Will there every be a test that can really measure IQ

[Tom McCurdy]

Do you think it will ever be possible to ever gain a true grip on someones intelligence. I mean I know people who get 4.0 (gpa scale) who get good scores on ACT and SAT who are just plain stupid in many areas. My personal feeling is that there will for sure never be a written test alone that could ever measure someone's smarts. The only way would to really understand the human mind, which we are still somewhat far from doing and actually perform some sort of test on the brain itself.

 

[MathematicalPhysicist]

it's like asking can we prove a definition of a word that we have defined by ourself.
there cannot be a direct measure of iq because we still don't know a whole lot about our brain.

 

[master_coda]

I think that even the idea that intellgence can be mapped to some sort of number scale in a meaningful way is absurd, no matter how much we learn about the human brain.

Even if there was a case where you were really sure someone is smarter than someone else, what good would that information be?

 

[hitssquad]

Learning disabilities and IQ

Even if there was a case where you were really sure someone is smarter than someone else, what good would that information be?

Assessment of general mental ability is useful in any situation where an assessment of relative general mental competence would be useful, such as juror selection, vehicle operations endorsement, or the diagnosis and accomodation of learning disabilities.

 

[master_coda]

Assessment of general mental ability is useful in any situation where an assessment of relative general mental competence would be useful, such as juror selection, vehicle operations endorsement, or the diagnosis and accomodation of learning disabilities.

So if someone is less intelligent, they must be less competent? Or less able to make objective judgements? Or be more difficult to teach?

If you are trying to determine if someone is capable of something, then if all you have is an intelligence test then you have almost no information at all. On the other hand, if you have a test specific to the capability you are testing, then what do you need an intelligence test for?

Any test on a very general category isn't particularly useful for making specific judgements, since there is generally far too much variation within the category to really be certain of a specific ability.

 

[wuliheron]

One of the problems with IQ tests is that it is impossible to separate our ability to emote from our ability to think. We decide what is and is not meaningful, what is and is not intelligent, and this changes over time. Hence IQ tests will always be improving and changing as well. Where you draw the line and say, "This is accurate" is a subjective decission.

 

[hitssquad]

How much g enters into 'specific' tasks

If you are trying to determine if someone is capable of something, then if all you have is an intelligence test

...IQ test...

then you have almost no information at all.

You have a relative measure of the person's general mental ability. Arthur Jensen makes the point on pages 271, 282, 283, etc., of The g Factor that g enters into "specific" tasks more than may commonly be made account for.

• Meta-analyses of hundreds of test validation studies have shown that the validity of a highly g-loaded test with demonstrated validity for a particular job in a particular organizational setting is generalizable to virtually all other jobs and settings, especially within broad job categories.

• Virtually every type of work calls for behavior that is guided by cognitive processes. As all such processes reflect g to some extent, work proficiency is g loaded. The degree depends on the level of novelty and cognitive complexity the job demands. No job is so simple as to be totally without a cognitive component. Several decades of empirical studies have shown thousands of correlations of various mental tests with work proficiency. One of the most important conclusions that can be drawn from all this research is that mental ability tests in general have a higher success rate in predicting job performance than any other variables that have been researched in this context, including (in descending order of average predictive validity) skill testing, reference checks, class rank or grade-point average, experience, interview, education, and interest measures. [22] In recent years, one personality constellation, characterized as "conscientiousness," has emerged near the top of the list (just after general mental ability) as a predictor of occupational success.

• One study, [24a,b] based on 24,000 subjects in training for thirty-seven diverse jobs, used the method of correlated vectors, in which the vector of g loadings of the ten ASVAB subtests was correlated with the vector of ten validity coefficients of each of the ASVAB subtests for predicting training success. The rank-order correlation between the two vectors was +.75; this correlation increased to +.98 when the effect of the subtests' differing reliabilities was statistically controlled. In brief, the larger a test's g loading, the better it predicts training success. The study was replicated [24b] on a sample of 78,000 subjects across 150 different job training courses, yielding a correlation of +.96 (controlling the subtests' reliabilities) between the subtests' g loadings and their validity coefficients. 25

On the other hand, if you have a test specific to the capability you are testing, then what do you need an intelligence test for?

A test specific to the capability you are testing may be highly loaded on experience. It has been consistently shown that the value of previous experience in any job fades over time with respect to the value of g. In other words, any given person with a master's degree in Library Science will tend to be outperformed as a reference librarian by any given person with no education beyond a high school diploma who also has a higher IQ, as long as the job lasts long enough for the latter person to crystallize his intelligence in the library milieu.

• Residual Validity of Amount of Education. Some employers use number of years of education or other educational credentials as a basis for selecting workers. These measures are usually valid predictors, though seldom as valid as tests of general ability, except for a specialized job where specific educational qualifications are intrinsic and essential. Educational credentials derive almost all of their predictive validity from their substantial correlation with g. In general, the number of years of education, for example, is correlated .60 to .70 with IQ. Since the applicants for many jobs are self-selected in terms of educational qualifications, the true correlation of educational level with ability is even higher. It may seem surprising, but in most selection situations the validity coefficient for years of education is typically not more than .10 to .15 versus a g validity of .40 to .50 in similar situations. Further, the incremental validity of education over a measure of g is practically nil. This is largely because the variance in educational level (measured by the highest school grade completed) is less than the variance in actual ability in the applicant pool and also because there is a wide range of ability at every level of education. On tests of general ability, such as IQ, and even on tests of scholastic achievement, there is considerable overlap between the score distributions of high school graduates and college graduates [36]

Ibid. p291.

Any test on a very general category isn't particularly useful for making specific judgements, since there is generally far too much variation within the category to really be certain of a specific ability.

This hypothesis is testable. It has been tested, and it remains unsubstantiated. See The g Factor.

 

[master_coda]

Well, I certainly don't have the experience or motivation to argue with that kind of effort. Automation is part of what I do, so any kind of useful statistics are good statistics.