This repository is the knowledge base for the MindTrails Managing Anxiety Study (R34) data. For more information about the study, see the Managing Anxiety page of the MindTrails wiki.
Previous Names
Goal
Main Outcomes Paper
Credibility Paper
Subsequent Data Cleaning
Contact
Until 5/18/2020, this repository was named R34-Data.
The initial goal of this repo was to describe data cleaning for the study and list projects that analyze the study's data. However, the current goal is to house the initial data cleaning and analysis scripts used for the main outcomes paper and an analysis script used for the credibility paper. Subsequent, centralized data cleaning done after these papers were published is described on the separate repo MT-Data-ManagingAnxietyStudy-Cleaning.
Lead: Julie Ji
See the Data Cleaning folder for drafts of three initial data cleaning scripts used in the main outcomes paper (Ji et al., 2021). However, the final cleaning script was lost, and the exact version of the raw dataset that was cleaned for that paper also seems lost. The scripts have issues, which will not be fixed now but are left open for reference.
R34_cleaning_script.R(author: likely Claudia Calicho-Mamani, but uploaded by Sonia Baee)Claudia - cleaning script.R(author: likely Sonia Baee's revision ofR34_cleaning_script.R)R34.ipynb(author: Sonia Baee)
These seem to be separate drafts (i.e., except for notes.csv, which is exported by
Claudia - cleaning script.R and imported by R34.ipynb, one script does not import files
exported from another). Given that R34.ipynb exports a data file whose name (FinalData-28Feb20.csv)
resembles that of the file (FinalData-28Feb20_v02.csv) imported by Script1_DataPrep.R (on the
main outcomes paper's OSF project), R34.ipynb seems to be the latest draft.
But the final version of R34.ipynb is unavailable (Sonia wrote on 11/22/2021 that the final
script was lost switching laptops).
R34_cleaning_script.R and R34.ipynb were uploaded to the present repo directly by Sonia,
whereas Claudia - cleaning script.R was obtained from Sonia's R34-Data
fork of the present repo on 10/26/2025 (the script was last updated 2/29/2020, as of commit
87e8df6 on Sonia's fork). The other cleaning scripts on Sonia's fork are identical to other scripts already
on the present repo (see Issue 4 on the
MT-Data-ManagingAnxietyStudy-Cleaning repo).
For more details about the initial cleaning done for the main outcomes paper, including more details on these scripts and the clean data exported from the initial cleaning pipeline, see the Initial Cleaning section of the README on the separate centralized data cleaning repo MT-Data-ManagingAnxietyStudy-Cleaning, which is introduced below.
See the Main Outcomes folder for drafts of analysis scripts used in the main outcomes paper. The final analysis scripts are on the main outcomes paper's OSF project.
Lead: Nicola Hohensee
See the Credibility folder for the analysis script used in the credibility paper (Hohensee et al., 2020).
Given that the final cleaning script and the raw data used to generate the clean data for the main outcomes paper are lost, and that the clean data on the main outcomes paper's OSF project has item-level data only at baseline (and only scale-level data over time), the separate MT-Data-ManagingAnxietyStudy-Cleaning repo seeks to obtain clean item-level data on key measures over time for the 807 participants in the main outcomes paper's intent-to-treat sample.
The repo does so by redacting two raw datasets and then comparing them to the clean datasets used in the main outcomes paper. Although neither raw dataset seems to be the exact version cleaned for the main outcomes paper, the repo's code is able to reproduce most of the scale-level data used in that paper from a combination of data drawn from these two raw datasets (and from the baseline item-level data used in the main outcomes paper).
After reproducing most of the scale-level data used in the main outcomes paper, the repo deviates from that paper in the cleaning of the demographics data (i.e., cleaning additional values for birth year and education; handling of blank values) and the OASIS data (i.e., recoding session values to be consecutive). The repo also outputs clean data for more measures (i.e., credibility) than are in the datasets used in that paper, and the new cleaning pipeline is reproducible.
For the raw datasets and cleaning scripts used by the new pipeline and the clean data it exports, see the MT-Data-ManagingAnxietyStudy-Cleaning repo. For differences between the initial data cleaning used in the main outcomes paper and the new pipeline, see the Differences section of the new pipeline's README.
If you would like to contribute to this project, contact Bethany Teachman at (bteachman@bvirginia.edu).