Abstract:
|
Giving the increasing cost and resource constraints of traditional surveys, web surveys have been more frequently used as a flexible alternative in the field. However, past literature has well documented the general limitations of using web surveys. There also exists vibrant research on improving their uses from both applied and methodological perspectives. The National Center for Health Statistics has conducted a pilot study of assessing the utility of using web surveys to accurately measure and estimate important health outcomes, in conjunction with the National Health Interview Survey (NHIS). This talk will presents the background and some initial study results. Specifically, estimates for several key health variables were obtained from the NHIS and from a web panel, where NHIS estimates were calculated using standard survey techniques and web panel estimates were weighted using the sample weights provided by the vender but not further calibrated. Relative differences and t-statistics were used to compare estimates between sources overall and for key NHIS domains (e.g. race and ethnicity, sex).
|