Measuring charisma: Dimensionality and validity of the Conger-Kanungo scale of charismatic leadership
Jay A Conger
As part of an ongoing research program on charismatic leadership, Conger and Kanungo (1994) developed and tested a questionnaire to measure the perceived behavioural dimensions of charismatic leadership as proposed in their model (1987, 1988). This article reports on further analyses of the dimensionality and validity of the Conger-Kanungo scale. It begins with a reanalysis of the data from the 1994 study. It then reports on three separate studies conducted in the United States using managers (N=103), in Canada using political party delegates (N=71), and in India using clerical staff (N=98), to examine convergent and discriminant validity. The results of the dimensionality analyses suggest a five-factor model consistent with the Weberian notion of charismatic leadership. The results of the U.S., Canadian, and Indian studies provide further evidence that the CongerKanungo scale has acceptable reliability and validity as a diagnostic tool in diverse contexts.
In its simplest form, the leadership process in organizations can be conceptualized around several distinct stages of activity (Conger, 1992; Conger & Kanungo, 1988; Kotter, 1990; Yukl, 1994). The first stage is an ongoing assessment of opportunities and constraints in both the internal and external environments to determine the organization’s direction. The second stage is the formulation of the direction into formal strategies and goals and their articulation for the organization’s membership. The final stage involves gaining commitment and mobilizing the organization’s membership to implement the strategies and goals and undertake initiatives to achieve them. To summarize, the three stages, then, are: a) environmental assessment, b) direction formulation and communication, and c) membership alignment and implementation. While the stages are presented as a linear process, in reality they are stages of activity that are ongoing and often overlap. In addition, environmental changes require that the stages be continually repeated.
Conger and Kanungo (1987,1988,1992) developed a model of charismatic leadership within organizations that proposes distinguishing behavioural components along these three distinct stages of the leadership process. In the assessment stage, for example, the behaviours of charismatic leaders are perceived to focus on followers’ needs and on environmental opportunities that challenge the status quo of the organization. In contrast, other leadership forms will be more intent on building on or undertaking incremental improvements to the status quo. In stage two, charismatic leaders are more likely to be seen as conveying futuristic visions for their organizations in an inspirational manner. Conger and Kanungo proposed that, in the implementation stage, charismatic leaders rely on unconventional means and exhibit behaviours of self-sacrifice and personal risk-taking to align commitment from followers and to empower them to act. On the other hand, noncharismatic leaders rely on more transactional approaches such as the exchange of extrinsic rewards (Burns, 1978) to gain support from followers.
In essence, charismatic leaders differ from other leaders by their ability to formulate and articulate an inspirational vision and by behaviours and actions that foster an impression that they and their mission are extraordinary. As such, individuals choose to follow such leaders in management settings, not only because of their formal authority, but also out of perceptions of their extraordinariness. Thus any measurement of charismatic leadership must be based on the followers’ perceptions of specific behavioural attributes of the leader that engender such outcomes. As such, the Conger-Kanungo model posits that charismatic leadership is an attribution based on followers’ perceptions of their leader’s behaviour. The leader’s observed behaviour is interpreted by followers as the expression of charisma in the same sense as a leader’s behaviours reflect that individual’s participative, people, and task orientations.
This paper will present a brief review of recent theoretical and empirical developments on charismatic leadership, an overview of the Conger-Kanungo model and its development, and new analyses of the dimensionality and validity of a questionnaire instrument designed to measure behavioural attributes of charismatic leadership as proposed in the Conger-Kanungo model.
The term charisma is derived from an ancient Greek word meaning “gift.” Later adopted by the early Christian church, charisma described gifts or charismata from God that enabled the receiver to carry out extraordinary feats such as healing or prophecy. Its application to secular and leadership contexts came much later in the pioneering work of the German sociologist Max Weber (1925/1968). He developed a typology of three ideal types to describe the forces of authority in a society: the traditional, the rational-legal, and the charismatic. Weber was intrigued by the forces that stabilized and ordered society, on the one hand, and those that brought change and disorder, on the other. In contrast to the other two forms, charismatic authority or leadership did not gain its legitimacy from laws and rules or titles and traditions, but rather from a faith in the charismatic’s exemplary character. Charisma, in essence, was Weber’s umbrella term for the forces of change and innovation in society.
Due in part to the change-agent aspect of Weber’s conceptualization, organizational researchers in the 1980s and 1990s have turned their attention to charismatic business leaders as a principal focus in understanding organizational transformation and innovation in what was seen as a largely unadaptive and bureaucratic corporate climate in North America (e.g., Bass, 1985; Conger, 1989; Howell & Higgins, 1990; Tichy & Sherman, 1993; Trice & Beyer, 1991).
While Weber’s writings largely concerned political and religious leaders, he did retain elements of the original sacred character of the term, noting that authority was vested in an individual’s personal gifts and abilities. In addition, charismatic leaders were revolutionary forces who offered a prophetic picture or vision of the future. They were perceived as possessing a power to “heal” the wrongs of the previous order, whether it be traditional or rational-legal. Moreover, charismatic authority operated informally through human relationships, while the other forms of authority were organized around formal and permanent structures and systems, so charisma relied, at the very least, on a perceived sensitivity and ability to minister to the needs of followers.
Weber’s work sparked several decades of extensive investigations by sociologists and political scientists, but it was not until the late 1970s and mid-1980s that researchers in the disciplines of organizational behaviour and organizational psychology turned their attention to charismatic leaders. Their models were less influenced by literal interpretations of Weber’s conceptualization of charismatic leadership than were those of previous scholars, and instead drew upon the disciplines of psychology, political science, and sociology to enhance conceptualizations (e.g., Burns, 1978). In large part, the interest in charismatic leaders reflected certain trends, particularly in the business world. The rise of global competition had heightened pressure on North American corporations to reinvent themselves, by means of significant organizational change, in order to regain their competitiveness. It was felt that leadership played a pivotal role in change processes, and that charismatic leaders were among the more exemplary types of change-agent leaders. Their sense of vision, their innovativeness, and their motivational capabilities seemed ideal for facilitating large-scale and highly adaptive organizational change. For these reasons, charismatic leadership has been and will continue to be a topic of great interest for management and organizational researchers.
While several theoretical models have been proposed (Bass, 1985; Conger, 1989; Conger & Kanungo, 1987; House, 1977; House & Shamir, 1993; Roberts, 1985; Zaleznik & Kets de Vries, 1975), few have been studied empirically in a concerted way with extensive measurement development and validity testing. The Bass (1985) and House (1977) frameworks have received the greatest attention (e.g., Bass, 1985; House, Spangler, & Woycke, 1991; Howell, 1985; Howell & Higgins, 1990; Smith, 1982), yet this body of research has certain shortcomings. For example, in the Bass model, charisma is only a subscale under another measure called “transformational leadership.” The measure itself also includes the following subscales: inspirational leadership, individualized consideration, and intellectual stimulation. The item relating to vision, however, is treated as indicative of inspirational rather than charismatic leadership (Bryman, 1992). This is in sharp contrast to the majority of conceptualizations, which treat vision as a component of charismatic leadership. In addition, the Bass measure weights as minor Weber’s notion that the charismatic leader is perceived to be exceptional. Only two items could be said to tap this dimension. Finally, Bass implies that charisma is a product or outcome of transformational leadership and at the same time a component of it. As Bryman (1992) has noted, it is difficult to see how it can be both.
While also receiving some measure of attention, House’s model has undergone less intensive testing than Bass’s. House’s test (e.g., House et al., 1991) of his 1977 theory relied upon secondary sources-the biographies and speeches of U.S. presidents-for its data, rather than survey or interview data from actual respondents such as leaders, followers, and outside observers.
Overall, it is fair to say that little attention has been directed towards empirically identifying the specific distinguishing behavioural attributes of charismatic leaders in organizational settings. Given the complexity of the phenomenon, measuring models of charismatic leadership is also fraught with problems of construct ambiguities and validity issues. This necessitates research on two fronts: further identifying the behavioural dimensions of charismatic leaders and developing valid and reliable measures of such leadership behaviour for use in future research.
The Conger-Kanungo Model of Charismatic Leadership The Conger and Kanungo (1987, 1988, 1992) model of charismatic leadership focuses on several behavioural dimensions of charismatic leadership perceived by followers in the three stages of the leadership process. The model is predicated on the assumption that the perceived demonstration of these behaviours by the leader will result in an attribution of charismatic leadership by followers.
Specifically, in stage one (the environmental assessment stage), in the Conger-Kanungo model the charismatic leadership role of managers is distinguished from other leadership roles by means of followers’ perception of their manager’s heightened sensitivity to environmental opportunities and constraints, and followers’ needs. Managers in a charismatic leadership role are also more likely to be perceived as both critics of the status quo and agents of radical reform. These qualities increase the likelihood that the organization’s direction will, in future, address unexploited and futuristic opportunities that will be perceived as more visionary and meaningful to followers in the vision formulation stage.
In stage two (the vision formulation stage), environmental assessments result in the formulation of a strategic direction for the organization’s future. The CongerKanungo model posits that followers’ perceptions of their manager’s formulation of a shared and idealized future vision and the effective, inspirational articulation of that vision distinguishes the manager’s charismatic leadership role from other roles. In contrast, noncharismatic managers will formulate more mundane, less futuristic goals that are communicated in a less inspirational manner.
In stage three (the implementation stage), the leader is charged with building trust in and commitment to the chosen strategic direction as well as with developing tactics to achieve the desired goals. Managers who are perceived as charismatic will be seen to be engaging in exemplary acts that subordinates interpret as involving great personal risk and sacrifice. Through these actions, such managers are able to build trust in their leadership. After all, if leaders are willing to make great sacrifices and take personal risks, then they must believe in the potential outcome of the vision and be highly committed. The risks become symbols of the leader’s personal investment. In addition, managers in a charismatic leadership role are seen to be deploying innovative and unconventional ways of achieving their vision. This heightens perceptions of the leaders’ uniqueness and effective creativity and reinforces notions of their exceptional character. In addition, their initiatives are disarming for competitors who are accustomed to more conventional responses, thus increasing the likelihood that the initiative will be successful.
Based on their model (1987, 1988), Conger and Kanungo (1994) developed and tested a 25-item questionnaire measure (henceforth called the “C-K scale”) of six perceived behavioural dimensions of charismatic leadership (environmental sensitivity; sensitivity to members’ needs; does not maintain the status quo; vision and articulation; personal risk; and unconventional behaviour). Using a sample of 485 respondents, principle component analysis with varimax rotation was performed on the 25 items of the C-K scale to determine if the behavioural dimensions could be empirically verified. An analysis of the total sample yielded the six clear interpretable factors of charismatic leadership. This was confirmed with a confirmatory factor analysis (CFA) using the LISREL VII program (Joreskog & Sorbom, 1989). Adequate test-retest reliability was also demonstrated with a separate sample of 75 respondents. Finally, convergent and discriminant validity was assessed by a correlational analysis of the C-K scale of charismatic leadership and its subscales with four other leadership role behaviour measures: task-oriented, participative, people-oriented, and Bass charisma. These results were confirmed by LISREL estimates of the relationships between the C-K subscales and the four leadership measures (Conger & Kanungo, 1994, pp. 450-451).
Although the analysis revealed sound psychometric properties of the measure with adequate reliability, convergent and discriminant validity coefficients, and a stable six-factor structure, several concerns remain to be addressed. The first set of concerns deal with the issues of dimensionality parsimony and scale brevity. In the principal component analysis described earlier, the six factors were selected on the basis of the eigenvalue > 1 rule. However, an examination of the wording of individual items and associated factor loadings (Conger & Kanungo,1994, Table 2, pp. 448-449) suggest some item redundancy and fewer factors.
The other major issue is the validation process. The 1994 study sought to establish convergent and discriminant validity using items from Bass’s (1985) charisma scale and from the Ohio State leadership behaviour scales. More recently however, Yukl (1988) has developed more standardized scales to measure different management practices associated with leadership roles. It is therefore desirable to get further evidence of the convergent and discriminant validity of the C-K scale by relating it to these additional measures, which are more standardized and better established.
The present paper addresses each of the above concerns in order. First, the data from the original study conducted in the United States and Canada (Conger & Kanungo, 1994) was used to further analyze the dimensionality of the C-K scale of charismatic leadership. Then, the resulting model was tested for convergent and discriminant validity in three independent studies conducted in the United States, Canada, and India.
Refining the C-K Scale: Reanalysis of the 1994 Study In the 1994 study, Conger & Kanungo administered a questionnaire to 750 managers in four different corporations in the United States and Canada to describe their superiors’ behaviour. One part of the questionnaire contained the 25-item C-K scale. Another part contained the Bass charisma scale (1985) and three other measures of task, people, and participatory leadership orientation. The study also reported that the C-K scale measures six charismatic leadership dimensions: 1-vision and articulation; 2-environmental sensitivity; 3-unconventional behaviour; 4-personal risk; 5-sensitivity to members’ needs; and 6-does not maintain status quo.
However, as suggested earlier, the classification of scale items into these six dimensions requires further investigation. Analysis of the factor loadings (Conger & Kanungo, 1994, Table 2, pp. 448-449) reveals several items with significant loadings on more than one factor. For example, the item “readily recognizes new environmental opportunities” is classified under factor two (environmental sensitivity) based on a factor loading of .45; however, this item has a loading of .53 on factor one (vision and articulation). In addition, both items constituting factor 6 (does not maintain status quo) are reversecoded, raising the possibility that the emergence of this factor is an artifact.
The 25 items constituting the C-K scale in the 1994 study were first scrutinized for possible redundancies. Based on the wording of items and the factor loadings in the original principal component analysis, it was decided to drop the two reverse-coded items forming the factor “does not maintain status quo.” Given that these were the only items in the scale that were reverse-coded, there is a strong possibility that the emergence of these two items as a separate factor was an artifact. Moreover, the import of the two items is captured to some extent by other items. For example, the significance of the reversecoded item “tries to maintain the status quo or the normal way of doing things” is reflected to some extent in the item “uses nontraditional means to achieve organizational goals .” Three more items were dropped using this logic. For example, the item “appears to be a skillful performer when presenting to a group” was dropped, since this notion is largely captured by the item “exciting public speaker.” This exercise resulted in a shortened scale of 20 items.
Based on their wording, these 20 items were then reclassified into five subscales (see the appendix) based largely on the behavioural subdimensions suggested by the Conger and Kanungo model described earlier. Seven items were hypothesized to constitute the first factor, “strategic vision and articulation” (SVA). The first five of these items were classified under “vision and articulation” while the latter two items were classified under “environmental sensitivity.” This grouping reflects the fact that environmental assessment and the formulation of a strategic vision are intimately related. The leader’s vision arises from an assessment of environmental opportunities that either are yet to be exploited or have only been partially explored. The vision is in essence a strategic targeting of these opportunities. Four items were hypothesized to constitute the second factor, “sensitivity to the environment” (SE). These items were originally classified under the factor “environmental sensitivity.” Three items each were hypothesized to constitute the factors sensitivity to member needs (SMN), personal risk (PR), and unconventional behaviour (UB). These classifications are in line with the original classification of these items under “sensitivity to member needs,” “personal risk,” and “unconventional behaviour,” respectively.
The five-factor model proposed above was then tested through a CFA using LISREL VII. Convergent and discriminant validities of the subscales were then assessed by testing a structural model using LISREL VII, with the subscales as exogenous latent constructs and the four other leadership scales (task oriented, participative, people oriented, and Bass charisma) as endogenous latent variables.
Analysis and Results
CFA of the 20-item, five-factor model yielded a X^sup 2^^sub 160^ = 458.06 (p=.000). Comparison to an absolute null model with X^sup 2^^sub 190^ = 4118.54 (p=.000) yielded a normed fit index (NFI) of .89 and a nonnormed fit index (NNFI) of .91 (see Medsker, Williams, & Holahan, 1994 for the method used to calculate these indices). This was a very good fit, considering the large sample size of over 400. The Cronbach’s alpha reliabilities of the individual subscales were: .87 (SVA), .77 (SE), .84 (SMN), .85 (PR), and .74 (UB).
Convergent and discriminant validity. Since the five subscales of the C-K scale are hypothesized to be tapping different behavioural dimensions, they can be expected to have differing relationships with the other leadership scales included in the original study. For example, given the nature of the two constructs, the subscale “strategic vision and articulation” (SVA) should be positively related to the Bass charisma scale (BASS). By the same token, given the context and scope of leader behaviours associated with formulating and articulating a strategic vision, SVA should not be strongly related to leadership scales such as “task oriented,” “people oriented,” and “participative,” which are more concerned with the day-to-day, routine organizational matters. Similarly, assessing the internal environment of the organization is part of successful task leadership. Therefore, the subscale “sensitivity to the environment” should be strongly and positively related to the task-oriented leadership scale (TASK). On the other hand, the subscale “sensitivity to members’ needs” (SMN) should be strongly and positively related to the people-oriented leadership scale (PPL) and the participative leadership scale (PART). Given the often-suggested dichotomy between task and relationship orientation, one can then expect a negative relationship between SMN and TASK leadership. There is no a priori reason to expect relationships between the subscales “personal risk” (PR) or “unconventional behaviour” (UB) and each of the other leadership scales.
The above hypotheses were tested by a structural latent variable model with the five subscales treated as independent exogenous variables and the other leadership scales (BASS, TASK, PPL, and PART) treated as dependent, endogenous variables. The alpha reliabilities for the latter scales are: .87 (BASS), .72 (TASK), .81 (PPL), and .81 (PART). Table 1 shows the LISREL (gamma) estimates.
As hypothesized, the subscale “strategic vision and articulation” (SVA) is positively and significantly related to the Bass charisma scale (BASS). Further, it is not significantly related to any of the other leadership scales. As expected, the subscale “sensitivity to the environment” (SE) is strongly related to the task leadership scale (TASK). The hypotheses involving the subscale “sensitivity to members’ needs” (SMN) are also substantiated. As hypothesized, SMN is strongly related to the leadership scales “people oriented” (PPL) and “participative” (PART). It is also negatively related to the TASK leadership scale. The subscales “personal risk” (PR) and “unconventional behaviour” (UB) were not expected to be strongly related to any of the leadership scales. Table 1 reveals that, in general, this expectation is supported. PR is significantly related to the BASS scale, while UB does not have any strong relationships with any of the leadership scales. These results provide evidence of convergent and discriminant validity for the C-K subscales.
The reanalysis of 1994 study resulted in a revised 20-item scale of charismatic leadership. Although the analysis supports the presence of five distinct subscales, the alpha reliability of the total scale is .88, justifying its use as an overall measure of charismatic leadership as proposed in the Conger and Kanungo model.
While the data in Table 1 provide evidence of the convergent and discriminant validity of the subscales, there is some evidence of the convergent and discriminant validity of the overall scale. The overall C-K scale has a correlation of .69 with the Bass charisma scale. Table 1 reveals that this degree of convergence is attributable to the positive relationship between four of the CK subscales and the BASS scale. On the other hand, the dimension of “unconventional behaviour” differentiates the C-K scale from the Bass scale as is revealed by the slight but significant negative relationship between the subscale UB and the BASS scale.
Although the CFA supported a five-factor solution with adequate fit, bear in mind that the CFA was performed on the same data as was the original principal component analysis. Therefore, it is necessary to replicate this result using other independent samples.
Study 1: Relating the C-K Scale to the Managerial Practices Survey
In order to reconfirm the five-factor structure and gather additional evidence of the convergent and discriminant validity of the C-K scale, it was decided to conduct a second independent study to correlate the C-K scale and the four subdimensions with another more recent and widely used leadership measure, namely, Yukl’s (1988) managerial practices survey (MPS).
The MPS assesses managerial behaviour on 14 different dimensions representing various managerial functions. For the purposes of this study, the dimensions chosen are supporting (five items), recognizing (six items), rewarding (three items), delegating (three items), consulting (five items), problem-solving (first four items), planning (first four items), monitoring (six items), inspiring (six items), and mentoring (four items). The original 4-point response format was retained.
The study involved administering a questionnaire containing the C-K scale and the above leadership measure to a sample of organizational employees. To minimize possible method bias, the questionnaire was divided into two parts. Part 1 contained the 25 item C-K scale and demographic variables. Part 2 included selected items from the MPS. The questionnaire was administered to employees attending the company training program of a large multinational corporation based in the United States, with the cooperation of the human resources department. Participants were asked to describe their superior in terms of the questionnaire items. Participation was voluntary and respondents were instructed to allow a minimum of 24 hours to elapse between completing the two parts.
The mean age of the total sample (N = 103) was 39.6 years. Two thirds were male (66% male) Ninetyseven percent had at least a college degree and 87% were married. The respondents were predominantly from the middle (69%) and senior (30%) levels and were drawn from all functional departments in the organization. Their salary levels ranged from $45,000 to over $85,000. The average number of years worked was 10.8, and the mean job tenure was 2.8 years.
The C-K scale was first subjected to a CFA to reconfirm the five-factor structure. The subscales were then related to the MPS dimensions through a structural latent variable model to test specific convergent and discriminant validity hypotheses.
Analysis and Results
The CFA on the 20-item C-K scale was conducted using the LISREL VII program. Given the somewhat small sample size (N=103) and the relatively large number of items (20), it is possible that the number of estimated parameters in the model is too large for the available sample size. The effects of a poor sample-size-toestimator ratio can be minimized by using mini-tests of composite items. It may be recalled that the subscale “strategic vision and articulation” has seven items. In line with Anderson and Gerbing’s (1988) suggestion, three indicators were constructed to represent this subscale by combining items through random assignment (see Shore, Barksdale, & Shore, 1995 for a recent example of this technique). For the subscale “sensitivity to the environment” with four items, one composite indicator was constructed by combining two randomly selected items. The remaining two items were retained as indicators, unaltered. The three other subscales have three items each and these were retained as the respective indicators. Thus, all five subscales were represented by three indicators each.
The CFA resulted in a model with X^sup 2^^sub 105^= 103.27 (p = .041). Comparison with an absolute null model with X^sup 2^^sub 105^ = 698.77 yielded an NFI of .85 and an NNFI of .95. These results lend support to the five-factor structure that emerged from the reanalysis discussed earlier. The alpha reliabilities of the five subscales are .86 (strategic vision and articulation), .72 (sensitivity to the environment), .75 (sensitivity to member needs), .85 (personal risk), and .81 (unconventional behaviour). The alpha reliability of the overall 20-item C-K scale was .87.
Convergent and discriminant validity. Evidence of convergent and discriminant validity can be assessed by relating the C-K subscales to the MPS subscales, since the C-K subscales should be differentially related to the MPS subscales. For example, the subscale “strategic vision and articulation” (SVA) should correlate highly with the MPS subscale “inspiring” (INS) and have a low correlation with the MPS subscale “monitoring” (MON) This hypotheses is based on the expectation that the intent of leader behaviours classified as “strategic vision and articulation” is to inspire employees, while monitoring is a more mundane, managerial function. On the other hand, the C-K subscale “sensitivity to the environment” (SE) should correlate highly with MPS subscales associated with activities involved in assessing the internal environment of the organization such as “monitoring” (MON), “problem-solving” (PRB), and “recognizing” (RCG). With regards to the other C-K subscales, one can expect the subscale “sensitivity to member needs” to correlate highly with the MPS subscale “supporting” (SUP). There is no a priori reason to expect significant relationships between the subscales “personal risk” (PR) or “unconventional behaviour” (UB) and any of the MPS subscales. As suggested in earlier discussions, these two C-K subscales constitute the distinguishing feature of the C-K scale.
The above hypotheses were tested by a structural latent variable model using LISREL VII, with the five subscales treated as independent exogenous variables and the MPS subscales (INS, MON, PRB, RCG, and SUP) treated as dependent, endogenous variables. The alpha reliabilities for these latter scales are: .87 (INS), .75 (MON), .81 (PRB), .87 (RCG), and .87 (SUP). Table 2 shows the LISREL (gamma) estimates.
As hypothesized, the C-K subscale, “strategic vision and articulation” (SVA) is significantly and strongly related to the MPS subscale “inspiring” (INS). Further, it is not significantly related to MPS subscales such as “monitoring” (MON) or “problem-solving” (PRB). On the other hand, the C-K subscale “sensitivity to the environment” (SE) is strongly related to the MPS subscales “monitoring” (MON), “problem-solving” (PRB), and “recognizing” (RCG), as hypothesized. In line with expectations, the C-K subscale “sensitivity to member needs” (SMN) is strongly and significantly related to the MPS subscale “supporting” (SUP). Finally, as hypothesized, the C-K subscales “personal risk” (PR) and unconventional behaviour” (UB) are not significantly related to any of the MPS subscales, with the exception of “supporting” (SUP) which shows a mild positive relationship with “unconventional behaviour” (UB).
The results of the analysis provide support for the five-factor structure of the C-K scale using an independent sample. The subscales as well as the overall C-K scale have good to excellent alpha reliabilities that range from .72 to .87. Convergent and discriminant validity was demonstrated by the distinct pattern of relationships between the C-K subscales and the subscales of the MPS.
Study 2: Discriminatory Power of the C-K Scale: The Canadian Study
The purpose of this study was to assess if the C-K scale of charismatic leadership can discriminate between charismatic and noncharismatic leaders. The general hypothesis being tested was that if the C-K scale does measure charismatic leadership, it should be able to differentiate between leaders who have been identified as charismatic or noncharismatic as determined by an independent source or assessment.
The study was conducted in situ at the leadership convention of a major political party in Canada. The convention was held to elect one of five leadership candidates. The outcome of the leadership contest was highly significant to party members since the new leader would lead the party campaign in the forthcoming general election. Given the then low popularity ratings of the party, members were expected to elect a charismatic leader who could reverse the fortunes of the party and lead them to victory in the elections. The charisma of the leader was a salient attribute in the minds of party members, so it was an ideal setting for a test using perceived charisma as a criterion variable.
The subjects who participated in the study, which was carried out in two stages, were party members attending the leadership convention. Ten members (four women and six men) were included in the initial panel. Their average age was 51 years, and they had been in the party for an average of 19 years. All of them had followed, in person or on TV, at least one of the four nationally televised leadership debates that were held before the convention. In stage one, these 10 panel members were asked to rate the charisma of each of the five leadership candidates on a 6-point scale ranging from least charismatic to most charismatic (1 representing least to 6 representing most). This data was used to identify the most and least charismatic of the five leadership candidates.
In stage two, respondents were chosen at random from among party members attending the convention. This stage included 71 party members (22 women and 49 men) with an average age of 42 years. They had been in the political party for an average of 17 years, and 92% of them had followed at least one of the leadership debates. They were given a questionnaire that asked them to compare any one of two leaders identified as charismatic in stage one with any one of two leaders identified as noncharismatic. Respondents were not aware of the classification of charismatic versus noncharismatic established in stage one. The questionnaire contained the C-K scale as well as a global, single-item measure of charismatic leadership: “overall, he/she is a charismatic leader.” The hypothesis being tested was that the C-K scale could find significant differences in the charismatic behaviours attributed to the two leaders, and the candidate identified by the panel as most charismatic would receive higher scores on the C-K scale than the candidate identified as least charismatic by the panel.
Analysis and Results
In stage one, the panel data yielded mean scores for each of the five candidates. In order, these were 3.0, 5.5, 5.1, 3.9, and 2.5. A one-way analysis of variance resulted in a significant F ratio (F^sub 4,45^ = 406.45, p
On the basis of the above results, in stage two, respondents were asked to choose either candidate 2 or 3 and compare him/her with either one of candidates 1 or 5. As a manipulation check, the single-item global measure was first examined to see if the classification in stage one is replicated. Comparison of scores on the single-item global measure revealed that the mean score of charismatic leaders (candidates 2 or 3, M = 5.09) was indeed significantly higher (t = 8.42, p
The data was then subjected to a multivariate analysis of variance (MANOVA) with the five subscales of the C-K scale as dependent variables, the classifying variable being charismatic versus noncharismatic. The resulting significant F ratio (F^sub 5,136^ = 8.42, p
As can be seen in Table 3, the F ratio is significant for all the subscales except “unconventional behaviour,” indicating that each of these subscales can differentiate between charismatic and noncharismatic leaders. With regard to the dimension of “unconventional behaviour,” the absence of significant differences between charismatic leaders (M = 3.78) and noncharismatic leaders (M = 3.75) could be due to the fact that this study was conducted in the context of a political leadership race, where all candidates campaigned on the motto of changing the status quo and exhibited unconventional behaviour.
These results indicate that the C-K scale can successfully distinguish between charismatic and noncharismatic leaders in accordance with designations of charisma by an independent measure. This result can also be interpreted as evidence of convergent validity, since there is convergence between charisma as measured by two independent measures: the C-K scale in stage two and the assessment of the panel of “judges” in stage one.
Study 3: Further Evidence of Convergent and Discriminant Validity: The Indian Study
The purpose of this study was to gather further evidence of the convergent and discriminant validity of the C-K scale using another research method. This study uses independent assessments by two subordinates of the same leader using the C-K scale and the Bass scale. In this sense, the present study is patterned on the classical multitrait multimethod approach to convergent and discriminant validity advocated by Campbell and Fiske (1959).
This study was conducted in India using independent assessments by subordinates of the same manager. The design of the study required that the charismatic leadership roles of a given manager be assessed by two randomly selected subordinates directly working under his or her supervision. A total of 49 pairs of male subordinates (each pair reporting to a different manager) participated in the study. These respondents belonged to a large national corporation. The questionnaire, which contained both the C-K scale and the Bass scale items, was administered to the respondents in their work place. The respondents were asked to assess their immediate supervisor by filling in the questionnaire.
Analysis and Results
Convergent and discriminant validity hypotheses were tested by constructing the multitrait multimethod (MTMM) matrix. The two independent assessments were treated as two separate methods, while the Bass scale and the C-K scale (and subscales) were treated as different traits measured by the two methods. The relevant part of the MTMM matrix is shown in Table 4.
For convergent validity, correlations of the different measures of the same trait should be statistically significant and sufficiently large (Campbell & Fiske, 1959). As can be seen from Table 4, the coefficients along the convergent validity diagonal (in bold) are very large and statistically significant. The correlation between independent measures of the C-K scale is .84, while the correlation between independent measures of the Bass scale is .80. The convergent validity coefficients for the C-K subscales are also large and significant.
Discriminant validity can be established through two separate comparisons as outlined by Campbell and Fiske (1959). First, the convergent validity coefficients should be greater than the correlations between one variable and any other variable with which it shares neither method nor trait. In the upper left hand corner of Table 4, one can see that the convergent validity coefficients of .84 and .80 are greater than both the correlations (circled) between C-K scale (subordinate 1) and Bass scale (subordinate 2), which is .51, and the correlation between C-K scale (subordinate 2) and Bass scale (subordinate 1), which is .59. The second comparison requires that the convergent validity coefficients be greater than correlations of different traits measured with the same method. From Table 4, one can see that the correlation (boxed) between the C-K and Bass scales is .60 as assessed by subordinate I and .72 as assessed by subordinate 2. Both these correlations are less than the convergent validity coefficients .84 and .80, thus satisfying the second criterion for discriminant validity.
This MTMM approach provides conclusive evidence of the convergent and discriminant validity of the 20-item C-K scale. Table 4 also shows the correlations between the Bass scale and the C-K subscales. As can be seen, the largest points of divergence between the C-K and Bass scales are the two subscales of “personal risk” (PR) and “unconventional behaviour” (UB), which have lower correlations with the Bass scale than the three other C-K subscales “strategic vision and articulation” (SVA), “sensitivity to the environment” (SE), and “sensitivity to member needs” (SMN). As we have suggested, the subscales of “personal risk” and “unconventional behaviour” constitute the unique features of the C-K scale of charismatic leadership.
The object of studies reported in this paper was to establish the dimensionality and validity of the C-K scale of charismatic leadership. The results of the reanalysis of the data from the 1994 study (Conger & Kanungo, 1994) and of the second study using Yukl’s (1988) MPS scales support a stable five-factor structure for the C-K scale. These studies also provide evidence of the convergent and discriminant validity of the C-K scale. The Canadian and Indian studies provided further evidence of convergent and discriminant validity in both political and organizational contexts.
It may be noted, first, that the five-factor solution representing the five dimensions “strategic vision and articulation,” “sensitivity to the environment,” “unconventional behaviour,” “personal risk,” and “sensitivity to members’ needs,” closely parallels Max Weber’s (1925/1968) conceptualization of the charismatic leader (see Conger & Kanungo, 1994, for a detailed account of the Weberian model). To Weber, the charismatic’s power was vested in the exceptional nature of the individual’s personal gifts and abilities. In the present formulation, followers perceive the exceptional nature of the leader through the leader’s behaviour that corresponds to the dimensions of unconventional behaviour and personal risk. Second, according to Weber, charismatic leaders were individuals with a prophetic vision or vision of the future. In the five-factor formulation, this would correspond to the dimension of “strategic vision and articulation.” Finally, Weber described the charismatic leader as an individual who would “minister” to the needs of the followers. This would require charismatic leaders to be aware of their environments as well as the needs of the followers. This aspect is captured by the dimensions of “sensitivity to the environment” and “sensitivity to members’ needs” in the present formulation.
The results of the above studies suggest that a fivefactor model provides a significantly better fit than other models to both the empirical and the theoretical notions of charismatic leadership advanced by Weber. They also provide evidence of the construct validity of the CongerKanungo scale. In future studies the scale could be used not only as a diagnostic tool but also to reveal the effects of the perceived behavioural components of charismatic leaders on followers’ attitudes and behaviours.
Anderson, J.C., & Gerbing, D.W. (1988). Structural equation modeling in practice: A review and recommended twostep approach. Psychological Bulletin, 103, 411-423.
Bass, B.M. (1985). Leadership and performance beyond expectations. New York: Free Press.
Bryman, A. (1992). Charisma and leadership in organizations. London, England: Sage.
Burns, J.M. (1978). Leadership. New York: Harper & Row.
Campbell, D.T., & Fiske, D.W. (1959). Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105.
Conger, J.A. (1989). The charismatic leader. San Francisco: Jossey-Bass.
Conger, J.A. (1992). Learning to lead. San Francisco: JosseyBass.
Conger, J.A., & Kanungo, R.N. (1987). Towards a behavioral theory of charismatic leadership in organizational setting. Academy of Management Review, 12, 637-647.
Conger, J.A., & Kanungo, R.N. (1988). Behavioral dimensions of charismatic leadership. In J.A. Conger & R.N. Kanungo (Eds.), Charismatic leadership (pp. 78-97). San Francisco: Jossey-Bass.
Conger, J.A., & Kanungo, R.N. (1992). Perceived behavioral attributes of charismatic leadership. Canadian Journal of Behavioural Science, 24, 86-102.
Conger, J.A., & Kanungo, R.N. (1994). Charismatic leadership in organizations: Perceived behavioral attributes and their measurement. Journal of Organizational Behavior, IS, 439-452.
House, R.J. (1977). A 1976 theory of charismatic leadership. In J.G. Hunt & L.L. Larson (Eds.), Leadership: The cutting edge (pp. 189-207). Carbondale, IL: Southern Illinois University Press.
House, R.J., & Shamir, B. (1993). Toward the integration of charismatic, transformational, inspirational, and visionary
theories of leadership. In M. Chemmers & R. Ayman (Eds.), Leadership theory and research perspectives and directions (pp. 81-107). New York: Academic Press.
House, R.J., Spangler, W.D., & Woycke, J. (1991). Personality and charisma in the U.S. presidency: A psychological theory of leader effectiveness. Administrative Sciences Quarterly, 36, 364-396.
Howell, J.M. (1985). A laboratory study of charismatic leadership. Paper presented at the annual meeting of the Academy of Management, San Diego, CA.
Howell, J.M., & Higgins, C.A. (1990). Champions of technology and motivation. Administrative Sciences Quarterly, 35, 317-341.
Joreskog, K.G., & Sorbom, D. (1989). LISREL VII: A guide to program and applications (2nd ed.). Chicago, IL: SPSS. Kotter, J. P. (1990). A force for change: How leadership differs from management. New York: Free Press.
Medsker, G.J., Williams, L.J., & Holahan, PJ. (1994). A review of current practices for evaluating causal models in organizational behavior and human resource management research. Journal of Management, 20 (2), 439-464.
Roberts, N.C. (1985). Transforming leadership: A process of
collective action. Human Relations, 38, 1023-1046.
Shore, L.M., Barksdale, K., & Shore, T.H. (1995). Managerial perceptions of employee commitment to the organization. Academy of Management Journal, 38, 1593-1615.
Smith, B.J. (1982). An initial test of a theory of charismatic leadership based on the responses of subordinates. Unpublished doctoral dissertation, University of Toronto, ON.
Tichy, N.M., & Sherman, S. (1993). Control your destiny or someone else will. New York: Currency-Doubleday.
Trice, H.M., & Beyer, J.M.. (1991). Cultural leadership in organization. Organizational Sciences, 1, 149-169.
Weber, M. (1968). Economy and society (3 vols.). G. Roth & C. Wittich (Eds.). New York: Bedminster. (Original work published 1925).
Yukl, G.A. (1988). Development and validation of the managerial practices questionnaire. (Technical Report). State University of New York, Albany, NY.
Yukl, G.A. (1994). Leadership in organizations. Englewood Cliffs, NJ: Prentice-Hall.
Zaleznik, A., & Kets de Vries, M.F.RR. (1975). Power and the corporate mind. Boston: Houghton Mifflin.
Jay A. Conger University of Southern California
Rabindra N. Kanungo McGill University
Sanjay T. Menon Clarkson University
Purnima Mathur Indian Institute of Technology (New Delhi, India)
Address all correspondence to Rabindra N. Kanungo, Faculty of Management, McGill University, 1001 Sherbrooke Street West, Montreal, QC, Canada, H3A 1GS.
Copyright Administrative Sciences Association of Canada Sep 1997
Provided by ProQuest Information and Learning Company. All rights Reserved