Because the study that generated that statistic could have used a very broad, inclusive definition of gamer. So, the average age of someone who plays a certain game online could be lower since that is a much more exclusive, narrow definition.
Plus, there is something called a "confirmation bias," meaning that you will only notice incidents that reaffirm a certain idea or perception you have (you only notice what confirms your bias). The average age of people on a certain online game COULD be higher, but you only notice young people playing it because you believe the average age to be much lower. Thus your mind is tricking you into believing the average is, in fact, lower because you only pay attention when something confirms this suspension.
Plus, there is something called a "confirmation bias," meaning that you will only notice incidents that reaffirm a certain idea or perception you have (you only notice what confirms your bias). The average age of people on a certain online game COULD be higher, but you only notice young people playing it because you believe the average age to be much lower. Thus your mind is tricking you into believing the average is, in fact, lower because you only pay attention when something confirms this suspension.