This is a question for an econometrician or anybody else who has a good way to estimate a continuous value for income when respondents select a category, such as 0-5000, 5001 - 9999, 10000 - 25,000, etc. Note categories have a different width. We have many variables that could predict a score (education, gender, marital status, number of children, etc. and need the prediction to be a value within the range of the category the respondent selected. There is, of course a problem with the "Over 125,000" category and help on that would me a bonus!
Chandra Bhat wrote an article, Imputing a cntinuous income variable from grouped and missing income observations, 1994, Economics Letters 46: 311-319. Questions:
1. Has anybody implemented this as a user written command in Stata?
2. Are there alternatives to this implemented in Stata?
Years ago somebody did this in TSP and I know nothing about that program. The program isn't commented, so I can't make any sense out of it.
Thanks,
Alan Acock
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
