Start: Nov. 1, 2009
End: Nov. 30, 2009
Quantitative Methods in the Social Sciences
In terms of Total Survey Error (TSE) - sampling error plus non-sampling errors – error due to the design and low quality of surveys is much greater than sampling error. However, assessing the contribution of the major sources of non-sampling error (frame, non-response, measurement, or processing) plus sampling error is a major challenge inherent in the concept of TSE. If TSE is not known then substantive analyses are open to serious methodological criticism. Therefore an understanding of TSE should be part of the basic knowledge of all survey researchers and analysts.
For comparative surveys, the number of TSE-sources is a multiple of the populations involved. With a 20-nation survey such as the European Social Survey (ESS) there are many instances where serious errors may damage the overall comparability of the survey, but there is no body of previous quality control evidence. There is therefore an urgent need for an exchange between methodologists and capacity building via research seminars and training. The seminars will bring together specialists from such diverse disciplines as statistics, linguistics or quality engineering and will include experienced as well as young researchers. The seminars will feed-back into surveys (ESS, SHARE and new comparative surveys). The research group comprises members of ESS, ISSP, official statistics, and national surveys and will thus promote collaboration with other projects, such as the ESS Infrastructure project. The research seminars will target pressing research needs and identify the topics which need to be covered in the training workshops in order to create a critical mass of young European researchers with expertise in survey design and quality.