This "natural selection" argument presupposes that women are sufficiently prevalent on boards that the competitive advantage they provide is apparent and obvious to all other companies. This assumption is problematic in a context of low prevalence.
The way State Senate leader Toni Atkins stated it - direct reference by the OP - "better decisions and businesses outperform the competition" means she believes that this is a 100% fact, and the competitive advantage is apparent to everyone but the foolish judges (quote: "legalities don’t match our realities"). She doesn't seem fazed by this assumption being problematic in a context of low prevalence.
If they aren't sufficiently prevalent to demonstrate the advantage, then on what basis is the claim being made?
When you do research, you select your sample judiciously, and you never really sample the whole population. Depending on the research question, you sometimes want to select a sampling technique that is not representative of the population.
Article says 26% met the criteria already, so it's not that out there to expect to see an effect