In the 2019 SIPP data, some respondents received duration of residence values (TRESDUR) that were one month shorter than they should have been. This error affected 2018 Panel Wave 2 records that met the following three conditions: (1) those with a residence spell start month of 1 (ERH_BMONTH), (2) those with not-in-universe values on year (TEHC_MVYR) and month moved to the current wave month 1 residence (EEHC_MVMO), and (3) those who remained in the same residence (ERESIDENCID) from month 12 of the 2018 SIPP reference period to month 1 of the 2019 SIPP reference period. While remaining in the same residence, these records should have durations of residence one month longer than those reported in the 2019 SIPP data (without exceeding the respondent’s age in months).

Because these respondents were returning from the 2018 SIPP, it is possible to calculate the correct duration for these respondents by merging 2018 SIPP and 2019 SIPP data together and identifying the time between the current year and month and the year and month of the last reported move. Within any given month in the 2019 SIPP data, residence duration in months should never exceed the respondent’s age in months.

Data users using or computing duration of residence estimates using the TRESDUR variable from the 2019 SIPP data should note these errors, and potential effects on comparisons with other SIPP data. The impact on TRESDUR values and the overall distribution is minimal, with less than 1% of person-month records estimated to be affected.