data study examined the trend in homicides
without knowing precisely when the Weed and Seed operations actually began; therefore it is
impossible to conduct a pre/post comparison examining the timing associated with Weed and
Seed implementation. The following assessment combines the homicide time series data from
the 274 sites included in the draft 2005 GPRA Weed and Seed homicide data study with the
expanded Weed and Seed start date information found in the CCDO administrative records. It
was anticipated that there would be a strong likelihood of matches between these two data sets
and that it would provide a more effective test of the hypothesis that following the
implementation of a Weed and Seed strategy, homicides should decrease.
In 2004, JRSA did perform a four-year homicide data comparison using the OR award and first
grant award date information that was available to it at that time. In this first effort to capture
Weed and Seed “start dates”, significant data collection issues were encountered. Consequently,
the 2004 pre/post implementation homicide crime analysis was based on data for only 54 Weed
and Seed sites. For these sites, a 30% reduction in homicides occurred three years after
implementation of their Weed and Seed strategies. One of this project’s objectives was to further
expand the number of useable site records for studying Weed and Seed’s effects on community
crime levels.
Local Site Strategy Implementation and Other Key Event Dates
Key strategy implementation dates collected for the 2005 expanded CCDO administrative
records data set include the Official Recognition date, date of the first Weed and Seed grant
award, and the site “start date”.
review before a jurisdiction can be invited to submit a Weed and Seed grant proposal.
The first award date is the date CCDO gives to a site’s first approved grant proposal.
The “start date” is the day that a Weed and Seed site reports actually beginning Weed
and Seed activity.
Of these three initial Weed and Seed strategy implementation dates, the OR date and the actual
activity start date reported by the local sites were considered the two most important dates for
this analysis.
Analyzing and Improving Data Resources
information regarding the Weed and Seed sites was collected. For instance, whether or not the
grant application was initially accepted or denied is recorded. Often when the initial grant is
denied, the site will have a delayed start. Whether or not a denied site eventually does begin
operations can be verified by confirming if the site has ever provided first and subsequent years’
progress reports. A site with a denied first grant and no subsequent progress reports was judged
as not ever starting. The lack of GPRA reporting for these sites provides additional verification
that these sites never really began operations.
The availability of the expanded CCDO administrative records for 321 sites raised the possibility
that there could be significant overlap with the 274 sites that have valid homicide data sets and
were used in the draft 2005 GPRA Weed and Seed homicide data study. This would allow a
more authoritative analysis using the various strategy implementation dates. In the end, it was
possible to link the expanded CCDO administrative records with 100 sites in the draft 2005
GPRA Weed and Seed homicide data study, which was nearly a doubling of the number of sites
for the previous pre/post analysis. The details regarding the record linking process and reasons
that information for many local sites could not be linked can be found later in this report in
Appendix C - Summary of CCDO Administrative Findings.
Official Recognition Versus The Site Reported “Start” Date: Which is the Real Beginning for
Strategy Implementation?
As previously stated, for a community to receive the federal Weed and Seed designation or OR
status, it has to pass an extensive pre-screening process. This requires the submission and
approval of an OR plan that is reviewed by CCDO. The review process has multiple criteria,
which require the interested jurisdictions to describe their plans and organizational structure for
implementing a Weed and Seed strategy that complies with the national model. A failing score
on any of the major criteria can result in not being invited to submit an application for funding.
The process of creating an OR plan requires substantial interagency cooperation and
organizational development which lead some to believe that achieving this by itself should be
viewed as the date that the Weed and Seed strategy actually begins. Others argue that while the
OR process is an important initial stage in creating a Weed and Seed site, most sites’ strategies
are not meaningfully implemented until federal funding and CCDO support are guaranteed and
the site staff are hired and working.
At first glance, there appears to be some validity to the question as to whether the OR date or the
“start date” for Weed and Seed sites is the most appropriate measure indicating the beginning of
Weed and Seed. For instance, there is an average period of 270 days between the OR date and
the actual program start date. However, of the 100 Weed and Seed sites where the CCDO
administrative records and the 2005 GPRA Weed and Seed homicide study data sets overlap, the
records for 40 (while having an OR date) do not have a start date.
Moreover, in 52 of the 60 other sites, the OR and start dates actually occur within the same
calendar year. Since homicide statistics are reported annually, this means that in the vast
Analyzing and Improving Data Resources
substantially different for the purpose of measuring the effects of strategy implementation.
Of the eight cases, in which the OR and start date are not in the same calendar year, the OR
versus start date analysis becomes moot for three, because when the OR date is used as the
initiation point for a Weed and Seed site, it coincides with the first year of homicide reporting
⎯thus there is no baseline history⎯and therefore no “pre” time period for the pre/post analysis.
In the remaining five cases, in which the OR precedes the start date by at least one calendar year,
the comparative results are only substantially different in one case.
Finally, while there is generally about a nine-month lag between the OR and the Weed and Seed
start dates, this amount of time is not long enough to create a significant difference for the annual
Weed and Seed homicide statistics. Therefore, in the remainder of this report, where the Weed
and Seed start date is available, it is used as the pre/post demarcation. Where the actual start date
is missing from the CCDO administrative records, which includes 40% of cases in the final
analysis, the OR date is used as a surrogate start date with the assumption that if the actual start
date was known, it would not be significantly different from the OR date.
Pre/Post Strategy Implementation Analysis Using 2005 Weed and Seed Homicide Data Set
In an effort to maximize the number of Weed and Seed sites in the pre/post analysis, both the
official recognition and start dates were considered as possible “start dates.” This method of
measurement is used because, as shown above, the analytical difference between the two is
minimal.
Using this type of measurement for Weed and Seed start dates, there were 100 sites with a match
between the expanded CCDO administrative date data set and the draft 2005 GPRA homicide
data set. This means that the pre/post analysis includes only 37% (100 out of 271) of the GPRA
homicide database’s sites. The majority of these matched sites (60 percent) have both a start date
and an OR date. For the remaining 40 percent of the sites only the OR dates were used. This
degree of matching with CCDO’s administrative records, while less than ideal, still results in an
improved data set for “exploratory” pre/post analysis.
Further attrition in the number of matched Weed and Seed sites occurred because some of the
CCDO start dates are not the original site start dates – but appear to be start dates for an interim
award or grant revision that took place after the actual site start date. When the full draft 2005
GPRA homicide database was compared to the CCDO “known” start or official recognition date,
it was clear that GPRA forms were far more complete and included more historical reporting
than the CCDO start date database. Sites where CCDO start dates seem to be an interim date
instead of the actual start date were removed from the analysis. Increasing the number of “valid”
sites in for future pre/post analyses would involve further searches of CCDO’s files to identify
the actual site start dates and may require reviewing archived records for some of the pre-1999
local sites.
Analyzing and Improving Data Resources
strategy implementation. This same start year was also used for homicides reported in the
remainder of the host jurisdiction.
Because of the variability of homicide statistics for a single year, the average number of
homicides was calculated for the pre/post periods for each Weed and Seed site and each host
jurisdiction. Because of the high percentage of mismatches between the CCDO administrative
records and the GPRA homicide data, most of the jurisdictions with multiple active Weed and
Seed Sites were reduced to only one site where a start date was known, thus minimizing the
validity of the “jurisdictional” analysis used in the draft 2005 GPRA homicide report.
Furthermore, it is not uncommon for jurisdictions with multiple Weed and Seed sites to have
different start dates for each of the sites. Therefore, in this analysis the multiple sites are treated
separately, each with its own unique start date without consideration of the impact of other active
Weed and Seed sites within the host jurisdiction.
The first analysis involved a comparison of the average number of homicides per site prior to the
start date compared to the average number of homicides following the start date.
Findings: Analysis of Pre/Post Weed and Seed Strategy Implementation Changes in
Homicides
For the 100 sites where the draft 2005 GPRA homicide database could be matched to the
expanded CCDO administrative records “start date” database, the average number of homicides
prior to Weed and Seed was 4.93. Following the implementation of Weed and Seed in these
sites, the average number of homicides decreased to 3.57 – an average decrease of 1.4 homicides
per site – representing a 28% decrease in homicides.
This may appear to be very small until the change in homicides in the remainder of the host
jurisdictions is considered. In the rest of the host jurisdictions, which have much larger
populations and volume of homicides than the Weed and Seed sites, there was also a decrease in
the number of homicides after the implementation of Weed and Seed. The average decrease in
the remainder of the host jurisdictions was, however, only .5 homicides. The decrease in
homicides within Weed and Seed sites is three times that which occurred in the remainder of the
host jurisdictions. In other words, the pre/post analysis shows that there was a significant
decrease in homicides following the implementation of Weed and Seed in the sites (even when
combining the very successful and the unsuccessful sites). During this same time period, there
was very little change in the host jurisdictions’ homicide rates.
Comparing the Pre/Post Strategy Implementation Analysis Results to the Draft 2005 Weed
and Seed Homicide Data Analysis
The pre/post analysis for the 100 sites, for which the draft 2005 GPRA homicide database and
expanded CCDO administrative records database can be matched, provides a more valid analysis
of Operation Weed and Seed. While this method provides promising results, it is still limited by
Analyzing and Improving Data Resources
100 sites also does not consider the sites’ performance relative to comparable trends for their
host jurisdictions. These limitations raise the question as to whether the simpler analysis used
for the GPRA homicides studies, which cover 274 sites, might still be of value.
To help address this question, a more detailed review of the similarities and differences of the
pre/post strategy implementation analysis and the draft 2005 GPRA homicide data study was
undertaken. The bridge between the two analysis methods is the Weed and Seed Homicide
Relative Change Scale. The results from both the pre/post strategy implementation analysis and
the draft 2005 Weed and Seed homicide data analysis can be translated into scores using the
Homicide Relative Change Scale. This analytic approach and the differences in the finding are
discussed below.
The Homicide Relative Change Scale
Because the sites’ target areas were selected due to their persistent and severe crime problems, it
is reasonable to assess their performance as a function of change relative to the surrounding
jurisdictions. This alternative approach adopts the viewpoint that, given this history, these
communities would be expected to have crime trends equal or worse than their jurisdiction
without the implementation of a Weed and Seed strategy. This was the method used for the draft
2005 GPRA Weed and Seed homicide report. The pre/post average number of homicides for
sites and the remainder of the host jurisdiction can easily be translated to the Homicide Relative
Change Scale used in the draft 2005 GPRA homicide report. It is just a matter of switching from
a linear measure (i.e., employing an arithmetic average) to a ranked order score as was used for
the draft 2005 GPRA homicide data study.
The Weed and Seed Homicide Relative Change Scale assesses homicide trends for local sites in
relation to similar changes occurring in their host jurisdiction. The Relative Change Scale ranks
all of the possible combinations of changes in both the sites and the jurisdictions according to the
degree of success for a site. The most successful category or ranking are those local sites that
have experienced substantial decreases in homicides while their host jurisdictions have had
increasing homicide trends. The least successful category or ranking is when homicides
increased in the Weed and Seed site but decreased in the remainder of the host jurisdiction. The
different combinations of homicides decreasing, increasing or remaining stable results in a 13-
step scale from the most successful to the least and is represented numerically from a +7 to a -5.
For example, in the draft 2005 GPRA Weed and Seed homicide data study, if the number of
homicides in a Weed and Seed site data series decreased by five homicides and the remainder of
the jurisdiction only declined by three homicides during the same time period, the site would be
represented as a site where its decrease in homicides was greater than the decrease in homicides
in the remainder of the host jurisdiction – and would be scored as a + 6.
Translating the pre/post average number of homicides for the same site and host jurisdiction was
done in the following manner. The average number of homicides in the Weed and Seed site
preceding strategy implementation was 12 and for the post-implementation period, the average
Analyzing and Improving Data Resources
Weed and Seed. At the same time, the number of homicides in the remainder of the jurisdiction
decreased from an average of 75 prior to Weed and Seed implementation to 70 for the time
period afterwards. This means that homicides only decreased by 5 in the remainder of the host
jurisdiction and so this would be classified as a site where homicides decreased more in the
Weed and Seed site than in the remainder of the jurisdiction – and would also be scored as a +6.
Weed and Seed Homicide Relative Change Scale
6 Site Decreases > Jurisdiction Decreases
5 Site Decreases = Jurisdiction Decreases
4 Site Decreases < Jurisdiction Decreases
3 Site Decreases & Jurisdiction Stable
Findings: Comparison Between the Pre/Post Analysis Results and the Draft 2005 Weed and
Seed Homicide Analysis Using the Relative Change Scale
How do the average relative change scores for the pre/post analysis data set compare to the draft
2005 GPRA homicide study results? Applying the Homicide Relative Change Scale to the 100
sites used in the pre/post strategy implementation analysis, the average score was +2.0. This was
just a little bit less than the result of +2.15 for the draft 2005 homicide data analysis.
While the difference between the two techniques does not change the direction or the magnitude
of success, it is very important to note that the more accurate results from the pre/post strategy
implementation data set differ significantly from the Relative Change cores from the 2005
GPRA homicide report in 28 cases (28%). Interestingly, 14 of the 100 sites switch from a
positive to a negative Relative Change Scale score – when compared to the original GPRA
homicide study. On the other hand, the number of CCDO sites that changed from a negative to
positive score also number 14. Therefore, it is to some degree serendipity that the pre/post
average relative change scores are so similar to the draft 2005 homicide study results. However,
the most important thing is that despite the methodological differences and issues, the Weed and
Seed homicide studies both indicate strong positive outcomes for the Weed and Seed strategy.
The following summary of findings indicates how the pre/post strategy implementation data set
differs from the larger draft 2005 homicide analysis data set. As noted above, while the results
for the two data sets are very similar, the distribution across the possible range of scores varies.
The results from the pre/post strategy implementation data set tend to push sites to the extremes,
Analyzing and Improving Data Resources
strategy implementation data set, there are more very successful sites (23% versus 16%) and
very unsuccessful site (24% versus 13%: using -4 and -5).
2005 Homicide Data Study Pre/Post Analysis
The pre/post strategy implementation data set provides a more precise measure for assessing the
success of the Weed and Seed strategy across sites. This approach’s main shortcoming is that
the results are limited to about only a third of the sites for which valid homicide data are
available because of incomplete site start date records. While the site/host jurisdiction homicide
trend comparison results may not be as precise as the pre/post strategy implementation analysis,
they provide useful information for almost three times as many sites. Until more complete
administrative information is available for site start dates, a combination of both methods to
track and assess the performance of local sites can provide a more comprehensive picture of the
National Weed and Seed Initiative.
Analyzing and Improving Data Resources
___ Supply
___ Demand
___ Firearm Crime Reduction Strategies
___ Gang Crime Reduction Strategies
___ Hotspot Policing
___ Increased Police Patrol/Visibility
___ Juvenile Crime Enforcement
___ Known/Repeat Offender Strategies
___ Public Order Crime Enforcement
___ Task Forces
___ Target Crime Enforcement________________
(Crime type)
___ Traffic Violation Enforcement
___ Use of Technology
___ Other (Describe): ______________________
___ Community Awareness
___ Community Education (rights, LE efforts, etc.)
___ Crime Watch/Resident Patrol Programs
___ Foot Patrol
___ Hotlines/Anonymous Reporting
___ Improving Public Perceptions of Safety
___ Senior Citizen-Focused Efforts
___ Substations/Storefronts/Mobile Command Centers
___ Training for Law Enforcement Personnel
___ Youth-Focused Efforts
___ Other (Describe): _________________________
___ Boys & Girls Club
___ Conflict Resolution
___ Community/Social Service Access
___ Community Service/Volunteer Programs
___ Domestic Violence Prevention/Intervention
___ Drug/Alcohol Education/Prevention
___ Drug/Alcohol Treatment
___ Employability/Job Training
___ ESL Programs
___ Gang Prevention Programs
___ GED/School Completion Programs
___ Health-Related Programs
___ Housing Assistance
___ Internship Programs
___ Juvenile Justice Programs
___ Life Skills/Family Programs
___ Literacy Programs
___ Mentoring Programs
___ Recreational Programs
___ Re-entry Initiatives
___ Resource Directory
___ Safe Havens
___ Teen Pregnancy Programs
___ Truancy Programs
___ Victim Services
___ Other (Describe): __________________________
___ Code Enforcement
___ Community Involvement Initiatives
___ Community Space Restoration/Improvement
___ Entrepreneurial Training
___ Home Ownership Programs
___ Housing Improvement/Rehab
___ Loan Programs
___ New Home Development
___ Public Infrastructure Restoration/Improvement
___ Small Business Development/Expansion
___ Other (Describe): __________________________
Analyzing and Improving Data Resources
Loitering
Noise
Nuisance abatement/crime enforcement
Trespassing
Human/individual rights
What the department is doing in the community
Homework help
Tutoring
Community service centers
Transportation to/from services
Domestic violence prevention
Sex education
At-risk youth programs
Youth council/court
Youth offender programs
Counseling
Leadership
Parenting
Clean-ups
Creating community gardens
Murals
Cleaning up vacant lots/public places
Improving public transportation system
Improving/creating sidewalks & roads
Lighting
Analyzing and Improving Data Resources
This CCDO/JRSA effort resulted in a more complete documentation of 321 separate Weed and
Seed sites. One site was duplicated (site 435 Athens GA) which reduced the count from 320. In
contrast, the GPRA homicide database has 274 sites that have sufficiently complete and accurate
historical data that are used for a “crime change” analysis. These data are used to assess the
change in homicide patterns in Weed and Seed sites, which are then compared to the change in
homicides for the sites’ host jurisdictions. In addition, the GPRA homicide database has partial
information for an additional 160 sites for a total of 434 sites. Of these, 51 sites have extensive
historical site data, but that information is unreliable with inexplicable differences across the
annual GPRA reports. The other 109 Weed and Seed sites are so new that they have little actual
Weed and Seed activity (none to one year) with which to justify a change analysis of their
homicide data.
It was hoped that there would be a significant overlap between the 320 sites with information
regarding implementation start dates for Weed and Seed sites and the 274 sites with complete,
multi-year homicide data. The first review of the CCDO start date database showed that:
317 out of 320 sites (99%) had an Official Recognition date, and
221 out of 320 sites (69%) had a first grant award and site activity start dates.
3 out of 320 (1%) did not have an Official Recognition date. These were Site ID No. 471 -
Minneapolis MN, Site ID No. 057 - Des Moines IA, and Site ID No. 473 - St. Paul MN.
First Award Dates and Start Dates
The relationship between first grant award dates and the site activity start dates is complex. On
average, the reported site activity start date for Weed and Seed sites occurs 88 days before the
first award date. But this average is somewhat misleading because some sites activity “starts”
before the first award date and some start after.
Of the 221 sites with both first award and start dates before the first award, 167 (76%) start
before the first award day, on average 125 days before the first award. This means that about
three-fourths of the sites actually begin their local Weed and Seed activity prior to, in
anticipation of the receipt of their first grant award.
Of the 221 sites with both first award and start dates, 54 (24%) start after the first award day, the
average delay between the first award date and the site start date is 30 days.
Official Recognition versus the Start Dates
There are 219 sites out of 320 (69%) with Official Recognition dates and start dates and first
award dates.
Analyzing and Improving Data Resources
For these 219 sites, the Official Recognition dates precede the start dates by an average of 270
days.
Also for these 219 sites, the Official Recognition dates precede the first award dates by an
average of 359 days.
Actually as described above, the relationships between Official Recognition dates and either the
start date and the first award date can go to another level of definition where the Official
Recognition dates are compared to those sites where the first award dates precede the site start
date and visa versa.
Chronology of Sites with Official Recognition (OR) But Missing Start Dates
There are 98 sites with Official Recognition dates but are missing both the first grant award and
the site activity start dates. It would be natural to expect that the missing dates would be heavily
weighted to the older sites. While there is a significant number of missing start dates for older
sites, the table below shows that the missing information is by no means limited to these old sites
1999 17
2000 25
2001 10
2002 8
2003 7
2004 9
2005 21
M
Out of the 320 CCDO sites, 284 (89%) can be matched to the 2005 homicide analysis database
which has 434 sites going back to 1996. However, the 2005 homicide database has data quality
limitations that result in only 274 sites being analyzed as to their level of success at reducing
homicides within the Weed and Seed sites. These 274 sites have GPRA homicide reports that
have reliable reporting over the years.
There also 51 sites in the 2005 homicide database, while having multi-year information, are not
included in the homicide analysis because are large differences across the GPRAs submitted for
these sites that cannot be easily reconciled.
There are another 63 sites in the GPRA 2005 homicide database that have 3 years of history and
either a single year or a partial year of Weed and Seed site activity and are not yet included in the
homicide change analysis.
Analyzing and Improving Data Resources
Finally, there 43 new sites that have only historical homicide data being reported as part of the
2005 homicide database and cannot be used as part of the homicide change analysis.
The following table indicates how well the GPRA homicide database and the CCDO official
recognition and site start data match for each of the GPRA homicide database groups.
Category of GPRA
2005 Homicide Data GPRA Sites CCDO OR Sites CCDO Start Date Sites
As discussed earlier in the CCDO start date database, there are many more records with only
Official Recognition dates than those with both Official Recognition and start dates. This
situation limits the ability to match these records to the GPRA homicide database. While 54
percent of the sites with OR dates match to the GPRA homicide database’s records with
complete and valid data; only 35 percent with both dates overlap. While there may be some
expectation that older sites may be more difficult to match, it is probably even more
disconcerting that the overlap between the newest Weed and Seed sites (i.e., those initiated
within the past two years) and the GPRA homicide database is lower than expected. These new
sites are still missing 20 to 24 percent of the start dates. This may be due to the carryover of
funding for these sites to the next fiscal year.
Analyzing and Improving Data Resources
Dunworth, T., Mills, G., Cordner, G., & Greene, J. (1999). National Evaluation of Weed and
Seed Cross-Site Analysis [NCJ 176358]. Washington, DC: National Institute of Justice.
Retrieved February 22, 2006 from
General Accounting Office. (2004). Grants Management: Despite Efforts to Improve Weed and
Seed Program Management, Challenges Remain (GAO 04-245). Washington, DC: Author.
Retrieved February 22, 2006 from
O’Connell, J., Perkins, M. & Zepp, J. (2004). Weed and Seed Crime Pattern Data Analysis.
Washington, DC: Justice Research and Statistics Association. Retrieved February 22, 2006 from
O’Connell, J., Perkins, M. & Zepp, J. (2003). A Comparison of Homicide Trends in Local Weed
and Seed Sites Relative to Their Host Jurisdictions, 1996 to 2001. Washington, DC: Justice
Research and Statistics Association. Retrieved February 22, 2006 from
Perkins, M. & Zepp, J. (2004). Weed and Seed local Evaluations Meta Analysis. Washington,
DC: Justice Research and Statistics Association. Retrieved February 22, 2006 from