Schneider, W. J., Mayer, J.
D., & Newman, D. A. (2016). Integrating
Hot and Cool Intelligences: Thinking Broadly about Broad Abilities. Journal of Intelligence, 4(1), 1.
This is an interesting, complex and condensed paper
.
Schneider, Mayer and Newman present the idea of
incorporating "hot" abilities alongside the "cold"
abilities already present in the CHC model.
What are
"cold" and "hot" cognitive abilities?
"Cold" cognitive abilities deal with
solving problems and performing tasks that are impersonal. Examples are tests in intelligence batteries –
"number recall", "picture completion"," vocabulary"
etc. – they all require impersonal information processing. "Hot" abilities, on the other hand,
concern
abilities involving the processing of highly charged and personally significant
information such as emotions, personality, and social relations.
Schneider and his colleagues examine two putative
abilities "emotional intelligence" (the ability to perceive, understand,
manage, and use emotions adaptively) and "personal intelligence" (the ability to
perceive, understand, and use information about one’s own and others’
personality adaptively).
In order for these abilities to be included in the CHC model they need to meet a
number of conceptual and empirical criteria. The
authors present the criteria and discuss the extent to which these two
constructs meet them.
I focus here on conceptualizations of broad
abilities, and on the criteria for the inclusion of a proposed broad ability in
the CHC model. To note, Schneider and
his colleagues believe that these two proposed abilities have the potential to meet
these criteria. Those who want to read
more about the ways "emotional intelligence" and "personal
intelligence" meet these criteria – should read the paper (link above).
The following contain extensive quotes from the paper with small
additions of my own.
The authors write that in adding more and more abilities
to the model (any model, not only CHC)' there is danger that the model will include “more and more factors of less and
less importance".
When we attempt to classify animals, it is perfectly reasonable to talk about animals
that walk, animals that swim, and animals that fly. We can group animals by
continent, by social structure, by mating habits, by diet, or by the climate in
which they thrive. We can make an extremely large number of perfectly reasonable
distinctions. However, taxonomists have mostly focused on classifying animals
by their theorized evolutionary lineages, because doing so usefully and
succinctly organizes many phenomena about animals that would otherwise be
unexplained.
In a similar fashion, scholars have made all
sorts of distinctions among cognitive abilities. Abilities can be classified
according to sensory modalities (visual, auditory, olfactory, gustatory,
tactile, proprioceptive, and so forth), according to particular kinds of symbolic
content (verbal, numerical, and figural), by the rate at which a task can be
done, by the difficulty level of the task (speed vs. power). Some abilities are defined by
giving a single correct response, others by generating many correct answers (convergent vs. divergent
production). Some abilities require serial processing, others parallel
processing (simultaneous vs. successive
processing.
Thus the same test cluster can be classified in
different ways. For example, the test
cluster "hand movements", "number recall" and "word
order" in the Kaufman test can be classified as assessing sequential
processing, but it can also be classified as assessing short term memory
range. Another test cluster (for
instance, "triangles", "face recognition", "gestalt
closure", "spatial memory", "picture series" in the Kaufman
test) can be classified as assessing simultaneous processing. From another point of view it's possible to
classify it as assessing visual processing.
The authors write that most of these distinctions are
rationally derived and have intuitive appeal. In addition, they are often
grounded in robust findings from neuropsychology and/or cognitive psychology.
However, at some point, the number of distinctions in the taxonomy makes the
system too difficult to use and impossible to verify empirically. Theorizing,
like all forms of creativity, thrives under properly balanced constraints.
In the factor-analytic
tradition of intelligence theories, the constraints come in the form of
correlation coefficients that must vary according to a particular structural
model. However, obtaining a well-fitting structural model is only one step in a
long process of construct validation research. Although factor analysis can
tell us that broad abilities exist, it does not
easily reveal their number and nature.
What exactly
is a broad cognitive ability?
The authors suggest a number of ways to conceptualize
a "broad ability":
1. Broad abilities are widely applicable
Broad abilities are useful in various aspects
of life and in various situations. For instance,
fluid ability, the ability to apply abstract reasoning and to solve new problem, is broad in the
sense that it is useful in many life situations. The ability to knit is an analogy for a
narrow ability, in the sense that its usefulness is not that broad.
2. Broad abilities are essential
precursors to other abilities
The authors argue that the ability to see is broad in the
sense that it is an essential ingredient in many other abilities. The ability
to whistle is a precursor to far fewer abilities. Working memory capacity is a
broad ability in this sense, because it is a relatively unified domain-general
capacity that is an essential ingredient in many other abilities—fluid
reasoning in particular. Fluid reasoning is supported by attention and memory as well as
the various perceptual processing capacities. That is, if one can hold one’s
attentional focus on a wide array of information in working memory (whether it
be external sensory information shaped by perceptual processing or internal
knowledge retrieved from long-term memory) one is better able to discern useful
patterns within this information
3. Broad abilities are conceptual
categories.
Factor analysis alone
cannot distinguish between variables that load together because of an
underlying ability or variables that load together due to
arbitrary reasons. For
example, a factor analysis of physical abilities among Norwegians might reveal
a small factor consisting of rifle shooting accuracy and cross-country skiing
endurance. Before we theorize about an ability construct that explains the link
between these two disparate skills, we would do well to consider the popularity
of the Biathlon event in Norway.
This is an example of a social habit that may cause undue classification
of tests and narrow abilities to a broad ability. It's better to define a broad ability as a
cluster of tests that group together due to a real ability underlying them; a
cluster of tests that are causally related.
4. Broad Abilities are
Collective Properties
A collective property is a feature that emerges from the measurement
of many objects. For example, the sample mean is not in the sample, it is a collective property of the sample. Even though some numbers in the
sample might have the same value as the sample mean, they are not conceptually
the same thing as the sample mean. Racial diversity in the classroom is not found in any
particular student but is a collective property of the group. To say that a
particular student adds to the racial diversity of a classroom is not so much a
statement about the student but about the current composition of the class as a
whole.
Broad abilities may be
collective properties in the sense that the cluster of tests measuring them has
qualities that do not necessarily exist fully in each of the specific tests. That is, the cluster has an emergent quality beyond
what each specific test comprising it has.
5. Broad abilities are causal
systems.
Traditional
hierarchical factor-analytic models imply
that narrow abilities have no relationships with each other after controlling
for broader abilities. In a causal system model, narrow abilities influence
each other directly in sometimes complex networks of association. Thus, broad
abilities are not so much “things” (i.e., discrete entities) as they are relatively
integrated portions of a large network of abilities.
The
ability to store information in working memory is a broad ability in this sense
because it refers to a collection of interdependent narrow abilities. For example,
it is reasonably clear that there are different subsystems of working memory
that handle different kinds of sensory codes (the phonological loop, the
visuospatial sketchpad) and different mechanisms for updating the contents of
short-term memory. The various components of
working memory form a functional unity in that they facilitate complex mental
functions such as planning and reasoning
6. Broad Abilities are Sets of “Entangled” Narrow Abilities
Test in intelligence batteries measure narrow
abilities. But a test is never a
"clean" measure of a narrow ability.
This is simply impossible, since the items must be presented in some
modality (visually, aurally,
visually+ aurally etc.) This doesn't mean that all tests that present
items visually measure visual processing, or that all test that present the
items aurally
measure auditory processing. It means
that these tests have a visual or auditory ingredient. For example, number recall has an auditory
ingredient, but it does not assess auditory processing. It assesses short term memory, since the
auditory processing component is small
compared to the short term memory component.
Thus, factors that influence a child's
functioning on one broad ability can affect his functioning on another broad
ability to a lesser degree. In this
sense, the abilities are "entangled" and the boundaries between them
are fuzzy.
What considerations
do researchers make when they decide to include a new broad ability in the
model?
The authors suggest that a suggested broad ability
has to meet these criteria:
1. The Content Domain of the Putative
Broad Ability Must Be Clearly Laid Out
Researchers
should be able to communicate about the ability sufficiently well that their
findings can inform subsequent research about the ability. In particular,
different teams of researchers should be able to create new measures of the
construct based on a shared understanding of what it is.
I think
the term "dyscalculia" suffers from this problem. Dyscalculia' of course' is not a cognitive
ability but a syndrome. Due to lack of a
unitary, agreed upon definition of dyscalculia, different studies assess it
using tests that are essentially and conceptually different. Different studies classify people to groups
of dyslcaulics and non-dyscalculics based on widely different criteria. This makes it difficult for research to progress.
2. The New Construct Can Be Measured With Performance Tests
Performance tests are tests in which one
performs an activity or a task (defines a word, constructs a model out of
blocks), as opposed to tests that require a person to answer theoretical
questions about his ability in the area assessed (rank your vocabulary on a
scale of 1 to 5).
3. The New Construct Has Properties Similar to Established
Cognitive Abilities
All well-established cognitive abilities thus far
identified develop with age and correlate positively with each other. It is
extremely likely that newly identified abilities will do likewise.
4. The New Construct Should Be Measurable With Tests That
Exhibit Convergent and Discriminant Validity.
Convergent
validity refers to the degree to which two measures of
constructs that theoretically should be related, are in fact related. Convergent validity can be established if two
similar constructs correspond with one another, while discriminant validity
applies to two dissimilar constructs that are easily differentiated. Tests measuring the same ability are supposed
to correlate higher with each other than with tests measuring other abilities.
The psychometric
model of intelligence, that includes broad abilities, each measured by a number
of narrow abilities, is the result of factor analysis of hundreds of cognitive
tests. Such analysis yields clusters of
tests that behave in the same way (a child who scores high on one of the tests
in a cluster usually scores high on other tests in the cluster, but not
necessarily on tests that belong to a different cluster). Each such cluster is conceptualized as a
"cognitive ability" and is named after the common feature of tests
that belong to it. But factor analysis
is not entirely objective. Its results
are affected by decisions the researchers make:
a. The number of tests included in the factor analysis. If the analysis is performed, say, on 10
tests, 2 factors may manifest. But if the analysis is performed on 100
tests, more factors may manifest. The number
of tests included in the analysis affects the number of factors that result
from it.
B. The variety of tests included in the factor analysis: if the analysis includes, for example, tests
measuring motor and visoomotor skills, such an ability may manifest. If no such tests are included in the factor
analysis, no such ability will manifest.
5.
The New Construct Predicts Important Outcomes Even After Accounting for More
Established Cognitive Abilities
By
including groups of overly similar tests in a study (for instance, number
recall, letter recall, color recall, picture recall etc.), it is possible to
make it appear that a test battery has whatever broad abilities one wants it to
have. Abilities “discovered” this way are not true abilities but what Cattell
called “bloated specifics”. A bloated specific differs from a true ability in
that it does not correlate with any external criteria worth studying. Thus,
before a new ability is to be admitted into a theoretical taxonomy, it is not
enough that a factor analysis supports its existence. That new factor must
predict something that matters, above and beyond the other facets of
intelligence.
6. The New Construct
Should Be Linked to Specific Neural Modules That Evolved to Help Humans Survive
and Reproduce
Without
this criterion, almost any useful skill could be dubbed a new intelligence. For
example, computer
skills are becoming increasingly important . Is there a digital intelligence?
Answering yes stretches the meaning of intelligence far beyond its traditional
boundaries. Computer skills have no plausible evolutionary history. Natural
selection has not had time to evolve special cognitive functions that help
humans use computers to survive and reproduce. The intelligent use of computers
relies exclusively on repurposing cognitive functions that evolved previously
to solve other problems of survival and reproduction.
No comments:
Post a Comment