Heterogeneous heterogeneity by default: Testing categorical moderators in mixed-effects meta-analysis

Categorical moderators are often included in mixed-effects meta-analysis to explain heterogeneity in effect sizes. An assumption in tests of categorical moderator effects is that of a constant between-study variance across all levels of the moderator. Although it rarely receives serious thought, there can be statistical ramifications to upholding this assumption. We propose that researchers should instead default to assuming unequal between-study variances when analysing categorical moderators. To achieve this, we suggest using a mixed-effects location-scale model (MELSM) to allow group-specific estimates for the between-study variance. In two extensive simulation studies, we show that in terms of Type I error and statistical power, little is lost by using the MELSM for moderator tests, but there can be serious costs when an equal variance mixed-effects model (MEM) is used. Most notably, in scenarios with balanced sample sizes or equal between-study variance, the Type I error and power rates are nearly identical between the MEM and the MELSM. On the other hand, with imbalanced sample sizes and unequal variances, the Type I error rate under the MEM can be grossly inflated or overly conservative, whereas the MELSM does comparatively well in controlling the Type I error across the majority of cases. A notable exception where the MELSM did not clearly outperform the MEM was in the case of few studies (e.g., 5). With respect to power, the MELSM had similar or higher power than the MEM in conditions where the latter produced non-inflated Type 1 error rates. Together, our results support the idea that assuming unequal between-study variances is preferred as a default strategy when testing categorical moderators.     

Auditory evoked brain responses: Intensity functions from bipolar human scalp recordings

Average evoked brain responses were computed from four simultaneous recordings during binaural tone stimulation at 10 intensity levels. Amplitudes and latencies were measured for the prominent components, and linear regression coefficients were computed. The results indicate that the sensitivity of the evoked response to stimulus intensity is dependent upon the selection of the appropriate measure. Linear amplitude functions were demonstrated. Additional research relating these responses to psychophysical measures is suggested.     

Functional equivalence between same-different classifications and judged similarity of Markov patterns

This experiment represented an initial at tempt at comparing same-different classifications and similarity judgments of Markov histoforms sampled from three different schema families. A measure of individual stimulus variability was more strongfy related to classification responses than to judged similarity. Dichotomous responses, derived from the similarity data by imposing threshold criteria, were found to be highly similor to real classification responses when the threshold values were estimated directly from the latter; little correspondence, however, was obtained using optimal classification criteria. The results support the notion that somewhat different features are used in making classification responses and similarity judgments, and indicate the need for developing more sensitive tasks involving controlled scanning of pattem information.     

Visual and tactual form discrimination: Psychophysical comparison within and between modalities

Intermodal perceptual equivaience was investigated, using visual and tactual oddity discrimination and cross-modal matching-to-sample tasks. Four groups of Ss were presented with 100 problems made up of randomly derived forms from five levels of sidedness. The same quadratic relationship was observed for all tasks as a function of complexity, with optimal performance at eight sides. Comman information utilization, as indicated by intercorrelations of performance variables and by correlational analyses with differences in form measures, was not as pronounced within side classes. Evidence concerning the nature of pattern-feature usage in discrimination was obtained from graphical analyses.     

Schematic concept formation: Feature measures and constraint redundancy as predictors

Constraint redundancy (R_c) was varied in order to manipulate the statistical structure of sets of patterns. Measures of pattern features and R_c were related to the categorization by Ss who had experienced different types of preexposure to sets of patterns from the same stimulus domain. Both R_c and pattern measures were related to the categories produced by Ss in a schematic-concept-formation (SCF) task.     

Recognition assessed by D' and by a non-parametric alternative (the A-index) as a function of the number of choices

Freedom from restrictive assumptions and ease of psychological interpretation are two basic criteria for a satisfactory measure of recognition performance. The A-index (Brown, 1965a) comes close to fulfilling these criteria. It is based on the proportion of wrong choices rejected in a multi-choice test as revealed by the number of choices required to select the correct choice. It can be transformed into the d' measure of signal detection theory (SDT) if the standard assumptions of this theory hold, but not in general. Two experiments on the recognition of words-in-noise are described. In the first there were 3, 5, 8 or 16 choices. They were typed on a card which was displayed to the listener during the presentation of a word. On both measures 5-, 8- and 16-choice recognition did not differ significantly, while 3-choice recognition was slightly, but significantly, superior. The second experiment concerned the effect of delaying the display of the card until 2 sec. after presentation. No evidence for an effect was obtained. In both experiments, there was suggestive but not conclusive evidence that the d' measure tended to overestimate recognition efficiency.     

JOB DEVELOPMENT FOR THE MENTALLY RESTORED PSYCHIATRIC PATIENT1

The experiences of the placement department of the Mental Health and Manpower Project have demonstrated that an effective team can be created to develop job opportunities for the mentally restored psychiatric patient. Composed of a vocational counselor and a placement counselor, the team is effective in securing meaningful employment and developing re-education programs in the community. While the vocational counselor provides the clinical information concerning the restored patient-client, the placement counselor serves as a more personnel oriented individual in finding the appropriate kind of employment position. The team points to the potential that the mental health center possesses in involving the general community in its programs.     

Predicting the Success of Freshman Engineers

An awareness and commitment rating was assessed for a freshman engineering class to determine whether this type of non-intellective measure could help improve the predictability of academic success. The total sample was broken into subsamples based on the interview condition of the student: staff interview; alumni interview; or no interview. The existence of a factor defined as awareness/commitment was supported. For the total sample a significant correlation was found, but at a very low level. For the subsample interviewed by staff, a correlation of .31 was determined. This relationship suggests the possibility of real usefulness for the admissions interview in identifying the factor of awareness and commitment, and thereby improving the prediction of academic success in a professional curriculum such as engineering.     

Prior stimulation and the perception of causality

An attempt was made to discover whether, in conditions similar to those of Michotte, the number of causal responses given by subjects was influenced by stimulation during the period just before the main experiment. Time-intervals between the arrival of object A and the departure of object B were systematically varied within the range 60 to 390 msec. Those who had earlier been exposed to shorter time-intervals (60 to 210 msec.) gave fewer causal responses than those who had been exposed to longer ones (150 to 300 msec. and 240 to 390 msec.). A possible explanation is offered in terms of the modification of existing “schemata.”