/workplace_gender_bias

Primary LanguageJupyter Notebook

Introduction

This is the beginnings of the gender simulation in the workplace code

Running the model

  • The model runs using python 2.7 and requires only standard libraries, pandas and numpy

  • To run the model from the command line, you need to provide several arguments:

    python model.py [path to experiment file] [path to default params file] [path to desired output folder] [n_replications_per_condition] [n_cores]
    • path to experiment file - Path to a file describing experimental parameters. There is an example in experiment.yaml
    • path to default params file - Path to the default parameters for the model. An example is default_params.yaml
    • path to desired output folder - Where the output is located. See R/plot_results.R for an example of how to plot the results
    • n_replications_per_condition - How many replications per simulation?
    • n_cores - how many cores you want to use to run the simulation (1 if you're not sure)

  • To replicate the results in Figure 1,2, please run following codes to generate results for simulations without any empirically-validated biases (1), with all of these (2) , or with each individually (3-8):

    1. python model.py ./parameters/nobias.yaml ./parameters/default_params_fig12.yaml ./results/NoBias 100 1
    2. python model.py ./parameters/allbias.yaml ./parameters/default_params_fig12.yaml ./results/AllBias 100 1
    3. python model.py ./parameters/rewardless.yaml ./parameters/default_params_fig12.yaml ./results/RewardLess 100 1
    4. python model.py ./parameters/penaltymore.yaml ./parameters/default_params_fig12.yaml ./results/PenaltyMore 100 1
    5. python model.py ./parameters/mixedrewardless.yaml ./parameters/default_params_fig12.yaml ./results/MixedRewardLess 100 1
    6. python model.py ./parameters/mixedpenaltymore.yaml ./parameters/default_params_fig12.yaml ./results/MixedPenaltyMore 100 1
    7. python model.py ./parameters/complain.yaml ./parameters/default_params_fig12.yaml ./results/Complain 100 1
    8. python model.py ./parameters/stretch.yaml ./parameters/default_params_fig12.yaml ./results/Stretch 100 1

  • To replicate the results in Figure 3, please run the following code to generate results for simulations:

    • python model.py ./parameters/weight.yaml ./parameters/default_params_fig3.yaml ./results/weight 100 6

  • To replicate the results in Figure 4, please run the following code to generate results for simulations:

    • python model.py ./parameters/intervention.yaml ./parameters/default_params_fig4.yaml ./results/intervention 100 6

  • To visualize the results, please run plot_paper.ipynb to generate figures of simulations.