Second-order General Factor Research Articles

Is there a unitary neuro-cognitive basis of mental speed?

There is ample evidence that speed of information processing is related to mental abilities, as more intelligent individuals have shorter reaction times in elementary cognitive tasks (ECTs). The aim of this study is to identify neuro-cognitive processes that might be responsible for this association, and to investigate their factor structure and its relation to general intelligence. 40 participants drawn from a community sample completed a choice reaction time task and two binary decision tasks while an EEG was recorded. A diffusion model approach was applied to the RT data, which decomposes different cognitive processes involved in speeded binary decisions. Behavioral and electrophysiological results indicate that ECT conditions vary in different neuro-cognitive components of information processing. Additionally, factor analysis shows that all mental speed variables have substantial loadings on a second-order general factor marked by intelligence. These results suggest that there is a common neuro-cognitive speed factor associated with general intelligence.

Confirmatory Factor Analysis of the Behavior Rating Inventory of Executive Function-Adult Version in Healthy Adults and Application to Attention-Deficit/Hyperactivity Disorder

The Behavior Rating Inventory of Executive Function-Adult Version (BRIEF-A) is a questionnaire measure designed to assess executive functioning in everyday life. Analysis of data from the BRIEF-A standardization sample yielded a two-factor solution (labeled Behavioral Regulation and Metacognition). The present investigation employed confirmatory factor analysis (CFA) to evaluate four alternative models of the factor structure of the BRIEF-A self-report form in a sample of 524 healthy young adults. Results indicated that a three-factor model best fits the data: a Metacognition factor, a Behavioral Regulation factor consisting of the Inhibit and Self-Monitor scales, and an Emotional Regulation factor composed of the Emotional Control and Shift scales. The three factors contributed 14%, 19%, and 24% of unique variance to the model, respectively, and a second-order general factor accounted for 41% of variance overall. This three-factor solution is consistent with recent CFAs of the Parent report form of the BRIEF. Furthermore, although the Behavioral Regulation factor score in the two-factor model did not differ between adults with attention-deficit/hyperactivity disorder and a matched healthy comparison group, greater impairment on the Behavioral Regulation factor but not the Emotional Regulation factor was found using the three-factor model. Together, these findings support the multidimensional nature of executive function and the clinical relevance of a three-factor model of the BRIEF-A.

Multigroup Confirmatory Factor Analysis of the Cognitive Dysfunction Questionnaire: Instrument refinement and measurement invariance across age and sex

The study adopted Confirmatory Factor Analysis (CFA) to investigate the factorial structure and reduce the number of items of the Cognitive Dysfunction Questionnaire (CDQ). The analyses were based on data for a total of 1,115 participants from population based samples (mean age: 63.0 ± 14.5 years, range: 25-95) randomly split into a refinement (N = 569) and a cross-validation (N = 546) sample. Equivalence of the measurement and structural portions of the refined model was demonstrated across the refinement and cross-validation samples. Among competing models the best fitting and parsimonious model had a hierarchical factor structure with five first-order and one second-order general factor. For the final version of the CDQ, 20 items within five domains were selected (Procedural actions, Semantic word knowledge, Face recognition, Temporal orientation, and Spatial navigation). Internal consistency reliabilities were adequate for the total scale and for the subscales. Multigroup CFAs indicated measurement invariance across age and sex up to the scalar level. Finally, higher levels of cognitive dysfunction as reflected by CDQ scores were predicted by advancing age, fewer years of education, and with deficits in general cognitive functioning as reflected by scores on the Mini-Mental State Examination. In conclusion, the CDQ appears to be psychometrically sound and shows the expected relationships with variables known to be associated with cognitive dysfunction and dementia. Future studies should apply it among clinical groups to further test its usefulness.

Hierarchical Factor Structure of the Turkish Version of the Posttraumatic Growth Inventory in a Normal Population

Background: Negative consequences of traumatic experiences, such as depression, anxiety, or dissociative symptoms, etc. have been reported by many researchers. However, it is proposed that stressful events not only lead to poor psychological outcomes but also may trigger positive changes. Several instruments gauging posttraumatic benefits have been developed to examine the effects of factors that may promote positive psychological outcomes in the aftermath of stressful events. The Posttraumatic Growth Inventory (PTGI) is one of the prominent instruments that assess posttraumatic positive changes. In this study we aimed to assess psychometric properties of the Turkish version of the PTGI. Method: This study differed to an extent from previous studies concerning the features of the PTGI. We used a dispositional form of the instrument in a sample recruited from high school and university students. Our data were collected from 723 volunteers. 367 subjects were males (50.76%) and 356 subjects were females (49.24%). Also we did not specify any selection criteria in recruiting subjects owing to their adverse life experiences with a presumption that stressful life events are not uncommon in normal population. We administered to participants a dispositional form of the Posttraumatic Growth Inventory and the Personal View Survey-III. Item statistics for the measure were computed. We performed an explanatory factor analysis by using principal components with promax rotation and a confirmatory factor analysis by using structural equation modeling. Since the factor inter-correlations were higher than .40 we computed Schmid-Leiman transformation to obtain second-order general factor loadings. Inner consistencies and 15-day test-retest intracorrelations were calculated. Results: Item discrimination indexes ranged from .28 to .72. Promax rotated principal components analysis pointed out a three-factor structure. It was found in model testing with structural equation modeling that three-factor structure was valid for the Turkish version of the PTGI. After computing second-order factor loadings, we detected that general factor accounted for 64 percentage of the total variance. Three subscales of the measure were the Changes in Self-Perception, Changes in Philosophy of Life, and Changes in Relationship. Internal consistency for the Changes in Self-Perception subscale was 0.88, for the Changes in Philosophy of Life subscale was 0.78, for the Changes in Relationship was 0.77, and for the overall items was 0.92. 15-day test-retest intra-correlation for the composite scores was 0.83 and intra-correlations for the subscale scores ranged from 0.70 to 0.85. Conclusion: Turkish version of the measure revealed a three-factor first-order structure. However, it seems that the concept of posttraumatic benefits measured by the PTGI has a tendency to represent a unidimensional psychological construct in Turkish sample. Since the three-factor structure was validated, three-su

Confirmatory factor analysis of Project Spectrum activities. A second-order g factor or multiple intelligences?

This paper compares different theoretical models of the structure of intelligence, based on the analysis of data obtained in a series of measured abilities corresponding to the Spectrum assessment activities (Gardner, Feldman & Krechevsky, 1998) in a sample of 393 children enrolled in kindergarten and first grade. The data were analyzed using confirmatory factor analysis. The models compared were: a) a model with six first-order uncorrelated factors, b) a model with six first-order factors and one second-order general factor, g; c) a model with two correlated second-order general factors, in which the cognitive intelligences load on a “cognitive” general factor and the non-cognitive intelligences load on a “non-cognitive” general factor, and d) a model with six first-order correlated factors. The percentage of variance in measured abilities due to g and to first-order factors was also estimated. Overall, the results indicate that the Spectrum activities are not as separate from g as proposed by the defenders of multiple intelligences theory, nor as unitary as argued by the defenders of g factor models.

The Peer Attitudes Toward Children who Stutter (PATCS) scale: an evaluation of validity, reliability and the negativity of attitudes

Persistent calls for school-based education about stuttering necessitate a better understanding of peer attitudes toward children who stutter and a means to measure outcomes of such educational interventions. Langevin and Hagler in 2004 developed the Peer Attitudes Toward Children who Stutter scale (PATCS) to address these needs and gave preliminary evidence of reliability and construct validity. To examine further the psychometric properties of PATCS and to examine the negativity of attitudes. PATCS was administered to 760 Canadian children in grades 3-6. Measures included reliability, a confirmatory factor analysis (CFA), a known groups analysis, convergent validity with the Pro-Victim Scale of Rigby and Slee, and the negativity of attitudes. PATCS appears to tap a second-order general attitude factor and three first-order factors representing the constructs of Positive Social Distance (PSD), Social Pressure (SP), and Verbal Interaction (VI). In the known groups analysis, participants who had contact with someone who stutters had higher scores (more positive attitudes) than those who had not, and girls had higher scores than boys. PATCS correlated moderately (0.43, p<0.01) with the Pro-Victim scale. Finally, one-fifth (21.7%) of participants had scores that were somewhat to very negative. Results provide evidence of the validity and reliability of PATCS and confirm the need for school-based education about stuttering. The PSD and SP factors suggest that education include discussions about (1) similarities and differences among children who do and do not stutter in order to increase acceptance, and (2) making personal choices and handling peer pressure in thinking about children who stutter. The VI factor suggests that open discussion about stuttering may alleviate frustration experienced by listeners and provide the opportunity to give strategies for responding appropriately. Results also suggest that education involve contact with a person who stutters.

Activities, self-referent memory beliefs, and cognitive performance: Evidence for direct and mediated relations.

In this study, the authors investigated the role of activities and self-referent memory beliefs for cognitive performance in a life-span sample. A factor analysis identified 8 activity factors, including Developmental Activities, Experiential Activities, Social Activities, Physical Activities, Technology Use, Watching Television, Games, and Crafts. A second-order general activity factor was significantly related to a general factor of cognitive function as defined by ability tests. Structural regression models suggested that prediction of cognition by activity level was partially mediated by memory beliefs, controlling for age, education, health, and depressive affect. Models adding paths from general and specific activities to aspects of crystallized intelligence suggested additional unique predictive effects for some activities. In alternative models, nonsignificant effects of beliefs on activities were detected when cognition predicted both variables, consistent with the hypothesis that beliefs derive from monitoring cognition and have no influence on activity patterns.

Multidimensionality of intrusion and avoidance: alternative factor models of the Impact of Event Scale

Previous research using exploratory factor analysis of the Impact of Event Scale (Horowitz et al., 1979) has suggested structures that are not consistent with the original two-factor model. IES data from 485 emergency service personnel who had experienced traumatic events during the course of their work were submitted to confirmatory factor analysis to test seven alternative factor models of the IES. This study is one of the most comprehensive analyses of the IES to date, investigating both theoretically and empirically derived models. The model of best fit was a four-factor model with a single second-order factor. This model comprised four first-order factors including intrusion, avoidance, numbing and sleep disturbance factors all subsumed by a second-order general distress factor. These results provide support for the multidimensionality of both intrusion and avoidance symptoms following traumatic experiences. Researchers and practitioners in the field of traumatic stress should be made aware of the assessment and treatment implications of this multidimensional structure. (C) 2003 Elsevier Ltd. All rights reserved.

Sex differences in general knowledge

A general information or knowledge test, which was shown to measure 19 domains of general knowledge, six first-order factors and one second-order general factor, was constructed. Data obtained from 469 female and 167 male undergraduates were tested for sex differences using Student's t and Hotelling's multivariate t. It was found that males obtained significantly higher means than females on the second-order general factor and on four of the six first-order factors identified as information about Current Affairs, Physical Health and Recreation, Arts and Science. Females obtained a significantly higher mean than males on the first-order factor identified as Family. There was no sex difference on the remaining first-order factor identified as Fashion. The results confirm the findings in a number of standardisation samples of the Wechsler tests that males obtain higher average scores than females on the Information subtests and that this is not attributable to a bias in favor of males on these tests.

Independent Examination of the Factor Structure of the Cognitive Assessment System (CAS): Further Evidence Challenging the Construct Validity of the CAS

This study is the first to examine independently the factor structure of the Cognitive Assessment System (GAS; Naglieri &amp; Das, 1997) with a primary dataset not collected by its authors. Participants were 155 students (59 boys, 96 girls), ages 8 to 11 (M = 9.81 years, SD = 0.88), in Grades 3 to 6. Confirmatory factor analysis (CFA) was used to compare the fit provided by the planning, attention, and simultaneous-successive (PASS) model, the theoretical model underlying the CAS, with alternative models of cognitive ability suggested by previous research. Results of this study indicated that the PASS model did not provide a better fit to the data than did alternative hierarchical and nonhierarchical models. Not only were the Planning and Attention factors of the PASS model virtually indistinguishable (r = .88), but they demonstrated inadequate specificity for meaningful interpretation. The model reflecting the actual hierarchical structure of the CAS was found to fit the data no better than alternative models based on different theoretical orientations. Of the hierarchical models examined in this study, the best fitting was a hierarchical (PA)SS model with one second-order general factor, psychometric g, and three first-order factors reflecting Fluid Intelligence/Visual Processing (Simultaneous), Memory Span (Successive), and Processing Speed (Planning/Attention). In sum, results of this study support Kranzler and Keith's (1999) conclusion that the CAS lacks structural fidelity, which means that the CAS does not measure what its authors intended it to measure. Results of this study, therefore, provide further evidence challenging the construct validity of the CAS.

The independent prediction of general intelligence by elementary cognitive tasks: genetic and environmental influences.

Current theories of intelligence have, in some cases, begun to include elementary cognitive tasks. Behavioral genetic studies of intelligence have not taken these theories into account. The current study includes 135 MZ and 128 DZ twin pairs from the Western Reserve Twin Project. The 11 WISC-R subtests as well as 6 elementary cognitive tasks were employed. Using a Schmid-Leiman (1957) transformation, analyses indicate a four-group factor model, supported by a second-order general factor at both phenotypic and biometric levels. Results indicate that the general factor, group factors, and specific residuals are necessary when examining additive genetic variance. Common environmental variance can be collapsed into a single general factor.

Structure of problem and positive behaviors in African American youths.

The association of 21 problem and positive behaviors was explored in a sample of 283 inner-city, African American youths of ages 8 through 12. Data reduction yielded 4 factors, 3 representing different types or levels of problem behavior, labeled Interpersonal-Minor Deviance, School Problems, and Covert-High Deviance, and 1 representing positive behaviors. The 3 problem behavior factors, although not the positive behavior factor, were significantly correlated with an underlying second-order general deviance factor. The problem behavior clusters identified differed by the settings in which they occur as well as their inherent magnitude of deviance. Discriminant validity analyses confirmed that deviance was not a unitary phenomenon. Limitations as well as other implications of the data are discussed.

Assessing inter- and intrapersonal resources: social support and coping among adults with a disability.

A promising model for addressing community reintegration and adjustment following disability is the stress-outcome model. In spite of demonstrated applicability, this model has found little support among clinicians and researchers in the rehabilitation field. One reason for this may be the lack of explicit conceptual and operational definitions for the central constructs of the model. This paper presents theoretical and measurement models for two such constructs which are considered buffers or mediators in the stress-outcome relationship: social support and coping. Issues associated with each construct are explored, with reference particularly to people with disabilities. Following a synthesis of the literature, models for each construct were proposed that represented the two constructs as each having a parallel three-factor structure. Using a sample of 120 spinal cord-injured adults from across Ontario, data were collected on existing instruments which measure the two constructs of interest (Interpersonal Support Evaluation List, Ways of Coping Questionnaire). Face-to-face interviews were conducted in participants' homes at 1, 4, and 12 months post-discharge from rehabilitation, in order to capture the period of initial community readjustment and reintegration. Psychometric evaluation of the measurement models included item analysis, factor analysis, and reliability assessment. These analyses provided empirical support for the three-factor structure for both constructs, as well as a second-order general factor for social support. Recommendations for revisions of the scales and further development of the measurement models were made, and the revised measures were discussed in terms of theoretical and research implications.

Gender-related differences in intelligence: Application of confirmatory factor analysis

Confirmatory factor analysis was used to test first- and second-order factor models on cognitive abilities and their invariance across male and female samples. Four first-order factors were found underlying the 23 subtest scores of two group intelligence tests, the AH 4 Test of Intelligence and the Advanced Progressive Matrices, and two group Piagetian tests, Science Reasoning Tasks and Arlin Test of Formal Reasoning. Slight differences found in the formal operations, spatial, numerical, and verbal factors of the male and female groups suggested some gender-related differences in these factors.There were two alternative second-order general factor models for the male group, one with three first-order factors and the other with four first-order factors. Only a general factor model with three first-order factors fit the data of the female sample; this female group model contained some differences in the loadings of the formal operations and spatial factors when compared with that of the male group.

Measuring Adolescent Depression: Factorial Validity and Invariance of the Beck Depression Inventory Across Gender

Extending the earlier work of Byrne and Baron (in press), the factorial invariance of the Beck Depression Inventory (BDI) was tested across nonclinical adolescent males (n = 351) and females (n = 334) using confirmatory factor-analytic procedures. Despite the differential loading pattern of Item 20 for females, a three-factor structure that comprised one second-order general factor of depression and three first-order factors representing negative attitude, performance difficulty, and somatic elements was found to be invariant across gender. Furthermore, except for two additional items (8 and 10), the remaining BDI measurements were group invariant. Results are expected to be of substantial interest to both researchers and clinicians whose concerns focus on depression as it bears on this population.

No demonstration that g is not unitary, but there's more to the story: Comment on Kranzler and Jensen

Because they employed a statistical procedure that was inappropriate for the purpose, Kranzler and Jensen (1991) have not demonstrated that a factor of general intelligence, g, depends on a number of independent processes. It is more parsimonious to interpret their results in terms of a g that is unitary. A factorial reanalysis of their data, including some that they did not present (cross-correlations between psychometric variables and variables from elementary cognitive tasks, or ECTs), shows that many ECT variables, especially those reflecting decision time rather than movement time, have substantial loadings on a second-order general factor. This factor is orthogonal to a second-order factor representing general speed of movement from a “home button” to the response button in ECTs. Possible implications for interpreting g in terms of information-processing efficiency are discussed.

Psychometric Properties of the Intrinsic Motivation Inventory in a Competitive Sport Setting: A Confirmatory Factor Analysis

The present study was designed to assess selected psychometric properties of the Intrinsic Motivation Inventory (IMI) (Ryan, 1982), a multidimensional measure of subjects' experience with regard to experimental tasks. Subjects (N = 116) competed in a basketball free-throw shooting game, following which they completed the IMI. The LISREL VI computer program was employed to conduct a confirmatory factor analysis to assess the tenability of a five factor hierarchical model representing four first-order factors or dimensions and a second-order general factor representing intrinsic motivation. Indices of model acceptability tentatively suggest that the sport data adequately fit the hypothesized five factor hierarchical model. Alternative models were tested but did not result in significant improvements in the goodness-of-fit indices, suggesting the proposed model to be the most accurate of the models tested. Coefficient alphas for the four dimensions and the overall scale indicated adequate reliability. The results are discussed with regard to the importance of accurate assessment of psychological constructs and the use of linear structural equations in confirming the factor structures of measures.

Factor structure of the Counselor Rating Form—Short.

The relative goodness of fit of four competing factor models of the Counselor Rating FormShort was examined across client (N = 191) and nonclient (N = 111) samples by confirmatory factor analysis. The four models tested were (a) the single, general, positive-evaluation factor, (b) the three orthogonal factors (expertness, trustworthiness, and attractiveness); (c) the three oblique factors; and (d) a two-step hierarchical-factor model consisting of the three independent firstorder factors and an independent second-order general factor. Results of the confirmatory factor analyses supported the validity of the two-step hierarchical-factor model for both the client and nonclient samples. The factor structure of this two-step hierarchical-factor model was found to be invariant across the two samples.

Between- versus within-family factor analyses of cognitive abilities

Confirmatory factor analyses of the 15 cognitive abilities tests from the Hawaii Family Study of Cognition were conducted on between-family (BF) means and within-family (WF) differences for 370 AEA (Americans of European ancestry) and 116 AJA (Americans of Japanese ancestry) sibling pairs. Difference chi-square significance tests, goodness-of-fit indices, and congruence coefficients for the estimated loadings on four specific cognitive abilities factors and on the second-order general intelligence factor nearly all indicated that the between-family factor structures were not significantly different from the within-family structures for both AEA and AJA siblings (the AEA and AJA structures were also not significantly different). The similarity of the BF and WF structures suggests that the genetic and environmental influences underlying cognitive abilities are “intrinsic” in nature, that is, not just due to between-family differences in culture, status, values, and fortuitous cross-assortative mating.

A Confirmatory Factor Analytic Investigation of time Sharing Performance and Cognitive Abilities

Eighty-one male subjects performed four information processing tasks and six dual task combinations, and completed a battery of psychometric ability tests selected to define three first-order factors and a second-order general ability factor. Confirmatory maximum likelihood factor analyses of the performance data provided no support for a general time-sharing factor, but a model with factors corresponding to the four single tasks provided a good fit to the data. The Grammatical Reasoning factor was highly correlated with the Verbal and second-order General Ability factors, suggesting that this task may be a good single index of total attentional resources.