Contact: billy.mitchell@temple.edu
Data and scripts from the manuscript: Emotion regulation strategy use and forecasting in response to dynamic, multimodal stimuli
All of the data and products for the project are organized into four self-explanatory directories: 1) Analyses, 2) Data, 3) Products, and 4) Tasks.
The analyses directory is likely the most interesting to folks, I'd assume. The analysis scripts for each study are contained within separate directories. I'd also attached a numbered prefix to the scripts to make the order that they were ran in more clear. Using the here function from the here package, which will automatically identify filepaths for the scripts and data currently being used, should allow users to be able to download our repository and run the scripts as they are without input from the user, providing no changes to file structure are made. Each of the first three studies has two scripts associated with it: 1) a dataframe creation script (identifiable by the "_DFcreation" suffix), and 2) an analysis script (identifiable by the "_analysis" suffix). All scripts are heavily commented to explain what each step is accomplishing and, when applicable, how we are accomplishing it. The scripts have been optimized to generate minimal warnings, but in some cases, such as with stargazer()'s well-known NULL length message, warnings could not be avoided. In each case, an explanation for what the warning is telling us is provided, including a link to external documentation, and I explain why the warning is of minimal concern towards the goal of our analyses. If anything is unclear, please contact the author; I'd be happy to walk through any issues.
The data directory divides the data by study, so data used for the preliminary study is in a separate directory from the data we'd used for Study 01, Study 02, or Study 03. Each study directory itself contains directores for source data, raw data, and, in some cases, derivative data. The source data is the data pulled directly from qualtrics, completely unedited and unmodified. It may also include manually coded data, such as the responses that the raters gave for strategy classifications. Raw data is data that has been cleaned up and manipulated into a format that's ready for analysis, but usually not specialized or subsetted in anyway. Derviative data includes dataframes built for a very specific purpose (e.g., person-level dataframes, subsets, etc.). If you'd like to recreate our original analyses using out scripts, you likely will not need to make any modifications to the structure of the data directory or really to access it all, but obviously it is here for folks to explore and use as they wish.
The products directory contains two subdirectories: Manuscripts and Presentations. Manuscripts contains every version of the manuscript for this project that has been shared as a preprint or submitted for review. It also contains a directory, plots_tables, containing the raw .tiff, .png, .ai, or .xlsx file for all visualizations and tables present in the most recent version of the manuscript. Presentations includes all powerpoint slides or presentations associated with this project that I have ever produced. This has been added in the interest of transparency to highlight how this information has been disseminated, and also as a reference, as I imagine these could be useful likes a visual I generated or style of presenting.
The tasks directory is once again divided into four subdirectories; one for each study. The Preliminary Study, Study 01, and Study 02 directories all contain the actual questionnaires participants responded to in the form of .QSF files exported from qualtrics. In order to view these, you would need to import them into qualtrics. These are likely only of interest if you'd like to replicate this study or use our questionnaires as a foundation for your own study. Study 03 contains the .HTML file hosted on Pavlovia to collect this data, as well as the jsPsych distribution that I had used. Please note that while I used the jsPsych.v7 tools to start, I did modify many of these functions quite significantly to get them to do what I needed them to do (things like showing a numeric representation of the slider that dynamically updated), so if you were interested in replicating or using our study design, I'd recommend looking at the distribution contained in this repository. I believe I have every modification highlighted with comments, but I'd also be happy to walk you through what we did if you have any questions or concerns.
It should be noted that I'd opted to format each script as an R Markdown file (.RMD) rather than a vanilla R script (.R). R Markdown is a notebook-style scripting format (similar to a Jupyter Notebook for Python) which allows the programmer greater freedom in how the functions they enter are presented and conveniently centralizes all of the output to the viewer window. It operates the same as a vanilla R script, but adds additional functionality making it ideal for communicating the analytic process to others. Users unfamiliar with R Markdown might be interested in this helpful .PDF summary of R Markdown and its functions hosted by the R Core Team. The author also has written a tutorial on how to use R Markdown that they will link to in the future, but which I will refrain from including at the moment due to the blinding that must occur for the review process.
It should also be noted that scripts for Study 01 and Study 03 utilize custom functions sourced from GitHub which will make datawrangling easier. Sourcing these scripts requires an active internet connection. In the event that you would like to run these scripts without an internet connection, you would need to download said functions and either add them to the script directly, or save them as an .R (not .RMD) file and replace the URL under source() with the filepath leading to the .R file containing the function.