Dispersion in models of categorical perception

A signal detection theory model of auditory discrimination with a nonlinear mapping from stimulus continuum to perceptual continuum can account for the enhanced discrimination at the category boundary found in categorical perception. Properties of this transformation are specified by a unimodal “dispersion function”. Furthermore, it is shown that a system consisting of two acoustic feature detectors with an associated decision function is also a dispersive system, which models categorical perception of a stimulus continuum as well as boundary shifts under adaptation. The effect of detector adaptation on discrimination is discussed in view of three different types of decision variable and different types of detector noise.     

Response pattern analysis of the Hand Test: age differences

Extended the literature on age differences on the Hand Test using a multivariate model to examine absolute and relative differences in response. Participants were 47 adults (M age = 22.47 yrs.), 24 males and 23 females; and 45 older adults (M age = 64.87 yrs.), 21 males and 24 females. Data were analyzed in terms of percentage of response and absolute frequency of specific responses, between the age groups. Results indicated similar findings to those previously reported for the Hand Test, though magnitude of personality deterioration or withdrawal was lessened, for the percentage analysis. Results indicated the importance of using both absolute frequency of response and percentage of response in the interpretation of projective test data, especially for older adults.     

Boundary disturbance in borderline and psychotic states

This study examines evidence for boundary disturbances in depressive, borderline, and schizophrenic hospitalized inpatients. Certain thought disorder signs on the Rorschach are reconceptualized as indices of boundary formation, which is defined as the capacity to create particular distinctions along wome bipolar coordinate of experience where previously no distinction was possible, and which defines each cluster of experience through a maintenance of differential functioning. The borderline group scored significantly high on the indices of laxness and moderately severe inner-outer boundary formation. The schizophrenic group scored significantly higher on indices of self-other and somewhat higher on severe inner-outer boundary disturbance. Schizophrenics may be organized at a level of self-other boundary formation which subsumes inner-outer boundary formation, while the borderline group is organized at a level of inner-outer boundary formation.     

Measurement and structure of pathological gambling behavior

The serious effects of pathological gambling on both gamblers and their families warrants research on the disorder. This study aimed at identifying components of pathological gambling with the goal of developing methods for early detection and preventive strategies. A factor analysis of behaviors hypothesized to relate to pathological gambling resulted in a multidimensional model of the disorder. Eight factors were identified, including distinct neurotic, psychopathic, and impulsive components, and used in a comparative analysis of 83 admitted compulsive gamblers and 61 nongamblers. Five of the factors differentiated the two groups even when sex, age, and education were controlled. Implications of these findings are discussed.     

Amount of information about the attitude object and attitude-behavior consistency

The synthesis of two separate lines of inquiry--research on information integration and longitudinal studies of attitudes--prompted the hypothesis that the degree of consistency between attitudes and behavior will increase as a function of the amount of information available about the attitude object. The hypothesis was tested in three separate longitudinal studies, ranging in length from 4 days to 4 months, that investigated the following behaviors: voting for candidates for political office, voting for two social policy election initiatives, and having an influenza vaccination. In support of the hypothesis, in each study, amount of information moderated the consistency between attitudes and behavior; and the significance of this relation remained even after controlling for the effects of a number of other potential moderators, including prior direct behavioral experience with the attitude object and attitude certainty. Consistent with previous research, direct behavioral experience was also a determinant of attitude-behavior consistency, and for the behavior of having an influenza vaccination, this relation was independent of the effect of amount of information. The discussion focuses on the interrelation among moderators of attitude-behavior consistency and on the theoretical implications of the findings.     

On the self-serving function of social anxiety: shyness as a self-handicapping strategy

We tested the hypothesis that socially anxious or shy individuals use their anxiety symptoms as a strategy to control attributions made about their performances in social-evaluative settings (i.e., self-handicapping strategies). Specifically, we predicted that trait-socially anxious or shy persons would report more symptoms of social anxiety in an evaluative setting in which anxiety or shyness could serve as an excuse for poor performance than would individuals in (a) an evaluative setting in which shyness was precluded as an excuse or (b) a nonevaluative setting. Furthermore, we predicted that this self-protective pattern of symptom reporting would not occur for individuals who were not trait-socially anxious because these persons would not commonly use such symptoms as a self-handicapping strategy. Results supported these predictions for male subjects, but not for female subjects. Sex differences in the strategic use of shyness are discussed in relation to other research on sex differences in the etiology and correlates of social anxiety.     

Statistical consistency and hypothesis testing for nonmetric multidimensional scaling

The properties of nonmetric multidimensional scaling (NMDS) are explored by specifying statistical models, proving statistical consistency, and developing hypothesis testing procedures. Statistical models with errors in the dependent and independent variables are described for quantitative and qualitative data. For these models, statistical consistency often depends crucially upon how error enters the model and how data are collected and summarized (e.g., by means, medians, or rank statistics). A maximum likelihood estimator for NMDS is developed, and its relationship to the standard Shepard-Kruskal estimation method is described. This maximum likelihood framework is used to develop a method for testing the overall fit of the model.