Improving Prelicensure Healthcare Students' Confidence in Managing Aggression Through Interprofessional Simulation

Statistical analysis supporting a quasi-experimental pre/post study of prelicensure nursing, physician assistant and respiratory therapy students at West Chester University. The repository holds the SAS programs, the executed SAS output, the analysis write-up and the manuscript submitted to the Journal of Interprofessional Education & Practice. By John Fisher, statistical analyst and listed author.

🌐 Live site: https://johnfisher-ai.github.io/fisher-wpv-aggression-study-sas/

---

The study

Ten Likert items measuring confidence in managing patient aggression were administered to students before and after an interprofessional high-fidelity simulation. The same ten items were asked both times, so each student contributes a matched pre/post pair. The primary analysis is a set of paired t-tests on those ten items. Secondary analyses describe the sample and model comfort against demographic and clinical-experience covariates.

Participants are identified only by a self-created ParticipantID. The merged analysis dataset holds 57 records, and the paired tests run on 54 matched pairs (53 on item 4).

Corresponding author: Julie Kurkowski, PhD, RN, CEN.

---

What is in this repository

The paper

File	What it is
Submission PDF to Journal of Interprofessional Education.pdf	The manuscript as submitted to the journal. 32 pages.

The statistical analysis write-up

File	What it is
Pre and Post Survey Statistical Analysis.docx	Narrative write-up of the pre and post survey analysis, with the figures. July 2, 2025.

The SAS programs

Source of truth for every number reported. Each program has a formatted PDF listing alongside it.

Program	What it does	Listing
Survey202506.sas	PRE survey. Imports the Qualtrics export, builds the Exp1–Exp3 clinical-experience dummies, runs descriptives, distributions and frequencies, then a sequence of PROC REG models predicting comfort, dropping collinear predictors using COLLIN.	PDF
SurveyPost202506.sas	POST survey. Same shape as the PRE program.	PDF
SurveyPairedTTest202506.sas	Merges PRE and POST on ParticipantID, keeps only the ten items, runs the paired comparisons. This is the program the paper's headline result comes from.	PDF

The SAS output

Results as executed on June 27, 2025.

File	What it is
202506_Survey_Paired_TTEST_Results.pdf	Paired t-test results for all ten items. 25 pages.
202506_Survey_Pre_Results.pdf	PRE survey descriptives, frequencies and regression sequence. 86 pages.
202506_Survey_Post_Results.pdf	POST survey descriptives, frequencies and regression sequence. 95 pages.

---

Live site

Four pages. All are live.

Page	What it will hold	Status
Overview	Plain-language summary of the study, the design, the sample and the headline result, with links to everything else.	✅ Live
The paper	The manuscript as submitted, in full, with tables, figures, and references.	✅ Live
The analysis	Full statistical write-up: descriptives, paired comparisons with effect sizes, assumption checks, multiplicity control, and regression diagnostics.	✅ Live
The code	The three SAS programs, syntax highlighted, with a Colab notebook that reproduces the analysis in Python.	✅ Live

The site is built by .github/workflows/pages.yml, which publishes the public/ folder and nothing else. Files under output/, docs/ and sas/ are reachable through this repository but are never served from the site.

---

Project structure

.
├── public/                             # The live site. This folder, and only this folder,
│   ├── index.html                      #   is published to GitHub Pages.
│   ├── analysis.html
│   ├── code.html
│   ├── paper.html
│   └── assets/
├── notebooks/
│   └── wpv_aggression_analysis.ipynb   # Python equivalent, runs in Colab on synthetic data
├── docs/
│   ├── Submission PDF to Journal of Interprofessional Education.pdf
│   └── Pre and Post Survey Statistical Analysis.docx
├── sas/
│   ├── Survey202506.sas                # PRE survey
│   ├── SurveyPost202506.sas            # POST survey
│   ├── SurveyPairedTTest202506.sas     # Paired comparisons, the headline result
│   └── *_SasProgram.pdf                # Formatted program listings
├── output/                             # SAS results, as executed
├── scripts/push_to_github.sh           # Commit and push
├── .github/workflows/pages.yml         # Builds and deploys public/ to Pages
└── CLAUDE.md                           # Working brief for this repository

Raw Qualtrics exports are not in this repository and never will be. They live outside it.

---

How the analysis runs

SAS is not installed on the authoring machine. The programs are written and edited here, then run on the West Chester University Apporto virtual desktop, and the results and the .log are saved back into output/. The proc import paths are Windows UNC paths into that desktop and will not resolve anywhere else, so the path is the first thing to change if a program is ever run elsewhere.

Anything that reaches the manuscript is cross-checked in Python (pandas, statsmodels, scipy) against the same data. That check has caught real errors.

---

Data protection

This is identifiable human-subjects research data.

• The raw Qualtrics exports stay outside this repository. They carry respondent IP addresses, geolocation and Qualtrics response keys.
• 202506_Survey_Pre_Results.pdf and 202506_Survey_Post_Results.pdf were produced before the programs were corrected to drop the Qualtrics identifier columns, so their first page printed the Qualtrics metadata. Respondent IP addresses and Qualtrics response IDs were redacted from both files before this repository was published. The text was removed, not covered over, and both files were verified page by page.
• The programs now keep only ParticipantID, the analysis variables and the covariates in use, and drop everything else before any proc print, proc contents or ods pdf. All new output follows that pattern.
• Participants are identified only by a self-created ParticipantID.

See CLAUDE.md section 1 for the full rule and the decision record.

---

Authors

John Fisher, statistical analyst and listed author. Julie Kurkowski, PhD, RN, CEN, corresponding author.

© 2026 John Fisher

---

License

The code in this repository is released under the MIT License: the SAS programs, the Python notebook, and everything under tools/ and public/.

Three things in this repository are not covered by that licence, and MIT should not be read as granting rights over them:

• docs/Submission PDF to Journal of Interprofessional Education.pdf. The manuscript is a co-authored work. It is included here with the authors' knowledge, but one author cannot license it on behalf of the others, and the journal's agreement governs its reuse.
• docs/Pre and Post Survey Statistical Analysis.docx and the executed output under output/. These are the analyst's own work and are shared for transparency, not offered for redistribution.
• The raw survey data, which is not in this repository and never will be. See Data protection.

If you want to reuse the analysis approach, the code is the part to take.