Disability estimates produced using the variable RDIS_ALT from 2019 SIPP are not directly comparable with those from the 2014 SIPP Panel.

In the 2014 SIPP Panel, RDIS_ALT indicated whether an individual had a value of 1 for one or more of the six core disability items (ESEEING, EHEARING, ECOGNIT, EAMBULAT, ESELFCARE, EERRANDS) OR had a value of 1 for at least one of three child disability variables (EDDELAY, EPLAYDIF, or ESKOOLWK). It should be noted that this is different from how RDIS_ALT is described in the 2014 SIPP Users’ Guide and the 2014 Metadata, which mistakenly stated that two other variables, EFINDJOB and EJOBCANT, were also used to construct RDIS_ALT. For more information, see this 2014 Panel Wave 1 User Note.

Beginning in 2018 SIPP, the construction of RDIS_ALT changed. Namely, in addition to using data from the six core disability items (ESEEING, EHEARING, ECOGNIT, EAMBULAT, ESELFCARE, EERRANDS) and from the three child disability variables (EDDELAY, EPLAYDIF, or ESCHOOLWK), data from three work-limiting disability variables (EFINDJOB, EJOBCANT, and EDISABL) were considered when constructing RDIS_ALT. That is, RDIS_ALT=1 if an individual had difficulty with at least one of six daily tasks (ESEEING, EHEARING, ECOGNIT, EAMBULAT, ESLEFCARE, EERRANDS) OR if the individual had a value of 1 for at least one of the child or work-limiting disability variables mentioned above.

Since research suggests that estimates of disability prevalence generally increase as the number of items used to construct the disability measure increases, disability estimates produced using RDIS_ALT in 2019 SIPP are likely higher than corresponding estimates in the 2014 SIPP Panel. These higher estimates should not be interpreted as evidence of higher disability rates in 2019. If data users are interested in a 2014 disability recode variable comparable to the 2019 version of RDIS_ALT, a recode must be made in the 2014 data that includes “yes” responses to at least one of the three work disability questions (EDISABL, EFINDJOB, EJOBCANT). Notably, however, as discussed in the 2019 SIPP User Note on Estimates of Disability Prevalence, caution should be exercised in using SIPP 2019 data for population estimates of disability, including interpreting differences between 2014 and 2019 SIPP estimates as evidence of population-level change in disability prevalence.