People love seeing themselves on scales. We all want to get to know ourselves a little better, want to understand whats going on in our heads. Testing offers an easy way to achieve that, just answer some questions and an algorithm will give you what you want to know.
Lets take the autism test from the other thread for example: http://archive.wired.com/wired/archive/9.12/aqtest.html
( t/102245&start=0 )
ASD is a disorder, that primarily affects information-processing, which results in more or less severe impairments in all sorts of social settings.
Now, since the test was made to identify autistic traits, one thing it checks for is competence in social situations. When you take a look at the test, you notice that many questions are targeted at this specific field, namely:
1), 10), 11), 13), 15), 17), 18), 22), 26), 27), 31), 33), 35), 36), 38), 39), 44), 45), 47), 48), 50)
thats 21 questions.
As it happens to be, there are other things than autism that can cause incompetence in social settings, the most relevant example being social anxiety with its various causes.
Seeing how many questions were targeted at social competence, I strongly suspected that anyone with social anxiety will automatically score a lot higher, regardless of whether he/she has autism or not.
A quick trial-run where I simulated social anxiety (answering all aforementioned social questions with either strongly agree/disagree, depending on their wording), while neglecting other questions (answering with slightly agree/disagree in random manner), gave me a score of 30.
Take a look at this:
here it tells me how the questions were weighted and how the final score was computed. As it turns out, 17 of the social competence-questions were counted and added to my final score (surprise). I gained 17 points by simply pretending to be socially anxious.
This test is extremely confounded. One thing that can be and most of the times is completely unrelated to autism, and in addition to that happens to occur quite often, distorts the result by an entire order of magnitude.
This is bad, very very bad. Even worse because it wouldn't be even that hard to design the test in a way that it circumvents this problem.
It's a shitty test.
And thats not the only example of shitty tests making their way to the public and to this forum. The INTJ-Whatever test-thread comes to mind.
( t/216342&start=0 )
The worst thing is, that people take tests like this seriously, because they don't know how to tell when a test is badly designed. They then identify themselves with the results, which can have a grave impact on their lives (self-fulfilling prophecies).
Don't fall for that, and don't support shitty tests. Don't support pseudo-science by spreading it to unassuming people.
Discuss
Lets take the autism test from the other thread for example: http://archive.wired.com/wired/archive/9.12/aqtest.html
( t/102245&start=0 )
ASD is a disorder, that primarily affects information-processing, which results in more or less severe impairments in all sorts of social settings.
Now, since the test was made to identify autistic traits, one thing it checks for is competence in social situations. When you take a look at the test, you notice that many questions are targeted at this specific field, namely:
1), 10), 11), 13), 15), 17), 18), 22), 26), 27), 31), 33), 35), 36), 38), 39), 44), 45), 47), 48), 50)
thats 21 questions.
As it happens to be, there are other things than autism that can cause incompetence in social settings, the most relevant example being social anxiety with its various causes.
Seeing how many questions were targeted at social competence, I strongly suspected that anyone with social anxiety will automatically score a lot higher, regardless of whether he/she has autism or not.
A quick trial-run where I simulated social anxiety (answering all aforementioned social questions with either strongly agree/disagree, depending on their wording), while neglecting other questions (answering with slightly agree/disagree in random manner), gave me a score of 30.
Take a look at this:
here it tells me how the questions were weighted and how the final score was computed. As it turns out, 17 of the social competence-questions were counted and added to my final score (surprise). I gained 17 points by simply pretending to be socially anxious.
This test is extremely confounded. One thing that can be and most of the times is completely unrelated to autism, and in addition to that happens to occur quite often, distorts the result by an entire order of magnitude.
This is bad, very very bad. Even worse because it wouldn't be even that hard to design the test in a way that it circumvents this problem.
It's a shitty test.
And thats not the only example of shitty tests making their way to the public and to this forum. The INTJ-Whatever test-thread comes to mind.
( t/216342&start=0 )
The worst thing is, that people take tests like this seriously, because they don't know how to tell when a test is badly designed. They then identify themselves with the results, which can have a grave impact on their lives (self-fulfilling prophecies).
Don't fall for that, and don't support shitty tests. Don't support pseudo-science by spreading it to unassuming people.
Discuss