Case ID 0357 should be dropped from the 2004prepost study. The interview was conducted with input from both the selected respondent and someone other than the selected respondent, without any way to identify the source of individual responses.

The total N for the 2000prepost study is now 1,211 cases.