A little tidbit that slipped my mind:
The Black/White (and White/Asian) gap can be
erased by controlling for "stereotype threat" (the fear of confirming a negative stereotype that subsequently hurts cognitive abilities while test-taking). In terms of experimental variables, this includes the ethnicity of the experimenter, the purported purpose of the test (ie. intelligence test versus a generalized test), and the knowledge of a future comparison with another group.
http://www.uiowa.edu/~grpproc/crisp/crisp.10.13.html (look for the page headers for pages 192 through 194)
This article provides a summary, though particular attention should be paid to Steele and Aronson's 1995 experiment, which you can find more about on google.