Comparing AQ test with Aspie-quiz

Post Reply New Topic

A comparison can be found here:
http://www.rdos.net/eng/aspeval/aq.htm

The validity of various questions in the AQ test can be found here (look from question 154 and onwards):
http://www.rdos.net/eng/aspeval/quizr4.htm

The main link to the evaluation is here as usual:
http://www.rdos.net/eng/aspeval

Thanks.

.81 is a significant correlation.

I did the math, and if I plugged in my AQ into the formula, the result is quite a bit higher or lower on the Aspie/NT tests. Like I'm much closer to 100 on both of them than the result I get when I plug my AQ into those two formulas.

I don't get what your getting at wolfpup - you do realise, its all statistical analysis stuff there, the regression curve essentially gives you an idea of what the typical line would look like.

Very interesting stuff, should have read the AQ paper properly though, I never realised the control group scored so low on it

Maybe I don't understand it then. I thought it meant that if you plugged your AQ score into those two formulas, the results should be your Aspie and NT score...but mint isn't. The Aspie and NT score where higher and lower respectively, than what those two formulas come out to.

Last edited by Wolfpup on 25 Apr 2007, 1:12 pm, edited 1 time in total.

Yes, the formula is generally only good for averages. I didn't post the data-points of the regression diagram, but there were a quite large span for the higher AQ scores, while the lower generally always corresponded to low Aspie-scores. That probably reflects that the AQ test is much narrower than the Aspie-quiz and therefore a high AQ score might not always correspond to a high Aspie-score although it usually does.

How much different were they?

I guess I'm pretty low on both-34 for the AQ, and and 116 on the Aspie one, 98 on NT.

So, your expected scores are 125/84. This is probably a difference that is not that uncommon given the distribution of AQ-scores in this range. Perhaps I should post the composite from DataFit, but I'm not sure how to convert it into an image.

Another factor that could skew individual scores is that males score higher on the AQ test while females score higher on Aspie-quiz. For that reason males (on average) should expect lower scores on Aspie-quiz and females higher based on the above calculation.

A couple of reflections why males score higher than females on AQ test.

A large difference between NT males and females, but no difference in ASD population:
"I find social situations easy"
"I find making up stories easy"
"I find myself drawn more strongly to people than to things"
"I find it easy to 'read between the lines' when someone is talking to me."
"I am good at social chitchat."
"When I was young, I used to enjoy playing games involving pretending with other children."

Males generally score higher:
"I am fascinated by numbers"
"When I'm reading a story, I find it difficult to work out the characters' intentions"
"I don't particularly enjoy reading fiction."
"I don't usually notice small changes in a situation or a person's appearance."
"I am not very good at remembering people's date of birth."

These probably are sufficient to explain the gender differences.

RDOS have you observed in people whom have taken the test a coralation between their scores and their observable AS severity and behaviour? I appreciate that many of your subjects are sourced on the internet and perhaps you have not the opportunity of testing people in observable conditions. Have you approached proffesional bodies and institutions with your tests and research? I gather that you have devised this test based on your own experiences and research and perhaps single handedly, if so i must say im impressed.

Going by correlation between self-diagnosed and scores vs. professional diagnosis and scores, I doubt there is any correlation to severity. Severity is probably related to general IQ (it is at least in DSM criteria), and because score is positively correlated with traits that are related to high IQ (g axis in factor analysis), I would expect a negative correlation to severity in the sense of functional level. However, if severity is defined in terms of social & communication differences, there is definitely a positive correlation.



I hope that somebody will eventually do something like that, but there is no plans for it right now.

BTW, here is the regressed data from DataFit (AQ score vs Aspie-score):

I think I might have mentioned this elsewhere, but I don't like the AQ test.

I scored much higher than I expected on that (45) - I think that is because the slightly and strongly options are given the same score. A lot of my answers were slightly in favour of Aspie characteristics, which I think skewed my results.
I scored around 140 on the rdos test.

Took a couple versions of rdos's Aspie Quiz, and scored as expected (AS).
IRL I'm officially dx'd, FYI.
Statistics are not my thing. Anecdotally, my reaction to the AQ test (the 30 or 40 question thing by Baron-Cohen, which is what I assume we're discussing) was not positive. Didn't get at my particular sorts of problems, it was alienating. B-C's theory is more for (what I consider) a certain type-but that doesn't cover all types of people w/ASD's.
Am neither much of a systematizer (high SQ score="male brain" aka "person w/ASD) nor much of an empathizer (high EQ score="female" brain aka NT), according to another couple of B-C's quizzes in "The Essential Difference". His tests seem to predict stereotypes instead of the challenge of sorting out variables of what's unique/identifiable about AS in particular. IMHO.