Variations in explicitly unpaired training are differentially effective in producing conditioned inhibition

To evaluate a contingency interpretation of conditioned inhibition (CI), rats were given “explicity unpaired” training in which the locus and duration of a CS within the inter-US (shock) interval were systematically manipulated for different groups. Summation and retardation tests in Experiment 1 indicated that stronger CI resulted from both a backward and a trace CS than from a midlocus CS of equal or greater duration. Complementing these findings, the same tests in Experiment 2 showed that, by comparison with novel-stimulus controls, CI developed to a trace CS but not to a mid-locus CS, nor to a trace CS that was accompanied by an immediate signal for the US. These findings argue against a contingency interpretation of CI and favor a contiguity interpretation stressing the short-term rehearsal of stimulus events. Such rehearsal of the US allows a backward CS, but not a mid-locus CS with an extended US-CS interval, to be discriminated as a signal for nonreinforcement, and thus to develop as a conditioned inhibitor. Similarly, excitatory conditioning to the memory trace of a CS allows the nominal trace CS to develop as a signal for nonreinforcement, and thus as a conditioned inhibitor, but not when its memory trace is overshadowed by another CS that immediately precedes the US. In short, the development of CI is facilitated when excitation is mediated by the memorial processing of either the US or a discrete CS for the US rather than by contextual cues.     

Time reallocation in a multiresponse environment: Effects of restricting response classes

Four adult male rats were each placed for three hours daily into an apparatus that provided individual compartments for six separate location-defined responses. The available responses consisted of: (1) the opportunity to turn off room lighting, producing darkness; (2) the opportunity to view a female rat; (3) the opportunity to turn off white noise; (4) the opportunity to drink; (5) the opportunity to eat; and (6) "other," representing time in the hallway between compartments. Each subject underwent a series of conditions characterized as an A-B-A-C-A design. Manipulations consisted of the removal of a low-probability response (darkness) and of a high-probability response (escape from noise) in a counter-balanced manner across subjects. The dependent measure for all subjects was the percentage of total session time spent in each compartment. Four predictive rules concerning the redistribution of behavior after response restriction were tested, including the constant-ratio rule, equal time redistribution, the most probable alternative, and the sequential-dependency rule. The results indicate no support for any of the four predictive rules and suggest that empirical assessment of restriction effects is necessary in reinforcement studies involving temporally extended responses.     

The validity of Mobley's (1977) model of employee turnover

Unlike earlier tests of an oversimplified version of this model, the validity of W. Mobley's (1977, Journal of Applied Psychology, 62, 237-240) original turnover model was fully investigated. Constructs that were neglected in prior studies were assessed and previously examined constructs were operationalized with more reliable measures. Measures of all constructs in Mobley's model were obtained from a survey of 192 hospital employees. Turnover data were collected a year following survey administration. Following the theoretical causal ordering of Mobley's constructs, each construct was regressed on all causally prior constructs. In general, each construct was accurately predicted by the linear combination of predictors representing its causal determinants. In the majority of instances, the best predictor of a model construct was the construct's immediate causal antecedent. Further, an alternative model was evaluated and compared with Mobley's model using path analysis.     

QUESTIONING THE EFFECTIVENESS OF BEHAVIOR MODELING TRAINING IN AN INDUSTRIAL SETTING1

The intention of this study was to improve behavioral modeling's effectiveness by substituting managers for professional trainers and to evaluate the effect on 44 male supervisors using Kirkpatrick's four levels of evaluation: reaction, learning, behavior on the job, and performance. Twenty-two supervisors were trained with six behavior modeling modules and the effect was compared to a control group consisting of 22 supervisors. The research also examined the effects of trainees’ self-esteem and the perceived power of the trainers. The results showed that behavior modeling resulted in favorable reactions and an increase in learning, but did not produce behavior change on the job, or improved performance results. Power and self-esteem did not moderate the training effectiveness. The findings are compared with previous behavior modeling research. The discussion concludes with a recommendation for researchers to identify more complete theoretical models which explain behavioral change on the job (e.g., Maltz's theory of psycho-cybernetics) as opposed to relying solely on Bandura's social learning theory.