Jeremy Edwards, "Did Protestantism Promote Prosperity via Higher Human Capital? Replicating the Becker-Woessmann (2009) Results", Journal of Applied Econometrics, Vol. 36, No. 6, 2021, pp. 853-858. All ASCII files are zipped in the file je-files.zip. They are in DOS format. Unix/Linux users should use "unzip -a". The binary file BW_data.dta is zipped in the file je-dta.zip. This paper mainly uses the dataset employed by Becker and Woessmann in their article "Was Weber Wrong? A Human Capital Theory of Protestant Economic History", Quarterly Journal of Economics, 2009, 124 (2): 531-596) is posted at the ifo Prussian Economic History Database (https://www.ifo.de/en/iPEHD). Appendix 1 of the Becker-Woessmann article describes the sources of their data. I use a subset of the Becker-Woessmann data, supplemented with some additional variables. My data are in the file BW_data.csv and also in the Stata data file BW_data.dta. These files contain 452 observations and 29 variables. The units of observation are the Prussian Kreise (counties) as of 1871, and these are the rows in the data files. The columns are the variables, which are defined as follows. Variable Names and Descriptions: kreiskey1871 Number identifying county county1871 County name rbkey Number identifying Prussian district in which county located kmwittenberg Distance of county from Wittenberg in kilometres hhsize Average household size gpop Population growth 1867-71 in percent f_prot Percentage of population Protestant f_jew Percentage of population Jewish f_rw Percentage of population literate f_miss Percentage of population missing education information f_young Percentage of population aged below 10 f_fem Percentage of population female f_ortsgeb Percentage of population born in county f_pruss Percentage of population of Prussian origin f_blind Percentage of population blind f_deaf Percentage of population deaf-mute f_dumb Percentage of population insane lnpop Logarithm of county population inctaxpc Income tax revenue in 1877 per capita (There are only 426 observations for this variable: see Becker and Woessmann (2009), p. 586) lnyteacher Logarithm of average annual income of male elementary school teachers in 1886 county Number identifying county in a particular Prussian district (used for panel-data estimation, see below) zupnew_1741 First component of linear spline in year of becoming Prussian zupnew_1814 Second component of linear spline in year of becoming Prussian zupnew_rest Third component of linear spline in year of becoming Prussian lat_proper Latitude in radians lon_proper Longitude in radians kmmainz Distance to Mainz in kilometres (calculated by Haversine formula for great circle distance using lat_proper, lon_proper, and latitude and longitude of Mainz) Bairoch Dummy variable taking value 1 if county contains cities in Bairoch et al. data set and 0 otherwise nonagric Share of labour force in nonagricultural occupations in 1882 To replicate the tables in the main text and the Appendices, the following Stata user-written commands must be installed: ivreg2, moransi, mundlak, weakiv, and xtivreg2. Each of these commands can be installed from within Stata by typing (for example) "ssc install ivreg2". The results in Tables 1 and 2 of the main text can be replicated by running the Stata do files Table 1.do and Table 2.do using the Stata data file BW_data.dta. Similarly the results in Tables A1-A6 in the Appendix can be replicated by running the Stata do files Table A1.do-Table A6.do using the Stata data file BW_data.dta. The regressions in these tables which allow for district fixed effects are estimated using panel data techniques, and hence the data in BW_data.dta are sorted by rbkey (the panel variable) and county (the "time" variable).