One assumption shared by many contemporary models of leadership is that situational variables moderate the relationships between leader behaviors and subordinate responses. Recently, however, R. J. House and J. L. Baetz (1979 in B. Staw & L. Cummings, Eds., Research in Organizational Behavior (Vol. 1), Greenwich, Connecticut, JAI Press) have suggested that the effects of some leader traits and behaviors may be relatively invariant; that is, have the same effects in a variety of situations. One possible class of leader behaviors which may have relatively consistent effects across situations are those known as leader reward and punishment behaviors. The first goal of the research reported here was to increase our understanding of the relationships between leader contingent and noncontingent reward and punishment behaviors and subordinate responses. Contingent reward behavior was found to have the most pronounced relationships with subordinate performance and satisfaction, followed by noncontingent punishment behavior. Neither leader noncontingent reward nor contingent punishment behavior were found to be related to either subordinate performance or satisfaction, with the exception that noncontingent reward behavior was negatively related to subordinates' satisfaction with work. The second goal of the research was to examine the effects of a variety of potential moderators on the relationships between leader reward and punishment behaviors and subordinate responses. The results of this study suggest that the relationships between leader reward and punishment behaviors and subordinates' performance are relatively free of moderating effects.
In a test of hypotheses derived from the literature surrounding the Vertical Dyad Linkage (VDL) Model of Leadership, 45 supervisor-subordinate dyads were studied in a setting which differed from much of past VDL research. The results generally confirmed previous findings in that IN-group status was associated with higher performance ratings, reduced propensity to quit, and greater satisfaction with supervision. Also, objective measures of actual job performance yielded results which were in a predicted direction. The need for conceptual and operational revisions of VDL constructs is discussed.
With extended practice on a task a shift seems to occur from controlled to automatic processing. If differential factors associated with automatic processing exist, they could only be observed after this shift occurs. Hence, any such factor would be late-appearing, in the sense that it could only be identified late in practice. The present paper reports two tests of the existence of late-appearing factors. Both tests involved extended practice on five video games; the two tests were carried out in two different populations approximately 1 year apart. The results of the two experiments were in complete agreement. In both cases all factors, with one possible exception in the second experiment, were identified by content exclusively and not by stage of practice. The results, therefore, are negative. Other studies using other materials, other subjects, or other conditions of practice may reach different conclusions; but the studies reported in this paper offer no support for the existence of late-appearing factors.
Most studies comparing the relative effectiveness of multiattribute weighting techniques have found few differences. These studies have, however, failed to systematically vary attribute properties such as number of attributes and distribution of correct attribute weights. It was hypothesized that any weighting technique would be more effective at arriving at the correct weights the smaller the number of attributes because of lessened information processing requirements. In addition, it was hypothesized that the relative effectiveness of different weighting techniques would depend on the peakedness of the distribution of correct attribute weights because different weighting techniques should generate more peaked distributions than others. Two experiments were conducted to test these hypotheses. The first experiment focused on the accuracy of the weights assigned to attributes by individuals; the second on the accuracy of groups. Both experiments confirmed the first hypothesis regarding the number of attributes, but only the first experiment confirmed the second hypothesis regarding the peakedness of the distribution of attribute weights.
The judgmental accuracy of group majority decision making in binary choice, multiple-item prediction tasks as a function of group size, average competence of group members, and the shape of the overall distribution of judgmental accuracy in the group are discussed. For the case where the addition of an (N + 1)th member to a group of size N has a known cost function and the value of a correct group decision can be specified, the optimal group size is calculated.
Entrapping conflicts are characterized by the tendency for individuals to make increasing commitments to some failing course of action, in large part to justify the appropriateness of previous investments made in that situation. It has been suggested that (1) these dilemmas frequently arise in organizational settings, and (2) whether decision makers escalate their commitment or withdraw from a failing course of action can be affected by a variety of factors. The present four experiments were designed to determine if an individual's degree of entrapment (i.e., tendency to escalate) may be affected by the behavior of a model in a similar situation. In Experiments 1–3, considerable generality was obtained for the proposition that decision makers in entrapping conflicts can be directly influenced by a model. Individuals who witnessed an entrapped model became significantly more entrapped than those who did not. This effect was obtained (1) for subjects and models of both sexes, (2) on measures of both process and outcome, (3) across two different experimental procedures, and (4) when the model was viewed either during or before the time that subjects were called upon to make their decisions. Experiments 3 and 4 delineated several limiting conditions of the modeling-entrapment relationship. Participants did not invest more of their resources in the presence than in the absence of an entrapped model if the model expressed regret rather than pleasure about his behavior. In fact, under such conditions a significant “reverse modeling” effect was obtained, in which subjects became less entrapped in the presence of the entrapped model. The modeling-entrapment relationship was also significantly reduced when the model was unlikable and unintelligent, and thus not an appropriate person for comparison. Theoretical and practical implications are discussed, including the possibility that models may be skillfully employed to help decision makers stop “throwing good money after bad.”
In this study two training procedures designed to reduce the adverse influence of irrelevant information were tested for their efficacy. One training procedure involved a lecture (paralleling typical classroom experience), while the other involved interaction and practice (paralleling typical laboratory experience). In a pretest, irrelevant information was shown to influence the judgments of 12 experienced student soil judges. The judges were then given lecture training; this was found to be of minimal help in reducing the influence of irrelevance. Next, the judges received the interactive training and that had a significant impact. In addition to reducing the influence of irrelevance, the interactive training also improved the accuracy of the judgments. In a followup study, five of the soil judges were reevaluated over a year later; the training appeared to have continued impact. Implications and extensions of the training procedures to other areas are discussed.
The relative effectiveness of individual and group performance was investigated by employing three decision making strategies. Ignoring this consideration, group decisions were superior (p ⩽ .01) to individual averages, equal to the selected best individual decisions, and inferior to the actual best individual solutions (p ⩽ .05). The results indicated that the degree of process loss/gain and the relative performance of groups and individuals was significantly influenced by the dependent variable utilized and the decision making strategy employed.
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.
In recent years, some degree of controversy has arisen over the methods that researchers should employ in the detection of moderating effects. More specifically, both M. R. Blood and G. M. Mullet (1977, Where have all the moderators gone?: The perils of type II error, Tech. Rep. No. 11, College of Industrial Management, Georgia Institute of Technology) and H. J. Arnold (1982, Organizational Behavior and Human Performance, 29, 143–174) have challenged the use of “conventional” moderated regression (e.g., S. Zedeck (1971, Psychological Bulletin, 76, 295–310) as an appropriate method for the analysis of moderating effects. Blood and Mullet, for example, have argued that conventional moderated regression is an overly “conservative” technique that is generally incapable of detecting moderating effects—even in data bases “constructed” so as to have strong interaction components. To remedy this problem, they suggest a “backward entry” regression analysis in which the interaction term is the first variable entered into the regression. Also critical of conventional moderated regression, Arnold argues that the same analytic strategy is inappropriate in instances where the researcher's concern is to demonstrate differing “degrees” of correlation between two variables for moderator variable based “subgroups.” The purpose of the present paper is to show that both the arguments of Blood and Mullet and those of Arnold are incorrect. The difficulties associated with the backward entry procedure are demonstrated through the use of Monte Carlo simulation methods. Results of the simulations revealed that the moderated regression analytic procedure is well suited to the detection of statistical interactions (i.e., moderating effects)—even in data bases constructed so as to have (a) very strong main effects for both the independent variable and the moderator variable, (b) dependent variables having large error components, (c) independent and moderator variables having only modest reliability levels, and (d) partially redundant (multicollinear) independent and moderator variables. The errors inherent in the recent arguments of Arnold are shown to result from (a) an unduly restrictive definition of the “degree of relationship” concept, and (b) a seeming belief that differences in correlation coefficients have necessary implications for the accuracy with which scores on one variables can be predicted on the basis of knowledge of scores on another variable. Implications of the present study's analyses are offered.
This paper reports results from a series of studies designed to directly test a categorization-based model of leadership perceptions in three areas: specifying the internal structure of leadership categories, investigating the accessibility and importance of various individual behaviors in making leadership judgments, and explaining leadership perception in terms of categorization. In Study 1, data provided by 263 subjects were used to calculate three measures of category representativeness: family resemblance scores, cue validity scores, and prototypicality ratings. Results showed that leader family resemblance, cue validity, and prototypicality were all strongly correlated. In Study 2, accessibility was investigated by administering to 64 subjects a specially constructed questionnaire containing leader behaviors which varied in prototypicality and then measuring the reaction time to rate the behavior as prototypical of a leader. Results showed that there was a significant negative correlation between prototypicality and reaction time to behavioral items, suggesting that more prototypical items were more easily accessed. Study 3 manipulated prototypicality of leadership behaviors for an experimental group where leadership had been primed extensively and for a group of subjects who had been given no primes. The results showed the prototypicality manipulation strongly affected leadership ratings, behavioral expectations, and causal ascriptions to the target person, but the priming factor had no significant effects on dependent variables. Suggestions for future research and practical implications are also discussed.
The purpose of this study was to examine how individual differences in repression—sensitization moderate the effects of negative feedback in task groups. Seventy-eight research participants, organized into 26 simulated task groups, were asked to work as a team on a series of anagram problems. Participants received either unfavorable individual or team feedback concerning performance. It was predicted that the direct threat associated with individual feedback would increase defensiveness among repressors. This hypothesis was supported; that is, repressors, in comparison to sensitizers, denied the group task as a measure of ability and effort and reported lower attraction to the task and group. It was predicted further that the indirect threat of team level feedback would increase defensiveness among sensitizers. This hypothesis was also supported. Here sensitizers, in comparison to repressors, denied the task as a measure of ability and effort and indicated lower attraction to the task and group. Additional analyses of covariance revealed that the cognitive processes associated with self-attributions (i.e., ability and effort) mediated the reports of task and group attraction. Theoretical and practical implications of these findings and directions for future study are also discussed.
In a series of experiments subjects estimated the risk of dying during a year for different hypothetical persons. The risk to die varied across persons and was always presented as three different risks during three time periods (e.g., for person x: first period 8 weeks with risk 13 deaths per thousand persons and year; second period 20 weeks with risk 1.5 deaths per thousand persons and year; third period 24 weeks with risk 6.0 deaths per thousand persons and year). The results indicated that when the risk level information correlated higher with statistical risk than the exposure times, the judgment heuristics applied were quite efficient. However, when the exposure times explained more of the variation in statistical risk than the risk level variation the heuristics no longer worked so well. When the statistical risk was held constant the judgments were very highly correlated with the risk during the shortest time period having the greatest variation in risk level and greatest mean. Retrospective verbal reports indicated a number of different judgment heuristics. A common characteristic of these was a two-stage sequential information processing starting with a search for an anchor on the judgment variable of risk. In the second stage the anchor was adjusted in accordance with the remaining information considered important. An initial search of the longest exposure time for using the corresponding risk as the initial anchor was the most common strategy. After this, the remaining two time periods and their risks were considered when adjusting the anchor to obtain the final judgment. Analyses of the results in terms of the weights given the risk levels indicated proportionality with exposure time for the risk during the longest time period. To summarize, man seems to be a good intuitive statistician in the task studied here if the exposure times of different risks do not explain more of the statistical variance than the risk levels themselves.
This research examined the influence of perceived competence in moderating the relationship between role clarity and the job performance of employees. Data from 340 lower level managers were used to test two competing hypotheses. Results indicate that perceived competence moderates the relationship between role clarity and the job performance of employees. The job performance of employees is a joint function of their perceived competence and the role clarity they experience. Consistent with a situational-constraints perspective on work performance, the job performance of more competent employees is more strongly affected by the role ambiguity they face than is that of their less competent counterparts. The alternative, ability/adaptability, hypothesis did not receive support. Implications of the findings for future research and for remedial action by organizations are discussed.
: The broad question addressed by this research is: How good are humans at balancing the costs and benefits of their information acquisition? Do they buy those, and only those, sources of information whose acquisition cost is outweighed by the improvement in decision quality that their use makes possible? The evidence reported here, together with that reviewed earlier, suggests that the answer is not encouraging. Specifically, the present findings extend those noted earlier in suggesting: (1) That the pattern of overpurchase for low-consequence decisions, and underpurchase for high-consequence decisions, is robust to variation in overall cue validity, as well as to procedural modifications such as manual versus computer-interactive transactions (Experiment 1). (2) That overpurchase is frequently coupled with mispurchase (Experiments 2 & 3). That is, subjects, in addition to buying overall more information than was normatively justified, frequently bought expensive cues when cheap, equally-valid ones were available (Experiment 2), or low-validity cues when higher-validity, equally-costly cues were available (Experiment 3). (3) That subjects perceive equally-valid cues as of differential validity (Experiments 1 & 2), and are able to detect real validity differences between cues reliably only when the differences are large (Experiment 3). Purchase behavior is generally shaped by these perceptions of validity, whether well-founded or not, though the relationship disappears when equally-valid cues are offered at different costs (Experiment 2).
The predictive utility of three mathematical models of psychological inequity was competitively tested with survey data obtained from a sample of employed graduate students. Index values of inequity, which were generated from the three models, were examined for curvilinear association with affective outcomes. In terms of satisfying the criterion of displaying an invertedU association (such that overcompensation and undercompensation were associated with greater negative affect relative to equitable compensation), a power-function weighting scheme was identified as a modestly successful algebraic representation of psychological inequity. The study represents the first attempt at explicitly examining mathematical models of individual psychological inequity.
Previous research has demonstrated that performance cues can significantly influence raters' responses on leader-behavior rating questionnaires. The purpose of the present study was to explore the cognitive mechanisms that mediate this influence. Subjects watched a videotape of a problem-solving group and later rated the group leader's behavior. Perceptions of the group's performance were manipulated both before (preobservation) and after (postobservation) subjects watched the tape. Both manipulations affected the leader-behavior ratings. Analytic procedures borrowed from signal detection theory suggest that the preobservation performances cue effects were mediated by selective encoding of leader-behavior information in memory, while the postobservation effects were mediated by probabilistic response biases. Selective memory retrieval, on the other hand, did not seem to play a role. The implications of these results for developing strategies to overcome the confounding influence of performance cues on leader-behavior ratings are discussed.
