OK! Well, I think the best way to answer these posts is simply to present the results. I've spent all yesterday and this morning (all yesterday and this morning. It took a long time) carrying out the analysis. Here's what I found.
GROUNDWORK.
The scales worked as expected. They met minimum criteria for alpha reliability and showed no cross contamination. (I’ll go into more detail if anyone wants a more statistically complete discussion.)
HYPOTHESIS I.
First it was necessary to construct a “liking for RPGs” scale. This was done by combining three things (through factor analysis): 1. the respondent’s average rating for all RPGs listed, 2. the respondent’s highest rating for all RPGs listed, and 3. the amount of overall RPG knowledge, as measured by the number of non “no-opinion” answers given for the RPGs listed. Controls had no knowledge of RPGs and were assigned scores of 3 for all roleplaying games, putting them well below the 1st percentile on the “liking for RPGs” scale.
No significant correlation appeared between this scale and Openness, or Schizotypy. Thus, there is no support for a claim that the roleplayers are more imaginative than others, either in terms of depth of thinking, or in terms of vividness or bizarreness of experience. (On a positive note, Schizotypy often presents in clinical settings; roleplayers may take comfort in being able to say that there is no evidence supporting rumors that the hobby makes them mentally ill!)
HYPOTHESIS II.
To test Hypothesis II, a measure for liking of D&D needed to be constructed. Factor analysis of the five D&D items (OE through 3e) found a clear common factor; liking for these five games was simply averaged to arrive at a strong “Liking for D&D” scale. Interestingly, 4e showed no relationship with liking for the other D&D games, and was therefore treated separately.
There was some question as to whether a respondent's raw preference for D&D, or relative preference for D&D over other games, would correlate negatively with Openness. To measure the latter, the average score given to all non-D&D games was subtracted from the D&D score to arrive at a D&D-NET score. Both this and the raw score were compared to respondent Openness scores. The correlations were r = -.28 and r = -.29; both were highly significant, with one-tailed p-vaules of .0044 and .0036, respectively. Whether one considers absolute or relative preferences for D&D, there is evidence to suggest that greater liking for D&D games appears to go along with lower Openness, and all that this entails.
It should be stressed that these correlations are not particularly large, and that many individuals will easily fall outside of this trend; it is only at the large scale that this relationship becomes clear. (As a final note, no correlations were observed between liking for 4e and any personality trait.)
HYPOTHESIS III.
Factor analysis was carried out on the RPG preferences to determine whether a recognizable factor pitting powergaming against story and aesthetic experience could be identified. Such a factor was readily found, pitting items such as I like the chance to be creative, I like a role-playing game with mystery, I like simple rules, I like plausibility or realism, I play for the story, and I like to be immersed in a different setting against I like the gamebooks, I like many kinds of dice, I play for the combat, I like lots of weapons, spells, and abilities, I like a detailed rule system, I like getting treasure, I like building up my character's abilities, I like collecting powerful items, I like to play characters who are extremely powerful, and I like amazing powers.
Thus, it appears that one can speak of a broad preference for powergaming vs. creativity, aesthetics, and immersion in another world.
HYPOTHESIS IV.
Since Hypothesis III wasn’t falsified, we may go on to Hypothesis IV - this would not have been possible if the factor of powergaming vs creative immersion were not found. But checking Hypothesis IV was a simple matter; the Powergaming vs. Creativity factor correlates significantly with a liking for D&D (r = .226, p = .0175, all values one-tailed), D&D-NET (r = .229, p = .0165), and even with a liking for 4e, though to a lesser extent (r = .180, p = .047). Finally, as predicted, it also correlates significantly negatively with Openness (r = -.27, p = .0051).
This is consistent with the claim that people who play RPGs for the creative and immersive experiences are generally more imaginative and artistic (or as some might say, pretentious) relative to those who play roleplaying games more for the dice, the gamebooks, and the chance to build up amazing powers. Further, it helps to explain why D&D is the game that it is; although there are many reasons to play D&D, Dungeons and Dragons appears to cater better to powergaming than to creative experience. We know this simply because, if these findings are correct, those with a powergaming mentality generally like Dungeons and Dragons better than those who play RPGs for the chance to be creative.
HYPOTHESIS V.
Hypothesis V was easily checked, simply by creating a “Law” scale. Law was coded as 3 for an answer of "Lawful," 2 for "Neutral," and 1 for "Chaotic."
I’d like to note here that many respondents took the time to explain to me that alignment was meaningless. Such claims were amusing in light of the finding of a negative correlation between Law and Libertarianism: r = -.40, p = 8.04×10-5 (one tailed). The chance for a correlation of this magnitude to appear without there being any genuine relationship between libertarianism and chaos is one in over 10,000. This is like being shown a balanced, fair d10, and calling beforehand what it will roll four times in a row. Thus, it seems that the idea of an affiliation with Chaos has a real-world relationship to libertarian sympathy.
Although renaming the Law-Chaos dimension in terms of Liberty vs. Order might compromise the fantasy setting, it could be considered less ambiguous, and engender fewer debates, than the standard terms Gary Gygax used.
HYPOTHESIS VI.
To test Hypothesis VI, a “Good” scale was created analogously to the Law scale—Good was coded as 3 for an answer of “Good,” 2 for "Neutral," and 1 for "Chaotic." The Good scale was independent of the Law scale (r = .03). Running this against Machiavellianism, I found a significant negative correlation, as expected: r = -.34, p = 6.7×10-4 (one tailed). The chance for this to have occurred spuriously is similar to the chance for being shown a balanced, fair d10, and calling beforehand what a person will roll on it three times in a row. Therefore, there is evidence to support the idea that an affiliation with Evil has a real-world relationship with Machiavellianism. So while alignment may be a pain, at this point it seems highly doubtful that alignment is truly meaningless, since the self-descriptions of respondents’ alignment were useful predictors of their personal values.
SUMMARY.
The results of this survey are consistent with the following conclusions:
* Engagement in the roleplaying hobby does not appear to be related to imagination. For many, roleplaying is more likely to be a hobby of beloved game books, favorite dice, miniature figures, and good friends, than it is to be a game of imagination.
* Some people play for the story, setting, or chance to be creative; others prefer powergaming. A preference for powergaming goes along with a liking for Dungeons and Dragons, and both a liking for powergaming and for D&D goes along with greater conformity and lower imagination and depth (or arguably, with less pretentiousness). However, this trend was not very pronounced, so clearly there will be many exceptions.
* The classic alignment graph describes the degree to which one subscribes to notions of liberty rather than order, or to integrity over Machiavellianism. This does not mean that human values don’t include more dimensions, such as conservatism. Nevertheless, alignment is not meaningless.
MY SINCERE THANKS TO EVERYONE WHO MADE THESE FINDINGS POSSIBLE!
I will start on calculating and sending out your personal scores in a few days; you can look for them by next Wednesday.