Abstract
Household screening is common when information about characteristics of household members is needed for selection of survey respondents. When key characteristics have a low prevalence, or are oversampled, this can result in a large number of sampled households screened, many of which have no persons selected. For in-person surveys this can be inefficient and costly, especially in an environment of declining response. A multimode design using a mail, push-to-web approach is an attractive alternative due to lower cost and high internet penetration. However, little is known about the comparable data quality properties between in-person and web modes. While in-person screening is considered a gold standard approach, respondents may fail to report household members and interviewers may unintentionally screen out reluctant respondents. Similarly, those self-responding sometimes fail to report unrelated household members or young children. In this study we compared in-person and web screening in the National Health and Nutrition Examination Survey. Households were randomly selected to complete a self-administered web screener and subsequently be screened by an interviewer during an in-person visit. We report on the comparability of household characteristics between modes to determine if web screening provides data equivalent to in-person screening. We examine time between the web and in-person screening to see if true change can account for differences. In the presence of conflicting data, we examine selection criteria based on the screening responses to see how inaccuracies affect selection status, or if inaccuracies or person omissions are systematically related to a specific mode. Approximately 93% (80/86) of households agreed on selection status between the web and in-person modes. Household composition matched fully for 84% (72/86) of households. These results indicate that web screening is a viable option enumerating households in population surveys.
Publisher
Public Library of Science (PLoS)