Creating Survey Weights


Bio: I am a survey statistician for the UK Household Longitudinal Study. My current research focuses on improving quality and efficiency of survey data, specifically within three broad themes. The first theme is related to motivation and its role in improving quality of survey answers. The second theme is concerned with tackling nonresponse through fieldwork, including through motivation and adaptive design. And the third investigates improvements of statistical estimation and correction for nonresponse in complex sample design situations, including in longitudinal and cross-sectional surveys. My wider research interests also include work on satisficing, real-world eye-tracking in survey context, nonresponse bias, sample design, and social desirability.

This workshop will introduce survey weighting and will be especially useful for those who plan to conduct their own survey or want to understand better how survey weights work. Both; design weights (those that adjust for unequal selection probabilities) and nonresponse weights (those that adjust for nonresponse) will be covered. Theoretical part will use simple examples and simple words to explain what weights are; why they are important and how they are created. A practical part will let participants create their own weights. Access to Stata and knowledge of a logistic regression is recommended but is not a requirement for participation in the workshop.