Do you know how exactly is evaluated a "normal" IQ test? By "normal" meaning not the online IQ tests one can take for free and have them automatically scored.
It depends on the test. The original IQ tests were quite broad and meant to be predictive of performance in the military. A lot of tests are hangovers from those days.
The more recent ones that have been widely validated and used for research (WAIS and KAIT are the two I keep mentioning, but there are others) tend to have a pretty broad battery of sub tests, everything from recall to word play.
The questions themselves are suppose to test the ability to manipulate and store data, from the past and from the present. For example, a common example on word play is the "Two birds with one stone" or "eggs in one basket". These kind require a certain degree of knowledge beforehand but the tester is looking to see if the mind is able to connect abstractions into concrete examples, etc. Like the free tests - the patterns, etc - those test fluid intelligence a bit more, but are often not built correctly online. A decent example of a more layman's version of these, but have some validation, are the top tests
here. They are different than the aptitude tests that are down below.
What I do not understand is how an IQ test is related with openess to ideas. If I recall well it mesures abilities like memory, spacial and visual sense, etc. Aren't those abilities typically related with Ss?
Well, not to put it in too fine of terms... IQ
does correlate to openness and to N and is more or less anti-correlated with S/ -openness. It is better to start from there rather than ask the opposite.
The reason why, and this is all theory, is that those that are open to new ideas end up having two ramifications which might be related - they absorb information faster, better and broader and they are more willing to change bad information with good information. Same theory goes for the P>J, I>E.
The tests are suppose to be a battery of sub-tests in order to avoid the aptitude problem. For example, the difference between doing basic word-problems (arithmetic) might depend on all sorts of outside factors, like job/training/education... but the likelyhood of being able to do those at the 130 IQ level and the ability to memorize and repeat/transform a string a numbers is unlikely - most major tests will have 2-5 major groupings of tests, each with 2-5 under that. In the WAIS, those two are part of the same 'verbal' category, as I believe the contradictions and common sense parts are. (Can google WAIS IQ to get a boatload of information on the test.)
The ability to do all of these things well comes down to the rate at which you can absorb and transform information - both from your past and from the present. This generally prefers people who self-test to N.
As for the T I said before, thinking about it better, you are probably right, because the T/F axis does not examine the ability to think or feel, but the preference on a more thinking or more impulsive way of reacting to external stimulis. So it doesn't have to be related with either IQ or intelligence
It just depends on the factors being measured and the tests being used. Using normal distribution, ie: FFM/Big Five, the most significant trait/sub-trait is openness (to ideas, as a sub trait). MBTI does create a few artifacts on the other three dimensions, so while it is correct to say that T/I/P have higher IQs, it's not as clear how accurate it is (or relevent.)
However, it has been shown that it can be significant inside MBTI and outside... But it isn't exactly agreed upon.
A lot of it has to do with distribution, arguments over mean/average, blah blah. Outside of just being a curiousity, I don't consider the I/T/P (and equiv FFM traits, although N- shows influence on some sub-traits in IQ) very important in understanding it... this is very notable when you break down into the 25 or so sub-traits that FFM tends to use and only one sub-trait significantly stands out.
I should mention that despite how I may sound, I have real problems with IQ. They have become exceptionally linear (strong correlations to pure academic testing, which proceeds into college, then into good jobs based upon that alone) and I fear are heading stronger that way due to the need to validate the tests against external factors.
And of course, all of this is pretty topical. Self-directed tests should never correlate that strongly against something that challenges the person... and the correlations across the board tend to be weaker than people think. In a way, the personality-> IQ correlation is akin to asking a person "Are you smart?" in a round about way.
