Gigi Pro Certified Test

[Yerba Mate]

Hello, is there any consensus on here about the accuracy of the Gigi Pro Certified Test?  I scored a 142 on it, but it seemed rather easy (or, it just played into my strengths of pattern recognition and speed).

Has anyone taken both the Gigi and a professionally administered IQ test?

Thanks.

[polar]

Hi, I vaguely remember gigi overestimated, but I dont know how much. This one was more accurate http://giqtest.com/ , or this one, notoriously known in this forum (: http://iqtest.dk

Dne pondělí, 20. srpna 2012 1:28:32 UTC+2 Yerba Mate napsal(a):

[☉]

The ceiling for that test is low, marked at "151+": http://giqtest.com/highscore.html . Since Yerba already scored 142 on Gigi, GIQTest might not be the best recommendation. Depending on the number of problems, ceiling effects could be experienced up to five items below the maximum, and assuming time isn't a factor in scoring. Furthermore, GIQTest includes non-matrix problems, which is a definite confound for determining any improvements in the non-verbal domain.

Gigi is a speed oriented test and makes no bones about it. I'd say that, in general, it is most in keeping with the research on DNB, since Jaeggi et al. have been noted to use tests in ways they weren't designed to be used, that is, in a timed fashion and chopped into pieces. Hence, if the gains in raw score are genuine at all, then it is merely a matter of increased visual processing that led to them (a common point made in comments). Given that, Gigi would make a good pre-training and post-training assessment. It caps around 173, so there's plenty of room not to hit your head against, at least as far as timed matrix problems go.

If someone had scored over 150 on their first attempt on Gigi, I'd say they're probably wasting their time with Gigi. Since that isn't the case, you could probably take that as a pre-training measure, train on an n-back variant protocol, and take Gigi again afterwards to note improvement, regression to the mean, or no net effect. I think it's cheap enough for that kind of program.

argumzio

[ST]

Argumzio, if Gigi isn't a necessarily accurate test, then would you say Raven's progressive matrices are the best type of IQ test? It seems to be the one that is the "industry standard" in research at the moment I think.

[☉]

I'd say Gigi is in the same league as RAPM, in timed conditions. Note that the RAPM norms do not go much higher than two standard deviations in an untimed setting (which is what it was designed for).

I say this, at any rate, not knowing specific statistical details of the Gigi to formulate any worthwhile comparison. However, I do not think there is such a thing as a test that is the absolute "best". Much research on the Raven's (cf. "Uses and Abuses of Intelligence" ) suggests that the old RAPM is outdated as well as serving well enough for its purpose in countries which are naive to modern industrialized society with its increasing abstractions and familiarity with scientific praxis.

It's been a while since I read that, so that is only a cursory recollection of the book, which hardly counts as the entirety of research on the test itself.

argumzio