[rococode]

For what it's worth, in my experience tasks from university researchers tend to be more fair. The last few studies I did, we paid an equivalent of $8-10/hour while estimating generously on time (e.g. we did $4/30min, but actual time ranged from 15-20min). People in my lab and another lab I was affiliated with seemed to pay similarly fair wages for both studies and data labeling. We also usually accept everyone's submission, even obviously junk answers. It could be that the people I've worked with are more ethical on this than other researchers, but I don't think that's true.

I don't think the IRB actually enforces a minimum wage, but the feeling of oversight seems to discourage super exploitative behavior. The downside is that high payments means higher requirements - something like 98% acceptance, 500 HITs, Master's qualification, and US location is what I generally see used, and we still get junk answers sometimes (in my last study one person somehow submitted a bunch of HTML forms with empty answers despite required fields, still not sure how they pulled that one off without intentionally doing some sketchy stuff...). This makes the fairly paid tasks inaccessible to most new Turkers, leaving them to pick up the scraps of the $1/hour HITs.