The problem of attrition often arises in longitudinal studies when the researcher is unable to gather follow-up information for a sizeable percentage of the study participants at one or more of the planned follow-up stages. Since non-respondents frequently differ in important ways from respondents, the findings based solely on the respondents are not likely to be representative of the total participant population. After investigation of numerous attrition-adjustment procedures, a cross-tabulation approach was utilized, weighting respondents proportionately to non-respondents with the same basic characteristics. The Mountain-Plains Career Education and Employment Program--a residential, family-centered program for multiple-problem disadvantaged, rural, poverty-level families--is used to illustrate this general approach to data analysis. (Author/MV)