S H A R E D F I L E S
This page contains downloadable STATA do, log, and data files to replicate the analyses found in:
Schwartz, Christine R. and Robert D. Mare. 2005. "Trends in Educational Assortative Marriage From 1940 to 2003
Note that only the final data sets used in the analyses are included in this archive, although I have included all of the do and log files that generate them. The raw census data used in this project are available from the IPUMS website. The CPS data are available from UNICON and from NBER. For more information about the specific data sources used in this paper, see Appendix Table A1 in the paper and the files in the documentation section below.
Please see the documentation section below for information about (1) how to recreate the final data sets used in this paper, (2) how to replicate the tables and figures presented in the paper, and (3) how to replicate the supplementary analyses discussed in the paper, i.e., those analyses "available upon request." Supplementary analyses discussed in the text but not shown are included below.
STATA do and log files that generate data sets
Final data sets (prevailing marriages and newlyweds)
STATA do and log files that generate tables and figures
STATA do and log files that generate supplementary analyses
Results of supplementary analyses
Correction
Figure 5 and Table 4 in the paper are labeled incorrectly and are discussed incorrectly in the text. The numbers presented in this table and figure are odds rather than odds ratios. The correct title of Figure 5 is "Odds of Crossing an Educational Barrier, Relative to Homogamy, Among Prevailing Marriages..." rather than "Odds of Crossing an Educational Barrier, Relative to the Odds of Homogamy..." and the y-axis should be labeled "Odds of Crossing" rather than "Crossings Odds Ratios." The same correction applies to the title of Table 4. Similarly, the text on page 627 should read, "The crossings parameters are the log odds of marriage for couples in adjacent education categories relative to homogamy..." rather than, "The crossings parameters are the log odds of marriage for couples in adjacent education categories relative to the log odds of homogamy..." There are additional instances in the text where these quantities are referred to as if they were odds ratios rather than odds; in each case, they are odds.