I used to be a statistician and you're talking rubbish. It's not that hard to run and publish an unbiased study, and it's not that hard to prove it was unbiased to anyone who asks (which they should but don't always do).
We never had anyone come into our office and say "here's a bribe, now stop what you're doing".
We DID have people - both from big companies and from small business interests - approach the ethics committee and say "we want you to turn a blind eye to this" but they would be the ones who wouldn't get approval to continue.
If you try a bit harder, you can find the sources that give you a 5-10% figure and there'll be nothing wrong with believing that those are scientifically honest.