some tests of general intelligence
Four examples of general intelligence tests, from a Scientific American article by L. Gottfredson
(1998).
The gF factor emerges from such tests in the form of tightly correlated
performance: a subject who does well on one "high-g" test is
likely to do well also on others.
The usual degree of correlation between scores in high-g tasks — about
0.75 — implies that about 50% of the
variance in subjects' performance is explained by positing the existence of
a common general intelligence factor, gF.
on the value of the three-stratum model
"This so-called three-stratum model affords freedom from otiose arguments
about being 'for or against the general factor.'
The three-stratum account has been called a theory. It is not.
And it is not a model of the human cognitive architecture: rather,
it is a taxonomy or model of test variances and co-variances. The taxonomy
does not explain human intelligence differences, it describes them."
Human intelligence differences: a recent history,
I. J. Deary, Trends in Cognitive Sciences 5:127-130 (2001).
"The realization that general intelligence and working
memory might be closely linked brings together two concepts with
massive psychometric evidence on the one hand and massive cognitive and
neuroscience evidence on the other."
Human intelligence differences: towards a combined
experimental-differential approach,
I. J. Deary, Trends in Cognitive Sciences 5:164-170 (2001).
slide 4
the centrality of working memory
The cognitive architecture used to examine associations between working
memory and psychometric intelligence (from Deary, 2001b).
"General fluid intelligence and working memory both reflect 'the
ability to keep a representation active, particularly in
the face of interference and distraction'."
P. C. Kyllonen and R. E. Christal, Reasoning ability is (little
more than) working memory capacity?! Intelligence 14, 389-433 (1990).
slide 5
gF and working memory
slide 6
flexible interference avoiding (the 3-back match task)
The 3-back working memory task, in which
the subject must determine whether or not the present target was shown
precisely three items back.
Gray et al. (2003) used this task to demonstrate
that the importance of WM for general intelligence stems from its
support of flexible, interference-avoiding manipulation of
data.
slide 7
flexible interference avoiding (the 3-back match task)
Lure foils: matched a stimulus 2, 4, or 5 items ago in the
sequence, and thus presented substantial interference with the 3-back task
demand.
Non-lure foils were all other non-target trials (items never seen
before, or matches seen 1-back or 6-or-more-back).
Behavioral findings:
- Lure foils were much less accurately rejected by subjects than were
non-lure foils.
- Lure-foil accuracy correlated positively with gF, and remained
significant after partialling out individual differences in non-lure
foil accuracy or target accuracy.
slide 8
flexible interference avoiding (the 3-back match task)
Lure foils: matched a stimulus 2, 4, or 5 items ago in the
sequence.
Non-lure foils were all other non-target trials (items never seen
before, or matches seen 1-back or 6-or-more-back).
fMRI findings:
- High-gF subjects showed a markedly greater increase in fMRI
signal in left lateral PFC on
lure trials than did low-gF subjects.
- Although gF did predict behavioral accuracy also on
non-lure and target trials, its correlation with brain
activity during these trials was much weaker.
In sum, only
working memory processes that were tied directly to
interference control were supported by gF-related brain activity.
slide 9
prefrontal cortex: size comparison
 |
 |
| human |
monkey |
slide 10
analogy: the bull's eye
If working memory is the answer, what is the question?
Analogy.
A schematic summary of a meta-analysis of scores of results quantifying
the relationships among various intelligence tests.
Raven's Progressive Matrices and other analogy
tests turn out to be central in any numerical assessment of general
cognitive function.
slide 11
a computational analysis of a Raven's task
To solve matrix analogy problems such as this one, it is necessary to
carry out the following computations:
- identify the primitive elements;
- determine the relevant features;
- describe each of the given items in terms of these;
- infer the rules that hold for each row and column;
- deduce a description of the missing item;
- scan the candidate answers for an item that fits best the expected
missing item.
Much as in scene interpretation, the information that is critical to this
analogy task has two components: the roles of the elements,
and their positions.
is the general intelligence factor relevant to real life?
From the
Scientific American article by L. Gottfredson
(1998):
... Is there indeed a general mental ability we commonly
call "intelligence," and is it important in the practical affairs of
life? The answer, based on decades of intelligence research, is an
unequivocal yes.
... No matter their form or content, tests of mental skills invariably
point to the existence of a global factor that permeates all
aspects of cognition. And this factor seems to have considerable
influence on a person's practical quality of life. Intelligence as
measured by IQ tests is the single most effective predictor known of
individual performance at school and on the job.
... The effects of environment on intelligence fade rather than
grow with time.
... Environments shared by siblings have little to do with IQ.
the fundamental findings of behavioral genetics
The breakdown of the respective contributions to the variance in
g in a given population is:
| genes: |
~ 50% |
| shared environment: |
~ 0-10% |
| unique environment: |
~ 40-50% |
Source:
S. Pinker, The Blank Slate: the modern denial of human nature, Chapter 19 (Viking, Harmondsworth, UK, 2000)
slide 14
general intelligence
Correlation of IQ Scores with occupational achievement suggests that
g reflects an ability to deal with
cognitive
complexity.
Scores also correlate with some social outcomes (the percentages apply to
young white adults in the U.S.).
Scores also correlate with some social outcomes (the percentages apply to
young white adults in the U.S.).
Scores also correlate with some social outcomes (the percentages apply to
young white adults in the U.S.).
Scores also correlate with some social outcomes (the percentages apply to
young white adults in the U.S.).
exceptional achievers
Lubinski et al. (2006) found that SAT scores delivered early predict well both
long-term achievement and life satisfaction.
Their study, titled
Tracking Exceptional Human Capital Over Two Decades,
compared
380 SAT takers under age 13 who scored in the top 0.01% of their age
group and were surveyed 20 years later,
and
586 graduate students who had been enrolled in a top-ranked
engineering, mathematics, or physical science program, and given
the same survey 12 years later.
This chart describes the career choices of the members of the two groups.
slide 21
implications for policy?
slide 22
