Legomsky justifies racial profiling by invoking base rates of terrorists within the group

Evidence

Certainly, I agree that only a minuscule percentage of noncitizens who appear to be Arab or Muslim are involved in any way with terrorism. But that is not the point. The more relevant figure, I maintain, is the converse – the percentage of those noncitizens involved in terrorism who are Arab and Muslim. If there is credible evidence that this percentage is higher in this subgroup than in the general population, then, it seems rational for the government to focus particular attention on that group. It is simply a matter of channeling inspection resources to places where they are statistically most likely to detect real terrorists. — Sherene Razack, Casting Out, 87

Categorization

• QUESTION: The prosecutor’s fallacy flavored logic here — Legomsky swaps the probability a profiled person is a terrorist for membership in a group that contains most terrorists — is worth unpacking as a statistical error.
• XREF: Connects to Bayesian reasoning about false positives and base-rate neglect, and to the Gigerenzer work on the HIV-screening base-rate problem.