of family structure. Black families are much more likely to be female-headed than white families. In addition, despite the Civil Rights legislation of the 1960's, blacks are still likely to face discrimination in the work place. As a result, the analyses includes variables for percent black and percent Hispanic (only in descriptive statistics) in each city. Three measures are used to control for common indicators of female-headed families. They are area poverty, less than high school education. and sex ratio. Area poverty is measured by calculating the proportion of persons in poverty. This controls for the overall poverty level in a city and should be highly correlated with female-headed families. The percentage of persons aged 25 and over that have less than a high school education was also included. This is an indicator the percentage of persons with low education levels and indicates persons that would not make good marriage partners. The percentage of persons without a high school education is expected to have a positive effect on the percentage of female-headed families. Finally, sex ratio or the number of males to every 100 females was calculated to control for the number of males to females aged 15 to 59 available in each city. Presumably the more males there are available the fewer female-headed families there will be. It should be noted that the sex ratio is a fairly crude measure of available males because we are assuming than all males are available marriage partners and this is not always the case. The 100 cities are classified by region. This is important because the changes in the economic structure differ greatly by region. The four regions used are Northeast, Midwest, South, and West. The census definition of a region was used to assign cities to regions. South is used as the reference category. To examine the second hypothesis, that by 1980, employment dislocation was an important source of growth in female-headed fam...
