Data Collection Over the Net:
The online survey and questionnaire method are by far the most efficient and easiest method I've ever used. Even now, I'm still in awe that 225 people filled out the questionnaire over a period of 6 days. More over, most of my survey respondents also filled out the questionnaire. The combination of the two sets allowed me to make connections I would otherwise not have been able to make.
Double Gender Bias:
There is a double gender bias because there are fewer female role-players than male ones, and there are also fewer women computer/internet users than male ones. Admittedly, these numbers are balancing out nowadays, but this double bias makes the female sample I got fairly unrepresentative of women gamers in general.
P = 0.000
Getting one significant correlation at the p=0.000 level would have been unimaginable to me when I was still in my Psychology Methods class and we were using such small samples in our pseudo-experiments. The sheer number of p=0.000 significant correlations just amazed me when I first started to analyze the data. When you have 225 respondents, something will either be significant or insignificant. The gray area sharpens and almost disappears. If something is not significant with 225 people, it will almost never be significant.
Projection of Motives
Many survey respondents projected their own motives for playing or enjoying RPG's onto everyone else. This is what most of us do unconsciously to a certain extent and is why the survey responses were helpful. When you enjoy RPG's for specific reasons, such as escapism or creativity, it is sometimes hard to realize that your gamer friends who also enjoy RPG's are enjoying them for a completely different set of reasons.
Questions I Wish I Had Asked in Hindsight
1) Do you prefer to be a player or a GM? Why?
2) What is your favorite game genre or system? Why? What is your least favorite?
3) Compare the feelings you feel when role-playing to other activities you do. Where else do you feel those same emotions?