Discrimination and reproduction of patterns: Feature measures and constraint redundancy as predictors

The statistical structure ofsamples ofpatterns [histoforms and polygons) was varied by manipulating constraint redundancy (R_c) at three levels. Both R_c and seven measures of pattern features were evaluated as predictors of reproduction errors and discrimination latency using multiple linearregression techniques. While it was possible to account for a significant proportion of performance variance from knowledge of R_c levels, feature measures accounted for more performance variance under all conditions of testing. The results were interpreted as support for the notion that patterns are processed in terms offeatures.     

A spatial analysis of the discriminability of forms in noise

Multidimensional scaling (MDS) techniques were used to study the form characteristics which transcend noise in a discrimination task. Three dimensions of random polygons were identified and were psychophysically related to quantifted form measures.     

Simultaneous vs sequential discriminations of Markov-generated stimuli

This experiment required Ss to make same-different classifications of Markov-generated histoforms that were distortions of three different prototypes. Task memory requirements were varied by presenting stimuli either simultaneously or sequentially for comparison, and the effects of shifts in memory requirements were also assessed. Classification performance in all cases was found to be superior under the simultaneous condition. Postshift performance was strictly a function of the mode of stimulus presentation following the shift, rather than the memory requirements present in the initial trials. These results, coupled with the observed relationships between performance and a measure of individual stimulus variability under the two memory conditions, strongly suggest that prototype encoding has limited relevance to performance in the same-different classification task used here.     

The variability of client,therapist and clinic in psychotherapy outcomes: A three-level hierarchical model

This article investigates the percentage of variance in psychotherapy outcomes explained by the case-mix variables for individual cases, by the therapist (therapist effects), and what additional variance is explained by the clinic with which the therapist is affiliated. While there has been substantial recent research regarding therapist effects, very little has been published regarding clinic-level effects after controlling for therapist effects. The study utilised the largest sample reported to date, using data from 28 clinics with a minimum of 2,000 cases in a clinical range of severity of symptoms with pre-post change scores on an outcome questionnaire. Only cases treated by a therapist with at least 30 cases were included. These selection criteria resulted in a case count of 156,258 clients treated by 874 therapists located at 28 clinics. After controlling for differences in case mix using diagnosis and intake score (severity of symptoms), the resulting analyses indicated that 5.21% of the variance in treatment outcome was explained by the therapist, while another 1.13% of the variance was associated with the clinic. Findings are discussed with implications for practice and policy.     

Evaluating risk factors in relation to preschool social–emotional skills using a person-centred approach

Few studies have utilised a person-centred approach when it comes to understanding risk factors in young children. The present study uses this type of approach while examining risk factors in relation to social–emotional skills. Parent reports for 444 preschool-aged children disclosed each child's exposure to eight different risk factors, as well as their social–emotional abilities. A latent class analysis was utilised to examine subgroups of risk, rather than using a cumulative approach, to see whether specific risk factors differentiated participants. Subgroup differences in distal social–emotional skills were then examined. The subgroup with systemic, economic and parent stress risk factors showed the lowest social–emotional abilities, in comparison to subgroups with only systemic or a combination of systemic and economic risks. These results suggest that addressing parent stress may be particularly relevant when considering children's social–emotional development. Implications for interventions and clinical practice are discussed.