Siddhartha Chib and Liana Jacobi, "Bayesian Fuzzy Regression Discontinuity Analysis and Returns to Schooling", Journal of Applied Econometrics, Vol. 31, No. 6, 2016, pp. 1026-1047. Data for analysis were taken from UK general household surveys 1986 to 1995. These data are available from the UK government on the website http://data.gov.uk/ Access to data requires an online application for data use at http://data.gov.uk/data-request. See also http://data.gov.uk/dataset/general_household_survey for information on survey years available. Variables used in analysis are created in the Stata do file "sj-program.do" which is zipped in the file sj-program.zip. Unix/Linux users should use "unzip -a". Creation of earnings/wage (related) variables was based on Devereux and Hart (2010) lnearn log of gross weekly earnings lnwage log of hourly wage x indicator with value "1" if leaves school at age 15 ("0" if leaves school at age 14) age age in years married indicator for being married male indicator for male (analysis based on male sample) year survey year workhrs hours worked Variables related on RDD design and sample creation based on window around policy change dobyear birth year dobmonth birth month zstarm distance of birth from policy change measured in months (in paper define as "z") z indicator for turning 14 under new policy change, 1 if zstarm>=0 (tau=0), in paper defined as z>tau) File created by: Liana Jacobi, Department of Economics, University of Melbourne