Pythakoreas
Chef Tain
Not necessarily it is a little more nuanced than that. What you are talking about is the response bias, and it is not simply a bad thing or just because the potential exists does not mean the data is of poor quality, in fact most studies are designed around eliciting some response bias, that is how you target a subsample. As for other potential bias, without anyway of testing the data for potential bias then it is pure speculation about what may or may not bias the results and arguments can be easily made for both sides of your point. Although I have never seen anything that would suggest that Steam Reviews offer any systematic bias which would unfairly bias results. That being said the usefulness of a review is pretty debatable.
I* Didn't get into the details as this is not a stats forum, and your answer* is more precise. I was pointing out that the graph Victoria used was inaccurate because it relies on voluntary response which should be avoided as a general rule.
EDIT: changed pronouns to make comment clearer
Last edited: