| Course | PU525 Biostatistics for Public Health |
|---|---|
| Unit | Unit 9 |
| Paper type | Statistical analysis plan |
| Length | About 870 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 PU525 Unit 9
Decided Before the Data Arrive: A Statistical Analysis Plan Committing a Composite County Home-Visiting Program's Evaluation to Its Primary Outcome, Exclusive Breastfeeding at Six Months, Its Test, Its Assumption Checks and How It Will Handle the Mothers It Cannot Reach
[Student Name]
Purdue University Global
PU525: Biostatistics for Public Health
Unit 9 Statistical Analysis Plan
[Instructor Name]
[Date]
The program, sample sizes and assumptions are composite and bracketed for the writer's own data; the methods and research cited are real.
1. Evaluation Question
Do first-time mothers enrolled in the county's nurse home visits more often reach six months of exclusive breastfeeding than similar mothers with low incomes who are not enrolled?
2. Primary Hypothesis
The proportion exclusively breastfeeding at six months differs between enrolled mothers and matched comparison mothers. One primary outcome is named to avoid searching many outcomes for a significant one. Breastfeeding is chosen because of its well-documented benefits for infants and mothers (Victora et al., 2016).
3. Variables
Outcome: exclusive breastfeeding at six months, yes or no, by maternal report of no other liquids or foods except medicines and vitamins, coded one for yes. Exposure: program enrollment, coded one for enrolled. Covariates, fixed in advance: maternal age in years; education, three categories; nutrition program participation, yes or no; return to work or school by [12] weeks, yes or no; preterm birth, yes or no; and cesarean delivery, yes or no.
4. Design and Sample
Enrolled mothers: about [300]. Comparison mothers: about [300] eligible mothers who gave birth in the same period at the same hospitals but lived outside the program's service area, matched on age group and parity. Power: with [300] per group and a comparison proportion of [25] percent, the study can detect a [12]-point difference about four times in five testing in both directions at the usual [five] percent threshold, allowing for [18] percent missing outcomes.
Why the Comparison Group Is Built This Way
Randomization was not possible because the program serves everyone eligible in its area. Comparison mothers come from the same hospitals and time period but live outside the service area, so they share the hospital's breastfeeding support at birth and differ mainly in whether they received home visits. Matching on age group and parity balances two strong predictors of breastfeeding. Residual differences between areas, such as income or employer policies, cannot be fully removed, which is why the covariates in section 3 are adjusted for and why the results will be described as associations unless the design supports more.
Data Collection
Breastfeeding status will be collected by telephone interview at two and six months by staff who do not deliver home visits, using the same script for both groups. Mothers will be asked about the past [24] hours and since birth. Interviewers will not know a mother's group in advance where possible, though some mothers will mention their nurse. Contact attempts will follow a fixed schedule of up to [six] calls and one text message, so that effort to reach mothers is equal across groups.
5. Primary Analysis
A regression model with a log link estimating the risk ratio for enrollment, adjusted for the covariates in section 3, using robust standard errors, also known as modified Poisson regression, which gives risk ratios directly for a common binary outcome (Zou, 2004). Results will be reported as the adjusted risk ratio and risk difference with 95 percent confidence intervals. The model, the covariates and the outcome are fixed now, so the result cannot be shaped by choices made after seeing the data.
6. Assumption Checks
Covariate balance between groups will be described before modeling. Influential observations and model convergence will be checked. If the log-link model fails to converge, a logistic model with predicted probabilities will be used, and this fallback is stated here in advance.
7. Missing Data
About [18] percent of mothers are expected to be unreachable at six months. For the main result, missing outcomes will be filled in [20] times by multiple imputation, including all covariates and the outcome, and auxiliary variables such as breastfeeding status at two months. As a check, results restricted to mothers with complete data will also be shown. Multiple imputation assumes data are missing at random given the variables included; that assumption cannot be verified, and results will be interpreted with that limitation (Sterne et al., 2009).
8. Secondary and Exploratory Outcomes
Any breastfeeding at six months and duration of exclusive breastfeeding will be reported as secondary outcomes, labeled as such, without adjustment for multiple comparisons. Any further analyses suggested by the data will be labeled exploratory.
9. Prespecified Subgroups
One subgroup only: mothers who returned to work or school by [12] weeks, because the program includes workplace lactation support. An interaction term will test whether the effect differs; the subgroup result will be interpreted cautiously.
10. Deviations
Any change to this plan after data collection begins will be recorded with its date and reason and reported in the final evaluation.
11. Reporting
The final report will present the flow of mothers through enrollment and follow-up, baseline characteristics by group, the primary result with its interval, sensitivity analyses and secondary outcomes, in that order (Rosner, 2016).
References
Rosner, B. (2016). Fundamentals of biostatistics (8th ed.). Cengage Learning.
Sterne, J. A. C., White, I. R., Carlin, J. B., Spratt, M., Royston, P., Kenward, M. G., Wood, A. M., & Carpenter, J. R. (2009). Multiple imputation for missing data in epidemiological and clinical research: Potential and pitfalls. BMJ, 338, b2393. https://doi.org/10.1136/bmj.b2393
Victora, C. G., Bahl, R., Barros, A. J. D., França, G. V. A., Horton, S., Krasevec, J., Murch, S., Sankar, M. J., Walker, N., & Rollins, N. C. (2016). Breastfeeding in the 21st century: Epidemiology, mechanisms, and lifelong effect. The Lancet, 387(10017), 475-490. https://doi.org/10.1016/S0140-6736(15)01024-7
Zou, G. (2004). A modified Poisson regression approach to prospective studies with binary data. American Journal of Epidemiology, 159(7), 702-706. https://doi.org/10.1093/aje/kwh090
The Purdue Global PU525 Unit 9 instructions
Unit 9 in PU 525 typically asks students to write a statistical analysis plan before data analysis begins. Instructions usually require the research question, hypotheses, variables, the primary analysis, handling of missing data, secondary and subgroup analyses, power and reporting. Name one primary outcome to protect against searching many outcomes for a significant result. Fix variable coding and covariates in advance. Choose a model suited to the outcome and state a fallback if assumptions fail. Plan for missing data explicitly, with a sensitivity analysis. Label secondary analyses and limit subgroups to those with a reason. Describe how comparison groups are formed and how data are collected. Cite methods sources in APA style, and commit to recording any deviation from the plan.
How the PU 525 Unit 9 example is put together
A first margin note explains why decisions are made before the data arrive. It states the evaluation question and a single primary hypothesis, citing evidence on breastfeeding. Variables and covariates are fixed. Design and power follow. Sections on how the comparison group is built, flagged by a margin note, and on data collection precede the primary analysis, which names a modified Poisson model, and a highlighted line notes that the result cannot be shaped after the fact. Assumption checks, missing data with a margin note, secondary outcomes, one subgroup, a deviations policy and a reporting order complete the plan. Each section is numbered so reviewers can cite it when the evaluation is reported.
Where the points sit in the PU525 Unit 9 rubric
Analysis plans in PU 525 are judged on a single clear primary outcome, fixed variable definitions, an appropriate primary model, explicit assumption checks and fallbacks, a sound missing data strategy, restraint with subgroups, a power calculation and a commitment to report deviations. Naming one primary outcome earns early credit. Choosing a model that gives interpretable effects for a common outcome shows care. Missing data plans should state their assumptions. Subgroups need a stated rationale. Plans that describe how the comparison group is formed and how data collection effort is kept equal across groups tend to receive the highest marks, since those choices protect the result from bias.
PU525 Unit 9 help: mistakes that cost points
Analysis plans often lose points by listing several primary outcomes, by leaving covariates to be chosen later or by ignoring missing data. Others promise many subgroup analyses. Name one primary outcome. Fix variable coding and covariates. Choose and justify the primary model, with a fallback. Plan multiple imputation or another method for missing data and add a sensitivity analysis. Label secondary analyses and keep subgroups to one or two with reasons. Calculate power. Describe how groups are formed and how data are collected equally. Commit to recording and reporting any change to the plan. Date the plan and store it where reviewers can see it was written before the data were opened.
Get a PU525 Unit 9 example written to your instructions
Describe the study your PU525 Unit 9 plan covers, its outcome, groups and data source, then attach the prompt and rubric. The plan that comes back fixes every test, assumption and missing-data rule in advance, with a power calculation whose inputs are shown. Your first custom sample is free and arrives in 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 PU525 papers and MPH samples
- PU525 Unit 1: Skewed Average Drinking Survey Post
- PU525 Unit 2: Birth Data Description Exercise
- PU525 Unit 3: Public Health Probability Problem Set
- PU525 Unit 4: Choosing the Right Test Worksheet
- PU525 Unit 5: Survey and Rate Confidence Intervals
- PU525 Unit 6: Statistical Output Interpretation Memo
- PU525 Unit 7: Relative Risk Headline Reflection
- PU525 Unit 8: Dental Visit Logistic Regression
- PU525 Unit 10: Biostatistics Report
- PU520 Unit 9: Healthy User Confounding Memo
- PU508 Unit 9: Opinion Leader Diffusion Case Analysis
- PU501 Unit 9: Syringe Services Policy Scan
PU525 Unit 9 questions, answered
Where can I find a free PU525 Unit 9 sample paper?
The complete PU525 Unit 9 statistical analysis plan is here, for a composite county home-visiting program's breastfeeding evaluation.
Why name one primary outcome?
To avoid finding a significant result by chance among many outcomes.
What model does the plan use?
A modified Poisson regression giving adjusted risk ratios for a common binary outcome.
How is missing data handled?
By multiple imputation, with a complete-data analysis as a sensitivity check.
How long is the PU525 Unit 9 plan?
About six pages under numbered headings.