yes, I think I have data subject to nonignorable non-response. But it isn't a longitudinal study and no follow-ups are made. I have collected literature, but some of it is with follow-ups or is about longitudinal studies. I am indeed interested in the slides, but I understand that I will receive them. I read parts of statistical analysis with missing data and multiple imputation for nonresponse in surveys. I also have two articles about pattern-mixture models from Little. But I wonder if pattern-mixture models are appropriate when there are a lot of different patterns.
Laura
