Thank you so much for answering. I was able to run it and get my results. However, I have a two of questions: 1. Is there a way to ask amelia to truncate the values, such that variables such as population do not take negative values in the imputations? 2. I think Prof. Scheve had a MI procedure to average the regression results from the Amelia data bases into a single output, however, the link on his page for MI procedures (http://pantheon.yale.edu/~ks298/software.html) is not working. Has that been moved to somewhere else? Thanks a lot again Jose Merino