| Stephen Morgan
Cornell University
Educational achievement and disidentification among white and black students
FINAL REPORT
The black-white gap in achievement, as measured by performance on standardized tests, has received considerable attention from researchers in the past five years. Claude Steele's stereotype threat and disidentification mechanism is perhaps the most heralded of new explanations for residual racial differences which persist after adjustments for social background are performed. Analyzing data from the National Education Longitudinal Study, we find qualified support for portions of the disidentification explanation.
Our findings suggest that there is support for a limited form of disidentification -- black students exhibit a weaker relationship between their academic self-concept and their academic achievement. We interpret this weaker relationship as a sign that they discount performance evaluations more than whites. While it is important to stress that this relationship is still positive for black students, it is about 30 to 40 percent weaker than it is for white students. Thus, while blacks do not completely reject the evaluations they receive, on average they are less likely to think they provide an accurate measure of their abilities. We cannot say exactly why this is so from our survey data. It is consistent with the idea that black students feel that the tests themselves are culturally biased. It is also consistent with the idea that black students feel stereotype threat when taking the tests, and thus intuit that the tests are poor measures of their capacities.
However, as best as we can tell, this is not the first step down a slippery slope. White and black students similarly ground their overall self-esteem on their academic performance at the three different points (8th, 10th, and 12th grade for most students) that the NELS enables us to measure the relationship. Black students remain as fully identified with schooling as white students, a finding which stands in clear opposition to the disidentification and selective valuation hypotheses. It is consistent with other assessments using national survey data that show that the behavior of black students that may represent alternative indicators of identification/disidentification (i.e. absenteeism, homework time, etc.) is quite similar to that of white students after controlling for family background.
Back to Funded Research Grants Page |