Are you a secret super-matcher?

rnz.co.nz

2 points·Quinzel··5 comments

I very much enjoyed doing this test. But, I would posit my 16/16 score is next to meaningless. Discounting the obvious dissimilar instances, leaves 4 or less hard cases and treating each as an independent 50/50 I could have just rolled 4 snake eyes: it's unlikely but not impossible.

I hazard a rerun or different test would show I am less than super. I don't think repeating the test would help them with basic data capture.

I also thought the dichotomous format was a significant weakness of the test and would impact on reliability. However to get 16/16 correct it would likely be down to more than just chance. The more items, the more reliable the test becomes. I noticed that the article stated if you scored 15 or 16/16 you may be a super-matcher. Or, like you say you may just be super lucky to have guessed them all right.

There is a formula for correcting for guessing: Corrected score = R - W/(n-1)

R = right responses W = wrong responses n = number of choices for each item.

I found it super easy. There was only a couple that needed a thorough look. Fun though! I similarly was a little unsure whether they just give full marks to everyone, but looking into the background of the researchers a little further, seems legit.

Yeah it actually seems pretty legit. On the basis of that test, they invited me to participate in some more tests that are not publicly available. For the other tests 3/4 I met the threshold that “may” indicate I am a “super-matcher”. However, I doubt it!

I’m unlikely to find excitement as a forensic scientist - I’ll have to satisfy my craving for adventure elsewhere ;)

Apparently forensic scientists train for a long time to learn pattern matching skills…

Reply on news.ycombinator.com