ilies, two indicators of employment dislocation are included. The first measure is the male unemployment rate, which is used to represent unemployment. The male unemployment is defined as the percentage of male workers age 16 and over that are unemployed. Male unemployment should have a positive effect on female-headed families because unemployed males generally do not make good marriage partners. Along with increasing unemployment levels the changes in types of jobs available in urban areas has also led to underemployment. Underemployment is indicated by a measure of part-year male workers. That is, the percentage of male workers age 16 or older that were employed less that 26 weeks in the year preceding the census. This variable should also have a positive effect on the percentage of female-headed families.Return to Table of ContentsResults and DiscussionTo best understand the effects changes in the economy have had on family structure both descriptive statistics and multivariate regression models were calculated. The descriptive statistics for selected area characteristics at the city level are listed in Table 1. Descriptive statistics for industry by decade are listed in Table 2. These are calculated for all cities and by region. Table 3 contains regression models by decade, while Table 4 contains regression results for a stacked data file.Table 1 provides a descriptive picture of the characteristics of the largest 100 cities in 1980 by decade. As expected the statistics illustrate the change in family structure that occurred in the 1970's and 1980's. During each decade the percentage of families that were female-headed with children grew while the percentage of families that were married couple families declined. By 1990, female-headed families with children made up 17.24 percent of all families while married couple families fell to 67.02 percent of all families. In addition to the growth of female-headed families, poverty rates an...
