Analysis code for longitudinal CFA of decathlon performance across academic years in Japanese collegiate decathletes

<p dir="ltr"><b>This repository contains the R analysis code and supporting documentation used to reproduce the statistical analyses reported in the study “A three-factor representation of decathlon performance yielded an inadmissible longitudinal CFA solution across academic years in Japanese collegiate decathletes.”</b></p><p dir="ltr">The repository includes scripts for descriptive statistics and first-year record-availability analyses, the primary longitudinal confirmatory factor analysis (CFA), diagnostic CFA models, supplementary diagnostic tables, and sensitivity analyses.

It also includes a README file describing the analysis workflow and data-processing rules, and a codebook describing the variables required by the scripts.</p><p dir="ltr">The scripts are designed to be run with the de-identified analysis dataset associated with this study. Athlete names are not required by any of the analysis scripts; repeated observations are linked using study-specific athlete IDs.</p><p dir="ltr">The analysis workflow consists of the following scripts:</p><p dir="ltr"><code>00_run_all.R</code>: master script for running all analyses in sequence.<br><code>01_descriptive_missingness_logistic.R</code>: descriptive statistics, missingness analyses, Fisher’s exact tests, and logistic regression.<br><code>02_primary_longitudinal_CFA.R</code>: primary longitudinal configural CFA.<br><code>03_diagnostic_CFA.R</code>: cross-sectional, two-wave, and one-factor diagnostic CFA models.<br><code>04_supplementary_diagnostic_tables.R</code>: generation of Supplementary Tables S1–S6.<br><code>05_sensitivity_analyses.R</code>: complete-case, standardized-record, merged-factor, and official event-score sensitivity analyses.</p><p dir="ltr">See <code>README.md</code> for software requirements, data-cleaning rules, expected results, and instructions for reproducing the analyses.</p>