![IconGems-Print](images/IconGems-Print.png)
252 – SPEED: Improving Survey Data Quality with Multiple Data Sources, Administrative Data, and Nonresponse Bias Control, Part 2
Proper Variance Estimation When Adjusting for Both Unknown Eligibility and Unit Nonresponse
Dhuly Chowdhury
RTI International
Phillip S. Kott
RTI International
It is common practice to reweight for sampled elements with unknown eligibility within mutually exclusive weighting cells and then reweight among the eligible sampled elements in each cell for unit nonresponse. Even when the weighting cells are identical to the design strata under stratified simple random sampling, it is incorrect to treat the respondent sample as if it were a single-phase sample for variance estimation purposes. We will show how variances should be estimated and then, using realistic data, compare the proposed approach to what is often done.