Statistical Modeling | Dr. Schmank's Academic Website

Age Differences Among Psychometric Models of Intelligence

Mon, 20 Apr 2020 00:00:00 +0000

The positive manifold is one of the most frequently replicated findings in cognitive psychology and has often been explored using factor analysis (Spearman, 1904; Conway & Kovacs, 2013). Well-known models of cognition are typically organized such that manifest variables load onto respective latent factors representing specific cognitive abilities, and some posit a higher-order factor of general ability, g. It has been well established that cognitive abilities change across the lifespan (Craik & Salthouse, 2011), yet there is less work examining how factor models of cognitive ability compare across different age groups.

The current project used data from the Hungarian-Weschler Adult Intelligence Scale-fourth edition (H-WAISIV; Weschler; 2008) to compare models of cognitive ability for both young (18-40 years) and older adults (65 + years). An exploratory factor analysis conducted on the young adult data (n = 457) produced a four factor model that explained 67% of the variance in the data. This factor structure was then used to conduct a confirmatory factor analysis (CFA) on the older adult data (n = 305). The CFA produced good fit to the data, that was not significantly improved by adding a higher-order factor. Finally, in line with recent criticisms of using latent variable modeling (see Borsboom, Mellenbergh, & van Heerden, 2003) an exploratory psychometric network analysis was conducted on young adult data and a confirmatory psychometric network analysis was then applied to the older adult data, which produced similar, well-fitting results.

Intriguingly, the older adult network demonstrates stronger connections between all measures sampled by the H-WAISIV, as indicated by the thicker blue edges or lines connecting nodes or measures in the network. This implies that performance on these psychological tasks were more related for older adults compared to their young adult counterparts. More research is required to explain what these psychometric networks indicate, however, additional research has indicated that psychometric networks fit intelligence data as well if not better than latent variable models.

Psychometric Network Model of Intelligence

Tue, 20 Aug 2019 00:00:00 +0000

The positive manifold—the finding that cognitive ability measures demonstrate positive correlations with one another—has led to models of intelligence that include a general cognitive ability or general intelligence (g). This view has been reinforced using factor analysis and latent variable models. However, a new theory of intelligence, Process Overlap Theory (POT; Kovacs & Conway, 2016), posits that g is not a psychological attribute but an index of cognitive abilities that results from an interconnected network of cognitive processes. From this perspective, psychometric network analysis is an attractive alternative to latent variable modeling. Network analyses display partial correlations among observed variables that demonstrate direct relationships among observed variables. To demonstrate the benefits of this approach, the Hungarian Wechsler Adult Intelligence Scale Fourth Edition (H-WAIS-IV; Wechsler, 2008) was analyzed using both psychometric network analysis and latent variable modeling. Network models were directly compared to latent variable models. Results indicate that the H-WAIS-IV data was better fit by network models than by latent variable models. We argue that POT, and network models, provide a more accurate view of the structure of intelligence than traditional approaches.

At the outset of my doctoral program at Claremont Graduate University (Fall, 2017) my academic mentor, Andrew R.A. Conway, PhD. described psychometric network modeling as a complimentary and adventageous statistical tool developed that could be implemented in our research lab ( .

Psychometric Network Model of Cognitive Ability

Mon, 20 Aug 2018 00:00:00 +0000

The positive manifold is one of the most replicated findings in the psychological sciences. The positive manifold refers to the finding of all positive correlations among a number of cognitive ability measurements, such that participants who score above average on one test (e.g., vocabulary) also score above average on other tests (e.g., mathematics). In the field of psychometrics, factor analysis is the primary statistical technique used to investigate the underlying structure of intelligence. Charles Spearman (1904) formulated the original factor model of general intelligence (g), which resulted in a single, explanatory factor of cognitive ability. This one-factor view of g was met with criticism, which led to the generation of many different factor models over the years.

A recent example is the Cattel-Horn-Carroll (CHC) model of intelligence. The CHC model initially saw the fractionation of g into a continuum of fluid (Gf) to crystallized (Gc) intelligence factors (Cattel, 1941; Horn, 1965) but later led to Carroll’s three-stratum hierarchical theory (1993). Thus, the CHC model consists of three differentiated levels: 1. observed measurements (e.g., vocabulary, mathematics) are located at the lowest-level; 2. broad cognitive abilities (e.g., short-term memory, crystalized intelligence) explain variation at the next level; and 3. g is located at the highest-level and explains the variation among mid-level cognitive ability factors. The CHC model provides a good fit to data yet after a century of research since the original formulation of Spearman’s g, the debate over exactly what g is and how g is structured continues (c.f., Kovacs & Conway, 2016; Protzko, 2017).

Recently, a number of articles have been published employing an alternative statistical technique to factor analysis called network modeling (cf., Epskamp & Fried, 2017; McNalley, 2006; van der Maas et al., 2017). In network models, partial correlation coefficients are calculated to establish the association between pairs of observed variables (referred to as nodes). It is typical for clusters of nodes that load together on a particular factor to be located more closely in these models than two variables that load on separate or orthogonal factors, although no factors are actually generated when utilizing network models, so there is no g (for an example, see Figure 4 of van der Maas et al., 2017). The primary advantage of network models is that instead of attempting to interpret subjective factors researchers can shift their focus to specific measurements and the one-to-one associations that exist between them (Guyon, Falissard, & Kop, 2017).

The primary objective of the current project was to develop a network model of intelligence based on data from the Woodcock-Johnson test published in Carroll (2003). From a philosophical perspective, we believe that network models avoid the intrinsic disadvantages associated with factor analysis because the nodes represent observed data instead of unobserved and difficult to interpret factors. Additionally, we argue that network models are superior to factor models when examining changes in intelligence, either as a function of developmental (in children and the elderly) or as a function of cognitive training (e.g., working memory training).