I am analyzing a data set - analysis of newspaper failure- over a 55 year time span and also looking at a time span of the last ten years. My goal is to analyze the two periods (split, but not in the conventional sense, the observation period) and to also get descriptive statistics on variables by record and also within the individual time periods of interest. It's for my dissertation. I've been looking everywhere and can't find the answers. Any help you have is so appreciated.

1. Does it make sense to respecify the origin to get the analysis for the shorter time span - the last ten years? (this is what I've done and it seems ok)

stset t_1, id(caseid) failure(outcome==1) time(t_0) origin (time 45) scale(1)

Second question is how do I obtain summary data by record and not by observation

2. If you do a simple sum or tab command you get data on a variable by record - not ideal and most of the variables aren't tvc's so of course sum data by record is necessary and sufficient.

Third question is related to the first:

3. How do I obtain summary statistics on data for a shorter time interval - say I want to get summary statistics for PM papers starting at origin 45 or in the last 10 years.

Any help with ANY OF THESE is appreciated. STATA commands generate even greater gratitude. I work in Stata 10 but many commands seem to be the same.

Thank you and PS my dissertation draft deadline is in two days. . .

Frankie