I've seen an awful lot of either not understanding what a measurement means, or not caring, among management, including out of "top-tier" folks from e.g. management consulting firms you've heard of. Talking really obvious "experimental error" types of things. Taking a measurement as indicating something when there there are obvious confounders which, if you don't look into them, are probably making your number meaningless.

I don't think engineers are the only rational ones, but I think the rational ones in management have noticed that their numbers being meaningful or correct only rarely matters for getting a "yes" or a promotion, so have responded... rationally, since getting useful figures and proving that they're useful is both 1) usually a lot more work, and 2) often not even practically possible.

The other option is that management is lying about what is important to them. I can accept that as well.

It is not at all required for engineers to be acting rationally, let alone for them to be the only ones acting rationally, for workers—of any type—to have a better understanding of what will let them get a good job done for the company than upper management.

This is especially true in large corporations, where upper management is very insulated from the realities of line-level work. It is also especially true in most publicly-traded corporations, where upper management is almost invariably going to care more about the trendline in the company's stock price than in their actual productivity, or even in the well-being of their workers.

If engineers acted rationally, we would not have so many technology fads.

I would argue that is rational for the engineers who become less employable if they aren't doing the latest stuff and creating new libraries.

The interests of engineers are often extremely detached from whether development is cost effective and maintainable by others.

This rational behavior in the face of stupid incentives is the same reason the rational set of management-sorts may not bother to make sure their facts & figures actually demonstrate what they claim they do, even when they know they probably don't. The extra effort wouldn't help them get the answer they're looking for, anyway, and might even harm their chances if they're honest about the level of uncertainty.

The part that's stupid is everyone, even the ones who know better, pretending a bunch of bad statistics and such aren't bad, but if the suits in the room are nodding along you damn well better not point that out. People succeeding in the management track either actually are bad at understanding these things, or (plenty often) do understand but also know to STFU about it.