Implementing process innovations: the benefits of combining delegative-participative with consultative-advisory leadership.

Table 2 presents the means, standard deviations, and intercorrelations of the study variables. As expected, both leadership facets were significantly positively related to success regarding the implementation of process innovations. We had argued earlier that each of these two leadership patterns is connected with undesired secondary effects and that the respective other strategy is necessary as a countermeasure to compensate these negative effects. To test this assumption, we conducted a hierarchical regression analysis of implementation success. In the first step, we entered job tenure and degree of innovation as control variables. We added delegative-participative leadership and consultative-advisory leadership in the second step of the regression. Finally, in the third step, we entered the interaction effect of both leadership patterns. Table 3 summarizes the results.

The regression weights of the control variables were not significant. In support of Hypotheses 1 to 3, adding delegative-participative leadership and consultative-advisory leadership as main effects in the second step explained a significant portion of the variance. Entering the interaction term of both leadership components in the third step likewise yielded a significant change in the amount of variance explained (see Table 3). Specifically, the interaction explained an additional 3% of the implementation success variance over and above the variance explained by the main effects. Together, the main effects and the interaction effect explained 30% of the differences regarding the success of the process innovation implementation. This is fairly substantial when considering that effectiveness and efficiency also depend on the quality and the potential of the process innovation itself. This amount of explained variance is also high in light of the fact that a common method variance might have occurred in this study--an aspect that we will discuss in the section on limitations.

Nevertheless, although the interaction effect is statistically significant, its direction is somewhat different than we had hypothesized. This is a surprising result that merits closer scrutiny. On the one hand, it appears to be the case that the highest level of implementation is evident when the levels of both leadership patterns are high. This is in line with our reasoning underlying Hypothesis 3: Combining both leadership patterns enables each leadership style to neutralize the discussed risks of the respective other set of leadership behaviors and vice versa. In this case--namely, when the levels of both leadership patterns are high--the respective advantages of each leadership style can be brought to full fruition.

On the other hand however, the results show that each of the two leadership styles is also conducive to success when the level of the respective other leadership pattern is low (see Figures 2 and 3). This finding contradicts our assumptions underlying Hypothesis 3.

[FIGURES 2-3 OMITTED]

Discussion

Contributions

The present study makes two contributions to the research literature. First, with respect to successfully implementing process innovations, it underscores the importance of the interaction between delegative-participative and consultative-advisory leadership. Based on a literature review of single studies and meta-analyses and personal communication with colleagues who specialize in leadership, we think that this is one of the few studies that examine the contribution of the interplay of those leadership facets on implementation success. Previous studies of innovation have only investigated one of the two leadership components discussed here (e.g., Axtell et al., 2000) but not their interaction. We believe however that viewing either delegative-participative or consultative-advisory leadership in isolation neglects the complexity of organizational leadership in action. We therefore propose that a more complete understanding of the factors underlying implementation success can be gained by examining the interplay of these two sets of leadership behaviors.

Second, the results of the present study increase our knowledge of the implementation process as we examine the conditions of successful implementation regardless of the long-term effects of innovations. This is important because there are very few empirical studies of the implementation of process innovations. Nevertheless, some scholars have focused on the implementation of innovations in general. For instance, De Dreu (2006) showed that innovation (operationalized as the implementation of new procedures and methods and assessed in interviews of team supervisors) is a curvilinear function of task conflicts in management teams. Such studies do not explain however in what ways leadership may have either functional or dysfunctional effects on implementation success.

Examining the Form of the Interaction Effect

Examining the form of the interaction effect we identified promises to shed some light on important issues. Situations in which there is low delegative-participative leadership are hardly uncommon in organizational settings. During the implementation of process innovations, leaders are often themselves the recipients of nonnegotiable instructions issued by their superiors. Frequently, they pass the thus generated pressure on to their subordinates. In this constellation, high levels of consultative-advisory leadership may be particularly conducive to successful implementation phases. When a subordinate is presented with a fait accompli, providing background information and offering advice may make it possible to develop or retain his or her motivation to cooperate despite the clear-cut guidelines that offer little leeway in regard to how the work is to be done. Such leadership actions may be construed by subordinates as considerate behavior on the part of the leader and may thus counteract some of the negative motivational effects of centralized decisions. If this reasoning is correct, this would mean that consultative-advisory leadership does not act as a buffer against the risks of high levels of delegative-participative leadership but rather serves to counter the risks entailed by low levels of delegation and participation.

The situation is similar when there are low levels of consultative-advisory leadership, which may also occur frequently in organizations. Often, leaders may either not be present physically or not have the time to offer consultations and advice. Or they themselves may not be sufficiently informed about the details of the process innovation. Particularly in this constellation, a delegative-participative leadership style may enhance implementation success because the subordinate thus receives the opportunity and the legitimization to help himself or herself in any way that appears expedient. Again, this would mean that delegative-participative leadership is not a buffer against the risks entailed by high levels of consultative-advisory leadership but rather acts as a countermeasure against the risks of a lack of consultative-advisory leadership.

Limitations

It is important to acknowledge the limitations of the present study. First, as is the case in most empirical studies of innovation, all constructs were measured retrospectively and may thus be affected by hindsight bias. Because one can assume that the degree of retrospective distortions will increase over time, the participants were asked to answer the questions with respect to more recent process innovations. We thus aimed to at least reduce the degree of retrospective distortions. To completely prevent hindsight bias effects however, longitudinal follow-up studies would be needed.

Second, the self-report data may be affected by common method bias (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). To prevent this problem, it is typically recommended that studies include different groups of people (e.g., members of the steering committee, project directors, developers of innovations, and users of innovations), with each group evaluating different variables (e.g., the leadership style is rated by the subordinates, whereas the implementation success is rated by top managers or customers). However, Spector (2006) argued that the potential problem of common method variance may be exaggerated. He provided empirical evidence that calls into question the assumption that the self-report method by itself yields systematic variance that significantly inflates results.

Nevertheless, we have addressed this problem by calculating the structure across all items of the study. Confirmatory factor analyses with maximum likelihood extraction were used to assess the fit of the general factor (G factor) model to the data in comparison to the subfactor model. If the model test confirmed the G factor model, this would indicate the presence of a common method bias. (In the confirmatory model tests, a correlation matrix was used as the starting matrix. The variances of the latent variables were fixed. All loadings and residuals were estimated freely.) However, the data did not support a G factor model. Several fit indices show that the G factor model does not fit the data (Adjusted Goodness-of-Fit Index = .54, Goodness-of-Fit Index = .64, root mean square error of approximation = .25, [chi square] = 392.91, p < .0001). By contrast, the three-factor model fits the data reasonably well (Adjusted Goodness-of-Fit Index = .90, Goodness-of-Fit Index = .92, root mean square error of approximation = .04, [chi square] = 6380.01, p < .0001).

Conclusions

Theoretical Conclusions