New and Emerging Methods Possible discussion points

Automatic editing Which errors are more important? random errors? systematic errors? Do we need new methodology for detecting random errors? based on edit rules? based on statistical outlier detection? What are the experiences with respect to detecting and correcting systematic errors at other agencies? What kinds of systematic errors can we distinguish?

Imputation Is imputation under edit restrictions an important new topic? Is imputation preserving known totals an important new topic? Is multiple imputation the way of the future? to smooth variability of single imputations? to estimate (co)variances?

Detection of influential errors/selective editing In this session only one paper on selective editing Is selective editing no longer “new and emerging”? Are we near the end of the theoretical development of selective editing? Should selective editing be based on complicated methods or do simple methods suffice?

Impact of editing Should we focus more on measuring the impact of (changing) the edit strategy? What are the experiences at other agencies in this respect? Can we improve on the Snowdon-X approach?

Data Is the focus shifting to periodic data/time series? Is the focus shifting to longitudinal data?

Software development How should we develop software? Open source Commercial of the shelf Find commercial vendor to develop software for us Develop software together in international projects Each agency develops its own software

Edit and imputation Is error localization more important than imputation? Is imputation more important than error localization?