|
Activity Number:
|
83
- Contributed Poster Presentations: Health Policy Statistics Section
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2020 : 10:00 AM to 2:00 PM
|
|
Sponsor:
|
Health Policy Statistics Section
|
|
Abstract #314043
|
|
|
Title:
|
Estimating Sampling Weights for Convenience Samples Using NHANES
|
|
Author(s):
|
Olivia Bernstein* and Dan Gillen
|
|
Companies:
|
University of California, Irvine and University of California, Irvine
|
|
Keywords:
|
Sampling bias;
NHANES;
Sampling weights
|
|
Abstract:
|
The Consent 2 Contact (C2C) registry at the University of California, Irvine collects demographic and medical data on potential trial participants. The C2C is a local registry and subjects self-select into the registry, thereby leading to potential sample bias. Enrollees tend to be highly educated and are likely to be non-Hispanic white relative to the general US population. Salazar et al (2019) recently used the Research Attitude Questionnaire (RAQ) to assess racial and ethnic differences in research willingness in the C2C but questions remain regarding the generalizability of this result. To address this we estimated sampling weights for C2C participants using data from the National Health and Nutrition Examination Survey (NHANES), a nationally representative survey of the United States that collects medical data. We created a combined dataset of C2C and NHANES subjects and used a random forest model to estimate the probability of being in C2C. We used the estimated inclusion probability to estimate Horvitz-Thompson sampling weights for C2C subjects. We then repeated the analysis by Salazar et al. while including the estimated sampling weights to account for biased sampling.
|
Authors who are presenting talks have a * after their name.
