The last measure that I will discuss was also incorporated into the study to look at social and cognitive behaviors that one could consider as part of the stereotype of furry behavior. This survey actually has 5 subscales, each with 10 questions, as well as a total score (which is just the sum of the 5 subscales.) (For the IRI, empathy measure, the four subscales DO NOT get added together, in case you are wondering).
The subscales tap into different behaviors there is a subscale for social behaviors, attention switching, attention to detail, communication and imagination.
The scoring of the survey according to its author is rather odd in my opinion. The response format is "strongly agree, agree disagree, strongly disagree". But then the "agree and strongly agree" answers get scored as simply agree and the "strongly disagree and disagree" answers get scored as simply disagree. From an information standpoint that seems silly, so I have scored participants' responses in two different ways. First according to the author's original instructions and second using the full response range (as was done in a recent Dutch study which used the same survey).
Typically males and females score differently on this survey so we will continue to look at the effects of both furry status and sex of participant in the analysis.
Using the original scoring system, on NONE of the subscales do furry mean scores differ significantly from non- furries and similarly there are no significant differences on any of these subscales between male and female mean scores. For the total score, which is the sum of the five subscales, the furry group mean is statistically significantly different, but only slightly higher (meaning less social) than the group mean for non-furries. The p value of this finding is .047, and the group mean for furries is 19.51 and for non-furries is 16.48. Both of these group means are in the range that the survey author indicates is "average". (The possible range of scores on this measure is from 0 to 50, and the author indicates that "average" is from 11 to 22).
Using what I refer to as the "Dutch" scoring system, strongly disagree = 1, disagree = 2, agree = 3 and strongly agree = 4, I rescored participants' responses and reanalyzed the data. Using this scoring procedure, no significant differences appear in any of the subscales or the total scale. There are a few what we like to call "borderline" cases...that means p values slightly over .05, not officially significant but kind of interesting.
For the attention to detail subscale the furry mean is about one point higher than the non-furry mean (p =.066), for the communication subscale the furry mean is about 2 points higher than the non furry mean (p =.065), and for the total scale the furry mean is about 6.5 points higher than the non-furry mean (p = .063).
For statisticians in the audience you probably appreciate that these differences in group means are small, the p values are NOT STATISTICALLY SIGNIFICANT, and a serious issue with the analysis is the very unequal group sizes when comparing furries with non-furries and males to females. In short these differences are not really differences, just kind of interesting...the kind of result that makes one wish they had a larger sample of females and non-furries.
So at this point we can pretty safely say that compared to non-furries who participated at Anthrocon our participating furries are quite similar on this measure of social type behaviors and it does not matter which scoring system is used. Furthermore, when looking at the furry group mean of 19.51 on the total scale score it is within the range that the author has identified as "average".