Double click on images to enlarge
Showing posts with label factor analysis. Show all posts
Showing posts with label factor analysis. Show all posts
Sunday, November 3, 2013
Posted with BlogsySunday, February 20, 2011
Research brief: WAIS-IV US-Canadian factor and score comparability
The transportability of the meaning of an intelligence test batteries composite scores across countries/cultures is important when a test is originally developed and normed in one country and is then adapted and used in a second country.
Bowden et al (2010) recently investigated the factorial invariance of the WAIS-IV across US and Canadian samples. The results are summarized in the abstract below (click to enlarge). The WAIS-IV was found to measure the same theoretical constructs across the two countries. However, the reported difference in latent mean factor intercepts indicated that the WAIS-IV provides higher scores with Canadian subjects. The need for Canadian norms are suggested.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner
Bowden et al (2010) recently investigated the factorial invariance of the WAIS-IV across US and Canadian samples. The results are summarized in the abstract below (click to enlarge). The WAIS-IV was found to measure the same theoretical constructs across the two countries. However, the reported difference in latent mean factor intercepts indicated that the WAIS-IV provides higher scores with Canadian subjects. The need for Canadian norms are suggested.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner
Generated by: Tag Generator
Tuesday, February 8, 2011
Carroll-like exploratory factor analysis of the WISC-IV Integrated: CHC interpretation
An excellent Jack Carroll-like exploratory factor analysis of the WISC-IV Integrated at Joel Schneider's blog.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition WISC-IV factor analysis
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition WISC-IV factor analysis
Generated by: Tag Generator
Wednesday, January 26, 2011
Blending neuropsychological and CHC psychometric IQ approaches to psych testing
Here is more on my take on how neuropsychological and CHC-based psychometric approaches to assessment can be understood and potentially blended. This is material from my keynote presentation at the Australian Neuropsychology Conference (click here for more information and a link to the PPT of the entire presentation).
Note: Images are embedded in this post. It should be possible to enlarge them by double clicking on each. If that does not work I would suggest you go to the link above and see the entire PPT show, which you can download for free.
CHC-based neuropsychologists have integrated CHC theory into assessment practice based largely on theoretical, non-CHC empirical research, or logical analysis. Empirical CHC-based neuropsychological assessment research has been sparse. The CHC psychometric-neuropsychological assessment gap is understandable given the two respective assessment models and distinct historical backgrounds. Hoelzle (2008) noted that: (a) the psychometric concept of g has had minimal clinical utility in neuropsychological assessment and theory, (b) neuropsychological assessment has been primarily atheoretical while psychometric research has been searching for the “holy grail” theoretical structural model of intelligence since the days of Spearman, and (c) psychometric models have been primarily focused on internal structural validity while neuropsychological assessment has focused more on the practical questions regarding the ability to differentiate between neurological and normal conditions. McGrew (2010a) described these different approaches as a difference between vertical (factor analysis) trait-oriented psychometric models and more horizontal (multiple regression) functional/pragmatic external prediction neuropsychological models. These two approaches are demonstrated visually below in the two slides.
As result of its pragmatic functional heritage, many neuropsychological assessment tests are mixed measures of multiple CHC domains. In the language of psychometric CHC Theory, many neuropsychological tests are factorially complex measures with considerable construct-irrelevant variance (when measurement of pure CHC constructs is the criterion). This contrasts with the strong emphasis in contemporary CHC intellectual assessment on developing tests that primarily measure one single CHC construct, purged as much as possible of non-relevant construct variance.
In the absence of CHC-grounded neuropsyhological assessment research and a consensus neuropsychological assessment model, psychometric (e.g., Flanagan, Alfonso, Ortiz & Dynda, 2010; McGrew, 2010a) , neuropsychological (Strauss et al., 2006), and “blended” neuropsychological+psychometric oriented scholars (e.g., Hale & Fiorello, 2004; Miller, 2007, 2010) have, at times, presented significantly different interpretations of neuropsychological or intelligence tests. For example, more neuropsychologically oriented researchers have described the copy and recall components of Rey-Osterrieth Complex Figure (ROCF) test (Meyers & Meyers, 1995, 1996) as measuring planning and organizational abilities, visual memory, visual perception, constructional abilities, motor or visual-motor ability, episodic memory, and incidental learning (Hale & Fiorello, 2004; Straus et al., 2006). Hints of CHC terminology are present in some of these terms but there is no direct mapping to CHC Theory.
Conversely, starting with the results of Hoezle’s (2008) CHC-organized mini-Carroll like meta-analysis of the results of secondary factor analysis of 77 datasets that included neuropsychological and other ability measures (this is one of the best and most comprehensive dissertations I have ever read), McGrew (2010a) suggested that the copy component of the ROCF was likely a mixed measure of Gv-SR (narrow CHC ability of spatial relations in the broad Gv or visual-spatial domain), Gp-P2 (narrow CHC ability of finger dexterity in the broad Gp or psychomotor domain), with possible visual memory (Gv-Mv) involvement. As for the ROCF recall component, McGrew (2010a) suggests that it likely measures a mixture of Gv-MV (narrow CHC visual memory in the broad Gv domain), Glr-M6 (narrow CHC free recall ability in the broad Glr or long-term storage and retrieval domain), with possible involvement of working memory (Gsm-MW). Miller (2010) provides the “blended” interpretation of the ROCF as primarily a measure of visual-perceptual organization, visual-spatial ability and visual-spatial memory (Gv) in particular. Obviously the three sets of ROCF interpretations mention a number of similar abilities, yet practitioners are likely left confused given the varying terminology and ability descriptions.
My CHC neuropsych summaries for the ROCF are presented below:
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition ROCF test Ray-Osterrith Complex Figure Test
Note: Images are embedded in this post. It should be possible to enlarge them by double clicking on each. If that does not work I would suggest you go to the link above and see the entire PPT show, which you can download for free.
CHC-based neuropsychologists have integrated CHC theory into assessment practice based largely on theoretical, non-CHC empirical research, or logical analysis. Empirical CHC-based neuropsychological assessment research has been sparse. The CHC psychometric-neuropsychological assessment gap is understandable given the two respective assessment models and distinct historical backgrounds. Hoelzle (2008) noted that: (a) the psychometric concept of g has had minimal clinical utility in neuropsychological assessment and theory, (b) neuropsychological assessment has been primarily atheoretical while psychometric research has been searching for the “holy grail” theoretical structural model of intelligence since the days of Spearman, and (c) psychometric models have been primarily focused on internal structural validity while neuropsychological assessment has focused more on the practical questions regarding the ability to differentiate between neurological and normal conditions. McGrew (2010a) described these different approaches as a difference between vertical (factor analysis) trait-oriented psychometric models and more horizontal (multiple regression) functional/pragmatic external prediction neuropsychological models. These two approaches are demonstrated visually below in the two slides.
As result of its pragmatic functional heritage, many neuropsychological assessment tests are mixed measures of multiple CHC domains. In the language of psychometric CHC Theory, many neuropsychological tests are factorially complex measures with considerable construct-irrelevant variance (when measurement of pure CHC constructs is the criterion). This contrasts with the strong emphasis in contemporary CHC intellectual assessment on developing tests that primarily measure one single CHC construct, purged as much as possible of non-relevant construct variance.
In the absence of CHC-grounded neuropsyhological assessment research and a consensus neuropsychological assessment model, psychometric (e.g., Flanagan, Alfonso, Ortiz & Dynda, 2010; McGrew, 2010a) , neuropsychological (Strauss et al., 2006), and “blended” neuropsychological+psychometric oriented scholars (e.g., Hale & Fiorello, 2004; Miller, 2007, 2010) have, at times, presented significantly different interpretations of neuropsychological or intelligence tests. For example, more neuropsychologically oriented researchers have described the copy and recall components of Rey-Osterrieth Complex Figure (ROCF) test (Meyers & Meyers, 1995, 1996) as measuring planning and organizational abilities, visual memory, visual perception, constructional abilities, motor or visual-motor ability, episodic memory, and incidental learning (Hale & Fiorello, 2004; Straus et al., 2006). Hints of CHC terminology are present in some of these terms but there is no direct mapping to CHC Theory.
Conversely, starting with the results of Hoezle’s (2008) CHC-organized mini-Carroll like meta-analysis of the results of secondary factor analysis of 77 datasets that included neuropsychological and other ability measures (this is one of the best and most comprehensive dissertations I have ever read), McGrew (2010a) suggested that the copy component of the ROCF was likely a mixed measure of Gv-SR (narrow CHC ability of spatial relations in the broad Gv or visual-spatial domain), Gp-P2 (narrow CHC ability of finger dexterity in the broad Gp or psychomotor domain), with possible visual memory (Gv-Mv) involvement. As for the ROCF recall component, McGrew (2010a) suggests that it likely measures a mixture of Gv-MV (narrow CHC visual memory in the broad Gv domain), Glr-M6 (narrow CHC free recall ability in the broad Glr or long-term storage and retrieval domain), with possible involvement of working memory (Gsm-MW). Miller (2010) provides the “blended” interpretation of the ROCF as primarily a measure of visual-perceptual organization, visual-spatial ability and visual-spatial memory (Gv) in particular. Obviously the three sets of ROCF interpretations mention a number of similar abilities, yet practitioners are likely left confused given the varying terminology and ability descriptions.
My CHC neuropsych summaries for the ROCF are presented below:
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ testing IQ scores CHC intelligence theory CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology psychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition ROCF test Ray-Osterrith Complex Figure Test
Sunday, January 2, 2011
Hot and cold CHC intelligence abilities--Gf,Gc,Gv hot--Ga,Glr cold
Interesting article in the journal Intelligence reviewing the state-of-the-art of factor analysis practices for identifying the g (general intelligence) factors. Abstract is below. Of interest is the use of the CHC framework to classify the type of broad CHC factor indicators found in the research synthesis.
Not unexpectedly, Gf, Gc, and Gv were found most often in IQ factor analysis research, followed by Gq, Gs and Gsm. Abilities that appear underrepresented in IQ factor analysis g research are the domains of Glr and Ga.
However, a couple of major caveats. The literature review was primarily adult samples. There has been considerable factor analysis activity with tests in childhood and adolescent samples that might increase the proportion of Glr and Ga indicators. Also, the authors did not include journals favored by those doing research in school psychology, special education, and speech and language---fields of study that most likely have published more studies in the under-represented CHC domains.
Never-the-less......the general trends are not surprising.
Clicking on images should take you toe larger versions.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition Gq quantitative abilities quantitative knowledge quantitative reasoning Gf fluid intelligence Gc crystallized intelligence Gsm Short-term memory Gll long-term retrieval Ga auditory processing Gv visual processing visual-spatial Gs processing speed factor analysis general intelligence g
Not unexpectedly, Gf, Gc, and Gv were found most often in IQ factor analysis research, followed by Gq, Gs and Gsm. Abilities that appear underrepresented in IQ factor analysis g research are the domains of Glr and Ga.
However, a couple of major caveats. The literature review was primarily adult samples. There has been considerable factor analysis activity with tests in childhood and adolescent samples that might increase the proportion of Glr and Ga indicators. Also, the authors did not include journals favored by those doing research in school psychology, special education, and speech and language---fields of study that most likely have published more studies in the under-represented CHC domains.
Never-the-less......the general trends are not surprising.
Clicking on images should take you toe larger versions.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition Gq quantitative abilities quantitative knowledge quantitative reasoning Gf fluid intelligence Gc crystallized intelligence Gsm Short-term memory Gll long-term retrieval Ga auditory processing Gv visual processing visual-spatial Gs processing speed factor analysis general intelligence g
MDS analysis of WISC-IV
It is no secret that I'm a big fan of multidimensional scaling (MDS--especially Guttman's Radex) model as a supplement to factor analysis of cognitive tests. While going thru some of my e-files I found a recent 3D MDS analysis of the WISC-IV. Below is the abstract and final 3D model. Clicking on images should take you to a larger version of the image.
For those interested, the content/stimulus dimension of my proposed cognitive ability assessment design and interpretation matrix is due to my application of MDS to data from the WJ III and the various Wechsler batteries. The complete "beyond CHC theory" presentation can be found at a prior post.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition WISC-IV Wechsler intelligence batteries WJ III WJ III NU MDS Guttman Radex model
For those interested, the content/stimulus dimension of my proposed cognitive ability assessment design and interpretation matrix is due to my application of MDS to data from the WJ III and the various Wechsler batteries. The complete "beyond CHC theory" presentation can be found at a prior post.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition WISC-IV Wechsler intelligence batteries WJ III WJ III NU MDS Guttman Radex model
Wednesday, December 29, 2010
Research brief: WAIS IV factor structure article
Click on image to enlarge. A related publication (IAP AP101 Report #2) regarding the WAISIV factor structure can be found here.
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence intelligence testing Atkins cases ICDP blog psychology school psychology neuropsychology Forensic psychology IQ tests IQ scores IQs Corner WAIS-IV factor analysis CHC theory
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence intelligence testing Atkins cases ICDP blog psychology school psychology neuropsychology Forensic psychology IQ tests IQ scores IQs Corner WAIS-IV factor analysis CHC theory
Thursday, December 9, 2010
Research briefs: Factor analysis of SB5 and a neuropsychological test battery
As per usual when I make a research byte/brief post, if anyone would like to read the original article, I can share via email---with the understanding that the article is provided in exchange for a brief guest post about it's contents. :) (contact me at iap@earthlink.net if interested). Also, if figure/images are included in the post, they can usually be made larger by clicking on the image.
Doweling, N. M., Hermann, B., LaRue, A., & Sager, M. A. (2010). Latent Structure and Factorial Invariance of a Neuropsychological Test Battery for the Study of Preclinical Alzheimer's Disease. Neuropsychology, 24(6), 742-756.
Abstract
Objective: To examine the latent structure of a test battery currently being used in a longitudinal study of asymptomatic middle-aged adults with a parental history of Alzheimer's disease (AD) and test the invariance of the factor solution across subgroups defined by selected demographic variables and known genetic risk factors for AD. Method: An exploratory factor analysis (EFA) and a sequence of confirmatory factor analyses (CFA) were conducted on 24 neuropsychological measures selected to provide a comprehensive estimate of cognitive abilities most likely to be affected in preclinical AD. Once the underlying latent model was defined and the structural validity established through model comparisons, a multigroup confirmatory factor analysis model was used to test for factorial invariance across groups. Results: The EFA solution revealed a factor structure consisting of five constructs: verbal ability, visuospatial ability, speed & executive function, working memory, and verbal learning & memory. The CFA models provided support for the hypothesized 5-factor structure. Results indicated factorial invariance of the model across all groups examined. Conclusions: Collectively, the results suggested a relatively strong psychometric basis for using the factor structure in clinical samples that match the characteristics of this cohort. This confirmed an invariant factor structure should prove useful in research aimed to detect the earliest cognitive signature of preclinical AD in similar middle aged cohorts.
Williams, T. H., McIntosh, D. E., Dixon, F., Newton, J. H., & Youman, E. (2010). A CONFIRMATORY FACTOR ANALYSIS OF THE STANFORD-BINET INTELLIGENCE SCALES, FIFTH EDITION, WITH A HIGH-ACHIEVING SAMPLE. Psychology in the Schools, 47(10), 1071-1083.
Abstract
The Stanford–Binet Intelligence Scale, Fifth Edition (SB5), is a recently published, multidimensional measure of intelligence based on Cattell–Horn–Carroll (CHC) theory. The author of the test provides results from confirmatory factor analyses in the technical manual supporting the five-factor structure of the instrument. Other authors have examined this factor structure through EFA using the standardization sample, and have not found evidence of a five-factor model. The purpose of the current study was to examine the internal construct validity of the SB5 using an independent sample of high-functioning students. Participants included 201 high-functioning, third-grade students ranging in age from 8 years, 4 months to 10 years, 11 months. Five models of the SB5 were analyzed using Analysis of Moment Structures (AMOS). Our findings indicated that a hierarchical, four-factor, post-hoc model provided the best fit to the data. Generally, implications for school psychologists include a better understanding of the factor structure of the SB5, especially as it relates to high-achieving children. Directions for future research are also discussed
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition factor analysis confirmatory factor analysis neuropsychological test battery Stanford-Binet V intelligence battery
Doweling, N. M., Hermann, B., LaRue, A., & Sager, M. A. (2010). Latent Structure and Factorial Invariance of a Neuropsychological Test Battery for the Study of Preclinical Alzheimer's Disease. Neuropsychology, 24(6), 742-756.
Abstract
Objective: To examine the latent structure of a test battery currently being used in a longitudinal study of asymptomatic middle-aged adults with a parental history of Alzheimer's disease (AD) and test the invariance of the factor solution across subgroups defined by selected demographic variables and known genetic risk factors for AD. Method: An exploratory factor analysis (EFA) and a sequence of confirmatory factor analyses (CFA) were conducted on 24 neuropsychological measures selected to provide a comprehensive estimate of cognitive abilities most likely to be affected in preclinical AD. Once the underlying latent model was defined and the structural validity established through model comparisons, a multigroup confirmatory factor analysis model was used to test for factorial invariance across groups. Results: The EFA solution revealed a factor structure consisting of five constructs: verbal ability, visuospatial ability, speed & executive function, working memory, and verbal learning & memory. The CFA models provided support for the hypothesized 5-factor structure. Results indicated factorial invariance of the model across all groups examined. Conclusions: Collectively, the results suggested a relatively strong psychometric basis for using the factor structure in clinical samples that match the characteristics of this cohort. This confirmed an invariant factor structure should prove useful in research aimed to detect the earliest cognitive signature of preclinical AD in similar middle aged cohorts.
Williams, T. H., McIntosh, D. E., Dixon, F., Newton, J. H., & Youman, E. (2010). A CONFIRMATORY FACTOR ANALYSIS OF THE STANFORD-BINET INTELLIGENCE SCALES, FIFTH EDITION, WITH A HIGH-ACHIEVING SAMPLE. Psychology in the Schools, 47(10), 1071-1083.
Abstract
The Stanford–Binet Intelligence Scale, Fifth Edition (SB5), is a recently published, multidimensional measure of intelligence based on Cattell–Horn–Carroll (CHC) theory. The author of the test provides results from confirmatory factor analyses in the technical manual supporting the five-factor structure of the instrument. Other authors have examined this factor structure through EFA using the standardization sample, and have not found evidence of a five-factor model. The purpose of the current study was to examine the internal construct validity of the SB5 using an independent sample of high-functioning students. Participants included 201 high-functioning, third-grade students ranging in age from 8 years, 4 months to 10 years, 11 months. Five models of the SB5 were analyzed using Analysis of Moment Structures (AMOS). Our findings indicated that a hierarchical, four-factor, post-hoc model provided the best fit to the data. Generally, implications for school psychologists include a better understanding of the factor structure of the SB5, especially as it relates to high-achieving children. Directions for future research are also discussed
- iPost using BlogPress from my Kevin McGrew's iPad
intelligence IQ tests IQ scores CHC theory Cattell-Horn-Carroll human cognitive abilities psychology school psychology individual differences cognitive psychology neuropsychology special education educational psychology psychometrics psychological assessment psychological measurement IQs Corner neuroscience neurocognitive cognitive abilities cognition factor analysis confirmatory factor analysis neuropsychological test battery Stanford-Binet V intelligence battery
Subscribe to:
Posts (Atom)