Program Evaluation Questions Participants Can Answer
Sixteen questions participants can answer at the end of a program. Take it below, or open it in the editor. It reports how the program was delivered and how people describe their own change, and it is straight about stopping there.
- 16questions
- About 4 minto complete
- Then and nowboth rated at the end
- Self-reportnot an outcome evaluation
What these program evaluation questions can and cannot establish
The CDC framework draws the line this page is built on: "Evaluation questions are not the same as data collection questions (e.g., survey, interview, and focus group questions)".[1] What your funder asked sits a layer above what you put to a participant.
Evaluation splits in two. A process evaluation tells you how well you delivered the program: attendance, adherence, your work plan.[3] An outcome evaluation asks how well you met the goals you set, which is the half a funder wants.[4] A survey is legitimate for both,[5] and contributes very different amounts to each.
The confusion is normal. Self-reported items get filed under "outcome and impact" all the time, and a board reads that as though the survey settled it. It did not. RAND is blunt that a real outcome evaluation is complex, costly work.[4] So the honest move is to send these sixteen questions this week and label what they are.
| What you have been asked | What kind of question that is | What these 16 questions contribute |
|---|---|---|
| Did the program run the way it was meant to? | Process: implementation and service delivery.[5] | Most of the answer. Participants were there for it. |
| Did the people in it change? | Outcome: the effects on the target population.[2] | Their own account of it. That is evidence, and it is not proof. |
| Did the program cause that change? | Outcome, with a comparison designed into it.[4] | Nothing on its own. There is no comparison inside one survey. |
| What did it produce? | Outputs: the products of program activities.[1] | Nothing. Outputs are counted in your own records. |
Three readers are on the wrong page. One course, with one syllabus and one cohort, is a level below a program: the course evaluation questionnaire fits better. One delivered session, including a training program where a stakeholder expects the four-level Kirkpatrick model,[7] belongs on the post training survey template. A wellness program means asking about sleep, workload and stress, not delivery: the employee wellness survey.
The 16 questions in the program evaluation survey template
Every question, how it is answered, and what to do with the answer.
Block 1. Who is answering, and how much they got
Three questions that decide how to cut everything after them.
- How much of the program did you take part in?The dose. Cut every later figure by this. A result built only from people who finished describes finishers, not the program.
- How did you first find out that the program was open to you?Reach. One route supplying almost everybody tells you which door the next cycle has to open.
- When did your part in the program finish?Recall distance. Answers from more than six months back are memory; keep them separate from the current cycle.
Block 2. How the program was delivered
Four statements about things a participant was in the room for. This block a survey can settle on its own.
- The program ran the way it was described to me before I joined.The gap between the offer and the delivery. Read it before anything else in the survey.
- I was able to take part in the activities that were planned for me.Access to what was planned. Read it beside question 1: low dose plus low access is a scheduling problem, not a motivation problem.
- Taking part fitted around my other commitments.Practical fit. The one delivery item you can act on before the next cycle without redesigning anything.
- The people running the program answered my questions when I had them.Responsiveness of the staff. Report the count of Disagree and Strongly disagree, not the average.
Block 3. The then and now pair
One question asked twice, both at the end, on one scale.
- Think back to the week before you started. How confident were you then about [WHAT THIS PROGRAM SET OUT TO CHANGE]?The retrospective rating. It is a judgement made today about a past state, and it is labelled that way in the report.
- How confident are you about [WHAT THIS PROGRAM SET OUT TO CHANGE] today?The current rating. Report the per-person difference between the two, and the share whose difference is zero or negative.
Block 4. What changed, in words and in actions
One open line, one behaviour checklist, one worth-it statement. Behaviour here is reported, not observed.
- What can you do now that you could not do before the program? Write none if nothing has changed.The line that goes in the report. Answers naming something outside the program come out of the change figure.
- Since the program, which of these have you actually done?Reported action. A tall None of these yet bar next to a high confidence rating is confidence with nothing attached to it.
- The program was worth the time it took me.Perceived value of the time. The one item that catches a program people rate highly and would not repeat.
Block 5. Satisfaction, recommendation and the open lines
Four at the end, where somebody who is done answering still has something to give.
- Overall, how satisfied are you with the program?Satisfaction. Report the share picking each option; a mean of five words is not a number.
- Would you tell someone in your situation to join this program?Recommendation with a caveat slot, so a qualified yes is never counted as a plain one.
- What is the one thing you would change about the program?One change, forced. The constraint is the point: it turns a wish list into a candidate for the next cycle.
- Is there anything else the people who run or fund this program should know?The catch-all. Read every line: a safeguarding or an access problem has nowhere else in this survey to arrive.
Before you send. [WHAT THIS PROGRAM SET OUT TO CHANGE] is the only placeholder, in questions 8 and 9. Put the same words in both. Every statement is worded so that agreeing is the good answer; keep that polarity if you add items. The set is already built, and you can equally create an online survey and paste in the blocks you want.
The then and now pair: a pre and post design in one send
Two of the sixteen are one question asked twice, both at the end, on one scale. It ships with the condition for leaving it out.
A before-and-after comparison assumes the scale means the same thing on both days. Sibthorp and colleagues set out why that fails on a self-report measure: a program can move the scale itself, by changing what a participant takes the question to mean, so the difference understates what the program did.[6] Asking the "before" rating at the end, on the same scale as the "now" rating, removes that. Both are made at one sitting on one metric.[6]
The limit ships with it. Where participants already understood the thing you are asking about, or the program could not shift their self-knowledge, the pair adds nothing and is the worse choice; one plain rating is better.[6] Apply that test before you send.
It is also what to reach for when no pre-program rating was ever collected. A true pre and post design takes the first rating before anybody starts; the then and now pair recovers a version of it in one send. Say which you have.
How to score a program feedback survey without inventing a number
Four readings, in the order you should trust them, and what breaks each one.
| Reading | How to work it out | What it will not survive |
|---|---|---|
| Delivery | Average the four statements in block 2, and report the count of Disagree and Strongly disagree beside it. | A small number of responses. Print the count next to the average, always. |
| The then and now gap | Subtract each person's question 8 rating from their question 9 rating, then average those differences. | Being labelled an effect of the program. It is a self-reported change, rated at one sitting. |
| Satisfaction and recommendation | Report the share choosing each worded option. Do not average words. | Comparison with another program's number, which came off a different instrument. |
| The two open lines | Read all of them, tag them into themes, then publish the tag counts with the quotes. | Being summarised without the counts underneath. |
There is no index and no composite score: one number across four blocks would hide which block is low. Report the blocks separately, with the response count beside every figure. No data collection method gives a complete picture,[1] so set these numbers next to what your own records already count. If what you are evaluating is a compliance program, effectiveness is scored against a control framework instead, on the compliance program survey.
Sending it to the people who were in the program
Three decisions particular to a program. Sending, reminding and closing work the same on any survey.
Decide where the program ends
A program is a standardised set of activities run to a specific result,[5] so this goes out at the close of a defined cycle: a cohort, a grant year, a season. Send it to everyone in it, including people who dropped out.
Fill in the placeholder in their words
Write what the program set out to change the way a participant would say it, not the way the funding application does. Identical wording in questions 8 and 9.
Report what it does not answer
Send the findings with one line naming which of the funder's questions this survey does not settle. That line stops a self-report block being read as an outcome measure.
One facilitated day in one room is rated on a workshop evaluation form, not here. And before a program is designed at all, the training needs assessment questions decide what is worth building.
Program evaluation survey FAQ
What is the difference between a process question and an outcome question?
Process questions cover implementation and delivery, and map to your inputs, activities and outputs. Outcome questions cover the effects on the people in it, and map to short, intermediate and long-term outcomes.[2] These sixteen answer the first kind directly.
Can this survey prove the program worked?
No, and no participant survey can. Showing that a program met its goals is an outcome evaluation, which RAND calls complex, costly work.[4] This gives that work one clean input, from the people who were in it.
How many questions should a program evaluation survey have?
Sixteen here, fewer than the twenty-to-fifty lists elsewhere. The test on each item was whether a line in the report would be missing without it. A demographics block did not survive it.
Who should get it, and when?
Everyone in one defined cycle, at the end of that cycle, including people who left part way through. Collecting enough responses is a real constraint, so decide the reminder before you send.[5]
What counts as a program here?
A relatively standardised set of activities that aim to achieve a specific result.[5] A grant-funded service, a volunteer program, a mentoring or after-school program, an initiative reporting to a board. One activity on its own sits a level below.
A programme is judged against the aim it was funded to hit. When the thing has simply stopped and the question is what to do differently next time, the lessons learned questions are the ones to send.
Methods and sources
What this template is based on
The split between how well a program was delivered and whether it met its goals follows the CDC framework and the RAND Getting To Outcomes steps.[1][3][4] The definition of a program, and the position that a survey is a legitimate evaluation method with known limits, come from the Australian Institute of Family Studies guide for beginners.[5] The then and now pair follows the retrospective pretest literature on response-shift bias.[6]
Question licensing
All 16 questions were written for this page. Program evaluation has no licensed instrument this template needs, so there is nothing here to attribute and no licence to clear: copy, edit, translate and redistribute them freely. The Kirkpatrick Model is named once as a third-party framework and no Kirkpatrick material is reproduced.[7]
Limits and disclosure
This is a self-report instrument. It records what participants say about how the program was delivered and how they describe their own change. It does not establish that the program caused an outcome, it produces no cost or benefit figure, and no single data collection method gives a complete picture.[1] Question wording can bias answers, and a survey gives no opportunity to ask why an answer was given or how a question was read.[5] It states no minimum response count.
SuperSurvey builds the editor this template opens in. It sells no evaluation service and no respondent panel, so nothing above has a product behind it.
References
- Kidder DP, Fierro LA, Luna E, et al. "CDC Program Evaluation Framework, 2024." MMWR Recommendations and Reports 73, no. 6 (2024): 1-37. cdc.gov
- Centers for Disease Control and Prevention. Developing Evaluation Questions. Division of STD Prevention, CS249668. cdc.gov
- RAND Corporation. Getting To Outcomes, Step 07: Process Evaluation. TL-259. rand.org
- RAND Corporation. Getting To Outcomes, Step 08: Outcome Evaluation. TL-259. rand.org
- Australian Institute of Family Studies. Using a survey to collect data for evaluation: A guide for beginners. Child Family Community Australia. aifs.gov.au
- Sibthorp J, Paisley K, Gookin J, Ward P. "Addressing Response-shift Bias: Retrospective Pretests in Recreation Research and Evaluation." Journal of Leisure Research 39, no. 2 (2007): 295-315. health.utah.edu
- Kirkpatrick Partners. What is The Kirkpatrick Model? kirkpatrickpartners.com
Send the program evaluation survey at the end of this cycle
Sixteen questions, about four minutes, with the then and now pair in place. Open it in the editor, fill in the placeholder, share one link.
Use this template