Test prep accounts for a small portion of test score differences and minorities are more likely to use test prep.
I must disagree. In my brief experience with the "test prep industrial complex", I saw that most of the families putting their kids through it are immigrants - especially South Asian and East Asian immigrants. Minorities who arrived in the US with little money and no local support network.
I assume the answer is greater than 0% and less than 100%. In which case, the second question is, what signal other than test scores could the top 10 universities use to accurately distinguish between those in your particular cohort who would most likely succeed and those who wouldn't?