Raymond W. Preiss
Graduate Faculty, Masters of Servant Leadership Program
Professor and Chair
Department of Communication Studies
Copyright 2014. Raymond W. Preiss. All rights reserved.
The social scientific enterprise is both deductive and inductive. We consider the “nature” of a construct and deduce its operational indicators and relationships with other variables. When we test those relationships, we induce from the results to the theory as we assess the viability of the construct to account for findings. Because we know that sampling error will inflate and deflate scores, we never expect all tests to be consistent with a theory. In fact, it is a tautology to say that 5 in 100 experiments will produce false positive findings, as we set the “significance level” of our tests to p < .05. We are saying that the standard for a “significant” association between two variables (or difference between two groups) must be large enough that it only occurs due to random sampling error in 5 of 100 samples (of any number of participants). Thus, we expect (and accept) 5 false positives in 100 tests. There is no was to be sure that any single test is “true” because it may be one of the false positives.
In the same way that sampling error will always produce false positives, the design of a study can have features or deficiencies that introduce error into the measurement of variables. One type of error is “noise,” a random error that can obscure a “true effect” that is present. This can produce a “false negative” finding because the association (or difference) between variables is concealed by the noise. A well designed experiment and reliable measurement can reduce false negatives by controlling noise. There can also be bias in a scale that systematically distorts results. This can produce either false positives or negatives depending upon the type of bias produced. We are devoting attention to the scales used to assess SL to guard against false positives, false negatives, and bias.
We have been discussing several measures of the conceptual variable “servant leadership.” For our purposes, Greenleaf has been a primary source. We understand that there are other sources, and if you are in the Masters of Servant Leadership Program at Viterbo University you are well grounded in the religious and ethical roots of Greenleaf’s “theory” of servant leadership. Thus far, we have examined the Servant Leadership Scale (SLS; Barbuto & Wheeler, 2006), the Executive Servant Leadership Scale (ESLS; Reed, Vidaver-Cohen & Colwell, 2011), and the Servant-leadership Styles Inventory (SSI; Fridell, Belcher and Messner (2009). These are all self-report instruments, although some ask followers to rate a leaders’ SL and other scales ask leaders to rate their own SL qualities. In this entry, I will consider Ehrhart’s (2004) General Measure of Servant Leadership (GMSL).
All instruments measuring SL (or any other variable) are termed “operational variables.” These scales measure a “conceptual variable” that has been defined in the literature. This is why researchers are careful to design tests that measure each “indicator” of the conceptual definition. You noticed this process in the SLS and the ESLS, although there was some disagreement over what the dimensions or factors of SL were termed. Usually, the researcher will generate many items based upon the descriptions of the concept. These items are subjected to a factor analysis that determine how the items are associated or move together. As any one item is scored higher or lower, how do all the other items change (higher or lower)? The items that move together are said to “load” on a factor. Each factor should correspond with an established quality or characteristic of SL. That would suggest “content validity,” as the items that load together reflect a quality that the respondents agree upon.
Ehrhart (2004) briefly discusses the creation of the GMSL (p. 73). He identified seven categories of operational indicators of SL: Forming relationships with subordinates, empowering subordinates, enabling subordinate growth and success, behaving ethically, possessing conceptual skills, putting subordinates first, and creating value for those outside of the organization. He cites an earlier (unpublished) working paper as the source of these seven categories. Earlier in the manuscript (p. 69), he considers Graham’s (1991) distinctions between transformational leadership (TL) and (SL): Commitment to followers and stakeholders; providing a moral compass. He concludes his conceptual definition by contrasting SL with Leader-Member Exchange (LMX) and making connections between SL and Organizational Citizenship Behaviors (OCBs). This is some very fine work. After all, the year was 2004 and there was very little published experimental work on SL. I might argue that he needed more evidence from Greenleaf or that the seven categories do not adequately reflect the richness of SL seen in the lives of Jesus, Gandhi, or King. As a social scientist, the better question is: Did he give the hypothesis a fair test? To answer this question we will need to examine the scale.
The GMSL is composed of 14 statements, two for each of the seven categories. We are not told how the items were generated, if some items were discarded, or if experts looked at the items to confirm that they appeared to fit the seven categories. Ehrhart asks us to take many risks based upon our faith in his judgment. If the items are bad, they may inject noise into the experiment, producing false positives or false negatives. If the items create bias, they may systematically affect the results and prevent a fair test of the relationship. This is a good time to study the items used in the GMSL (Ehrhart, 2004, p. 93).
General Measure of Servant Leadership
- My department manager spends the time to form quality relationships with departmental employees.
- My department manager creates a sense of community among departmental employees.
- My department manager’s decisions are influenced by departmental employees’ input.
- My department manager tried to reach consensus among department employees on important decisions.
- My department manager is sensitive to department employees’ responsibilities outside the workplace.
- My department manager makes the personal development of department employees a priority.
- My department manager holds employees to high ethical standards.
- My department manager does what she or he promises to do.
- My department manager balances concern for day-to-day details with projections for the future.
- My department manager displays wide-ranging knowledge and interests in finding solutions to work problems.
- My department manager makes me feel like I work with him/her, not for him/her.
- My department manager works hard at finding ways to help others be the best they can be.
- My department manager encourages department employees to be involved in community service and volunteer activities outside of work.
- My department manager emphasizes the importance of giving back to the community.
Considering the year, there is much to applaud in this instrument. The 14-items are clear and reflect major aspects of SL theory. There are also many gaps. The items do not address issues such as stewardship, close listening, persuasiveness, foresight, or vision. However, these qualities, especially foresight and vision, might inject bias into the experiment. Thus, the scale is probably conservative, in that it underemphasizes potential sources in systematic error. This said, I wonder if the use of “department manager” limits the application of the scale. “Department” may be a dated term or an inappropriate description of many organizations, especially small businesses, service firms, or high tech companies. While these are validity concerns, Ehrhart (2004) reports strong reliability evidence (unit-level Cronbach’s alpha = .98). The scale may be conservative, but it is very reliable.
When you study the results, you don’t need to be a statistician or proficient in confirmatory factor analysis to follow the logic. In general, the follower-unit data supported the hypotheses better than did the manager-unit data. For followers, SL produced OCBs (helping and conscientiousness), especially when followers believed that the procedures were just. For managers, results were mixed. Recall the objective of this study. Ehrhart reasoned that units act collectively to display organizational citizenship behaviors. The follower-level data confirmed this outcome, while the manager-level data was less consistent. This suggests that the climate or culture associated with SL followers promotes OCBs, and that SL managers seem to underestimate the robust nature of this climate! This point alone makes the study very interesting. Take notice, servant leaders. The followers may be ahead of you! You may have been too busy being a visionary to notice! This finding reinforces the notion that servant leaders should be humble and stand aside while followers receive credit for accomplishments. Having faith in the SL climate may result in OCBs that leaders do not anticipate.
As you consider the complexity of this experiment, keep folding the results back upon the conceptual definition. With our understanding of Greenleaf, we are in a position to appreciate the power of SL organizational climates. While the GMSL was largely successful in producing a fair test, I would want to see other experiments using the instrument before I designed a GMSL study. I am suspicious of the seven categories and the hollow picture of SL conveyed by the instrument. However, if four or five studies used the GMSL successfully, I might well be persuaded to employ this conservative scale.
The good news is that those five studies (any many others) do exist. Wu, Tse, Fu, Kwan, and Liu (2013) employed the GMSL in a study of hotel worker’s service behavior. Ehrhart’s 14-item instrument was presented to 19 human resource managers, 110 supervisors, and 304 supervisor-subordinate dyads. The GMSL was converted into Chinese using the back-translation technique. A professor translated the items into Chinese, and a second professor translated the scale back into English. A bilingual management professor examined both versions to ensure their equivalence. Also, hotel employees were interviewed to ensure that the scale was appropriate for their industry. Results indicated that SL positively influenced customer-oriented behaviors as a result of LMX outcomes. This level of rigor is best-practice in cross cultural research and stands in sharp contrast to the procedures used to develop the GMSL.
Mayer, Bardes, and Piccolo (2008) examined the intersection of leader behaviors and follower needs. The logic of the study focused upon the ability of servant leaders to meet followers’ needs and promote job satisfaction. To this end, Mayer et al. (2008) posit that needs such as self-determination and control can be nurtured by features of SL such as moral orientation, personal development, and sensitivity to the needs and desires of followers. The belief that needs are being met (or are being addressed), was thought to be associated with organizational justice perceptions, leading to satisfaction with the job. To test this logic, Mayer et al. (2008) presented the 14-item GMSL (and measures of organizational justice, need satisfaction, and job satisfaction) to 187 business undergraduate students. GMSL scores were correlated with job satisfaction scores, and structural equation modeling indicated that a path from SL to organizational justice to need satisfaction to job satisfaction was the best fitting model. There was also a partial mediation path moving from SL to need satisfaction to job satisfaction. This suggests that both a justice path and a needs satisfaction path may be operating.
This important study has many implications, but for our purposes the GMSL functioned as Ehrhart (2004) described. When you examine this study, consider the Discussion section carefully. SL is analyzed in a rather casual manner. On one level, Mayer et al. (2008) examines organizational justice and needs satisfaction explanations (p. 193). The discussion does not address the complexities of SL, and leader characteristics such as trust, humility, vision, wisdom, and service to others are not explored. This is not to underestimate the hypothesis that was confirmed. Unfortunately, the authors spend more time summarizing the implications for organizational justice researchers than for servant leadership scholars.
One point I would add is that the composite GMSL score was the root of the direct and indirect paths to job satisfaction. This suggests that it is not necessary to explore each of the dimensions of SL individually because the total score is the best indicator of the SL construct. Ehrhart’s (2004) compact, parsimonious instrument is an elegant tool if the GMSL successfully predicts outcomes. There is evidence that it does. Hunter, Neubert, Perry, Witt Penny and Weinberger (2013) found that follower-rated GMSL was negatively associated with leader-rated extraversion and positively associated with leader-rated agreeableness. Follower-rated GMSL was associated with lower follower turnover intentions and lower follower disengagement. Jaramillo, Grisaffe, Chonko, and Roberts (2009) used structural equation modelling to detect a path between SL (using the GMSL), ethical level, person-organization fit, and commitment to the organization (low turnover intent). Walumbwa, Hartnell and Oke (2010) found that the GMSL was associated with organizational citizenship behaviors, with the relationship mediated by self-efficacy beliefs and commitment to the supervisor. Schneider and George (2011) observed that voluntary service club members were more likely to stay in the organization and were more satisfied and committed to the organization as GMSL scores increased. Kool and van Dierendonck (2012) used the GMSL to detect that SL mediated commitment to organizational change through interactional justice and optimism perceptions. Rivkin, Diestel, and Schmidt (2014) found that SL scores (measured by a German translation of the GMSL) were associated with short-term indicators of strain (ego-depletion; need for recovery) and long-term indicators of job-strain (emotional exhaustion; depersonalization).
This summary of GMSL studies is not exhaustive. It does illustrate the utility of Ehrhart’s (2004) scale in addressing important SL outcomes. All of this evidence has aggregated using a scale that was less concerned with identifying the dimensions of SL than testing key outcomes. A careful reader is left to wonder why so much time (and so many journal pages) has been devoted to fleshing out the dimensions. There is no doubt that construct validity is one reason. We want our conceptualizations and instruments to correspond with the phenomenon they purport to understand. The GMSL studies suggest that participants respond to the core characteristics. It may not be the most appropriate scale for every context or experiment, but it may be a good option for many tests.
Barbuto, Jr., J. E., & Wheeler, D. W. (2006). Scale development and construct clarification of servant leadership. Group & Organization Management, 31, 300–326. doi: 10.1177/1059601106287091
Ehrhart, M. G. (2004). Leadership and procedural justice climate as antecedents of unit-level organizational citizenship behaviors. Personnel Psychology, 57, 61–94.
Fridell, M., Belcher, R. N., & Messner, P. E. (2009). Discriminate analysis gender public school principal servant leadership differences. Leadership & Organizational Development Journal, 30, 722-736. doi: 10.1108/01437730911003894
Graham, J. W. (1991). Servant leadership in organizations: Inspirational and moral. Leadership Quarterly, 2, 105-119.
Hunter, E. M., Neubert, M. J., Perry, S. J., Witt, L. A., Penny, L. M., & Weinberger, E. (2013). Servant leaders inspire servant followers: Antecedents and outcomes for employees and the organization. The Leadership Quarterly, 24, 316-331.
Jaramillo, F., Grisaffe, D. B., Chonko, L. B., and Roberts, J. A. (2009). Examining the impact of servant leadership on salesperson’s turnover intention. Journal of Personal Selling & Sales Management, 29, 351-365. doi: 10.2753/PSS0885-3134290404
Kool, M., & van Dierendonck, D. (2012). Servant leadership and commitment to change, the mediating role of justice and optimism. Journal of Organizational Change Management, 25, 422-433. doi: 10.1108/09534811211228139
Mayer, D. M., Bardes, M., & Piccolo, R. F. (2008). Do servant leaders help satisfy follower needs? An organizational justice perspective. European Journal of Work and Organizational Psychology, 17, 180-197. doi: 10.1080/13594320701743558
Reed, L. L., Vidaver-Cohen, D., & Colwell, S. R. (2011). A new scale to measure executive servant leadership: Development, analysis, and implications for research. Journal of Business Ethics, 101, 415–434 doi: 10.1007/s10551-010-0729-1
Schneider, S. K., & George, W. M. (2011). Servant leadership versus transformational leadership in voluntary service organizations. Leadership & Organizational Development Journal, 32, 60-77. doi: 10.1108/ 01437731111099283
Walumbwa, F. O., Hartnell, C. A., & Oke, A. (2010). Servant leadership, procedural justice climate, service climate, employee attitudes, and organizational citizenship behavior: A cross-level investigation. Journal of Applied Psychology, 95, 517-529. doi: 10.1037/a0018867
Wu, L.- Z., Tse, E. C.- Y., Fu, P., Kwan, H. K., & Liu, J. (2013). The impact of servant leadership on hotel employees’ ‘servant behavior.’ Cornell Hospitality Quarterly, 54, 383-395. doi: 10.1177/1938965513482519