Multivariate Behavioral Research

Van der Kloot, Willem A.; Van Herk, Hester

doi: 10.1207/s15327906mbr2604_1pmid: 26751022

Dissimilarity measures (D) derived from sortings of stimuli can be submitted to multidimensional scaling (MDS) either directly, or after transforming them to profile distances (Δ) computed on the rows of the D matrix. The latter procedure was criticized by Drasgow and Jones (1979) who performed two simulation studies, which are criticized here in turn. In the present article two sets of real sorting data were used for comparing the results of MDS on D and Δ, both with each other and with the results of two other procedures: multiple correspondence analysis (by means of HOMALS) on the raw sorting data, and MDS on the pairwise similarity ratings of the same stimuli by the same subjects. The three procedures were compared both with respect to the final configurations and with regard to the fit of the corresponding distances to the data. These comparisons suggested that MDS on D is slightly superior to MDS on Δ. The latter analysis, however, yields results that are similar to those of the much more efficient HOMALS program. The differences, however, are on the average very small.

Byrne, Barbara M.

doi: 10.1207/s15327906mbr2604_2pmid: 26751023

The purposes of the study were: (a) to test for the factorial validity of the Maslach Burnout Inventory (MBI), for 543 teachers at the intermediate (n = 163), secondary (n = 162), and university (n = 218) levels, and (b) to test for the equivalence of factorial measurements and structure across groups. Initial confirmatory factor analysis of the hypothesized 3-factor structure yielded a malfitting model for each group of educators. With a view to improving the MBI for use with educators, subsequent exploratory and confirmatory factor analyses resulted in the deletion of four scale items. Tests for invariance revealed the equivalency of remaining items across intermediate and secondary teachers, and items measuring Emotional Exhaustion and Depersonalization across all three groups; the structure of burnout was only partially invariant across educators. The study has important implications for substantive studies focusing on multigroup comparisons across teaching panels.

Graham, John W.; Collins, Nancy L.

doi: 10.1207/s15327906mbr2604_3pmid: 26751024

Confirmatory factor analysis of multitrait-multimethod (MTMM) data has proven to be a useful tool for assessing convergent and discriminant validity. However, researchers have not made full use of the results of MTMM analyses in examining the relationship between MTMM factors and variables outside the MTMM. Often, researchers simply average the various measures of each trait. Alternatively, they estimate LISREL MTMM models, but estimate only relationships between MTMM traits and the outside variables. In the present article, we show that these two approaches to analyzing data outside the MTMM produce equally highly biased parameter estimates when the actual correlations between MTMM method factors and the outside variables are substantial. An algebraic explanation and a simulated data illustration are given for the bias due to misspecification. Also, the problem is illustrated with a brief empirical example. Implications for applied research are discussed.

Williams, John Delane

doi: 10.1207/s15327906mbr2604_4pmid: 26751025

A solution is shown for addressing the age x cohort x period issue in life-span research. Previous solutions have utilized only portions of the data in a given two-way layout. The proposed solution uses all data, which for at least two of the three two-way layouts involves missing cells. The method can be used for repeated measure designs, or designs in which new subjects are measured at each period. The design allows the assessment of each main effect and each two-way interaction. The hypotheses tested are explicitly shown, together with the linear models that accomplish this testing.

Grote, Gudela F.; James, Lawrence R.

doi: 10.1207/s15327906mbr2604_5pmid: 26751026

The Situation-Response (S-R) Measure of Achievement Motivation was developed to analyze the cross-situational consistency of achievement-related behavior. This measure was based on a conceptualization of achievement motivation that included the following three components: need to achieve, need to avoid failure, and perceived self-efficacy. Data obtained from 246 college students provided evidence for the validity of the new instrument. However, exploratory factor analyses performed on the items in the instrument indicated the presence of only two factors, namely Striving and Apprehensiveness. Regression analyses further indicated the possible inappropriateness of a theoretically-based difference score, which combined the striving and apprehensiveness composites into a resultant tendency. Confirmatory factor analysis was used to test behavioral consistency of the responses to the S-R measure. Although all of the models tested had relatively poor fits with the data, the results (a) provided evidence for item-specific covariation that inflated the cross-situational correlations among the achievement-related composites of the S-R measure, and (b) indicated, after this item-specific covariation had been partialled out, a lack of support for second-order factors representing general achieving tendencies across situations. A final set of analyses indicated the presence of three distinct types of response patterns. These types were tentatively called socially anxious, cynically motivated, and anxiously striving. It was found that membership in one of these subgroups was more informative of an individual's pattern of achievement motivation than conventionally used personality tests.

Cliff, Norman; Charlin, Ventura

doi: 10.1207/s15327906mbr2604_6pmid: 26751027

We generalize the formulas derived by Daniels and Kendall (1947) for the variance of the sample tau correlation. It is assumed that multivariate data are sampled from a population, and sample taus between pairs of variables are being used to estimate their population counterparts. Expressions for the variance of tau-a are generalized to allow for ties on either variable, and we further provide an expression for the covariance between two taus, including the special case where there is a variable in common. Unbiased estimators of the variance and the covariances are also derived for use in small samples. The variances and covariances of tau-a are used to provide asymptotic variances for tau-b and Somers' d.

Leutner, Detlev; Weinsier, Philip D.

doi: 10.1207/s15327906mbr2604_7pmid: 26751028

The present study addressed the question of whether computers and information technology constitute a uniform attitude object which can influence the study interests of students. Based on a facet design, an interest questionnaire with 72 university course descriptions was constructed in which computers and information technology was embedded as one of four item-design facets (Weinsier & Leutner, 1988). One hundred students from each of two universities responded to the questionnaire. The multidimensional interest structures of the two samples were nearly identical. The design facets constituted uniform attitude objects and the multidimensional scaling solution of the inter-item correlation matrix could be partitioned almost perfectly according to regional hypotheses derived from the facet design. Next to the discipline facet (i.e., academic discipline), computers and information technology constituted the most relevant facet of student interest. An analysis of variance supported the conclusions drawn from the multidimensional scalings.

Thompson, Paul A.

doi: 10.1207/s15327906mbr2604_8pmid: 26751029

The bootstrap is a relatively new technique. In using it, the analyst intensively examines the data actually gathered to estimate the precision of the sample statistic, rather than relying on a parametric theory. While this makes little sense when parametric theories are available (such as is the case with the correlation coefficient, the mean and most other common statistics), it is a useful adjunct to traditional statistical methods when these elegant methods cannot be used. An application of the bootstrap to a complex psychological analysis approach is demonstrated. The method provides variance estimates and allows the testing of nested competing models. Most importantly, it gives a preliminary idea about the variability of quite complex parameters.

De Ayala, R. J.; Hertzog, Melody A.

doi: 10.1207/s15327906mbr2604_9pmid: 26751030

The application of item response theory (IRT) models requires the identification of the data's dimensionality. A popular method for determining the number of latent dimensions is the factor analysis of a correlation matrix. Unlike factor analysis, which is based on a linear model, IRT assumes a nonlinear relationship between item performance and ability. Because multidimensional scaling (MDS) assumes a monotonic relationship this method may be useful for the assessment of a data set's dimensionality for use with IRT models. This study compared MDS, exploratory and confirmatory factor analysis (EFA and CFA, respectively) in the assessment of the dimensionality of data sets which had been generated to be either one- or two-dimensional. In addition, the data sets differed in the degree of interdimensional correlation and in the number of items defining a dimension. Results showed that MDS and CFA were able to correctly identify the number of latent dimensions for all data sets. In general, EFA was able to correctly identify the data's dimensionality, except for data whose interdimensional correlation was high.