Using Data That Matters

1. What do we mean by “performance measurement” vs. “program evaluation”?

Performance measurement is the routine tracking that a program does to understand whether it is reaching the right people, delivering services as intended, and making progress toward its goals. It answers questions like: How many people are we serving? How often are staff meeting with participants? Are referrals being completed? Are participants staying engaged?

Program evaluation is a more formal effort to understand how well a program is working and, in some cases, whether the program itself caused changes in outcomes. An evaluation might examine how a program was implemented, how participants experienced it, or whether violence was lower than it would have been without the program.

A simple way to think about the difference is: performance measurement helps you manage and improve the program as it operates; evaluation helps you study the program more systematically. Both are important, and strong performance measurement usually makes later evaluation much easier.

2. How are outputs different from outcomes?

Outputs are usually counts of activities or services. Examples include the number of participants enrolled, outreach contacts made, case-management meetings completed, conflicts mediated, or referrals provided. They are important because they show whether the program is operating as intended, but they do not tell you whether participants or communities are actually better off.

Outcomes capture the changes the program hopes to produce. These may include increased employment, improved school engagement, greater housing stability, reduced involvement in violent conflicts, fewer violent injuries or arrests, or reductions in shootings in the community.

A simple way to remember the difference is: outputs are what you deliver; outcomes are what you hope changes as a result. Strong performance measurement usually tracks both.

3. What is the difference between process measures and outcome measures?

Process measures tell you whether the program is being delivered the way it was designed. In a logic model, these measures usually come from the program’s activities and outputs. Examples include whether staff are reaching the intended participants, how often they meet with them, what services are provided under different program components, and whether referrals or other planned activities are completed.

Outcome measures come from the short-, intermediate-, and long-term outcomes in your logic model/theory of change and tell you whether the changes the program expects are actually occurring.

For more detailed definitions of terms related to evaluation and evidence, see Key Terms for Understanding CVI Evidence on ARCvi’s Evidence pages.

4. What’s the difference between evidence-informed and evidence-based?

An evidence-based program or practice has been formally studied and has research showing that it can produce the outcomes it is designed to achieve.

An evidence-informed program or practice uses the best available research to guide its design and decisions, even if the exact program has not itself been rigorously evaluated. It may combine research findings with practitioner experience, community knowledge, participant needs, and local conditions.

Basically, evidence-based means there is research showing that a particular approach works; evidence-informed means research is being used to shape how the work is designed and delivered. For more details, see our Research Brief: What Does the Research Say About Best Practices in Community-Based CVI?

5. How should program leaders assess community needs before designing or refining a CVI program?

A community needs assessment helps programs understand where violence is concentrated, who is most affected, what services and resources already exist, where gaps remain, and what residents identify as priorities. We like to use the Office of Juvenile Justice and Delinquency Prevention’s (OJJDP) Comprehensive Anti-Gang Model community assessment process. It includes a structured community assessment to help organizations and partnerships understand the nature and scope of violence and street gang involvement in their community before deciding how to respond.

If your program is interested in understanding local health and wellness needs, Community Health Needs Assessments (CHNAs) can be a useful source of information. Federal law requires nonprofit hospital facilities to conduct a CHNA at least once every three years and develop an implementation strategy addressing the significant health needs identified. These assessments must incorporate input from people representing the broader community and are made publicly available.

Depending on the hospital and community, CHNAs may include information on violence, injury, community safety, and related needs. CVI organizations can review CHNAs from hospitals serving their neighborhoods to supplement their own community assessments and identify health needs, service gaps, and potential healthcare partners. For example, see Temple University Hospital’s summary report from their CHNA for 2025.

6. How can a CVI coalition or partnership identify strengths, gaps, and priorities across participating organizations?

A needs assessment can help partners identify what is already in place, where important gaps remain, and which areas require additional attention or investment. ARCvi’s CVI Needs Assessment Tool can be used or modified by coalitions and partnerships to systematically review their collective capacity and identify priorities for strengthening the overall CVI strategy.

7. Is there a national effort to develop more consistent CVI definitions, data, and performance measures?

Yes. The national Community Violence Intervention Action Plan, developed through a multi-year process involving CVI leaders from across the country, identifies greater standardization as an important next step for strengthening the field. Its 2024 report, Mapping Transformation for the Field, identifies six priorities for the field, two of which are related to data and definitions: (a) “research, evaluation, quality improvement, and impact analysis” and (b) “standardization of essential CVI elements: definitions, data-driven performance management, and direct service coordination.”

The Action Plan calls for greater consistency in the terminology the field uses and stronger standards for program design, implementation, evaluation, and performance management. Greater consistency can help individual programs use data to improve their work while also making it easier to compare experiences across programs, identify promising practices, and build a stronger evidence base for CVI nationally. The Action Plan is now being translated into more concrete field-building efforts.

8. How can a logic model help us decide what data to collect and what performance measures to track?

A logic model should serve as the blueprint for the performance measurement system. Each major part of the model should point to something you may want to track. For more detailed information, see our Quick Guide: Using a Logic Model to Guide Performance Measurement.

Another great resource is Harry Hatry’s Urban Institute report Transforming Performance Measurement for the 21st Century. This report helps readers identify the performance indicators that matter most to your program and your constituents, and then determine appropriate data sources and collection procedures to track them reliably.

9. Can we just pick 5 or 6 core measures that fit our theory of change and logic model — and drop the rest?

Core measures are the ones that best show whether you are reaching the right people, delivering the core parts of your model, and making progress toward the outcomes in your theory of change. If your program is new, you may want to start with core measures, but your program should eventually capture a comprehensive set of performance measures that reflect your program components and activities, and the range of outcomes represented in your logic model. For more details, see our Quick Guide: Using a Logic Model to Guide Performance Measurement.

10. What does a simple performance measurement “scorecard” look like for CVI outreach and engagement?

Your program and logic model will be very context specific, so there is no one “scorecard” that works for every program. See our Quick Guide: An Example Data Scorecard for CVI Outreach and Engagement for example measures related to outreach and engagement.

11. What does a simple “scorecard” look like to capture our CVI case management component?

Your case management component and logic model will be very context specific, so there is no one “scorecard” that works for every program. See our Quick Guide: An Example Data Scorecard for CVI Case Management for example measures related to case management.

12. How often should staff collect data?

There is no single schedule that works for every measure. A useful rule is to collect data as close as possible to when the activity happens, but review different types of information on different timelines. Staff should usually record service and activity data right away, or at least within a day if possible. This includes participant contacts, outreach encounters, mediations, referrals, case-management meetings, missed appointments, and safety concerns. Waiting a few days increases the chance that details will be forgotten or recorded inconsistently.

13. We have the basics of data and performance measurement in place. How can we get better at actually using our data?

CVI organizations may also want to follow or contact the Black & Brown Collective for Community Solutions to Gun Violence, a national group of Black and Brown researchers focused on producing rigorous, culturally responsive, and community-engaged research on community violence. The Collective emphasizes working with practitioners and communities so that research and data are useful for improving policy and practice.

We also think the Urban Institute’s Measure4Change Playbook is a great resource for becoming more effective at gathering, managing, and strategically using data. You can pick and choose the strategies that fit your organization now, or identify others for a future wish list as your capacity grows.

14. What’s a low-cost path to a workable performance management database if we’re starting from scratch?

A program does not need an expensive case management system to begin collecting useful performance data. Start by identifying a small set of core measures based on the program’s logic model and the information needed to produce them. For a small program, a well-designed spreadsheet or shared data-entry form may be enough to track participants, services, referrals, and outcomes. Independent Variable Consulting, a firm founded by juvenile justice system and violence reduction expert Adam Serlin, helps organizations build rapid prototypes using off-the-shelf software as a fast, affordable way to establish basic data infrastructure. As needs become more complex, programs can transition to a more customized database such as Independent Variable’s PERFORM Case Management system.

15. When is monitoring enough, and when do we need an external evaluation?

Routine monitoring using your performance measures is often enough when the goal is to understand whether a program is reaching the intended population, delivering services as planned, and making progress on key outcomes. An external evaluation may be useful when you need a more independent assessment of implementation or want to determine whether the program itself caused changes in outcomes. Not every program is ready for an impact evaluation, and a process or implementation evaluation may be the right next step. See our Evidence Resource Collection for more on the different types of evaluation and how to decide what fits your program.

16. Does our program need to develop our own control/comparison group to show that the program is reducing violence?

Usually, no. Programs generally do not need to create their own control or comparison group as part of routine performance measurement. If the goal is to determine whether the program itself caused a reduction in violence, however, some form of comparison is typically needed — and that is usually part of a formal impact evaluation designed with an evaluator or research partner. See our Evidence Resource Collection for more on comparison groups and impact evaluation.

17. What is the most reliable and simplest way to gather updated data on gun violence?

The simplest place to start is with an existing gun violence or homicide dashboard. Many cities provide dashboards with frequently updated information on shootings, shooting victims, homicides, and trends over time and by geography.

Many of these dashboards are real-time or near-real-time, meaning they are updated as incidents are recorded rather than only through annual crime reports. In Philadelphia, several public dashboards provide updated shooting and homicide data; many allow downloading the raw data. See ARCvi’s Philadelphia Resources section for a list of area dashboards. Philadelphia organizations may also find Independent Variable Consulting’s PERFORM Reports especially useful. PERFORM turns public data into easy-to-use charts and reports, including neighborhood- and ZIP-code-level trends, without requiring users to analyze the raw data themselves.

For organizations that want to put local trends in a broader context, we also recommend the AH Datalytics Public Data Portal. This free tool compiles frequently updated crime data from hundreds of law enforcement agencies across the country, allowing users to examine trends in their own city, compare them with other jurisdictions and national patterns, and explore crime measures beyond firearm violence. Users can also create customized graphs and comparisons.