My colleague and I have been working with a data set that we believe is an
unbalanced panel, but we are not certain. Let me describe it and then ask how I
would use xtreg to estimate a regression model using these data.
The cross section consists of golfers which we observe over the time period 1992-
2001. We observe some of the golfers for 1 year, some for 2 (which may or may not
be consecutive), some for 3 years, and so on. There are only 39 golfers that we
observe for the entire time period (10 years). Some of the golfers that we observe in
the early years leave and are replaced by new golfers. The golfers that leave are
generally those that fail to make the cut (minimum money earnings) or those that
choose to move to some other tour. The total number of golfers is 370 and the total
number of observations is 1718.
Based on the discussion in Wooldridge, I think we satisfy the attrition criterion, which
states that the decision to drop a golfer is related to the earnings variable (the
dependent variable in our analysis). That is, a golfer is eliminated when he fails to
make the earnings cut.
How should we handle this in Stata?
Don Alexander
--
Donald L. Alexander, Ph.D.
Department of Economics
Western Michigan University
1903 W. Michigan Avenue
Kalamazoo, MI 49008-5330
Voice: 269-387-5526
Fax: 269-387-5637
E-mail: [email protected]
Web Page: http://homepages.wmich.edu/~alexande/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
