Refine
Document Type
- Article (5) (remove)
Language
- English (5)
Has Fulltext
- yes (5)
Is part of the Bibliography
- no (5)
Keywords
- fixed-links modeling (2)
- confirmatory factor analysis (1)
- executive processes (1)
- fluid intelligence (1)
- fluid reasoning (1)
- incomplete design (1)
- missing data (1)
- model of measurement (1)
- normal distribution (1)
- omissions (1)
Institute
Working memory capacity (WMC) and fluid intelligence (Gf) are highly correlated, but what accounts for this relationship remains elusive. Process-overlap theory (POT) proposes that the positive manifold is mainly caused by the overlap of domain-general executive processes which are involved in a battery of mental tests. Thus, executive processes are proposed to explain the relationship between WMC and Gf. The current study aims to (1) achieve a relatively purified representation of the core executive processes including shifting and inhibition by a novel approach combining experimental manipulations and fixed-links modeling, and (2) to explore whether these executive processes account for the overlap between WMC and Gf. To these ends, we reanalyzed data of 215 university students who completed measures of WMC, Gf, and executive processes. Results showed that the model with a common factor, as well as shifting and inhibition factors, provided the best fit to the data of the executive function (EF) task. These components explained around 88% of the variance shared by WMC and Gf. However, it was the common EF factor, rather than inhibition and shifting, that played a major part in explaining the common variance. These results do not support POT as underlying the relationship between WMC and Gf.
The paper reports an investigation on whether valid results can be achieved in analyzing the structure of datasets although a large percentage of data is missing without replacement. Two types of confirmatory factor analysis (CFA) models were employed for this purpose: the missing data CFA model with an additional latent variable for representing the missing data and the semi-hierarchical CFA model that also includes the additional latent variable and reflects the hierarchical structure assumed to underlie the data. Whereas, the missing data CFA model assumes that the model is equally valid for all participants, the semi-hierarchical CFA model is implicitly specified differently for subgroups of participants with and without omissions. The comparison of these models with the regular one-factor model in investigating simulated binary data revealed that the modeling of missing data prevented negative effects of missing data on model fit. The investigation of the accuracy in estimating the factor loadings yielded the best results for the semi-hierarchical CFA model. The average estimated factor loadings for items with and without omissions showed the expected equal sizes. But even this model tended to underestimate the expected values.
Developmental differences in the structure of executive function in middle childhood and adolescence
(2013)
Although it has been argued that the structure of executive function (EF) may change developmentally, there is little empirical research to examine this view in middle childhood and adolescence. The main objective of this study was to examine developmental changes in the component structure of EF in a large sample (N = 457) of 7–15 year olds. Participants completed batteries of tasks that measured three components of EF: updating working memory (UWM), inhibition, and shifting. Confirmatory factor analysis (CFA) was used to test five alternative models in 7–9 year olds, 10–12 year olds, and 13–15 year olds. The results of CFA showed that a single-factor EF model best explained EF performance in 7–9-year-old and 10–12-year-old groups, namely unitary EF, though this single factor explained different amounts of variance at these two ages. In contrast, a three-factor model that included UWM, inhibition, and shifting best accounted for the data from 13–15 year olds, namely diverse EF. In sum, during middle childhood, putative measures of UWM, inhibition, and shifting may rely on similar underlying cognitive processes. Importantly, our findings suggest that developmental dissociations in these three EF components do not emerge until children transition into adolescence. These findings provided empirical evidence for the development of EF structure which progressed from unity to diversity during middle childhood and adolescence.
This article reports an investigation of how inhibition contributes to fluid reasoning when it is decomposed into the reasoning ability, item-position, and speed components to control for possible method effects. Working memory was also taken into consideration. A sample of 223 university students completed a fluid reasoning scale, two tasks tapping prepotent response inhibition, and two working memory tasks. Fixed-links modeling was used to separate the effect of reasoning ability from the effects of item-position and speed. The goodness-of-fit results confirmed the necessity to consider the reasoning ability, item-position, and speed components simultaneously. Prepotent response inhibition was only associated with reasoning ability. This association disappeared when working memory served as a mediator. Taken together, these results reflect the inhomogeneity of what is tapped by the fluid reasoning scale on one hand and, on the other, suggest inhibition as an important component of working memory.
The paper outlines a method for investigating the speed effect due to a time limit in testing. It is assumed that the time limit enables latent processing speed to influence responses by causing omissions in the case of insufficient speed. Because of processing speed as additional latent source, the customary confirmatory factor model is enlarged by a second latent variable representing latent processing speed. For distinguishing this effect from other method effects, the factor loadings are fixed according to the cumulative normal distribution. With the second latent variable added, confirmatory factor analysis of reasoning data (N=518) including omissions because of a time limit yielded good model fit and discriminated the speed effect from other possible effects due to the item difficulty, the homogeneity of an item subset and the item positions. Because of the crucial role of the cumulative normal distribution for fixing the factor loadings a check of the normality assumption is also reported.