Merging NBER and IPUMS ASEC data for 1962-1967 surveys

I am using IPUMS data for early years of the CPS (1962-1967), and have questions about several issues.

  1. I would like to merge information from NBER CPS data into IPUMS CPS for these years for variables that IPUMS does not have, specifically the number of children under 14 in each family/household. However, for 1962 and 1964 they do not have the same number of observations. For example, in 1964 the raw NBER data has 5 more observations (54,548 vs 54,543). 4 of these are blank and can be eliminated, but I don’t know why the fifth is there. For 1962, the raw data have 4 more observations(71,745 vs 74,741). I am wondering if it is possible to determine which observations IPUMS dropped?

  2. From multiple posts, I’ve seen that IPUMS CPS and the raw NBER data share a common sort order. However, even ignoring the handful of extra observations in the raw data, this is not the case for 1964. This post is all I could find on the issue, and I’m hoping that someone has made progress on the issue since then? Help merging NBER to IPUMS March CPS data for 1964 and 1967

As a followup to this question, I am wondering if IPUMS has a way to merge in other identifiers from the original raw data? Specifically, the family variables PSU Number and Schedule Number. Those two combined with a handful of other variables allows me to uniquely identify observations in 1964, which would then allow me to merge. However, these variables are not listed among those available fro 1964.

  1. I would ideally like to do this for all the years 1962-1967, but NBER does not post the raw data for the years 1963 and 1965. I’m wondering if IPUMS is able to share the original raw data for those two years, or if it is available from any other source?


Unfortunately, documentation on these early ASEC samples is considerably lacking. These datasets were not officially released by the US Census Bureau as public use files. Instead researchers from the University of Wisconsin preserved these data and we have made this data available via IPUMS. Therefore, regarding your questions: (1) There is not a good way to determine which observations are included in the NBER files and not in the IPUMS CPS files. (2) There still isn’t a great way to get around the issues with sequentially merging the 1964 IPUMS ASEC file to the NBER file. (3) We are currently unable to share the raw data for these years. We are always working on expanding the capabilities of IPUMS and addressing some of these challenges are certainly areas were we are considering how to improve.