In this context, I think they mean, believe that the published data is true. You don't have to believe in science itself, but you do have to trust scientists when they report data (you can check the conclusions yourself, but you can't check if the data was accurately reported, if the published protocol was followed etc).

That said, I believe the legal repercussions of lying about these studies will be too high to stomach, so I do trust the companies to be truthful here. That doesn't mean that it is irrational to choose not to trust them.