**10 years of mixed models: It's (well past) time to set your contrasts.**
*Laurel Brehm & Philip M. Alday*
This is documentation and data from a project looking at how psycholinguists (fail to) appropriately describe contrast coding in their papers.
The subfolders are the two main efforts to present the results to a hopefully broad audience
1. **`AMLaP 2020`**: Slides from a talk on this given at AMLaP 2020. The `Rmd` file is what was used to generate them; the HTML and PDF files are the rendered slides. Supporting graphics are in `img.zip`. [Recording of talk](https://mediaup.uni-potsdam.de/Play/Chapter/223).
2. **`Current Analysis and Preprint`**: Snapshot of current version of manuscript (see files history for prior versions). This reflects the latest submission to a journal, currently the Journal of Memory and Language. The `Rmd` file is both the analysis code and the manuscript. The PDF is the rendered output. `Contrasts_Papers_Deidentified.tab` is a tab-separated file containing the de-identified annotations for all papers that were used in the analysis, i.e. the data. The `id` field in this file was generated by using the script `deidentify.R`. Using hashes to de-identify papers makes it easy for an individual to check the annotation of a single paper, but makes it more difficult for a casual inspection of the annotations to be used to shame or otherwise "call out" an individual author or manuscript. Finally, `renv.lock` is for tracking the version of R and associated packages used in generating the PDF from the `Rmd` file and thus aid in reproducibility.
