Keep It Simple, Señor: The KISS Approach in Spanish Parliament

This package provides the processed monthly data, Python code and reference results needed to reproduce the Bayesian analyses reported in the accompanying article on linguistic complexity in Spanish parliamentary discourse between 1993 and 2025. It includes eight monthly datasets, the model specification, fixed sampling settings and software dependencies. The code supports the 20 main model fits, with and without government membership controls, and six sensitivity analyses examining alternative priors and speech length thresholds. It also includes numerical reference results for posterior estimates, sampling diagnostics, conditional posterior predictive checks and PSIS-LOO model comparisons. The README explains the data structure, variable definitions and execution instructions. The package reproduces the Bayesian analyses from the processed monthly inputs; it does not reproduce the original data collection, text processing or descriptive figures. Original speech texts, individual speaker profiles and posterior samples are not included. This standalone reproduction package was derived from the researchers’ original analysis code with assistance from a large language model (LLM). The LLM assisted with adapting and organising the code into a compact package to facilitate reproducibility. The researchers reviewed the resulting code and checked its consistency with the original implementation, processed inputs and reference results.