Esteban M. Aucejo and Jonathan James, "Catching Up to Girls:
Understanding the Gender Imbalance in Educational Attainment Within
Race", Journal of Applied Econometrics, Vol. 34, No. 4, 2019,
pp. 502-525.

All files are in the zip file readme.aj.txt. This zip file contains
both ASCII and binary files. The ASCII files are all in DOS format. 

The programs required to reproduce the analysis are contained in the
following folders within readme.aj.txt:

-- nlsy97
-- progs
-- stata

(I) Creating the NLSY 1997 Data

The folder `nlsy97' contains the main set of programs that clean the
raw NLSY 1997 data.  The main program to run is a sas program called
`create data.sas'.  This creates the underlying data for the analysis,
which includes four .csv files and one .dta (STATA) file.  A
description of these output files follows:

-- meas.csv contains the measurements used for the factor model

-- panel.csv contains the history file of education decisions

-- pop.csv contains gender, race and populaiton weights for each
   individual in the sample

-- SumData.csv contatins summary data as of Age 24

-- RFdata,dta conatins the variables use in the reduced form analysis


(II) Reduced form analysis in STATA

The reduced form analysis is conducted in STATA.  The do file
`Tables_1_3_4.do' in the `stata' folder will create the summary
statistics and run the reduced form analysis.


(III) Programs for the main analysis

The main analysis is done in MATLAB.  There are three main steps to
run the analysis.  (a) create MATLAB data sets from the .csv files,
(b) estimate the factor model and take factor draws, and (c) create
the output tables.  The programs for all of these steps can be found
in the `progs' folder.

(a) Create MATLAB data sets

To create the data necessary to run the MATLAB code, go to /progs/data
and run the files `read data.m' and `read summary data.m'.  These
programs create the .mat files `Gender Gap Data.mat' and `summary data
age 24.mat'

(b) Estimate the model and take factor draws

The program `estimate_model.m' in the folder `/progs/full model'
estimates the factor model as well as 50 bootstrap samples.  The
estimates are stored in `/progs/results/factor model estimates'.

After the model has been estimated, run the program `factor_draws.m'
in the folder `progs/full model' to create simulated factor draws for
each of the bootstrap samples.  The factor draws are stored in the
folder `progs/results/simulated factors'.

NOTE: included in these files are only the simulated draws for the main
sample.  This program will have to be run to create the simulated
draws for the bootstrap samples.

Also in this same folder, the program `reduced form.m' creates the
results for the reduced form schooling model for the 50 bootstrap
samples

(c) Creating the output

To create the output for the tables in the paper, first run the file
`create_output.m' in the folder `progs/tables and figures', then run
the associated file corresponding to the desired table or figure in
the folder `progs/latex output'.