The Autism Spectrum Quotient Test (AQ) is constructed with a 4 levels possibility of answer: Strongly Agree, Slightly Agree, Slightly Disagree, Strongly Disagree. When it comes to calculate the score I find references for considering the levels "strongly/slightly" as equivalent. So for certain items they both either receive either the score "1" or "0". Would it be acceptable to consider a 4 levels type of score, that grants a different "weight" for "strongly/slightly" levels? For example "+2, +1" "-1,-2" depending on the item? Has someone done something similar? Could you suggest specific literature?
I fell sure this was looked at when the test was first constructed. The answer is almost certainly that it would make little overall difference.
According to Baron Cohen et al. (2001; attached), the scoring system does indeed call for awarding just 1 point for answers in the keyed direction, whether they are "strong" or "slight." This is an unusual approach, to be sure; usually when one uses a likert-type scale, each response option has its own weight, the idea being to increase the amount of information conveyed by the score. Maybe I didn't read the article carefully enough, but as far as I can tell the authors provide no explanation for this decision. My guess is that they figured some people would be reluctant to give a simple "yes" to some of the more pathological-sounding items, but might be willing to say "yes, sometimes a bit." But that's just a guess.
So can you change the scoring system? Well, yes, of course you can investigate the effects of so doing - but be very clear about what you're doing. If you want to study this, I'd say to score your protocols both ways (the original 1/0 system and your 4-point system) and see whether your approach actually increases the validity or clinical utility of the scale.
I agree with Stephen that this is unusual since it implies the strong and slight responses are identical. We faced this problem when doing http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0031661 and opted to treat them differently (as interval) and for the EQ and SQ also. Particularly in the old days test constructors would do things so that it was easy for those administering a diagnostic test to score it by hand. I am not sure why Baron-Cohen and colleagues decided to do this for the AQ. They stress it should not be used for diagnostic purposes. Maybe if they see it as just a screen device they felt binary scoring was okay.
Thank you for your answers.
I was thinking to try a 4 points likert scale as Daniel Wright did, but I wasn't sure it was correct. Thank you!
Have you also ever used the Interpersonal Reactivity Index as a measure for Empathic Concern? What do you think about it compared to the EQ scale? Would you expect it to behave similarly?
I haven't used the Interpersonal Reactivity Index, so I don't know. If you look at both it would be interesting to see how they are associated.
- Badges
- Science topic
- Similar topics
- Psychology
- Psychometrics