CAPL-2 Colombia: Reproducible analysis of inter-rater reliability
Description
Reproducible code and analytical documentation for the evaluation of inter-rater reliability of the Motivation and Confidence and Knowledge and Understanding questionnaires of the CAPL-2 Colombia version. The analyses were conducted in a paired sample of 322 Colombian schoolchildren and include exact agreement, quadratic weighted kappa, Cohen’s kappa, ICC(2,1), standard error of measurement (SEM), minimal detectable change at 95% confidence (MDC95), Bland–Altman analysis, and assessment of proportional bias.
Files
Steps to reproduce
Download the R Markdown file and the analytical dataset and place both files in the same working directory. Open CAPL2_Colombia_valores_referencia_reproducible_v2.Rmd in RStudio and install the required R packages if necessary: readxl, dplyr, tidyr, tibble, purrr, gamlss, gamlss.dist, capl, and knitr. Run the document using Knit → Knit to HTML. The workflow reads the analytical dataset, verifies the expected variables and analytical sample sizes, fits the definitive sex-specific GAMLSS models, evaluates model fit using GAIC, SBC, normalized quantile residuals and worm plots, and generates reference percentiles from P5 to P95. It also reproduces the correspondence between official CAPL-2 categories and the Colombian percentile distributions, as well as empirical percentiles for PACER, plank, CAMSA and average daily steps. The final section of the R Markdown document reports sessionInfo() to document the software environment used for reproducibility.