When I was a wee thing, compliance issues consisted of not blowing up the lab (too often). Or not using extension cords as clotheslines for hanging up wet X-rays to dry ("really, Dr. Theron, it wasn't like it was plugged in or anything"). I can't speak to what IRB's were like, as I wasn't doing any human research in those distant days. But we all have heard of the horrible excesses and unethical goings on of the times. If you haven't read Rebecca Skloot's excellent book The Immortal Life of Henrietta Lacks you should.
For those of us that work with animals, we all know and love and work with, the Institutional Animal Care and Use Committees that govern our research lives. I lived through the no IACUC through ineffective IACUC through the regulated situation in which we do our work today. I have sat on my University's IACUC in the past, and been the stats person for the IACUC. It was in no sense fun. Even free lunch did not help.
My rule: do not offend the IACUC. Follow the rules. yes they will seem arbitrary. Yes you will be irritated and forced to compromise things that, in your perception, are Best for The Science.
I know plenty of people who bend the rules. But one of the things that I have learned is that by and large IACUC rules are not good ones to be bent. I know people who live by "tis better to ask for forgiveness than permission". IME, this seldom works with IACUC's and their escalating scales of penalties. Some Big Dogs I know are beyond rules. The worst, when caught, blame people under them, oblivious to the culture they have created. Some PI's are quite compulsive and follow everything. I had a colleague whose benchtop had outlines in masking tape, with a number, of where every piece of equipment needed to be, and a cross-indexed master list of places and equipment. In the days before personal computers, this was a hand-typed list.
In my lab, IACUC, IRB and compliance are unbendable, rules which cannot be bent. I've got two lines of justification for this: ethical and practical. Practical involves the level of problems that occur when one is caught breaking rules. These problems can grow into a moratorium on doing research with animals, a situation which can be devastating to all of the people in the lab. The decisions on which rules to bend are essentially ethical ones. None of us would do something that we perceived as "hurting the animals". And even if this instance of breaking the rules wouldn't hurt the animal, another instance might. The trouble with rules is that you can't plan for every instance.
I have found that the temptations to bend the rules usually stem from poor planning. Something MUST be done, or the experiments will fail, and the means to do so are not within the protocol (people, drugs, etc). The other category of bending comes from laziness or obliviousness, though it is seldom seen that way: I just don't want to get out of bed to go check on the animals. Or I am doing something else and plain forgot. These situations can be addressed with more planning, and taking working with animals more seriously.
Working with non-rodents, large animals, expensive animals tends to encourage the planning necessary to avoid these problems. It's harder to persuade people, trainees, collaborators, that mice deserve the same respect as dogs. This is part of the PI's job. The time a PI invests in laying down the no-nonsense, we are serious about protocol compliance, and I expect it from everyone, is well worth it.
The IACUC and animal facility at my new place is wonderful. They are reasonable about violations (yes, my lab had one, yes we, me and the trainee, learned from the experience). But I've been places where it's not quite that way. I assure you, finding out that your IACUC has a sadistic streak is not anything anyone wants to do.