|Balanced and well-structered response set|
Imbi Traat, University of Tartu, Estonia
Monday, March 10, 2014
4:00pm, room 223 Weber Building
Nonresponse is an inevitable phenomenon in nowadays sample surveys. The resulting response set may differ far too much from the entire sample, and therefore, lead us to the biased conclusions about the population. Increasing response rate is not enough for the quality of the response set - its structure is important. In the presentation, results from the collaboration work with C.-E. Särndal are presented. He has initiated studies about monitoring and directing data collection procedure towards more balanced final response set (The 2010 Morris Hansen Lecture. Dealing with Survey Nonresponse in data Collection, in Estimation. Journal of Official Statistics, 2011, 27, 1-21). In the lecture, I will speak about how to measure balance/imbalance of the response set. What are further challenges here? Does balancing have effect on the estimation precision? Analytical as well as empirical results will be given.