Richard Bartle also published an 8-Type Model in his book, and I wonder if that has any effect at all on your study. (Essentially, it adds a new axis to the model, called Implicit/Explicit.)

The Socializer type was subdivided into the Networker and the Friend, for instance, creating the very split you found: Relationship/Chat. Killers into Griefer/Politician. The Competitor is not present.

Secondly, the lack of the Explorer type may be not because it never existed, but perhaps because it no longer exists. Bartle's original model was developed around a decade ago, which is ample time for cultural norms to shift drastically.

Today, a friend of mine (26) was telling me (19) about the good old days of RolePlaying on a TrekMUSH game that still exists, but is now in the dumps. There HAS been a major cultural shift, and the 90s are something of a focal point for it. The idea of someone who is curious and investigative is becoming more and more rare among RPGers. How many gamers (and notice I use the term "gamers"; a good portion of these people consider themselves nerds or geeks at worst, and know they are not hackers: most gamers are not programmers) would be interested in seeing the code underneath the game? Perhaps 5%, I'd guess, optimistically. It doesn't matter to them HOW it works or WHY it works as long as it DOES work. That's why their requests are so foolish: they don't think about how to improve the game itself, but rather their own experience in it. (Which is not, I admit, necessarily wrong.) But when they DO think about how to improve the game itself, regardless of themselves, they rarely have the understanding necessary to suggest worthwhile recommendations.

I'm aware of the Golden Age fallacy, but there seems to be more than enough merit in this application of it.

I think that Bartle's Type holds, at least abstractly.

One more question, though: You were testing for competitiveness vs. griefing, yet I don't see that. It looks more like you added it as another criteria for griefing, or rather, made that category a broader Competition/Grief category, which appears to say nothing at all.

Michael - The groupings are the result of the factor analysis, and are determined after the survey, and not the other way around. In other words, competitiveness was naturally associated with griefing according to the data itself, rather than imposed by an apriori theoretical framework. Having said that, I don't think enough statements were used for the competitiveness aspect. The problem is that it's hard to come up with general competitiveness statements ...

One distinction that the factor model makes is that people do not fall into single boxes, but rather, have a score on all motivation factors, and people who are used to Bartle's model may miss this important distinction. It makes sense that someone could score high on Relationship + Achievement, and you would find this sort of person in a sustained raid-oriented guild. But in Bartle's model, "Socializer" and "Achiever" are on opposite corners of the model. The "Type Progression Sequence" he discusses in his book further accentuates the assumption that Types are more or less distinct from one another.

The motivations model not only provides a reliable measure of a player's motivations, but it also attempts a more holistic assessment rather than dropping people into boxes. The reasoning of course is that the factors are configural - knowing that someone scores high on Relationship + Achiever allows you to say something more about the person - that they probably like serious, raid/loot-oriented guilds.

I've seen a number of players who pride themselves on knowing the locations of monsters, NPC's, and what monsters drop what items. Some of these people seem only interested in gloating, and others interested in helping people who have questions. I don't remember seeing a question about helping other players with information or introducing new players to the game. I think the description "helping other players" encompasses gameplay and talk, and could be separated.