Cognitive Skills and Survey Nonresponse-Evidence from Two Longitudinal Studies in the United States

This study examines the relationship between general cognitive skills (i.e., those measured by intendedly general tests of ability or aptitude) and cooperation with requests for continued participation in longitudinal studies. Using the Wisconsin Longitudinal Study (WLS) of 1957 high school graduates, we find a consistent monotonic relationship in which those with higher measured skills are less likely to refuse to participate. Analyses indicate that this result cannot be explained by family background measures and is only slightly attenuated by the inclusion of measures of subsequent educational and occupational attainment and social participation. Meanwhile, in the 1998 wave of National Longitudinal Study of Youth in 1979 (NLSY79), the opposite relationship is observed: those with higher cognitive skills are somewhat more likely to refuse to participate. Further analyses reveal that this is due to some aspect of refusal conversion, as the pattern of initial refusals in the NLSY79 is similar to that of ultimate refusals in the WLS. One important difference between the refusal conversion efforts in the two studies is that NLSY79 has used significant financial incentives and WLS does not, which may explain the difference.