Paul said:
|
I still think the numbers represent an inaccurate analysis of the survey they completed ... Suppose you broke down the PS3 into 5 equal groups based on how "hardcore" of gamers they were. While it would be difficult to gather, it would not be surprising to see the following statistics on buying intentions (first 18 months):
- Top 20%: 75% intend to buy PS-Vita
- Second 20%: 50% intend to buy PS-Vita
- Middle 20%: 25% intend to buy PS-Vita
- Bottom 40%: ~0% intend to buy PS-Vita
While it is simple to see that my (hypothetical) numbers add up to 30% of PS3 owners intend to buy a PS-Vita in the first 18 months. Now, being that these divisions would also impact a person's playing habits as well as their online activities surrounding games, it is likely that their participation in a survey would be disproportionate ... Suppose you had the following breakdown in participation rates:
- Top 20%: 70% of the survey
- Second 20%: 20% of the survey
- Middle 20%: 5% of the survey
- Bottom 40%: ~0% of the survey
The survey would end up reporting that 63.75% of PS3 owners intend to buy a PS-Vita in the first 18 months ...
I suppose what I said could be simplified by saying that statistical surveys are only meaningful when you can ensure a randomized cross-section of the target population is polled. If this isn't ensured, you're likely just going to have results which demonstrate the bias of your sampling.