A Statistical Timetable for the sub-2 hour Marathon

(To appear in MSSE, 2019)

Author: Simon D. Angus

Dept. of Economics, Monash University

Melbourne, Australia.

Requirements

You will need:

• MATLAB (tested with version r2018b); and
• Statistics Toolbox (provides fitnlm, predict, and fitgev).

To replicate the paper

Open a MATLAB prompt, navigate to the working folder, and to run for example, the male version of the code run:

main('wr_male.csv', [0.5 0.2 0.10 0.04 0.02 0.01], 1)

with arguments

• 'wr_male.csv': use the male CSV input file;
• [0.5 0.2 0.10 0.04 0.02 0.01], the ALPHA values to use for the run;
• 1: sets ISMALE=1.

To run the female version, one runs equivalently:

main('wr_female.csv', [0.5 0.2 0.10 0.04 0.02 0.01], 0)

When run, main will produce two gendergap files for later comparison with gender_gap. For instance, if ISMALE=1 the two files will be: out_expected_male.csv, and out_1in10_male.csv . After running both male and female variants, one can then conduct a long-run gender gap analysis as follows:

OUTNAME = 'gg_1in10.csv';
gender_gap('out_1in10_male.csv','out_1in10_female.csv',OUTNAME)

Help?

Please get in touch at simon.angus@monash.edu.