Chester A. Schriesheim

Controlling Acquiescence Response Bias by Item Reversals: The Effect on Questionnaire Validity

The prevailing conventional wisdom is that it is advisable to mix positively and negatively worded items in psychological measures to counteract acquiescence response bias. However, there has been virtually no unambiguous empirical evidence to support this recommendation. Thus, an experiment was conducted to evaluate the ability of subjects to respond accurately to both positive and reversed (negative) items on a questionnaire. Items from the LBDQ—XII Initiating Structure and Consideration subscales were used to create a written description of a fictitious manager. One hundred-fifty subjects, all upper-division business undergraduates, were given the written managerial description and then asked to complete a questionnaire containing the twenty Initiating Structure and Consideration items. The managerial descriptions were in two forms (to portray high and low Initiating Structure), and the questionnaires contained items in three forms (all positively worded, all negatively worded, and mixed). The data were evaluated using a one-way analysis of variance and post hoc t-tests. Significant differences in response accuracy were found between the item wording conditions. It was concluded that it may not be advisable to employ reversed (negatively-worded) items to control acquiescence response bias, as such changes may actually impair response accuracy.

The Similarity of Individual Directed and Group Directed Leader Behavior Descriptions

The similarity between individual directed and group directed leadership descriptions of subordinates was examined among 308 public utility employees. The correlations among the variables revealed ...

The Effect of Grouping or Randomizing Items on Leniency Response Bias

This study was concerned with the effects of item presentation mode on the degree of leniency bias inherent in responses to standard field research questionnaires. Two types of modes were examined : the first with items measuring the same dimensions grouped together and the second with items measuring the same dimensions distributed randomly throughout the questionnaire. Sixty respondents completed questionnaires containing items from the Leader Behavior Description Questionnaire (Stogdill, 1963) and the Michigan Four Factor Leadership Questionnaire (Taylor and Bowers, 1972); there were thirty respondents in each of the modes (grouped and random). The random relative to the grouped mode showed substantially less leniency response bias, as assessed by both correlational and factor-analytic procedures. However, the magnitude of these effects was still considerable for the random mode. There were no notable differences in leniency effects between the two questionnaires, but some differences were obtained for specific leadership dimensions. Based upon these and the total set of findings, implications for questionnaire validity and for future research are discussed.

The effect of leniency on leader behavior descriptions

Abstract This article reports the results from five studies designed to estimate the magnitude and effects of leniency on subordinate-provided descriptions of leader Consideration and Initiating Structure. Using a newly developed scale to measure leniency response bias (the tendency to describe others in favorable but probably untrue terms), the results of the studies indicate that subordinate reports of leader Initiating Structure are not particularly susceptible to the effects of leniency. The results for Consideration, however, showed that: (1) consideration items were not socially neutral and were susceptible to leniency; (2) consideration reflected an underlying leniency factor when applied in a field setting; (3) leniency explained a substantial proportion of the variation involved in Consideration-dependent variable correlations; and (4) this did not appear to result from conceptual overlap between Consideration and leniency, but rather from spurious correlation through the susceptibility of the measures to the effects of leniency. The importance of these findings for leadership research are discussed, and several alternatives for the control of leniency in leader behavior description are discussed.

The Great High Consideration—High Initiating Structure Leadership Myth: Evidence on its Generalizability

Summary This paper examines the purported superiority of the high-Consideration, high-Initiating Structure leadership style, where it is claimed that American leaders with this style have more satisfied subordinates than leaders with other Consideration-Structure combinations. For two different versions of the Ohio State leadership scales, regressions were calculated with the use of data collected from four different U.S. samples: (a) university maintenance workers (N = 230), (b) employed students (N = 178), (c) middle managers in an air-sea transport company (N = 96), and (d) employees in the operations division of a large public utility (N = 258). On the basis of the regression results, it was concluded that the superiority of the high-high leadership style is indeed a myth, and that Consideration alone explains almost all of the variance in subordinate satisfaction. The implications for testing complex leadership models are briefly discussed.

The Invariance of Anchor Points Obtained by Magnitude Estimation and Pair-Comparison Treatment of Complete Ranks Scaling Procedures: an Empirical Comparison and Implications for Validity of Measurement

This article examined the invariance of results from magnitude estimation and pair-comparison treatment of complete ranks scaling procedures and explored whether roughly interval measurement is achieved by these methods. Magnitude estimation and ranking questionnaires were administered to the same set of 116 college students, and their responses were used to scale sets of 20 and 10 frequency expressions, respectively. A comparison of the results obtained with those from two earlier studies indicated that magnitude estimation scaling produced fairly invariant anchor points which appear to attain at least roughly equal interval measurement, whereas pair-comparison treatment of complete ranks did not. Implications for validity of measurement are briefly discussed.

Item Presentaton as an Influence on Questionnaire Validity: a Field Experiment

This study was concerned with the effects of questionnaire format on convergent and discriminant validity. Two types of formats were examined: the first with items measuring the same dimensions grouped together and the second with items measuring the same dimensions distributed randomly throughout the questionnaire. Sixty subjects completed questionnaires containing items from the Leader Behavior Description Questionnaire (Stogdill, 1963) and the Michigan Four-Factor Leadership Questionnaire (Taylor and Bowers, 1972); there were thirty subjects in each of the conditions (grouped and random). The random relative to the grouped condition showed superior convergent and discriminant validity, as assessed by both multitrait-multimethod and analysis of variance procedures. Implications for questionnaire validity are discussed.

Contributions of the Ohio State Studies to the Field of Leadership

The Ohio State Leadership Studies have contributed in general to the fields of management and organizational behavior, and specifically to the field of leadership. These contributions can be summarized in several categories. Process contributions include the use of an interdisciplinary approach to the study of leadership phenomena as well as a consistent emphasis on performing high-quality research. Conceptually, the Ohio State studies helped to shift the focus of the field from a universal trait approach to a more situational, behavioral-based view. Methodologically, the Ohio State studies contributed a model of programmatic construct validation and investigation, and provided future researchers with useful research instruments. In sum, these contributions are substantial and are worthy of continued recognition.

Leniency Effects on Convergent and Discriminant Validity for Grouped Questionnaire Items: A Further Investigation

Two previous investigations revealed that grouping (rather than randomizing) questionnaire items measuring similar constructs (in subsections) resulted in impaired discriminant validity (Schrie-sheim and DeNisi, 1980) and that grouping also strengthened the impact of leniency response bias (Schriesheim, 1981). This study reanalyzed the data of the two earlier investigations to determine whether the impairment of the discriminant validity of the grouped questionnaire items might be due to spurious correlations attributable to leniency. The responses of thirty discount store employees to a questionnaire containing grouped items measuring leniency in leader behavior descriptions (Schriesheim, 1980) and four similar constructs from the Leader Behavior Description Questionnaire (Stogdill, 1963) and Four-Factor Theory Questionnaire (Taylor and Bowers, 1972) were examined for convergent and discriminant validity by using a traditional zero-order multitrait-multimethod correlation matrix analysis (Campbell and Fiske, 1959) and by statistically controlling the effect of spurious correlation attributable to leniency (by computing a first-order partial multitrait-multimethod correlation matrix, controlling for leniency as measured by the leniency scale). It was found that controlling for leniency resulted in a slight decrement in convergent validity, but that discriminant validity was substantially improved. Implications for questionnaire validity and further research are discussed.

Social Desirability and Leader Effectiveness

The effect of social desirability response bias on the Least Preferred Co-worker (LPC) scale used in tests of Fiedlers Contingency Model of leadership effectiveness was investigated in two studies. In the first study, 89 managers of a steel fabricating plant completed a questionnaire containing the Crowne and Marlowe (1) measure of social desirability and three scales used by Fiedler in tests of his model (including LPC). None of these measures correlated significantly with social desirability. In the second study, 16 first-line construction crew supervisors were randomly assigned to two conditions. In the first condition, as the supervisors arrived at the testing site, they were taken aside and told by the companys president to be supportive and considerate of subordinates. In the second condition, the supervisors were not taken aside nor told to be considerate. A manipulation check using the Crowne-Marlowe social desirability scale indicated a significant difference between the two conditions (as expected), but no differences were obtained in any of the three Contingency Model measures administered to the respondents (including LPC).