This repo contains code and data accompanying the paper "Psychological heuristics for portfolio decisions" by myself, Simon Algorta, Dieudonne Kantu, Konstantinos Katsikopoulos and Ozgur Simsek. The paper is currently under review and a citation will be added once available.
Code for the simulation experiment is in the simulation folder. Code for the behavioural lab experiment is in the behavioural_lab folder.
To reproduce our results from the paper,
- Run run_experiment.R,
- Run process_simulation_output.R,
- Run plot_simstudy_figures.R.
The first few lines of run_experiment.R create the few 100 datasets used in the simulations. The rest of the file, which re-runs the experiment, takes a long time to run!
The folders data and results are empty in the repo but are included as output from the code above will be saved into these directories.
To run a single or small set of simulations, see illustrative_example.R
All functions used to run the portfolio selection problem are contained in simulation/fns, with the exception of experiment_master_fns.R.
The two tasks are contained in behavioural_lab/tasks/task[12]. To view the tasks as they would have been viewed by study participants, open the ui.R file in those folders and run the app. The task instructions are in Instructions1.Rmd. Note the "submit" button will not work in the app as this linked to a server that is no longer in use. A instructional video used to walk participants through the task is available at https://youtu.be/5kzl6QcoOCo.
To reproduce our results from the paper,
- Run make_simulated_data.R and make_selection_data.R,
- Run infer_heuristics_task_selections.R and infer_heuristics_simulated_selections.R
- Run analysis_and_plots_for_paper.R
The folders output and results are empty in the repo but are included as output from the code above will be saved into these directories.