When a DTA (Department of Transitional Assistance) worker assesses whether a family's (T)AFDC (Temporary Aid to Families with Dependent Children) case will be closed, s/he decides which one of 67 different codes best describes the reason cash benefits for the household will be stopped. To carry out the analyses, we sorted all of the 67 codes into clusters of codes that logically grouped together: Cluster I, Increased Income; Cluster H, Sanctions; Cluster III, Eligible Persons Moved; Cluster IV, Fraud; Cluster V, Client Request; Cluster VI, No Longer Eligible; Cluster VII, Other or Multiple Meanings. The Appendix displays a description of the case closing codes in each cluster that provided a basis for our analyses. We used SPSS (Statistical Package for the Social Sciences) to calculate the trends in the number and percentages of case closings from October, 1995 to August, 1997 for each of the clusters mentioned above, and from October, 1993 to August, 1997 for the category of Earned Income only (a sub-set of Cluster I, Increased Income). We used four time periods for our analysis of earned income case closings, reflecting the progression of events surrounding the passage and full implementation of Massachusetts' welfare reform measures. These time periods are: October 1993 to September 1994, former AFDC phase; October 1994 to September 1995, legislative deliberation and passage phase; October 1995 to September 1996, initial implementation phase; and October 1996 to August 1997, the most recent implementation phase.