Transforming Pre–Post Binary Data into Analytical Evidence
Aim of the Project
This project demonstrates how binary pre–post assessment data can be transformed into reproducible and interpretable statistical evidence using R. The goal is to illustrate a structured workflow for analyzing training outcomes measured through correct/incorrect questionnaire responses.
Methods Used
The analysis follows a transparent and reproducible pipeline:
Construction of total and domain-level scores from binary items
Reliability assessment using Cronbach’s alpha
Distributional evaluation through graphical inspection
Normality testing using the Shapiro–Wilk test
Within-subject comparison using the Wilcoxon signed-rank test
Estimation of effect size to quantify magnitude of change
Why the Wilcoxon Signed-Rank Test?
Because the outcome variables are derived from summed binary items, the resulting scores are bounded and non-normally distributed. The Wilcoxon signed-rank test provides an appropriate non-parametric alternative for evaluating paired differences without assuming normality.
Key Findings
Significant improvement observed between pre- and post-training scores (p < 0.001)
Median gain of approximately 2.5 points on the total scale
Large overall effect size (r ≈ 0.80) indicating substantial practical impact
Strongest improvements observed in Statistical Reasoning, followed by Reporting & Publishing
Reproducible Report
View the full analysis on RPubs:
https://rpubs.com/Samadou_Tchakondo/prepost-binary-data-analysis
Tools
R
R Markdown
psych package
Non-parametric statistical methods