E-411 PRMA

Lecture 13 Intelligence and Factor Analysis

Christopher David Desjardins

Motivation for the CHC model

  • Cattell proposed a theory of intelligence that consists of crystallized and fluid intelligence

  • crystallized intelligence, Gc, ability to use skills, experience, & knowledge (e.g. retrieving and applying information)

  • fluid intelligence, Gf, capacity for solving novel problems

  • Horn added additional factors Gs

  • Carroll proposed a 3 layered hierarchical view with g on the top, then factors similar to Cattell's and Horn's, and specific factors depending on the second-level factors

Carroll's model

CHC model

McGrew and colleagues tried to reconcile these approaches

Proposed a modified Carroll model without a g

Omission of g based on the lack of utility for their needs

McGrew calls for adoption of the CHC model and made data available here

CHC model

There are 10 broad-stratum abilities

  • fluid, crystallized, quantitative knowledge, reading/writing ability, short-term memory, visual processing, auditory processing, long-term storage/retrieval, processing speed (perform automatic cognitive tasks, couple minutes), and decision time/speed (reaction time, couple seconds)

There are 70 narrow-stratum abilities

Information-processing framework

  • Rather than using factor analysis to derive "what" intelligence is, examine how information is processed

  • simultaneous processing, integration at once

  • successive processing, sequential integration

  • Extant tests do not take into account problem solving strategy

  • Sternberg proposed successful intelligence, how well we adapt, share, shape, select environments that confirm to personal and societal success

How to measure intelligence?

  • Myriad of tasks developed depending on age of testtaker
  • Infants focused on sensorimotor skills; shift towards verbal and performance abilities as children age
  • Mental age has fallen out of favor
  • For children, tests often used for school placement
  • Adults, test more diverse and typically used clinically

Problems in intelligence

  • Nature vs. nurture

    • Preformationism & predeterminism - slave to your genes
    • Twin studies
    • Interactionist view - unlimited potential

  • Stability

    • Young adult intelligence most important predictor of intelligence of older adults
    • Aging, physical/mental health, medications - confounders
    • "Early ripe, early rot"

More problems

Flynn effect - rise in intelligence test scores expected to occur on a normed intelligence test from the date the test was first normed.

TED Talk by Flynn

Personality \(\equiv\) Intelligence

Gender differences?

Family effects starting in the womb?

Cultural considerations? Culture-loading and creation of culture-fair tests

Stanford-Binet IQ test

  • Conceived by Binet to screen for children with developmental disabilities
  • Originally intelligence calculated as the ratio of mental age to true age
  • Deviation score, mean of 100 and standard deviation of 15
  • Measures fluid reasoning (fluid), knowledge (crystallized), quantitative knowledge, visual-spatial, and working memory (see Table 10 - 2)
  • Adaptive

Fifth edition

  • Full scale IQ from ten subtests

  • Subtest scores can be combined to get other composite scores (e.g. verbal score)

  • Standardized for the USA population aged 2 to 85

  • Manual reports high internal consistency, test-restest, and inter-rater reliability (though items with low reliability were pruned)

  • Criterion-Related VE from concurrent and predictive data

  • Use in a clinical population and factor structure unclear

Weschler intelligence tests

  • Age-appropriate tests for very young children, children, adults

  • All originally used a verbal, performance, and FSIQ, now only young children uses V and P scales

  • Many subtests and items specific just to Weschler's tests (Table 10 - 3)

  • Consists of core subtest and supplementals used to extract clinical information

  • Short forms exist, but discouraged

  • Good psychometric properties

Comparison and other tests

  • Both purport to measure intelligence

  • Highly correlated, differ by amount of g

  • Both work within the CHC model

  • Both represent gold standard

  • Different factor structures and definitions of intelligence

  • Kaufman tests focus on processing not structure

Group tests

  • USA army developed tests for recruits in WWI

  • Alpha, those who could read, Beta, those who couldn't

  • Assigned duty and service based on performance

  • Tests used in post-war because they were much cheaper

  • Later, Army General Cliassification Test and Armed Service Vocational Aptitute Battery

  • Also used in the schools in the USA for placement (not as much now)