# OptOut
# 6/25/2018

This respository holds the Python scripts that I wrote to create datasets for the OptOut project. This project is in collaboration with Professor Alexandra Killewald, in which we use sequence analysis (specifically optimal matching) and multinomial logit regression to identify common patterns of maternal employment of American mothers and key pre-birth characteristics associated with each pattern. 


File descriptions:

1sample_selection.ipynb - construct extended (18-year-seq) sample and 14-year-seq sample (the former is a subset of the latter)
2seq_construction.ipynb - construct 18-year-long monthly employment status sequences for extended sample
2seq_construction_seq14.ipynb - construct 14-year-long monthly employment status sequences for 14-year-seq sample
3IV_construction.ipynb - construct pre-1979 independent variables for extended sample
3IV_construction_seq14.ipynb - construct pre-1979 independent variables for 14-year-seq sample
3IV_construction_subsample.ipynb - construct independent variables for post-1979 sample, married sample, and work sample within extended sample
3IV_construction_subsample_seq14.ipynb - construct independent variables for post-1979 sample, married sample, and work sample within 14-year-seq sample
4childbirth_seq.ipynb - construct monthly sequences to show cumulative number of children in each month (for online supplement)
5age_seq.ipynb - construct monthly sequences between age 20-40 for mothers (in extended sample) and for childless women

Sequence analysis codes are in the R files: one for main output and one for online supplement. 

Regression analysis codes are in the Stata files. replication_package.do calls the other do files and runs all regression analysis. 


If you have any questions, please contact me at xiaolinzhuo at fas.harvard.edu