Hi,
I extracted data from the 2015 to 2023 CPS Longitudinal 1-Year apart data. In the extract there are many variables that are duplicates – e.g, (BPL_1, BPL_2; EDUC_1, EDUC_2). Sometimes the values are identical, but other times they are not. Why are there duplicates?
Thanks,
Sam
The CPS longitudinal data files provided by IPUMS CPS include data from multiple observations of the same individuals, at different points in time; this is what makes them longitudinal data files. The variable EDUC_1 reports the individual’s value of the variable EDUC in time period 1, and the variable EDUC_2 reports the individual’s value of EDUC in time period 2. Each individual’s first and second observation of each variable are included on a single person record, which allows users to easily analyze changes in variables over time. There are some instances in which we expect the value of a variable to change between time period 1 and time period 2 for a non-negligible share of respondents. For example, many people increase their educational attainment over time by completing additional years of education or earning degrees.
In some cases, you may see unexpected changes to variables over time. For example, you may see an individual’s educational attainment decrease between time period 1 and time period 2. Or you may see it increase by more than would be expected to occur in reality in the period of one year. In another example, you may see an individual’s race or ethnicity change over time. In some of these cases, it is possible that the longitudinal data files are linking different individuals over time. This can happen because of the CPS’s sampling method; entire dwellings, rather than individuals or families, are sampled for inclusion in the CPS. Whoever lives in the sampled dwellings during the CPS observation period is surveyed. When individuals or entire families or households move dwellings during the CPS observation period, they are not followed. The new residents of the sampled dwelling take their place and take on their values of CPS identifiers CPSID (for households) and CPSIDP (for individuals). The longitudinal data files make links based on CPSIDP. IPUMS CPS has created a validated longitudinal person identifier, CPSIDV, which links the same individuals over time. In instances where changes in race, sex, or age over time suggest that CSPIDP is not identifying the same individual over time, CSPIDV does not link the individuals.
Thank you very much! This is incredibly helpful.
Sam