A general approach to nonlinear factor analysis

The method presented attempts to allow for nonlinear, possibly nonmonotonic relations between manifest and latent variates. An attempt is made to provide a workable criterion for choosing between alternative models on the basis of observable data as well as for constructing the appropriate function. An idealized numerical example is given.The author is indebted to Mr. E. J. Burr of the Department of Mathematics, University of New England, and to Dr. J. A. Keats of the University of Queensland, for their advice and criticism.     

Articulatory mediation of speech perception: A causal analysis of multi-modal imaging data

The inherent confound between the organization of articulation and the acoustic-phonetic structure of the speech signal makes it exceptionally difficult to evaluate the competing claims of motor and acoustic-phonetic accounts of how listeners recognize coarticulated speech. Here we use Granger causation analyzes of high spatiotemporal resolution neural activation data derived from the integration of magnetic resonance imaging, magnetoencephalography and electroencephalography, to examine the role of lexical and articulatory mediation in listeners’ ability to use phonetic context to compensate for place assimilation. Listeners heard two-word phrases such as pen pad and then saw two pictures, from which they had to select the one that depicted the phrase. Assimilation, lexical competitor environment and the phonological validity of assimilation context were all manipulated. Behavioral data showed an effect of context on the interpretation of assimilated segments. Analysis of 40 Hz gamma phase locking patterns identified a large distributed neural network including 16 distinct regions of interest (ROIs) spanning portions of both hemispheres in the first 200 ms of post-assimilation context. Granger analyzes of individual conditions showed differing patterns of causal interaction between ROIs during this interval, with hypothesized lexical and articulatory structures and pathways driving phonetic activation in the posterior superior temporal gyrus in assimilation conditions, but not in phonetically unambiguous conditions. These results lend strong support for the motor theory of speech perception, and clarify the role of lexical mediation in the phonetic processing of assimilated speech.     

A general approach to confirmatory maximum likelihood factor analysis

We describe a general procedure by which any number of parameters of the factor analytic model can be held fixed at any values and the remaining free parameters estimated by the maximum likelihood method. The generality of the approach makes it possible to deal with all kinds of solutions: orthogonal, oblique and various mixtures of these. By choosing the fixed parameters appropriately, factors can be defined to have desired properties and make subsequent rotation unnecessary. The goodness of fit of the maximum likelihood solution under the hypothesis represented by the fixed parameters is tested by a large samplex ² test based on the likelihood ratio technique. A by-product of the procedure is an estimate of the variance-covariance matrix of the estimated parameters. From this, approximate confidence intervals for the parameters can be obtained. Several examples illustrating the usefulness of the procedure are given.This work was supported by a grant (NSF-GB 1985) from the National Science Foundation to Educational Testing Service.     

An intrasubject approach to causal attribution

The present investigation draws on the judgment research tradition in order to examine the causal attributions made by individual subjects in an often used attribution task. Formal empirical tests of Kelley's (1967) attribution theory have demonstrated that attributions are influenced by the interaction of consensus, distinctiveness, and consistency information. None of these studies, however, have separately examined attributions made by individual judges. Implicit assumptions about individual differences, for example, have been made by the template-matching model of causal attribution (Orvis, Cunningham, & Kelley, 1975) but have not been scrutinized at the intrasubject level. Log linear modeling of attributions in the present research showed that while subjects were influenced by the causal information in the task, the relation between this information and attributions was more importantly characterized by individual differences than by uniform patterning. The nature of these individual differences and the significance of an idiographic approach to causal analysis are discussed.     

A general approach to procrustes pattern rotation

A method is introduced for oblique rotation to a pattern target matrix specified in advance. The target matrix may have all or only some of its elements specified. Values are estimated by means of a general procedure for minimization with equality constraints. Results are shown using data from Harman and Browne.The author is indebted to K. G. Jöreskog and M. W. Browne for many helpful suggestions and comments throughout the course of this study. Thanks are also due P. H. Schönemann and H. Weiner.     

The response styles theory of depression: A test of specificity and causal mediation

This prospective study tested the diathesis-stress and causal mediation components of the response styles theory of depression. In addition, it examined whether rumination predicts increases in anxious as well as depressive symptoms. At Time 1, 87 college students completed measures of rumination, hopelessness, depressive symptoms, and anxious symptoms. Participants also completed measures of hopelessness, depressive symptoms, and anxious symptoms at three time points later in the semester: immediately after receiving their most difficult midterm exam grade (Time 2), 4–8 hours later (Time 3), and 4 days later (Time 4). Regardless of exam outcome, the tendency to ruminate in response to depressed mood was associated with: (1) increases in anxious symptoms between Time 1 and Time 3; and (2) increases in both anxious and depressive symptoms between Time 1 and Time 4. In addition, the relationship between rumination and increases in both depressive and anxious symptoms was mediated by hopelessness. In other words, individuals with a ruminative response style exhibited increases in both depressive and anxious symptoms because they exhibited increases in hopelessness.     

Employees' reactions to psychological contract breach: A moderated mediation analysis

This study examined the mediating role of acquiescent silence in the relationship between psychological contract breach and employee job satisfaction, as well as the moderating roles of perceived ethical climate on that relationship and on the mediation effect of acquiescent silence. Survey data were collected from a sample of 273 full-time employees from nine high-tech firms in Taiwan. The results showed that acquiescent silence partially mediated the relationship between psychological contract breach and job satisfaction. Moreover, the results from the moderation analysis showed that perceived ethical climate moderated the influence of psychological contract breach on acquiescent silence and the results from the moderated mediation analysis revealed that this mediation of acquiescent silence was moderated by perceived ethical climate such that at the lower level of perceived ethical climate, the mediation effect of acquiescent silence became stronger. Implications for managerial practices and suggestions for future research were discussed.     

A general approach to testing multifaceted personality constructs.

Some personality characteristics are composed of multiple, distinct subcomponents (e.g., Type A, hardiness, attributional style, self-monitoring). The advantages and disadvantages of 3 typical approaches to testing these constructs are reviewed. An alternative approach based on structural equation modeling is then offered. This approach has many advantages over its alternatives, including the provision of an explicit test of the structure of the multifaceted construct, the simultaneous test of the effects of this general construct and the unique aspects of its subcomponents, and the explicit consideration of measurement error. Although the modeling approach does have limitations, these limitations are equally applicable to all of its alternatives. Indeed, the principal disadvantages of the modeling approach seem to be its statistical complexity and the lack of education regarding its proper use.     

Reaction time analysis of simple mental tasks: A general approach

A general model for choice reaction time (RT) is presented which is based on a theory that describes the operational characteristics of a problem solving subject. The model is not a stage model neither in the way envisioned by Donders nor in the sense proposed by Sternberg. Rather, it is based on the assumption that RT is the result of an underlying process that continually switches between two states: a state of processing and a state of non-processing. In the state of processing the subject executes the so-called operations that are needed to solve a problem. However, from time to time the subject enters a state of non-processing. During the latter state, the subject either processes those operations that are not directly related to the execution of the task or uses this time to recover in order to prevent the system from becoming overloaded. It is shown that the model successfully describes the RT data of three different tasks: a simple choice task, a mental addition task and a mental rotation task. The model provides a means to study individual differences not only in cognitive psychology but also in the realm of mental testing. Special attention is given to the problem of parameter estimation and the statistical properties of the estimators.     

Verisimilitude: a causal approach

I present a new definition of verisimilitude, framed in terms of causes. Roughly speaking, according to it a scientific model is approximately true if it captures accurately the strengths of the causes present in any given situation. Against much of the literature, I argue that any satisfactory account of verisimilitude must inevitably restrict its judgments to context-specific models rather than general theories. We may still endorse—and only need—a relativized notion of scientific progress, understood now not as global advance but rather as the mastering of particular problems. This also sheds new light on longstanding difficulties surrounding language-dependence and models committed to false ontologies.