|
Abstract
|
Data from 30 surveys (46,356 cases from 10 Michigan Election Surveys and 20 Gallup polls) are analyzed to study the interrelations of cohort, sex, race, region, and education from 1952 to 1973.
A d-system model is developed that pools results from the various surveys into a set of parameters forming a flowgraph model. This technique allows one to see whether the surveys used are consistent and whether the results are consistent with outside sources such as census reports. (Generally speaking, the surveys were found to be consistent and the model showed generally good fit with census and NORC General Social Survey estimates, but significant differences in two or three parameters.) A computer model of the system is used to generate implications and extrapolations.
Substantively, the main results were: (1) a definite trend away from sex equality in education, especially for college attendance; (2) very little narrowing in race, region, and religious differences in education, until the most recent cohorts where Northern White Catholics and Southern White Protestants show virtual parity with "Yankees" (Northern White Protestant); and (3) a projection that shows discouragingly little narrowing of intergroup education differences in the general population for the next 5 years. |