Data, R code, and reproducibility outputs for "Hotel Water-Saving Appeals and Guest Heterogeneity: Dispositional Reactance, Nonlinear Evaluations, and the Limits of Message–Belief Matching"

This repository contains the de-identified analysis dataset, English-language R code, variable dictionary, validation benchmarks, and complete computational outputs supporting the manuscript “Hotel Water-Saving Appeals and Guest Heterogeneity: Dispositional Reactance, Nonlinear Evaluations, and the Limits of Message–Belief Matching”. The study used a 4 × 2 between-subjects experiment with 814 tourists who had stayed in hotels in Andalusia. Participants were assigned to one of four water-saving message appeals (self-efficacy, response efficacy, vulnerability, or severity) presented either as a single poster or as multiple posters located at different water-use points. The released CSV is a data-minimized version of the definitive analytic dataset. It contains only variables required to reproduce the analyses reported in the article and associated robustness checks and omits direct identifiers, locations, open-text fields, cleaning flags, replacement indicators, and variables not required for this research object. The R code reconstructs derived measures from item-level data and reproduces the factorial models, HC3-robust associations, multiplicity adjustments, reactance moderation analyses, quadratic functional-form analyses, bootstrap turning points, generalized additive models, natural-spline comparisons, repeated cross-validation, repeated split-sample validation, ordinal robustness checks, response-style sensitivity analyses, message–belief matching tests, behavioral robustness analyses, model diagnostics, and the manuscript figure. Version 1.1.0 additionally provides formal validation of the reported U-shaped relationships. All 11 manuscript outcomes are tested using a Lind–Mehlum/Sasabuchi-type U-shape test with HC3 covariance and Holm multiplicity adjustment, together with Simonsohn’s Two-Lines procedure. All eight outcomes described in the manuscript as U-shaped satisfy both formal criteria, whereas perceived threat, anger, and adjusted hypothetical water-saving behavior satisfy neither. These tests strengthen the statistical characterization of functional form but do not identify the psychological mechanism underlying the curvature. The definitive analytic sample contains N = 814. The main analysis and the formal U-shape validation were successfully executed using R 4.5.1. Complete execution logs, session information, detailed tables, diagnostic plots, manuscript-ready outputs, and numerical validation targets are included in the deposit. This record is intended to support transparent computational reproduction of the analyses reported in the associated article.