From jmccrary@econ.Berkeley.EDU Wed Mar 13 11:56:28 2002 Date: Tue, 12 Mar 2002 16:20:48 -0800 (PST) From: Justin McCrary To: jroth@nber.org Subject: Mare-Winship vs. Regular Dear Ms. Roth, Thanks for your earlier response. I was noticing that the 1964 and 1965 record counts for the Mare-Winship files are low, relative to the 1966 Mare-Winship files. It turns out the 1964-1965 files are missing respondents in the 4th and 8th month. I thought if you did not know this, you would appreciate being notified. Here is the frequency of "month in sample" for 1964, 1965, 1966, and 1967. All processing was done in Unix via commands of the form gzip -dc da9307.yr67.gz | gawk '{print substr($0,16,1)}' | sort -k 1,1 | uniq -c Justin 1964 9048 1 9211 2 9088 3 8877 5 9137 6 9173 7 1965 9162 1 9196 2 9217 3 8779 5 8836 6 9302 7 1966 13770 1 13786 2 13792 3 13832 4 13559 5 13576 6 13893 7 13826 8 1967 8548 1 8724 2 8626 3 8500 4 8695 5 8227 6 8842 7 8512 8 2 9