This is prompted by EcoGeoFemme's somewhat recent post on thesis mistakes.
When I worked in consulting, the standard procedure was to have someone else check your work, even if it was painfully simple. When I first started working, I spent probably 1/2 my office time checking numbers because nobody else wanted to do it. Labs would fax over 20-30 pages of data, and then someone would enter the numbers into a spreadsheet and I would check the two sets against each other. Now, of course, that data is sent electronically, saving some poor low-level employee a little work.
That didn't happen for my thesis. I did a bunch of complex calculations (complex in that it was a lot of manipulation, not that the math was difficult), and it was like pulling teeth to get my advisor to look over them. Sure, he'd look at the write-up and if something wasn't right, he'd point it out. But if I said "this value is multiplied by some other factor based on this assumption", he didn't actually check my math as long as the statement was technically correct. Nor did anyone else. When I first started my thesis, he was a lot more careful checking things. So maybe he decided I was trustworthy once he'd seen that my work was more or less error free?
So I'm hoping that I caught everything myself. I'd rather be in EcoGeoFemme's position and have it caught at some point before publication, though.