Section 1. A Framework for Program Evaluation: A Gateway to Tools

This section was originally adapted from the 1999 CDC report "Framework for Program Evaluation in Public Health," developed by Bobby Milstein, Scott Wetterhall, and the CDC Evaluation Working Group. It retains the original framework's organization while updating the discussion and identifying subsequent developments in evaluation practice.

Around the world, communities develop programs and initiatives to improve the conditions that shape people's lives. These efforts may prevent violence, expand access to safe and affordable housing, support student learning, or address other priorities identified by community members.

How do we know whether these programs are working, for whom, and under what conditions? How can we improve their effectiveness, accessibility, and fairness? And how can an organization decide which promising approaches are appropriate for its community?

Program evaluation provides a systematic way to explore these questions. By combining evidence with community knowledge, organizations can understand implementation, assess results, identify unintended consequences, and use what they learn to improve their work.

In 1997, the United States Centers for Disease Control and Prevention (CDC) convened an Evaluation Working Group to organize the essential elements of program evaluation. Its framework was published in 1999 and became the basis for this section. CDC updated the framework in 2024. The discussion below preserves the original six-step teaching structure and explains how it relates to the newer guidance.

Before beginning, it helps to define several terms used throughout this section.

By evaluation, we mean a systematic process for assessing the merit, value, or significance of an effort. Evaluation brings together questions, evidence, interpretation, and judgment to support learning and decisions. Its methods should fit the purpose, community context, and consequences of those decisions.

Throughout this section, the term program describes the effort being evaluated. It may include activities intended to improve outcomes across a community, within settings such as schools or workplaces, or among particular populations, such as young people, people affected by violence, or people living with HIV. This definition is intentionally broad.

Examples of different types of programs include:

• Direct service interventions, such as a program offering free school breakfasts to support children's access to nutritious food
• Community organizing efforts, such as a campaign led by farmworkers to improve wages and working conditions
• Research initiatives, such as a community partnership examining approaches to reducing racial inequities in access to health services
• Monitoring and surveillance systems, such as a system that tracks housing conditions to inform inspection and repair priorities
• Advocacy efforts, such as a campaign seeking policies that reduce exposure to secondhand smoke
• Social marketing campaigns, such as a locally developed campaign helping families find available nutrition and infant-feeding support
• Infrastructure and capacity-building projects, such as an initiative strengthening agencies' ability to support community-led development
• Training programs, such as a workforce development partnership that combines job training with support for transportation and other employment barriers
• Administrative systems, such as a revised appointment process intended to reduce waiting times and improve access to services

Program evaluation, the focus of this section, supports learning and accountability in community health and development work. It can examine a specific project, a set of related activities, or a broader initiative. The scope should be explicit so that the evaluation's questions and conclusions match the effort being assessed.

Community members and other interested or affected parties include people who participate in a program, people affected by its decisions, those who implement or fund it, and those who may support or question its approach. Earlier evaluation literature often uses the term stakeholders; newer CDC guidance uses interest holders. Useful questions include: Who is affected? Whose knowledge is needed? Who makes decisions? Whose perspectives have been excluded?

This section presents a practical framework for developing a shared understanding of evaluation. Its goal is to help people involved in community health and development assess their efforts, learn together, and use findings responsibly.

Why evaluate community health and development programs?

Evaluation can be closely connected to everyday program operations. Our emphasis is on practical, ongoing evaluation that involves program staff, community members, and other partners, with specialized expertise when needed. This approach can support both program improvement and accountability to the people a program is intended to serve.

For example, evaluation complements program management by:

• Clarifying program plans, assumptions, and intended results
• Improving communication and shared understanding among community members and partners
• Gathering evidence and feedback to improve effectiveness, address unequal outcomes, and account for the use of resources

People working to improve their communities already evaluate their work informally when they ask questions, consult partners, reflect on feedback, and adjust their approach. For low-stakes decisions, these practices may provide enough information. When decisions involve substantial resources, sensitive information, or significant consequences for people, a more systematic and documented evaluation can make the basis for those decisions clearer and more defensible.

How do you evaluate a specific program?

Before beginning an evaluation, work with the people involved to answer these questions:

• What will be evaluated, and how will the findings be used?
• What criteria will be used to assess program performance, and who will help choose them?
• What levels of performance would indicate success, including fair access and outcomes?
• What evidence is needed to assess performance against those criteria?
• What conclusions are justified by the evidence, and what uncertainties remain?

Consider Drive Smart, a hypothetical program designed to reduce alcohol-impaired driving. The following targets illustrate how an evaluation might be planned; they are not recommended benchmarks for every community. A real program would establish its targets using baseline information, available resources, community priorities, and a realistic timeline.

• What will be evaluated?
  • Drive Smart's public education and "Safe Rides" transportation service, including their reach, accessibility, implementation, and relationship to alcohol-impaired driving.
• What criteria will be used to assess program performance?
  • The percentage of community residents familiar with the program and its goals
  • The number of people using "Safe Rides," including whether the service reaches residents facing transportation barriers
  • The percentage of respondents reporting alcohol-impaired driving during a consistently defined period
  • The number of recorded crashes involving alcohol impairment, interpreted alongside changes in traffic volume, reporting, and other relevant conditions
• What levels of performance would indicate success?
  • 80% of community residents will know about the program and its goals after the first year.
  • The number of people using "Safe Rides" will increase by 20% relative to a defined baseline during the first year.
  • The percentage reporting alcohol-impaired driving will decrease by 20% relative to baseline during the first year. For example, a decline from 10% to 8% is a 20% relative decrease, or a decrease of 2 percentage points.
  • Recorded crashes involving alcohol impairment will decrease by 10% relative to a comparable baseline period during the first two years, with attention to uncertainty when counts are small.
• What evidence will indicate performance against these targets?
  • A community survey using a defensible sampling approach and accessible participation options will assess awareness and reported behavior at baseline and follow-up. The evaluation will examine nonresponse and the limitations of self-reported information.
  • "Safe Rides" records will document service use, distinguishing trips from unique users where this can be done without collecting unnecessary identifying information.
  • Crash records will provide information on documented alcohol involvement, with consistent definitions and attention to missing information or changes in reporting practices.
• What conclusions are justified by the available evidence?
  • What changed, for whom, and how plausibly did Drive Smart contribute? Could other policies, services, or community conditions explain the findings?
  • If intended changes have not occurred, should the program adjust its approach, address implementation barriers, allow more time, or gather better evidence?

The following framework provides an organized approach to answering these questions.

A framework for program evaluation

Program evaluation helps people understand and improve community health and development efforts. The original CDC framework presented here organizes evaluation around connected steps and standards for quality. It is a practical guide that can be adapted to a program's purpose, resources, and setting.

The original framework contains two related dimensions:

• Steps in evaluation practice
• Standards for evaluation quality

The illustration shows the original 1999 framework. Its six connected steps provide a useful teaching sequence, but evaluation is iterative: questions about evidence or use may require revisiting earlier decisions. Planning for how findings will be used should begin at the outset, and engagement should continue throughout the process.

CDC's 2024 revision uses six steps: assess context; describe the program; focus the evaluation questions and design; gather credible evidence; generate and support conclusions; and act on findings. Its accompanying guidance emphasizes collaboration, fair and just evaluation practices, and ongoing learning across every step. The original sequence below remains the organizing structure for this section, with those considerations incorporated throughout.

• Engage community members and other interested or affected parties
• Describe the program
• Focus the evaluation questions and design
• Gather credible evidence
• Develop and justify conclusions
• Use findings and share lessons learned

Use these steps as a connected learning process, adapting decisions as the program and its context change.

The original framework also drew on 30 evaluation standards organized into four categories. Those historical categories are retained below to explain the original model:

• Utility
• Feasibility
• Propriety
• Accuracy

These categories help assess whether an evaluation is useful, practical, ethical, and technically sound. CDC's 2024 framework uses five standards: relevance and utility; rigor; independence and objectivity; transparency; and ethics. When planning an evaluation, identify the applicable guidance and document how its standards will inform your decisions.

Engage Community Members and Other Interested or Affected Parties

People and organizations may be affected by both an evaluation's findings and the decisions made from them. Evaluation therefore requires attention to relationships, power, and differing perspectives. Community members, program participants, staff, partners, funders, and decision-makers may each contribute knowledge that others lack. Meaningful engagement gives people opportunities to shape questions, interpret findings, and influence decisions rather than simply provide data.

Participation can strengthen relevance, trust, and shared responsibility, but it does not guarantee agreement. People should be able to question methods, identify harms, or disagree with conclusions without pressure to endorse the program or the evaluation.

Although engagement appears first in the original framework, it belongs in every step. Begin by understanding the community context, identifying whose participation is needed, and agreeing on how their contributions will affect the work.

Three overlapping groups are important to involve:

• People or organizations involved in program operations may include community leaders, volunteers, collaborators, coalition partners, funders, administrators, managers, and staff. Their knowledge can clarify how the program actually operates and what resources it needs.
• People or organizations served or affected by the program may include participants, families, neighborhood organizations, advocacy groups, institutions, public officials, and other residents. Include people who face barriers to participation and those who question the program. Listening to different perspectives and engaging people who may oppose the program can identify blind spots and strengthen the evaluation.

Also consider people who could experience unintended consequences from decisions based on the evaluation. For example, changes to service locations, eligibility, hours, or funding may affect people who are not currently participating. Their perspectives can help identify consequences that program records alone will miss.

• Primary intended users of the evaluation are the specific people expected to act on its findings. They may include community representatives, participants, program staff, organizational leaders, and funders. They are not necessarily the same people as the program's intended participants. Identify these users early, clarify the decisions they face, and maintain communication without allowing their preferences to override evidence or the interests of affected communities.

The form and extent of community and partner involvement will vary. People may help govern the evaluation, develop questions, collect information, interpret results, or decide how findings are shared. Offer accessible participation options, language support, reasonable compensation where appropriate, and clear explanations of the time and responsibilities involved.

Agree on how decisions will be made, how power will be shared, and how the group will resolve conflicts. Document different views and create ways for people to raise concerns safely. These practices help prevent the priorities of the most influential participants from automatically determining the evaluation's direction.

Describe the Program

A program description explains the effort being evaluated: what it intends to accomplish, how its activities are expected to contribute, who is involved, and what resources and conditions shape its work. It should distinguish the program's planned approach from what is actually being implemented.

The description establishes the frame of reference for evaluation. For example, a program focused on enforcing alcohol sales restrictions has different activities and immediate outcomes from a program providing safe transportation for young people. A clear description helps the group choose appropriate questions and examine how particular activities may contribute to results.

People may also have different understandings of the program's purpose. In a youth development initiative, some may emphasize educational support, others access to employment, and others young people's leadership and decision-making. Discuss these expectations openly and include young people's own priorities when defining the program.

A shared description makes evaluation more useful, even when some differences remain. Document areas of agreement, unresolved questions, and assumptions about how change will happen. Developing this description can improve program planning before outcome data are available.

Include the following aspects when describing a program.

Statement of need and opportunity

A statement of need and opportunity describes the issue, goal, or possibility the program addresses, together with community strengths and existing efforts. Explain the nature of the issue, who is affected, its scale, and how it is changing. Identify relevant social, economic, and institutional conditions rather than treating a community's challenges as individual shortcomings.

Expectations

Expectations describe the program's intended results and what would count as meaningful progress. Organize them into short-term, intermediate, and longer-term outcomes, explaining how one is expected to contribute to another. A program's vision, mission, goals, and objectives express these expectations at different levels of detail. Check whether expectations are realistic and shared by the people affected.

Activities

Activities are the actions a program takes to contribute to change. Describe its components, their sequence, how they relate, and who is responsible for each. Distinguish activities delivered by program staff from those led by community members or partner organizations. Also identify external developments, such as changes in transportation services, employment opportunities, or public policy, that may influence results.

Resources

Resources include time, skills, relationships, facilities, equipment, information, funding, and other assets available to the program. Reviewing resources helps reveal whether planned activities are adequately supported. Include contributions from community members and partners, especially work that is unpaid or easily overlooked. Clear cost information is also needed when assessing efficiency, cost-effectiveness, or costs and benefits.

Stage of development

A program's stage of development reflects how established its activities and supporting systems are. Programs change over time, and their components may develop at different rates. A newly funded initiative may require different evaluation questions from a long-running program that is expanding or changing its approach.

Three useful phases are planning, implementation, and assessment of effects or outcomes. During planning, evaluation can test assumptions and refine the approach. During implementation, it can examine delivery, participation, accessibility, and adaptations. Once sufficient time has passed, it can assess intended and unintended outcomes. These phases may overlap, and planning for outcome assessment often needs to begin before implementation.

Context

A description of the program's context considers community history, geography, culture, politics, economic conditions, institutional practices, and related efforts. A responsive evaluation is sensitive to these influences and to differences in power and access to resources. Context helps explain findings and assess whether an approach may transfer elsewhere. For example, a housing initiative may need substantial adaptation across communities with different housing markets, transportation systems, or local regulations.

Logic model

A logic model summarizes how a program is expected to work. It connects resources and activities with outputs and intended outcomes, making assumptions about change visible. A flowchart, map, or table can show these relationships, including feedback loops or external influences when a simple sequence would be misleading.

Developing a logic model together can clarify program direction and identify questions for evaluation. The model expresses a theory about how change may occur; it does not establish that the program caused an outcome. If a model is used to estimate outcomes that were not directly measured, report those results as estimates and explain the evidence, assumptions, and uncertainty behind them.

The detail needed in a program description depends on the evaluation's purpose. Combine relevant documents, discussions with participants and staff, and observations of activities to create an accurate account. Check whether written plans match actual practice. Include contextual factors such as staff turnover, resource constraints, policy changes, community leadership, and partnerships that may affect implementation and results.

Focus the Evaluation Questions and Design

Focusing the evaluation means deciding what you most need to learn, why it matters, and how the findings will inform action. An evaluation cannot answer every possible question. Agreeing on priorities helps the group use its time and resources well while keeping community concerns visible.

Different questions require different designs and types of evidence. Plan before collecting information, including how you will analyze it and protect participants. Some design choices are difficult to change later, but the plan should allow justified adaptations when conditions change or important questions emerge.

Consider the following issues when focusing an evaluation:

Purpose

Purpose describes the overall reason for the evaluation. A clear purpose guides the questions, methods, level of detail, and intended use of findings. It also helps the group distinguish learning and improvement from other purposes, such as accountability or decisions about expansion.

There are at least four broad purposes for a community program evaluation:

• To gain insight. Evaluation can help a group explore a proposed approach, such as whether a resident-led neighborhood safety initiative fits local priorities and resources. Findings from similar efforts, together with local knowledge, can clarify assumptions, feasibility, and adaptations needed before implementation.
• To improve how work is carried out. Evaluation can examine what was delivered, who participated, what barriers arose, and how implementation differed from the plan. This information can guide improvements in quality, accessibility, coordination, and the use of resources.
• To assess program outcomes and contributions. Evaluation can examine changes associated with a program, such as improvements in school completion or access to stable housing. Assessing whether the program caused those changes requires a design that can address alternative explanations. Outcome findings can support accountability and decisions, provided the conclusions match the strength of the evidence.
• To support learning among people involved. Participating in evaluation can build skills, strengthen relationships, and help people reflect on their work. These benefits should arise through transparent and voluntary participation. They may:
  • Strengthen community decision-making, for example by giving participants meaningful authority in choosing questions and interpreting findings;
  • Support reflection on the program, for example through a follow-up conversation that helps participants identify useful experiences and unmet needs;
  • Build staff and community evaluation skills, including collecting, analyzing, and interpreting information; or
  • Contribute to organizational learning, for example by clarifying how activities align with the organization's mission and community priorities.

Users

Intended users are the people expected to apply the findings. Their information needs, decisions, and timelines should help shape the evaluation, alongside the interests of people affected by those decisions. Discuss tradeoffs openly: a smaller evaluation may answer a focused question well but provide less information about variation across settings or populations. Participation in choosing the focus can help users understand both the value and limits of the eventual findings.

Uses

Uses describe what people plan to do with what they learn. These might include revising activities, allocating resources, improving accessibility, deciding whether to expand a program, or strengthening community oversight. The following examples connect possible uses with the four broad purposes described above.

Some specific examples of evaluation uses

• To gain insight:

  • Assess community needs, priorities, and resources
  • Identify barriers to accessing and participating in the program
  • Learn how to describe and measure activities and outcomes meaningfully

• To improve how work is carried out:

  • Refine plans for introducing or adapting a practice
  • Determine how closely implementation followed the plan and why changes occurred
  • Improve the clarity and accessibility of educational materials
  • Strengthen cultural responsiveness and ongoing learning
  • Assess whether participants' rights, privacy, and well-being are protected
  • Set priorities for staff training and support
  • Make and document adjustments during implementation
  • Improve communication among participants, staff, and partners
  • Identify ways to improve participants' experiences and satisfaction
  • Compare costs with outcomes and benefits using an appropriate approach
  • Examine who benefits, who is underserved, and whether anyone experiences unintended harm
  • Inform community decisions about supporting and sustaining the program

• To assess program outcomes and contributions:

  • Assess participants' development and use of skills
  • Examine changes in behavior and conditions over time
  • Inform decisions about where to allocate resources
  • Document progress toward program objectives
  • Assess whether accountability commitments have been met
  • Combine findings from multiple evaluations to assess what may work in similar settings and where uncertainty remains

• To support learning among people involved:

  • Create opportunities to reflect on program experiences and information
  • Encourage dialogue and deepen understanding of community issues
  • Clarify shared goals and document differences among partners
  • Build evaluation skills among staff, participants, and community partners
  • Gather accounts of success, difficulty, and unexpected outcomes with appropriate permission
  • Support organizational learning, change, and improvement

Questions

An evaluation needs specific, answerable questions. Invite people involved in and affected by the program to identify what they need to know, then prioritize questions according to their importance, feasibility, and likely use. Ask whose experience each question captures and whether important differences or unintended consequences could be overlooked.

Methods

Evaluation methods draw on social research, community knowledge, and other fields. Common design families include experimental, quasi-experimental, and observational approaches, including case studies. Experimental designs use random assignment to estimate effects by comparing groups; randomization supports comparability but does not guarantee identical groups or eliminate every source of bias. Quasi-experimental designs use approaches such as comparison groups, interrupted time series, or differences in changes over time without random assignment. Observational and case study designs can describe experiences, implementation, context, and possible explanations using documents, interviews, observations, surveys, or other sources.

No single design is best for every question. Designs differ in their ability to support causal conclusions, describe experiences, explain implementation, or examine context. Select methods that fit the questions and ethical constraints, and state what they can and cannot establish. Combining qualitative and quantitative methods can strengthen understanding when the approaches are well designed and their findings are meaningfully integrated.

Methods may need to change as conditions or information needs change. For example, an evaluation focused on improving delivery may later need to inform expansion decisions. Document why changes were made, how they affect interpretation, and whether additional expertise or information is needed. Avoid changing methods simply to obtain preferred findings.

Agreements

Agreements describe the evaluation's purpose, questions, methods, intended users, deliverables, responsibilities, timeline, and budget. They should also address decision-making, data access and stewardship, privacy, reporting, conflicts of interest, and the process for revising the plan. Clarify who can review findings and how disagreements will be documented.

An agreement may be a contract, protocol, memorandum of understanding, or another written document appropriate to the work. Its purpose is to make expectations explicit and confirm shared understanding. The agreement should protect the integrity of the evaluation, including the reporting of unfavorable findings, while allowing justified changes through an agreed process.

Focusing the evaluation may involve discussions with supporters and critics, review of possible uses, and interviews with intended users about their decisions and timelines. Compare options for scope, cost, accessibility, and strength of evidence. A timely, focused evaluation can be useful, but explain the limitations of any compromises rather than treating speed or convenience as substitutes for credible methods.

Gather Credible Evidence

Credible evidence is information that is relevant, trustworthy, and adequate for answering the evaluation's questions. What is needed depends on the question and the claim being considered. Assessing a program's causal effect may require a strong comparison design, while understanding participants' experiences may call for carefully conducted interviews or observations. Community acceptance matters, but credibility also depends on transparent methods and attention to bias and uncertainty.

Community knowledge and technical expertise can complement each other. People with lived experience may identify questions, barriers, or interpretations that external evaluators would miss. Specialists can help address issues such as sampling, complex analysis, measurement, or sensitive data. Bring these perspectives together according to the evaluation's needs.

All evidence has limitations. Using multiple relevant sources and methods can help identify consistent patterns, explain differences, and reveal missing perspectives. Involving community members in gathering and interpreting information can improve relevance and understanding, provided they receive appropriate support and the process protects confidentiality and independent judgment.

The following features affect the credibility of evidence:

Indicators

Indicators translate concepts about a program, its implementation, and its expected outcomes into observable or measurable features. Define each indicator clearly, including what is counted or described, for whom, over what period, and using which source.

Examples of indicators include:

• The program's capacity to deliver services, such as staffing levels and appointment availability
• Participation rates, with a clearly defined eligible or intended population
• Participants' satisfaction, experiences, and reported barriers
• Program reach and intensity, including who participated, what they received, and for how long
• Changes in participants' knowledge, skills, opportunities, or behavior
• Changes in community conditions, relationships, or norms
• Changes in organizational or policy environments, such as new practices or improved service access
• Longer-term changes in community health or well-being, such as housing stability or reported access to needed care

Indicators should reflect the criteria that matter for assessing the program, including community-defined outcomes. Complex initiatives usually require several indicators. Where appropriate and feasible, examine differences across populations or locations while protecting privacy and avoiding misleading conclusions from small numbers.

One approach is a "balanced scorecard": a small, complementary set of indicators covering different aspects of the program. For example, it might describe service delivery, participants' experiences, observed outcomes, resource use, and changes in surrounding conditions. The aim is a balanced view rather than a large collection of measures with no clear purpose.

A logic model provides another way to choose indicators. Identify measures along the pathway from resources and activities to outputs and outcomes. Quantitative indicators can describe amounts or rates, while qualitative information can explain experiences, processes, and meanings that numbers alone may not capture.

Indicators need not focus only on long-term outcomes. They can also describe service quality, community capacity, trust, or relationships among organizations. Work with people familiar with the context to define observable signs of these concepts, and check whether those signs have the same meaning across participants and settings.

Indicators may need revision as understanding improves; document changes and their effects on comparisons over time. Remember that tracking program performance is one part of evaluation. A change in an indicator does not by itself explain why it happened. For example, rising unemployment among participants may reflect broader economic conditions, changes in who enrolled, program limitations, or a combination of factors.

Sources

Sources of evidence may include participants, residents, staff, administrative records, documents, observations, and other relevant information. Different sources provide different perspectives. Program records may describe services delivered, while participants and people unable to access the program can explain experiences those records do not capture. Examine discrepancies rather than assuming one source is always more reliable.

Explain how sources were selected, who or what was excluded, and how those choices may affect conclusions. Numerical and narrative information can complement each other when their purposes and methods are clear. For example, participation records may reveal a decline in attendance, while interviews help explain transportation difficulties or changes in scheduling.

Quality

Quality concerns whether information is sufficiently accurate, consistent, relevant, and complete for its intended use. It depends on clear definitions, appropriate tools, accessible data collection, training, sampling, coding, data management, and error checking. Pilot procedures when feasible and examine missing data and possible bias. Discuss tradeoffs, but do not lower evidence requirements simply because a preferred conclusion would be convenient.

Quantity

Quantity refers to how much information is needed. Plan the amount and range of evidence required for the questions, methods, and intended decisions. Quantitative studies may need sample-size planning; qualitative work requires attention to the depth and diversity of perspectives. More data do not automatically mean better evidence, especially if collection is biased or burdensome. Gather information with a clear purpose and establish a reasoned stopping point.

Logistics

By logistics, we mean the people, timing, settings, technology, and procedures needed to gather and manage information. Ask participants about language, communication, accessibility, and privacy preferences rather than assuming that everyone in a community shares the same needs. Explain participation clearly, establish appropriate consent procedures, and specify who can access data, how they will be protected, and how long they will be retained.

Develop and Justify Conclusions

Evidence does not interpret itself. Conclusions require analysis, attention to context, and a clear explanation of how findings relate to the evaluation questions and agreed criteria. Include different perspectives and consider alternative explanations. Agreement can support use, but consensus is not proof: a conclusion must remain grounded in evidence even when some people disagree with it.

The principal elements in developing and justifying conclusions are:

Performance standards

Performance standards describe the basis for judging the program, such as expected levels of access, service quality, or improvement. These differ from the standards used to assess the quality of the evaluation itself. Agree on performance criteria and thresholds early where possible, explain whose values they reflect, and document any revisions so that judgments are not adjusted simply to fit the results.

Analysis and synthesis

Analysis and synthesis help organize and interpret the evidence. Analysis examines patterns within information; synthesis brings findings together to develop a broader understanding. In an evaluation using multiple methods, analyze each source appropriately and then examine where findings agree, differ, or leave gaps. Explain how information was classified, compared, and combined, including decisions about missing data, uncertainty, and potentially conflicting evidence.

Interpretation

Interpretation explores what the findings mean in context. In a hypothetical community survey, 15% of respondents report witnessing violence during the previous year. That finding has a different meaning if a comparable earlier survey found 50% than if it found 7%. Before interpreting either pattern, check whether sampling, questions, reporting periods, and participation were comparable. Even a credible change over time does not establish that the program caused it. Community perspectives can help explain possible influences and the practical significance of the findings.

Judgments

Judgments assess the program's value or performance by comparing interpreted findings with explicit criteria. Different criteria can lead to different judgments. For example, managers may see a 10% increase in outreach as progress, while residents may point out that services still do not reach neighborhoods facing the greatest barriers. Both perspectives can inform the assessment. Explain the basis of each judgment and distinguish improvement from meeting an agreed standard.

Recommendations

Recommendations identify actions to consider in response to the findings. They require attention to evidence, feasibility, resources, alternatives, and the priorities of affected people. For example, evidence that a program expanded access to services for survivors of domestic violence may support continued investment, while also revealing needs for safer access, different hours, or stronger coordination. Decisions about changing or ending services should consider continuity and the consequences for people who rely on them.

Recommendations should make their reasoning explicit and remain proportionate to the evidence. Explain expected benefits, possible harms, practical requirements, and uncertainties. Where values or priorities differ, document those differences rather than presenting a contested choice as the only possible conclusion.

Three practices can improve the relevance and usefulness of recommendations:

• Share draft recommendations with people expected to use or be affected by them.
• Seek responses from people with different experiences, roles, and perspectives.
• Present feasible options and tradeoffs, identifying when the evidence supports a clear recommended action.

Strengthen conclusions by actively examining alternative explanations and evidence that challenges the initial interpretation. Do not dismiss an explanation merely because it is inconvenient. Where multiple interpretations remain plausible, describe them and the information that could help distinguish among them. Plan major analyses in advance when possible and identify additional exploratory analyses clearly.

Use Findings and Share Lessons Learned

Findings do not automatically lead to action. Plan for their use from the beginning by identifying decisions, responsible people, timelines, and the resources needed for follow-through. Continue communicating throughout the evaluation, and make space to discuss findings that are unexpected or difficult to act on.

The following elements can help people use evaluation findings:

Design

Design connects the evaluation's questions and methods with the decisions it is intended to inform. As discussed earlier, clarify who needs the findings, when they need them, and what actions are possible. Involving community members and intended users in these decisions can improve relevance while helping everyone understand the evaluation's boundaries.

Preparation

Preparation means building the understanding and capacity needed to act on findings. Discuss possible results before they are available, including what would count as sufficient evidence for a decision and how uncertainty will be handled. This can reveal unrealistic expectations and identify support that users will need.

For example, present intended users and community partners with hypothetical findings and ask what actions they would consider. If the information would not help them make a decision, revisit the questions or reporting plan. Explore favorable, unfavorable, and mixed findings without encouraging people to choose a preferred result in advance.

Feedback

Feedback is the continuing exchange of information among people involved in the evaluation. Provide accessible opportunities to comment on decisions, emerging findings, and draft interpretations. Explain how feedback influenced the work. Review can identify errors and missing context, but it should not give participants, staff, or funders authority to suppress credible findings.

Follow-up

Follow-up provides support as people interpret findings and decide what to do. Completing a report is a milestone; learning and action continue afterward. Assign responsibility for agreed actions, identify needed resources, set review dates, and track progress. Someone may coordinate this work while ensuring that the evidence and its limitations remain visible in decisions.

Supporting use also means guarding against misuse. Findings are shaped by the design, participants, setting, and period studied. A single case study may offer valuable insights, but it usually cannot establish that every site in a large program will have the same experience or outcome. Explain what can reasonably transfer and what requires further investigation.

Anyone involved may selectively emphasize favorable or unfavorable findings. Reduce this risk by reporting the full pattern of results, explaining uncertainty, and correcting misleading interpretations. New uses of existing findings may be reasonable, but they require a fresh assessment of whether the evidence supports the proposed decision.

Dissemination

Dissemination means sharing the evaluation's methods, findings, and lessons with relevant audiences in a timely and accessible way. Plan language, format, communication channels, and timing with intended users and affected communities. Use plain-language summaries and other appropriate formats alongside technical detail. Report findings fairly while protecting confidentiality and honoring legitimate restrictions on sensitive information.

Benefits can also arise from participating in the evaluation itself. These "process uses" include clearer thinking about assumptions, stronger skills, improved relationships, and a shared understanding of the program. Notice and assess these benefits rather than assuming that participation will always have a positive effect.

Evaluation can help staff and community partners clarify goals, examine how their work connects, and make decisions using evidence and explicit values. It can also reveal disagreements or organizational barriers that need attention. A learning-oriented approach treats those discoveries as information for improvement.

Additional process uses for evaluation include:

• Developing indicators together can clarify what different participants value and which outcomes need greater attention.
• Connecting findings to planning and resource decisions can make learning part of routine practice. For example, a funder might support adaptations identified through evaluation, while avoiding incentives that reward only favorable results or discourage serving people facing greater barriers.

Standards for Evaluation Quality

The Joint Committee on Standards for Educational Evaluation developed The Program Evaluation Standards to guide sound and fair evaluation. The discussion below adapts the 30 standards reflected in the original CDC framework, retaining their historical four-category organization. The later third edition of the Joint Committee's standards reorganizes the guidance into five categories, including evaluation accountability, and should be consulted when selecting standards for a new evaluation.

Standards help groups assess the quality of their evaluation and make reasoned choices when resources or conditions create tradeoffs. For example, technically accurate findings may have little value if they arrive after a decision or cannot be understood by intended users. A useful design also needs to be feasible and ethically acceptable.

Apply relevant standards during planning, implementation, reporting, and follow-up. The historical summaries below use updated explanatory language; they are not a verbatim reproduction of the current standards. Document how you apply the guidance and how you address tensions among practical requirements without compromising participants' rights or the integrity of findings.

The original 30 standards were grouped into four categories:

• Utility
• Feasibility
• Propriety
• Accuracy

The utility standards address the evaluation's relevance and usefulness:

• Identification of Interested and Affected Parties: Identify people involved in or affected by the program and its evaluation, including those whose perspectives are often overlooked, so that relevant priorities can shape the work.
• Evaluator Credibility: Those conducting the evaluation should bring appropriate skills, demonstrate trustworthiness, understand the context, and explain how they will manage potential conflicts of interest.
• Information Scope and Selection: Collect information that addresses the agreed questions and the needs of intended users and affected communities. Avoid gathering information that has no clear use.
• Values Identification: Explain the values, perspectives, criteria, and reasoning used to interpret findings and judge the program.
• Report Clarity: Describe the program, context, evaluation purpose, methods, findings, and limitations in language and formats that intended audiences can understand and use.
• Report Timeliness and Dissemination: Share relevant findings in time to inform decisions, clearly identifying preliminary results and protecting sensitive information.
• Evaluation Impact: Plan and conduct the evaluation in ways that support learning, responsible action, and follow-through, while monitoring possible unintended effects of the evaluation itself.

Feasibility Standards

Feasibility standards address whether the evaluation can be carried out effectively with the available resources and within its setting. They encourage practical choices without treating convenience as sufficient justification for weak or unfair procedures.

The feasibility standards address:

• Practical Procedures: Use procedures that gather needed information while minimizing unnecessary burden and disruption for participants, staff, and community partners.
• Political and Contextual Viability: Anticipate differing interests, power relationships, and conditions that may affect participation or use. Establish fair ways to address disagreement and protect the evaluation from inappropriate interference.
• Cost-Effectiveness: Match the evaluation's scope and resource use to the value of the information needed, including the time and contributions required from community members.

Propriety Standards

Propriety standards address ethical conduct, fairness, rights, and the well-being of people involved in or affected by the evaluation. The original framework included the following eight areas.

• Service Orientation: Design the evaluation to help the program respond effectively and fairly to the people it is intended to serve, including those experiencing barriers to access.
• Formal Agreements: Document responsibilities, procedures, timelines, reporting arrangements, and the process for changing agreements so that expectations are clear and accountable.
• Participants' Rights and Well-Being: Protect people's dignity, privacy, and ability to make informed choices about participation. Establish appropriate consent and review procedures for the evaluation's setting and activities.
• Respectful Interactions: Treat people with respect, provide accessible ways to participate, and address coercion, discrimination, distress, or other risks arising from the evaluation.
• Complete and Fair Assessment: Examine strengths, limitations, benefits, harms, and unintended consequences fairly, with attention to differences in experiences and outcomes.
• Disclosure of Findings: Make findings and limitations accessible to relevant audiences while protecting confidential information and meeting applicable disclosure obligations. Explain any restrictions on access.
• Conflicts of Interest: Identify, disclose, and manage actual or perceived conflicts that could influence the evaluation's design, conduct, interpretation, or reporting.
• Fiscal Responsibility: Use resources responsibly, maintain appropriate financial records, and account transparently for evaluation expenditures.

Accuracy Standards

Accuracy standards address whether the evaluation provides dependable information and defensible interpretations. They support confidence in the findings while requiring explicit acknowledgment of uncertainty and limitations.

The original framework included 12 accuracy standards:

• Program Documentation: Describe the program clearly and accurately, including what was planned, what was implemented, and any important changes.
• Context Analysis: Examine conditions that may shape implementation, outcomes, interpretation, and the transfer of findings to other settings.
• Described Purposes and Procedures: Document the evaluation's purposes and methods in enough detail for others to assess the work, including justified departures from the original plan.
• Defensible Information Sources: Explain how information sources were selected and why they are appropriate, including relevant exclusions and possible sources of bias.
• Valid Information: Use measures and collection procedures that support the intended interpretations, checking whether they are appropriate for the people, languages, and settings involved.
• Reliable Information: Use procedures that produce sufficiently consistent and dependable information for the evaluation's purposes, and report relevant measurement limitations.
• Systematic Information Management: Organize, review, protect, and verify information systematically. Correct errors through documented procedures and preserve a clear record of important changes.
• Analysis of Quantitative Information: Analyze numerical information using methods suited to the design and questions, accounting for missing data, uncertainty, and relevant assumptions.
• Analysis of Qualitative Information: Analyze interviews, observations, documents, and other descriptive information systematically, preserving context and considering interpretations that challenge emerging themes.
• Justified Conclusions: Show how conclusions follow from the evidence and criteria, and distinguish observed changes from claims about what caused them.
• Impartial Reporting: Report findings fairly and transparently, including unfavorable or mixed results. Explain how potential bias and competing interpretations were addressed.
• Metaevaluation: Assess the quality of the evaluation itself using appropriate standards, during the work and after completion, to identify strengths, limitations, and improvements.

Applying the Framework: Conducting Useful and Credible Evaluations

Evaluation can support community learning, program decisions, and accountability to participants, partners, and funders. The appropriate scope depends on what people need to know and the consequences of the decisions ahead. Useful starting questions include:

• What evaluation approach fits our questions, context, and resources?
• What are we learning, whose experiences are represented, and what remains uncertain?
• How will we use what we learn to improve effectiveness, accessibility, and fairness?

The framework helps connect these questions with practical steps and standards. Its value lies in making evaluation decisions deliberate, transparent, and responsive to the people affected by the program.

Applying the framework requires judgment and a suitable combination of community knowledge and evaluation skills. Relationships may be complex, resources limited, and conditions subject to change. Develop an approach that fits those realities while protecting ethical standards and the integrity of the evidence. Seek specialized support for questions or methods that exceed the team's experience.

Cost is a common concern, but an evaluation's budget should be considered in relation to its questions and intended decisions. A modest evaluation can provide useful information about implementation or participant experience. More demanding claims, such as estimating causal effects or detecting small differences across populations, may require additional resources and expertise.

Build evaluation into routine work where appropriate, using existing information only when it is suitable and can be used responsibly. Coordinate collection and reporting with decision timelines, and avoid repeatedly asking community members for information that is already available.

Technical complexity can also be a barrier. Begin with clear, useful questions and select methods the team can carry out well. Provide training and support, and be explicit about what the resulting evidence can establish. A focused, well-executed evaluation is more useful than an ambitious design that cannot be implemented credibly.

Staff and participants may reasonably worry that evaluation will be punitive, exclusionary, or used to justify decisions already made. Address those concerns directly. Explain the purpose, share decision-making where appropriate, protect candid feedback, and demonstrate how findings will support learning and accountability. Trust depends on how the evaluation is conducted and used, not simply on assurances that participation is welcome.

In Summary

Evaluation helps communities understand how programs operate, what changes occur, who benefits, and what needs improvement. It can inform decisions about resources, adaptation, expansion, or discontinuation while identifying unintended consequences. Its conclusions should reflect both the evidence and its limits.

The original CDC framework remains a useful teaching model when its historical basis is clear and newer guidance informs current practice. Use the steps and applicable standards to develop an evaluation that fits the program, involves affected people meaningfully, and connects findings with action. The Magenta Book - Guidance for Evaluation is an additional evaluation resource; this link points to an older edition, so consult the latest HM Treasury guidance when using it to plan an evaluation.