| Course | PU605 Public Health Programming |
|---|---|
| Unit | Unit 9 |
| Paper type | Evaluation design |
| Length | About 900 words, 6 pages double-spaced |
| Format | APA 7 student paper |
| School | Purdue University Global |
| Program | Master of Public Health (MPH) |
| Updated | October 2026 |
Free sample paper for PU605 Unit 9
Nine Events in Four Years Cannot Fall Measurably in Two: A Unit 9 Evaluation Design That Says So on Its First Page and Then Measures What Should Change First
[Student Name]
Purdue University Global
PU605: Public Health Programming
Unit 9 Evaluation Design
[Instructor Name]
[Date]
Welbourne and its figures are composite; the evaluation logic, the measures and the research cited are real.
Why the Injury Outcome Cannot Be the Primary Measure
Nine falls in four years is roughly 2.3 a year across the four tracts. A program reaching 800 of about 2,300 eligible units might prevent one event in two years. National surveillance puts these injuries in the thousands a year across the country, which is what makes the mechanism well established and a single city's count tiny (Harris et al., 2011). No design can distinguish one prevented event from the ordinary variation in a count this small, and a before-and-after comparison of annual totals will show a decline or an increase with equal probability regardless of what the program does. Reporting such a comparison as evidence would be misleading in either direction, including when it flatters the program.
What the Design Measures Instead
1. Guards installed above the first floor in eligible units, by tract and by month, from the program's own register. Target 800 units and 2,400 windows.
2. Guard prevalence in the four tracts, from the same sidewalk survey used in the assessment, repeated each August by two people over four days. Baseline 4 percent. This is the only measure that does not depend on the program's own records.
3. Retention at one year: guards in place and functional at a sample of 200 installed units, inspected in person. Target 90 percent.
4. Consent rate: share of approached owners who agree, reported separately for associations, larger owners and owner-occupants.
5. Belief about screens: a two-question survey at installation and at the one-year check, reporting the share who say a screen would hold a child.
6. Referral pathway function: share of trauma center cases followed by an offer in that building within ten days.
The sidewalk survey is the measure that matters most, because it is the only one that would still show something if the program stopped reporting.
The Comparison the Design Does Use
Tract 19 is already covered by the existing code requirement for buildings with four or more units and has guard prevalence of 14 percent. It is not a control, since it differs in housing stock, but it supplies a benchmark for what a requirement achieves without a program behind it. The design reports the four tracts' prevalence against tract 19's each August and against their own baselines, and labels this as a benchmark comparison rather than a controlled one.
Design Type, Stated Plainly
A pre-post design with repeated cross-sectional exposure measurement, no control group, and implementation measures reported alongside. That is a weak design for causal inference and an appropriate one for this question, because the program's own logic says exposure has to change before injuries can, and exposure is measurable now. Where an outcome is rare and an intervention's mechanism is well established elsewhere, evaluating delivery and intermediate change is the defensible route rather than an underpowered outcome study (Victora et al., 2004).
What Would Make an Injury Outcome Measurable Later
1. Pooling across cities, which requires the register's data fields to match what other programs collect, so the design specifies them now.
2. A five to ten year follow-up using hospital discharge data, which the health department can run at low cost once the register identifies which buildings were treated.
3. A case-control study using the register as the exposure source, comparing guard presence between buildings where a fall occurred and matched buildings where none did, which becomes feasible once several hundred buildings are documented.
The design includes these as a deliberate handover rather than as a promise, because the two-year evaluation cannot deliver them.
How the Measures Map to the Program's Logic
Installations and consent test the delivery chain. Prevalence tests whether delivery changed the exposure in the population rather than only in participating units. Retention tests the assumption the logic model flagged as its second weakest. Belief tests the activity the program expects least from. Dimensions of reach, adoption, implementation and maintenance are all represented, which is what prevents the evaluation from reporting only the number that looks best (Glasgow et al., 1999).
Reporting, and What Will Not Be Claimed
1. Quarterly monitoring reports on installations, consent and fidelity.
2. An August prevalence figure each year, published whether or not it moves.
3. A final report that states the injury count for the period and explicitly declines to attribute it, in a sentence drafted now and included in the funder agreement.
4. A statement of what the program cost per unit installed, which is the figure another city will want.
Three Threats to Even These Measures
Self-report in the register, since the program counts its own installations and has an interest in the number; the mitigation is the independent sidewalk survey. Attrition in the retention sample, since the households hardest to revisit are likely the ones where guards came out. And seasonality in the belief survey, asked at installation in summer and at the check possibly in winter, which the design handles by scheduling checks in the same season rather than at twelve months exactly.
References
Glasgow, R. E., Vogt, T. M., & Boles, S. M. (1999). Evaluating the public health impact of health promotion interventions: The RE-AIM framework. American Journal of Public Health, 89(9), 1322-1327. https://doi.org/10.2105/AJPH.89.9.1322
Harris, V. A., Rochette, L. M., & Smith, G. A. (2011). Pediatric injuries attributable to falls from windows in the United States in 1990-2008. Pediatrics, 128(3), 455-462. https://doi.org/10.1542/peds.2010-2687
Victora, C. G., Habicht, J. P., & Bryce, J. (2004). Evidence-based public health: Moving beyond randomized trials. American Journal of Public Health, 94(3), 400-405. https://doi.org/10.2105/AJPH.94.3.400
PU605 Unit 9 instructions, in plain terms
Unit 9 usually asks for an evaluation design for your own program. If the health outcome is too rare to move measurably inside the funding period, say so first and offer the intermediate measures in the same breath. Include at least one measure independent of the program's own records, since self-counted output is the weakest evidence you have. Name your design type plainly, including that a pre-post comparison without a control is weak for causal inference and appropriate when the mechanism is established elsewhere. Describe any comparison as a benchmark rather than a control if that is what it is. Say what would make an outcome measurable later, and draft the sentence that declines attribution now.
Inside the PU605 Unit 9 example
The design on this page opens with the arithmetic that rules out the injury outcome, then lists six measures with baselines and targets. A highlighted line identifies the sidewalk survey as the measure that matters most, because it would still show something if the program stopped reporting. A margin note explains why the design names its comparison tract a benchmark rather than a control. The design type is stated plainly as pre-post with repeated cross-sectional exposure measurement. Three routes to a later outcome study follow as a handover. Reporting commitments include a sentence declining attribution, drafted in advance and flagged by a further note, and three threats close the design.
Where the points sit in the PU605 Unit 9 rubric
Scoring rewards honesty about power. The strongest designs state why the headline outcome cannot be evaluated and then supply measures that test the program's own logic chain. Credit goes to at least one independent measure, and to describing a non-equivalent comparison as a benchmark. Points follow for naming the design type without dressing it up, and for mapping each measure to a step in the logic model. Drafting the no-attribution sentence before data exist is explicitly rewarded, because a favorable count will otherwise tempt a claim. Threats to the measures themselves, including self-report and attrition in the retention sample, are expected, along with a cost per unit figure.
PU605 Unit 9 help from the desk
The commonest error is promising an outcome evaluation that the event rate cannot support, which produces either a false claim or an apparent failure. Do the arithmetic first. The second is relying entirely on the program's own register, since the program has an interest in the number. Third, do not call a dissimilar tract a control, because the comparison will be read as causal. Avoid scheduling the retention check exactly at twelve months if that moves it into a different season from the baseline question. And write down what you will not claim, so a good year cannot be converted into evidence. Send the program and we will design its evaluation.
Get a PU605 Unit 9 example written to your instructions
Send the objectives and logic model the PU605 Unit 9 evaluation must test, any required design, such as a comparison area or time series, and the rubric. Process and outcome are separated in the free first custom sample, a comparison is built in where feasible, and failure is defined before any data exist; expect it inside 24-48 hours. The paper above is an original model document written by our desk, not a submitted student paper and not an official Purdue University Global document.
More PU605 papers and MPH samples
- PU605 Unit 1: Program Framing Discussion Post
- PU605 Unit 2: Community Assessment
- PU605 Unit 3: Priority Setting Analysis
- PU605 Unit 4: Program Logic Model
- PU605 Unit 5: Intervention Design
- PU605 Unit 6: Implementation Seminar Reflection
- PU605 Unit 7: Stakeholder Engagement Plan
- PU605 Unit 8: Implementation Plan
- PU605 Unit 10: Program Plan Document
- PU580 Unit 9: Cancer Cluster Letter
- PU570 Unit 9: Prevention Strategy Brief
- PU565 Unit 9: Mixed Methods Matrix
- PU555 Unit 9: Measurement Plan
PU605 Unit 9 questions, answered
Where can I find a free PU605 Unit 9 sample paper?
The whole evaluation design, all six measures and the threats, is open here.
Why not evaluate the injury outcome?
Because about 2.3 falls a year across four tracts means the program might prevent one event in two years, which no design can separate from chance variation.
What is the most important measure?
The independent August sidewalk survey of guard prevalence, since it does not depend on the program's own records.
Is a benchmark the same as a control?
No. The comparison tract differs in housing stock, so it shows what an existing requirement achieves and cannot support a causal claim.
What would make an outcome study possible later?
Pooling with other cities using matched data fields, a five to ten year discharge-data follow-up, or a case-control study using the window register as the exposure source.