Analysis of the multitrait-multimethod matrix by confirmatory factor analysis.

The multitrait–multimethod (MTMM) matrix permits examination of the convergent and discriminant validity of psychological measures. Estimation using confirmatory factor analysis (CFA), the predominant analytical technique, has often resulted in severe difficulties, such as out-of-range estimates and convergence problems. This article shows that an important special case of one of the more frequently advocated CFA models is not identified and is therefore not estimable. Because most MTMM data are likely to conform closely to this special case, resulting analyses suffer from empirical underidentification. Alternative CFA models are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Situational specificity in assessment center ratings: A confirmatory factor analysis.

Assessment center ratings of eight abilities from each of five situational exercises were examined for their cross-situational consistency and discriminant validity. A series of confirmatory factor analyses revealed that the ratings were largely (if not totally) situation specific, and that assessors failed to distinguish among the eight target abilities. These results combined with previous research suggest that the assessment center method measures mainly situation-specific performance, not cross-situational managerial abilities. We suggest that the intended constructs might be better measured if more ability-related behaviors were elicited within each exercise and if the cognitive demands placed on assessors were reduced. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Utility of confirmatory factor analysis in test validation research.

The multitrait–multimethod ({mtmm}) design is often used in test validation research to disentangle problems due to shared method variance. However, {mtmm} research requires extensive data collection that may be prohibitive in clinical settings. Furthermore, interpretation of {mtmm} data can be ambiguous and misleading. In the current article, maximum-likelihood confirmatory factor analysis ({cfa}) is presented as a means for less ambiguous interpretation of complete and incomplete {mtmm} designs. Confirmatory Factor analysis is applied to four data sets that represent four designs: monotrait–monomethod, multitrait–monomethod, monotrait–multimethod, and multitrait–multimethod. In all four cases, {cfa} results provided more rigorous support of the original authors' positions and provided valuable supplemental findings as well. Additionally, {cfa} was applied to an artificial data set. Intuitive interpretations of the artificial data were compared with the results of the {cfa}. The results showed that the intuitive approach can lead to highly spurious conclusions regarding convergent and discriminant validity. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A confirmatory factor analysis of the structure of tacit knowledge in nursing

The purpose of this study was to examine the structure of tacit knowledge in the nursing profession. Using research in implicit learning and practical intelligence as a theoretical framework, a paper-and-pencil measure of tacit knowledge for nurses was developed and refined. Five models, each representing an established theory of traditional intelligence, were compared in terms of fit to the data. The results of the confirmatory factor analysis suggest that Model 3, a hierarchical model of intelligence, was the most plausible representation of tacit knowledge in nursing. The results indicate that the portion of managerial decision making learned implicitly on the job is mainly accounted for by managing tasks and others. Suggestions for nursing education include teaching effective strategies for managing tasks, such as handling increased workloads, establishing priorities, and delegating responsibility.     

Marital conventionalization as a measure of marital satisfaction: A confirmatory factor analysis.

It has been suggested that self-report measures of marital satisfaction are contaminated by a social desirability bias, which V. H. Edmonds (1967) termed marital conventionalization. Subsequent research showed that marital satisfaction measures were strongly correlated with marital conventionalization but not with social desirability scales. Because these measures tend to be so strongly correlated, this study examined whether marital satisfaction and conventionalization are best seen as 2 separate constructs or as a single construct. These measures were found to have 1 underlying factor based on confirmatory factor analyses with a sample of 2,109 individuals. Further analysis suggested that the relationship between marital satisfaction and conventionalization scales differs depending on the level of marital satisfaction. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Evaluation of a model of attention with confirmatory factor analysis.

Structural equation modeling (specifically, analysis of moment structures; J. L. Arbuckle, 1996) was used to evaluate the goodness of fit of a model of components of attention (A. E. Mirsky, B. J. Anthony, C. C. Duncan, M. B. Aheam, & S. G. Kellam, 1991) to neuropsychological test data from 2 samples. One sample consisted of psychiatrically normal persons with and without sleep-disordered breathing, and the other sample consisted of the adults studied by A. F. Mirsky et al. (1991), who gave rise to this model. That sample included psychiatric patients as well as normals. An exploratory data reduction procedure, principal-components analysis, suggested that attention might be conceptualized as composed of 4 independent elements or components: focus-execute, sustain, shift, and encode. Neither the proposed orthogonal model nor a model permitting correlated factors adequately fit either data set, suggesting that these 4 attention constructs are as yet not clearly validated in the measures used to assess them. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

The factor structure of Union Commitment: An application of confirmatory factor analysis.

M. E. Gordon et al. (see record 1980-33572-001) developed a Union Commitment scale and identified four underlying factors. L. Friedman and R. J. Harvey (see record 1986-28952-001), in a reanalysis of the Gordon et al. data, concluded that union commitment was best represented by two factors. Although several investigations have replicated the Gordon et al. factor structure, none of these have used confirmatory factor analysis (CFA). In the present study we used CFA to test the explanatory power of the two a priori factor structures. A sample of 465 blue-collar workers working for a large utility company in the midwestern United States participated in this investigation. The results suggest that, first, a modified version of Gordon et al.'s four-factor solution provides a better representation of the factor structure of union commitment. Second, the use of either the four-factor solution or commitment as a unidimensional construct may be justified, depending on the research focus. Finally, we found that unit weighting of the union commitment items was the most appropriate strategy for generating composites. The implication of these results and suggestions for future research are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Adherence in childhood diabetes: Results of a confirmatory factor analysis.

Tested the equality of the factor pattern of adherence to treatment of childhood diabetes identified by S. B. Johnson et al (see record 1988-05437-001) across 2 independent samples. Samples of 162 and 157 diabetics (aged 6–18 yrs) were studied. Factors 1–4, Exercise, Injection, Diet Type, and Eating/Testing Frequency, were confirmed. Factor 5, Diet Amount, proved to be too complex; the adherence measures comprising this factor (total calories and concentrated sweets consumed) are best treated as separate, single-indicator constructs. Results support a multivariate concept of adherence, offer insight into the nature of the components underlying diabetes adherence, and provide measurement information for reliable component estimation. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Construct validity of multiple job performance measures using confirmatory factor analysis.

The construct validity of job performance ratings obtained from self, supervisors, and peers was assessed relative to performance data obtained from an objective test of task proficiency. Data were gathered from a sample of 256 first-term jet engine mechanics in the U.S. Air Force. Widaman's (1985) approach to assessment of convergent and discriminant validity of multitrait–multimethod data using confirmatory factor analysis to contrast alternative models was used. As expected, all analyses revealed support for a model with correlated performance traits and uncorrelated methods. Significant convergence among the three rating sources was found. In addition, convergence was found between task ratings and scores from the objective proficiency test. Significant discriminant validity was found in all of the analyses. Results revealed that self, supervisors, and peers can be equally valid as sources of performance information, although performance of some tasks may be more validly evaluated by one source than another. Needs for further research on the effects of task characteristics on the validity of ratings are discussed. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Hierarchical confirmatory factor analysis of the Child Behavior Checklist/4–18.

Hierarchical confirmatory factor analysis was used to evaluate the 2-factor 2nd-order structure of T. M. Achenbach's ( 1991 ) Child Behavior Checklist in a sample of 577 children who had been identified as having serious emotional disturbance. An alternative 1-factor 2nd-order model also was tested. Results supported T. M. Achenbach's model in which the broadband Internalizing factor was represented by the narrowband Withdrawn, Somatic, and Anxious/Depressed syndromes, and the broadband Externalizing factor was represented by the narrowband Delinquent and Aggressive syndromes. Consistent with T. M. Achenbach's model, the remaining narrowband syndromes (i.e., Social, Thought, Attention ) loaded equally on both broadband factors and should not be included in scoring either Internalizing or Externalizing. Fit of the 1-factor model also was good and only slightly poorer than the 2-factor model. Therefore, an overall score would be appropriate as a measure of global problem behavior. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Measuring motivations for intercollegiate athlete alcohol use: A confirmatory factor analysis of the Drinking Motives Measure.

This study examined the psychometric properties of the Drinking Motives Measure (DMM) on a sample of 227 collegiate athletes. Confirmatory factor analyses indicated that the 4-factor structure of the DMM provided a better fit than either 2- or 1-factor models, but the overall fit of the 4-factor model was moderate at best. A revised 3-factor model consistent with prior research (M. L. Cooper, M. Russell, J. B. Skinner, & M. Windle, 1992) provided the best fit. Hierarchical multiple regression analyses indicated that the 3 DMM factors included in the revised model accounted for 17%-21% of the unique variance on alcohol consumption variables. Results provide preliminary evidence supporting the internal consistency, construct validity, and convergent validity of the revised 3-factor DMM with collegiate athletes. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Coping patterns of African American adolescents: A confirmatory factor analysis and cluster analysis of the Children's Coping Strategies Checklist.

The current study examined patterns of coping strategies in a sample of 497 low-income urban African American adolescents (mean age = 12.61 years). Results of confirmatory factor analysis indicated that the 4-factor structure of the Children's Coping Strategies Checklist (T. S. Ayers, I. N. Sandler, S. G. West, & M. W. Roosa, 1996) was not replicated in the current sample. The final model was a 3-factor model that was invariant across gender. The Active Coping factor and Revised Avoidant Coping factor were highly correlated in the present sample. Results of cluster analyses identified 2 coping groups differing on the frequency of coping use and preferred coping methods. The 1st group was more likely to use avoidant coping and less likely to use social support-seeking coping than the 2nd group, which showed more consistent use across coping strategies. There were no significant differences in the association between stressors and symptoms across the 2 groups. The results highlight the importance of examining factor structures of coping measures with underrepresented groups. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Specification of the criterion construct space: An application of hierarchical confirmatory factor analysis.

Used hierarchical confirmatory factor analysis to test alternative latent structures underlying performance in multiple dimensions as measured by work samples and 3 rating sources. The model receiving strongest support was one that specified multidimensional trait (performance dimension) and measurement method second-order factors (SOFs). This model exhibited significant convergent and discriminant validity, showed significant measurement method effects, and indicated that different measurement methods assess somewhat different aspects of the total criterion construct space. Significant correlations between measurement method SOFs and performance predictors suggest the need to reconsider the interpretation of method factors, not as bias, but as possibly representing different overall perspectives on performance. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Goodness-of-fit indexes in confirmatory factor analysis: The effect of sample size.

In confirmatory factor analysis, hypothesized models reflect approximations to reality so that any model can be rejected if the sample size is large enough. In the present article, we examine the influence of sample size on different fit indexes for both real and simulated data. Contrary to claims by Bentler and Bonett (1980), their incremental fit index was substantially affected by sample size. Contrary to claims by Joreskog and Sorbom (1981), their goodness-of-fit indexes provided by {lisrel} were substantially affected by sample size. Contrary to claims by Bollen (1986), his new incremental fit index was substantially affected by sample size. Hoelter's (1983) critical N index was also substantially affected by sample size. Of the more than 30 indexes considered, the Tucker-Lewis (1973) index was the only widely used index that was relatively independent of sample size. However, four new indexes based on the same form as the Tucker-Lewis index were also relatively independent of sample size. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Commentary on Keith and Witta's hierarchical and cross-age confirmatory factor analysis of the WISC-III.

Comments on T. Z. Keith and E. L. Witta's (1997) analysis of what is measured by the 3rd edition of the Wechsler Intelligence Scale for Children (WISC-III) by noting that a general factor comes through as a valid and replicable source of variance in every one of the analyses and that the statistical manipulations using the LISREL program were well chosen. J. B. Carroll also endorses Keith and Witta's contention that the WISC-III contains a factor that approaches being a quantitative reasoning factor. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

A new, more powerful approach to multitrait-multimethod analyses: Application of second-order confirmatory factor analysis.

The advantages of applying confirmatory factor analysis (CFA) to multitrait–multimethod (MTMM) data are widely recognized. However, because CFA, as traditionally applied to MTMM data, incorporates single indicators of each scale (i.e., each trait–method combination), important weaknesses are the failure to (a) correct appropriately for measurement error in scale scores, (b) separate error due to low internal consistency from uniqueness due to weak trait or method effects, (c) test whether items or subscales accurately reflect the intended factor structure, and (d) test for correlated uniquenesses. However, when the analysis begins with multiple indicators of each scale (i.e., items or subscales), second-order factor analysis can be used to address each of these problems. In this approach, first-order factors defined by multiple items or subscales are posited for each scale, and the method and trait factors are posited as second-order factors. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Measurement equivalence: A comparison of methods based on confirmatory factor analysis and item response theory.

Current interest in the assessment of measurement equivalence emphasizes 2 major methods of analysis. The authors offer a comparison of a linear method (confirmatory factor analysis) and a nonlinear method (differential item and test functioning using item response theory) with an emphasis on their methodological similarities and differences. The 2 approaches test for the equality of true scores (or expected raw scores) across 2 populations when the latent (or factor) score is held constant. Both approaches can provide information about when measurrment nonequivalence exists and the extent to which it is a problem. An empirical example is used to illustrate the 2 approaches. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Reporting practices in confirmatory factor analysis: An overview and some recommendations.

Reporting practices in 194 confirmatory factor analysis studies (1,409 factor models) published in American Psychological Association journals from 1998 to 2006 were reviewed and compared with established reporting guidelines. Three research questions were addressed: (a) how do actual reporting practices compare with published guidelines? (b) how do researchers report model fit in light of divergent perspectives on the use of ancillary fit indices (e.g., L.-T. Hu & P. M. Bentler, 1999; H. W. Marsh, K.-T., Hau, & Z. Wen, 2004)? and (c) are fit measures that support hypothesized models reported more often than fit measures that are less favorable? Results indicate some positive findings with respect to reporting practices including proposing multiple models a priori and near universal reporting of the chi-square significance test. However, many deficiencies were found such as lack of information regarding missing data and assessment of normality. Additionally, the authors found increases in reported values of some incremental fit statistics and no statistically significant evidence that researchers selectively report measures of fit that support their preferred model. Recommendations for reporting are summarized and a checklist is provided to help editors, reviewers, and authors improve reporting practices. (PsycINFO Database Record (c) 2010 APA, all rights reserved)     

Maximum likelihood confirmatory factor-analytic investigation of factors within Guilford's structure of intellect model.

K. G. J?reskog's (1969, 1970, 1974, 1978) confirmatory factor-analytic methodology was used to statistically test alternative hypotheses regarding the factor structure among tests of cognition, convergent production, and divergent production that underlies correlational data from the Aptitude Research Project. Large substantive improvements in goodness-of-fit indexes over random multifactor or general (single) factor models were not found, hence, the structure-of-intellect (SI) a priori models did not provide a superior fit to the data. The conclusion is that neither the high-dimensional SI model nor the general factor model provided a plausible level of overall fit to the data. Suggestions for future research and results of exploratory factor analyses are presented. (PsycINFO Database Record (c) 2010 APA, all rights reserved)