'Disparate impact' is the idea that, to assess fairness, you treat individuals not as individuals, but as members of identity groups (e.g., grouped by ethnicity, sex, sexual orientation). Then, if any group doesn't get a proportion of some desirable outcome (e.g., employment at a particular company) matching its proportion in the relevant population, this is presumed to be illegal discrimination.<p>The burden of proof then falls on the decision-maker (e.g. employer), i.e. guilty until proven innocent.<p>This seems crazy when you consider processes with any randomness.<p>There's no reason why every single company, for every role, would naturally have demographic proportions matching the relevant population for every single identity group.<p>If you toss a coin ten times and get 6 heads and 4 tails, that doesn't mean the coin is biased.<p>The EO is right: disparate impact doctrine has forced employers to discriminate based on race, so they can maintain proportions.
If you're trying to understand the context of this and other EOs, see here:<p><a href="https://www.nytimes.com/2025/04/23/us/politics/trump-executive-order-university-accreditors.html?smid=url-share" rel="nofollow">https://www.nytimes.com/2025/04/23/us/politics/trump-executi...</a>