Holtzman, T. and Saperstein, A. (2026). Formatting the two-step gender measure: Experimental insights from the United States. Demographic Research 54(35): 1125-1158.doi:10.4054/DemRes.2026.54.35.

SUMMARY

This repository contains files for reproducing the analyses found in Holtzman and Saperstein (2026). For questions about this repository, please contact Tessa Holtzman at tessamh@stanford.edu.

If you use any code from this repository, please cite the paper “Formatting the Two-Step Gender Measure: Experimental Insights from the United States”

Software versions: 
Stata: StataNow/SE 19.5. 
R (Figure 1): R Version 4.4.2 (figure produced using R Studio)
Last run: 2/21/2026

RESPOSITORY ORGANIZATION

/do
Contains the .do and .rmd files to replicate the analyses. Each .do file includes a global macro at the top that defines the working directory. You must update this path before running the code. 

LGBT Sample Analysis: .do files 01-03 must be run in order. These files reproduce the main analyses for the paper. 

File							Description
---							---
01_create_LGBT_dataset.do		Creates LGBT sample dataset for the replication package based on the public data and the randomization 							variables dataset. 
02_clean_LGBT_data.do			Cleans data from TESS/NORC.
03_LGBT_main_analysis.do			Runs the main analyses for the LGBT sample. 
04_LGBT_fig_1.r				Replication code for Figure 1. 

Supplementary General Adult Sample Analysis: Files 05-08 also must be run in order. These files contain the confirmatory analyses run on the general US adult sample (from Prolific). The Prolific data are NOT included in this replication package because they are not public. Full replication of these files requires access to these data (see /data below), however, log files are included in the replication package. 

File							Description
---							---
05_create_genadult_dataset.do		Creates general adult dataset from the Prolific data.  
06_region_genadult_data.do		Infers respondent region based on zip code (Prolific data).  
07_clean_genadult_data.do		Cleans the Prolific data. 
08_genadult_main_analysis.do		Runs the confirmatory analyses for the US general adult sample.  

/do/log
Contains the most recent log files from the last full run of all .do files.

/data
Contains necessary data files for replicating LGBT results. 

File										Description
---										---
9089.123_Saperstein_FinalWtd_20230621.dta		The cleaned data file from NORC. Can also be downloaded from: 
										https://tessexperiments.org/study/saperstein1530. 
9089.123_Saperstein_randomizationorder.dta		Variables on randomization order for the NORC data (not included in the public release) but 										made public in this replication file.  
Prolific_metadata.txt						Text file describing Prolific metadata. 

Files Not Included 
We do not include the data files from Prolific (the general adult sample) in this replication package. For details on data cleaning and processing steps see the log for 05_create_genadult_dataset.do. For details on analysis see 	the log for 07_genadult_main_analysis.do. If you would like to request access to the Prolific data in a de-identified format, please contact: Aliya Saperstein (asaper@stanford.edu).