Web Survey Bibliography
Survey data collection from volunteer Web panels is growing in popularity. While this practice has obvious advantages in terms of cost, flexibility, and speed, estimates from such data for characteristics of the general population may be seriously biased due to low web penetration in the general population, nonprobability samples, and high nonresponse rates. The propensity score adjustment was introduced as a possible solution to the selection bias in observational studies. Its modification, propensity score weighting using a reference survey, has been suggested as a remedy of problems in volunteer panel web surveys. We can derive a propensity score weighted estimator that is less biased, if not unbiased, of the parameter of interest under certain model assumptions. However, there is no evident solution for its variance estimation. In the present study, simple variance estimators are compared: (1) the estimator applicable under poststratification, (2) a variation of (1) available in SUDAAN, (3) a model-based estimator, and (4) the jackknife estimator. The study focuses on the performance of the estimators in a simulation designed from a real web survey.
JSM (abstract)
Web survey bibliography - Isaksson, A. (3)
- Simple Approaches to Estimating the Variance of the Propensity Score Weighted Estimator Applied on Volunteer...; 2005; Isaksson, A., Lee, S., de Rouvray, C.
- Simple Approaches to Estimating the Variance of the Propensity Score Weighted Estimator Applied on Volunteer...; 2005; Isaksson, A., Lee, S.
- On the Variability of Estimates Based on Propensity-score-weighted Data from Web Panels; 2004; Forsman, G., Danielsson, S., Isaksson, A.