The Face-to-Face Recruited Internet Survey Platform (FFRISP) is an internet survey panel that attempted
to collect information that is representative of the general American population, while drawing on the
advantages of working with such a panel, namely, obtaining data more quickly, more accurately, and less
costly than using interviewer administered survey modes.
The goal of the research described in this presentation is to examine the accuracy of the FFRISP by
comparing survey estimates generated from it to those of the National Health and Nutrition examination
supplement (NHANES) and from supplement surveys of the Current Population Survey (CPS). Survey
data from the CPS and the NHANES have been used as benchmarks for behavioral data. Both surveys
are conducted through interviewer-administered surveys on very large samples of individuals, with high
response rate outcomes.
The paper starts by comparing distributions of questions about health, physical activities, and mood
disorders—-such as general self-perceived health, biking to work, or depression feelings—-across the
NHANES and FFRISP surveys. It then compares distributions of identically worded items about shopping
behavior across the CPS Food Security supplement and FFRISP surveys. These comparisons will be
made looking at full sample distributions, as well as at distributions of subsamples (e.g., respondents with
lower education vs. respondents with higher education). The paper will then compare whether
associations between variables are similar across these surveys. Similarities and differences between the
findings generated with FFRISP data and those generated with CPS and NHANES data will be
discussed.