Predicting effectiveness of bayesian classification systems

A model is presented for evaluating potential effectiveness of a Bayesian classification system using the expected value of the posterior probability for true classifications as an evaluation metric. For a given set of input parameters, the value of this complex metric is predictable from a simply computed row variance metric. Prediction equations are given for several representative sets of input parameters.     

The effects of sex and sex-role identity on the attribution of maladjustment

This study examined the effects of a stimulus person's gender and sex-role identity and an observer's gender and sex-role identity on the observer's judgments of the stimulus person's personality and level of adjustment. After having been classified as androgynous or nonandrogynous, 141 males and females viewed videotapes of a case conference on a bogus client. In the videotapes the client's gender and sex-role identity were factorially manipulated. As predicted, male clients and clients with a masculine sex-role identity were seen as possessing less favorable personality characteristics and as being less mentally healthy than were female clients and clients with a feminine history. Sex-role incongruence (e.g., a masculine female) influenced the subjects' judgments only of a female client. A sex-role congruent female was seen as more attractive and better adjusted than a sex-role incongruent female. The subjects' gender and sex-role identity did not influence their judgments of the clients in any consistent or interpretable fashion. Finally, it was found that the subjects were most influenced by sex-role congruence/incongruence of opposite sex stimulus persons. The discussion of these results centered on: (a) the relative influence of gender and sex-role identity on people's perceptions of a stimulus person and (b) the need for further exploration of how subject characteristics might influence these perceptions.This study is based on a Masters thesis submitted to the University of South Florida in partial fulfillment of the requirements for the Master of Arts degree. Portions of the study were presented at the 1978 meeting of the Southeastern Psychological Association. All correspondence should be sent to the second author c/o Department of Psychology, University of South Florida, Tampa, Florida 33620.     

Effects of sexual attitudes and physical attractiveness on person perception of men and women

A sample of 180 males and 180 females completed person perception measures for stimulus women who varied in physical attractiveness and permissiveness of sexual attitudes. The results indicated that both physical attractiveness and sexual attitudes had strong effects on peson perception. Physically attractive stimulus women were rated higher than the unattractive stimulus women on a physical appeal factor and an item measuring interpersonal attraction. The sexually conservative women were rated higher on an evaluative factor, lower on a personability factor, and higher on the interpersonal attraction item than the sexually permissive stimulus women. It was also found that the effects of physical attractiveness were greater for males than for females and that women rated the sexually permissive stimulus women less favorably than did men.     

A survey of communicative disorders students' attitudes toward stuttering

Cooper's (1975) Clinician Attitudes Toward Stuttering (CATS) inventories were conpleted by 1,902 speech-language pathology and audiology students from 33 universities, each in a different state. The majority of respondents perceived stutterers as having psychosocial problems and believed that client and parental counseling were essential aspects of treatment. These views were relatively unaffected by student training and clinical exposure. In addition, most of the assumptions of Johnson's diagnosogenic theory of stuttering were held by the majority of the students surveyed. Most of the respondents believed that stuttering was difficult to modify and that speech clinicians were neither adept nor comfortable treating stutterers. These latter attitudes were significantly affected by training, i.e., the more advanced the students' training, the more pessimistic their views on clinician competence. Other professional attitudes and the effect of training on them are also discussed.     

CHANGING PERSPECTIVES IN COGNITIVE NEUROPSYCHOLOGY*

Neuropsychology, the study of behavioral and mental disorders resulting from brain damage, seeks to integrate observable mental and motor functions with knowledge of nervous system structure and function. The first part of this paper deals with the empirical and theoretical contributions that arise from the study of classical clinical syndromes, from cognitive psychology based on information processing, and from the computationally-based study of complex cognitive systems such as language. The second part of this paper deals with two areas of neuropsychological research, memory mechanisms and motor behavior, that exemplify some of the methodological and conceptual problems that arise in trying to bridge animal and human studies, or psychological and physiological levels of analysis. Suggestions are made for how greater progress can be made by considering fundamental processes, and, by focusing on levels of abstraction useful to both the psychologist and the biologist. The last part of the paper briefly discusses some of the major issues involved in an attempt to establish a neurobiologically-based theory of human behavioral and mental function.     

Helping as a function of empathic responses and sociopathy

The present study reports two experiments which investigated helping as a function of empathic anxiety (anxiety in response to modeled distress) and individual differences in sociopathic tendencies. The first experiment concerned the effects of empathic anxiety, sociopathy, and perceived similarity/dissimilarity to a distressed person on helping. As predicted, subjects who experienced high levels of empathic anxiety helped more than subjects who experienced low levels of empathic anxiety. High sociopathic subjects experienced lower levels of empathic anxiety and helped less than did low sociopathic subjects. However, a path analysis disclosed that sociopathy's relationship to helping was not due to (or mediated by) empathic anxiety. The similarity manipulation did not influence empathic anxiety or helping. In the second experiment, subjects were classified as primary or secondary sociopaths (based on their levels of trait anxiety). As in the first experiment, empathic anxiety and sociopathy were significantly related to helping. Primary sociopaths displayed less empathic anxiety than secondary sociopaths, but the two groups did not differ in their willingness to help. Thus, in neither study could the sociopath's behavior be explained by empathic anxiety. The discussion concerned empathy as a mediational variable in helping and the role of personality variables in helping. A post hoc explanation of the sociopath's behavior, based on Schwartz' theory of the characteristics of helpers, was proposed.     

The word order problem in agrammatism : II. Production

This study demonstrates that agrammatic aphasics have difficulty describing relations that are signified by the order of noun phrases around verbs or prepositions. The word order deficit is found in a Sentence Order Task, in which written words must be arranged linearly to form a sentence, as well as in oral production. Ordering tendencies reflect factors like animacy or potency, rather than the systematic use of syntactic rules. These results point to a fundamental syntactic disorder in agrammatism.     

Consultative services in the schools: A model

A model Child Evaluation Center was established in a regular public school to demonstrate alternative methods for providing psychoeducational services. In addition to its service/demonstration function, the center functioned as a psychoeducational field training clinic for the training of school psychologists. The center was successful in encouraging teachers to utilize consultative services, and, in the process, reducing the number of referrals for formal psychological evaluation.     

On qualitative axiomatizations for probability theory

In the literature, there are many axiomatizations of qualitative probability. They all suffer certain defects: either they are too nonspecific and allow nonunique quantitative interpretations or are overspecific and rule out cases with unique quantitative interpretations. In this paper, it is whown that the class of qualitative probability structures with nonunique quantitative interpretations is not first order axiomatizable and that the class of qualitative probability structures with a unique quantitative interpretation is not a finite, first order extension of the theory of qualitative probability. The idea behind the method of proof is quite general and can be used in other measurement situations.This research was partially supported by the national Science Foundation grant NSF BNS7702911 and by the joint NSF-NIE grant NSF SED 78-22271 to the University of California, Irvine.