There are also quite a few who never/rarely play games. Averages do things like that all the time. Which of course opens the question of is that an important factor?
Yeah, but a more useful analysis would be to look for a median instead of mean, or clean the dataset to remove outliers to say something like, of those who play video game, the average/median is... this would also lead itself to the next step of analysis about if it's an important factor like you mention by categorizing them into players vs non-players.
True. My main point is that almost everytime I read something about "average" hours spent playing games or whatever, it turns out to be a hack job. Like why use average if you have a bunch of zeros in it? What does that average actually tell us? Usually nothing useful. Metrics like what percentage of teens/men/etc play games now vs 10 years ago, or of the people who play at all, how many hours do they play now vs 10 years ago, etc would be so much more useful than what they gave because the increase in average hours played might only be due to fewer zeros being in the dataset.
