I love that you chose the trusty old NOLOCK hint as an example. I started work at a particular organisation many years ago and found the God-like NOLOCKs reverently placed against every table in every query. I'd never seen this behaviour before, so I just had to ask why. The dismissively authoritative answer (from the manager, no less) was "If you don't, it breaks replication". Wait... what? This didn't sound quite right to me. But I was new and this guy was the manager. I spoke to another newcomer (a DBA with some pretty good chops) and he raised his eyebrows. We did some googling and the respected SQL Server bloggers all seemed to be saying "first fix your query" before worrying about NOLOCK. It was the queries that were "breaking replication" (causing locking and blocking). The NOLOCK was (at best) treating the symptom, not the cause.

The bigger problem was that try as we might, we couldn't convince the manager to stop using NOLOCK hints (and writing queries so bad that they needed them). It became a petty professional jealousy thing. Any argument, article, or metric, was met with stonewalling like "Well that's just a matter of opinion", and "there's probably just as many articles arguing the other way." Essentially he was saying "I'm the manager and I'm not going to be told how to do my job by a couple of Johnny-Come-Latelies".

In our case the "...we've always done it that way…" became entrenched for utterly irrational reasons (hubris) and became exceedingly hard to shift. I hear they're still repairing the damage today.