This web page contains the c program to calculate adjusted D statistics (adjusted to account for small numbers in some categories). The usual interpretation of the D-statistic is how many people would have to change categories (e.g., change occupations, census tracts, etc.) in order to have equal distributions (e.g., of men and women) across categories. The adjusted D-statistic is how many people would have to change categories so that the two distribution differered by no more than chance.
See the appendix in Cotter et al., 1997, "All Women Benefit...", American Sociological Review 62 (October):714-734. JSTOR or Adobe Acrobat .pdf file (approx. 2.5 meg)
This is based on earlier work by Cortese, Falk, and Cohen, 1976, "Further Considerations on the Methodological Analysis of Segregation Indices", American Sociological Review 41: 630-637. JSTOR
Use at your own risk: I offer no guarantees for this program nor any computer support for its use. There is some documentation embedded within the c program, but that's all the help you will get.
The following files are available:
|dstat.c||The c program.|
|cps.dat||Input data (CPS occupational distributions)|
|dstats.dat||The statistics output by the program.|
|dstats.lst||The output print file.|
|Last updated May 1, 2003||
comments to: Reeve Vanneman. email@example.com