What are paradata? Paradata are data collected about interviews and the survey process. Some examples of paradata are how long the interview took to complete, whether the person used records to answer any survey questions, and who responded to the interview. While paradata can be seen as just a part of doing a survey, they can also give us a deeper understanding of patterns in survey data.

This project shows one interesting way of using paradata and survey data from Wave 1 of the 2014 Survey of Income and Program Participation. This blog looks at whether the average number of days sick reported differs by if a person reports their own information or has another member of the household report information for them in a proxy interview.

Figure 1 shows that younger adults (ages 15-59) have 6.1 days sick on average when they report their own data, while proxy reports have only 3.7 days sick on average. However, at older ages (60 and older), people who report by proxy have substantially more sick days when compared with people who self-report. At ages 60 and older, proxy reports have 14.2 days sick on average while self-reports have only 7.4 days sick.

Figure 2 summarizes the regression-estimated difference in reported days sick by whether a person reports their own information or has their information reported by proxy

At first glance, it may appear that there is no relationship between proxy status and days sick reporting for the overall sample. However, digging deeper, we see age differences in how proxy status relates to days sick reporting. Proxy reports for younger adults (ages 15-59) have 1.2 fewer days sick on average, even after we adjust for the probability that a person will have a proxy report. For older adults (ages 60 and older), the opposite is true; proxy reports have 4.7 additional days sick on average when compared with self-reports.

Why do proxy reports have fewer days sick at younger ages and more days sick at older ages? Future research will investigate this question. One possibility is that proxy respondents perceive younger adults as healthier and older adults as sicker. Another explanation is that people who report by proxy may be different from self-reports. At younger ages, healthy individuals may be less likely to be at home, while at older ages, individuals who report by proxy may be unable to self-report due to poor health.

This work will be presented at a poster session during the 2017 Annual Meeting of the Population Association of America. An online version of the poster can be found here. More information on the Survey of Income and Program Participation can be found at www.census.gov/sipp/.

-X-