This repository contains the analysis code for a study examining whether affirmative communication reduces absolute ideological bias expression compared with debate and no-treatment conditions.
Participants evaluated evidence related to migration and climate change. Absolute ideological bias expression was measured using the absolute signal detection theory bias parameter |b′|.
The study was preregistered on AsPredicted:
AsPredicted preregistration by4zm8
The present analysis focuses on the affirmative, debate, and no-treatment conditions. The counterfactual condition belongs to a separate research project and is not included.
scripts/analyse_p1.Rimports and prepares the data, applies the exclusion criteria, calculates the outcome variables, and conducts the statistical analyses.scripts/outputs_p1.Rgenerates the tables, figures, and analysis log.scripts/php/routing.phpcontains the survey-routing logic used to assign the experimental conditions.
The analysis was conducted in R 4.5.0.
Required R packages:
jsonlitereadrdplyremmeanssandwichlme4ggplot2flextableofficerwritexl
Run the scripts from the repository root in the following order:
Rscript scripts/analyse_p1.R
Rscript scripts/outputs_p1.R
The output script creates a timestamped directory under outputs/ containing the analysis log, CSV tables, figures, an Excel workbook with one worksheet per table, and a Word document with the generated tables.
Individual-level survey data are not included in this repository. Access to the data is subject to the data-protection and data-sharing procedures of the research project.
The analysis additionally requires the verified IBE item key and the final exclusion list. These project-specific files are not publicly distributed.
The code in this repository is available under the MIT License.