The AQ is Useless

Post Reply New Topic

One of the major problems with self testing is confirmation bias. Just think, how many times have you googled some medical symptoms and found out you could be seriously ill. You panic, all the symptoms seem to agree with a big problem. You fixate on the issues and schedule an appointment with your doctor. If you are lucky, your doctor calms you down, if not, they dismiss you and are annoyed.

With self testing many find what they want and not what is real. I've taken numerous self test. Depending on my mood, the result are all over the place. They are useful, but should be viewed with caution.

I am not diagnosed, but have many personality traits that are consistent with a spectrum disorder.

Why do you think they are useless?

The professionals don't think so. When I was being diagnosed, I was given paper versions of the AQ, EQ, and a few others as take-home questionnaires, in addition to various tests I was given in the office such as the ADOS, the Vineland, the MMPI, an IQ test, etc. So these self-report tests are considered, by some practitioners at least, to be a valid part of the diagnostic process, just not the whole thing.

In my official diagnosis, I had to do some tests that were very similar to those online as well. I did many many tests of all different kinds of things and these kinds were included.

I guess I misunderstood by the way the response was written.

I never had take-home questionnaires. My parents have always taken questionnaires for me, except for ADHD severity.

False positives lead to over self-diagnosis. That was my point.

Some of these tests may be of use, but the AQ test is particularly crude on two levels. Firstly, the response options may as well have been yes/no, because that is as far as the differentiation goes. Secondly, it reflects the innate bias of the author, insofar as it tries to portray being autistic as a series of “difficulties” in a very subjective manner. For instance, why should my disinterest in remembering phone numbers (I can remember passwords) be viewed as a difficulty (especially now that most phones have memories for this)? I’ve tried most of these tests over the last two decades or so, and the only two that impressed were the Jung Briggs Maier (not sure of the spelling) and the C&M 5 minute on line test (though it was only a Mickey Mouse version of the full duration Temperament and Aptitude psychometric test, which takes hours to complete!). It is quite ridiculous that no one in autism research would use this test, even though it was available or free (but no longer available at all, unfortunately)! viewtopic.php?t=570

My entire discord server took the same AQ test recently and the vast majority of them (who aren't autistic) scored below the "Likely mildly autistic" score, proving that they don't give positive results regardless of whatever answers or choices you pick.

Are there some tests out there that are completely randomly generated and take none of your answers into consideration? Absolutely.

However, I don't believe that the actual AQ tests are skewed in favour of "autistic". I was expecting most of my discord community to get scores indicative of autistic traits, but the non-autistic people scored non-autistic scores.

The tests are tools and far from useless.

Sorry, I did not mean to confuse you.

I also found that the AQ in particular had questions that did not make much sense. You have to be able to read through the lines and understand what the questions are really asking and that is hard for many Autistic people. I found the Aspie Quiz, the one with the spider web graph, to be most helpful to me.

The problem with this test is the answers are too black and white. I find some of these random or it depends on the situation so either answer would be honest if you put agree or disagree. Plus it has to do with interpretation. I have seen NTs score in the aspie range but they say their introversion skewed with the results or their anxiety or other.

I agree. You have to be able to explain yourself. That one question about whether you would rather go to a play at the theater of to a party was a doozie. I could go either way on that one. If the play is a play that I would want to see and that is of a subject matter that I want to see and that is not going to send me into overstimulation meltdowns and shock than and the party is going to have music that I can't handle and lots of people chit chatting about stupid things that is going to socially overwhelm and fatigue me to the point of collapse, I would absolutely prefer the theater. But if the party is a nice intimate dinner party with low lights, soft quiet music, and people having civilized, calm, soothing, intelligent conversation or just relaxing and not saying much of anything in a beautiful and natural setting and the play at the theater is going to be massively overstimulating, than I would much rather go to the party.

Or you have to know what they are really asking. Do you want to go to an overstimulating environment where the emphasis is socializing? (the party/indicates nt) or would you want to sit alone in a dark less stimulating place and focus on what you want to focus on by yourself? (theater indicating aspie/autie)

You don't have to apologize you did nothing wrong.

I guess you're talking about the AQ test? I agree. If the test was at least scored such that strongly agree is +1, slightly agree is +0.67, slightly disagree is +0.33, and strongly disagree is 0, that might help somewhat. I wouldn't say that the AQ is useless, but it's definitely inferior to the ADOS, IMO.

One of the strongest points of the RDOS test, that it gives both an NT and ND answer, and gives a decent breakdown of the scores. Not that I'd say that the test isn't flawed, i.e. too skewed towards romantic relationships for example, but it's probably the best test that you can take yourself.

To be truthful, though, if you'd really want an accurate answer, you'd have to have someone experienced in ASD's examine your (or whomever you're trying to diagnose) behaviour in a variety of situations.

Confirmation bias doesn't just occur with yourself; for example when diagnosing a child, the parents are usually (always?) consulted. If that child's behaviour were very close the cutoff, parents' responses to a psychiatrist's questions (whether they believe or deny their child's ASD) could easily sway it either way.

I agree. After I first encountered the AQ test, and got quite a high score (44), I asked some of my friends and colleagues to try taking it. Most of them (whom I assume to be NT) got scores in the 10-20 range. One or two, who do indeed seem to have some unusual traits that might fit with the asperger profile, scored in the low 30s. Based on my own observations, therefore, I would not say there was evidence of a skewing towards false positives. I was astonished, in fact, at how low the scores of "normal" people were. I was the only one in my circle of friends who got what I suppose could be called a high score.