Joanna Diong¹, Hopin Lee², Darren Reed¹

1. School of Medical Sciences, Faculty of Medicine and Health, The University of Sydney
2. Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford

Diong J, Lee H, Reed D (2022) The effect of face-to-face versus online learning on student performance in anatomy: An observational study using a causal inference approach. Discover Education (in press).

Protocol registration on the Open Science Framework (OSF): https://osf.io/ws8mv

OSF project repository: https://osf.io/xhs83/

Stored in doc:

• notes.pdf: Project study notes on the thinking process behind development and refinement of the causal graph
• causal_graph.txt: DAGitty code to generate the causal graph

De-identified processed data of examination marks for undergraduate and postgraduate cohorts are available in data -> proc:

• bios1168_proc_.csv
• bios5090_proc_.csv

De-identified raw data of introductory histology marks for the postgraduate unit are available in data -> raw:

• BIOS5090_histology.csv

Python (v3.9) and R (v4.2.2) code files were written by Joanna Diong.

script: Main script to run analysis.

proc, plot, utils: Modules containing functions used to clean data and plot figures.

A reliable way to reproduce the analysis would be to run the code in an integrated development environment for Python (e.g. PyCharm).

Create a virtual environment and install dependencies. Assuming you are running off Python on an Anaconda distribution or similar, using the Terminal (Mac or Linux, or PyCharm Terminal),

python -m venv env

Next, activate the virtual environment.

For Mac or Linux,

source env/bin/activate

For Windows,

.\env\Scripts\activate

Then, install dependencies,

pip install -r requirements.txt

Run script.py.

script: Main script to run analysis to obtain E-values.

Run script.R.

Python generated files, saved in data -> proc:

• bios1168_clean_.csv
• bios1168_result_.txt
• bios1168.pdf figure file
• bios5090_clean_.csv
• bios5090_result_.txt
• bios5090.pdf figure file

R generated file, saved in data -> proc:

• evalues.txt