This supplement will address the following three aims:
1) Develop best practice methods for translating PSID?s telephone instrument to the web. In particular, we would address issues related to instructions/probes; rostering and calendaring; capturing financial amounts; and how to characterize (and minimize) item non-response. We would then create technical specifications and program the 2015 core instrument for web.
2) Test the web program in 2015 with two samples:
a) a sample of approximately 1000 families (who are not currently in PSID). This step would involve recruiting a sample from an external (i.e., non-PSID) internet panel. These panels are comprised of experienced web respondents and can be used to evaluate competing design features (randomly assigned to respondents) but not uptake rates; and,
b) a supplemental sample of approximately 500 highly cooperative respondents from the PSID. This step would involve recruiting a sample of PSID families who have a track record of needing little follow-up and who express a strong preference for completing their core interview on the web. Although not representative, we believe this group would be similar to those most likely to initially uptake the web mode.
3) Conduct an evaluation of operational issues, data quality and cost. Operational issues relate to protocols for maximizing response (and completion) rates. Using both the external internet panel and PSID samples, we will analyze data quality, including rates of missing data and breakoffs. Using the external panel we will also examine distributions of responses under different design parameters. Using the PSID sample we will also examine comparability with key estimates obtained from telephone in the prior (2013-2015) waves. Finally, we will refine cost estimates for integrating web into the next cycle under a variety of informed uptake assumptions.
National Science Foundation
01/01/2013 to 12/31/2019