Progress in the quest for better information: a UCLA professor advises me that about 85% of UCLA grads overall take the California bar exam. So out of 18 African-American students, we'd expect 15 to sit for the Calif bar.

However, he also pointed out that at least a few students listed in the class of 2006 may still have one semester to go (because they took time off) and are not eligible to take the bar until February. So that may cut into the 18.

So today, I'll extend my unscientific sampling technique to white students listed in the class of 2006. While this number won't mean much in absolute terms, it will have been derived the same way as the African-American pass rate, making it more of an apples-to-apples comparison.

I started at the beginning of the facebook listings and took the first 30 students who were likely to be white. Since there's no correlation between the alphabet and academic performance, the first 30 would be as random as any 30. As before, I ran the names through the bar passage search engine.

23 out of 30 of the white students came up on the pass list = 76% pass rate. This number was derived the same way as the 44% number in the African-American sample: by taking the names on the pass list and dividing it by total names in the sample.

If we assume the overall UCLA pass rate is still around 89%, we can make a couple further inferences. Since 23 of the white students passed, probably 26 students in the sample sat for the exam, because 23 out of 26 = 88% pass rate. That would leave 4 who didn't take the exam for whatever reason. 4 out of 30 = 13%, which corresponds pretty well to the professor's estimate of how many UCLA grads take the Calif bar (85% do, which means 15% don't).

So going back to the African-American sample, let's use the same baseline and assume 85% took the bar, and 15% didn't. That means of the 18 students, 15 took the bar and 3 didn't. 8 out of 15 = 53% pass rate. (That's a more reliable baseline estimate than the previous 44%.)

Let's also have a more conservative hypothesis where, for whatever reason, only 70% of the African-American students took the bar. That would mean 12 out of 18 took it and 6 didn't. 8 passed out of 12 = 67% pass rate.

So, after this extra facebook experimenting, we can narrow the predicted bar pass range for African-American students to 53-67%. Compared to the white pass rate of 89%, there's still a gap of 22-36 percentage points. So it looks like UCLA tracks the statewide racial gap of 30 points pretty closely.

21 Nov 06