The field of statistics is kind of based on the idea that you can take a sample and it might represent a population. Maybe you have a more specific complaint about the Steam Survey methodology in mind that I'm not picking up on?
So, you do agree that it's not hard data? Not sure what made you think I am "complaining" about the survey. Just pointing out it does not represent the full picture. It's supposedly random and voluntary.
When considering whether or not the survey reflects reality, I invite you to look at other survey methodologies and the error bars that are possible with as many as, say, 1500 randomly sampled survey participants. The voluntary nature might skew the data a little bit, but not so much that it completely invalidates the trends shown. Sure, not the whole picture, but probably quite close.
Maybe you are right. I cannot tell because I have no way of verifying the results.
Would you personally use the survey for actual real world research? Not talking about phoronix articles and blogs. Say, you are a marketing guy and the company you work for asks you to make predictions.
Personally I would not. And that is the essence of what I said.
