Using survey data (n = 205) obtained from retail managers and executives of national retail chain store companies, we identified three leadership styles that were based on Quinn's theoretical model of competing leadership roles. Three leadership clusters, labeled loner/internal-focused, team builder/goal-oriented, and conceptual producer/external-focused, were identified through the use of a clustering technique. These three clusters were then compared on the basis of personal, organizational and managerial characteristics, using multivariate and univariate analyses of variance. The findings indicate that leadership styles are influenced by various factors such as personal values, job characteristics, job satisfaction, organizational commitment, career progression, and personal demographic characteristics. Theoretical and managerial implications are discussed.
Significant changes over the past 20 years have influenced and reshaped the retailing industry. Importantly, these changes have not only occurred in terms of logistics, technology, and supply chain management, but also in terms of human resource management. The once common practice of hiring non-college graduates for management positions is now being rapidly replaced with college recruitment and even recruitment of MBA students. One reason this shift has occurred is that US retailers increasingly need a well-educated and trained workforce to keep abreast of technological advances and sophisticated managerial concepts in the highly competitive global marketplace. At the same time, much attention has been paid recently to a potential future leadership crisis. Although it may make many traditional retail executives uncomfortable, it is becoming increasingly clear that those who will lead the retail industry must be competent not only in technical and managerial skills but also in dealing with human resources and leadership issues. In short, the performance expectations and leadership skills required for retail executives are rising. Consequently, it is a critical time to examine issues relevant to leadership among current managers and executives in the field. Such an examination, based on empirical research, could be even more compelling for both practitioners and academicians.
The objective of this study was to identify different leadership styles that may exist among today's retail managers by utilizing Quinn's (1984) model of competing leadership roles as a theoretical framework. The second objective was to examine these leadership styles based on their personal values, organizational, managerial, and demographic characteristics. To our knowledge, Quinn's model has not been applied to the retailing sector, although studies using this model have been widely published in the managerial, organizational, and behavioral science fields (Altschuld and Zheng, 1995; Belasen et al., 1996; Cooper and Quinn, 1993; Faerman et al., 1987; Giek and Lees, 1993; McCartt and Rohrbaugh, 1995; Pearce et al., 1997; Rogers and Hildebrandt, 1993; Singhapakdi et al., 1996; Yang and Shao, 1996). Therefore, it is our goal that this exploratory study of leadership styles in retail management provide important benchmark data to build a theoretical foundation for future studies in this area. This study is also important in helping to identify leadership issues in retailing that may warrant further study.
Quinn's theoretical model of competing leadership roles
Quinn's (1984) model of competing leadership roles is based on four quadrants that represent four different sets of leadership or organizational orientations created by two dimensions: structure (flexibility vs stability) and focus (internal vs external) (see Figure 1). On the vertical continuum of the structure, managerial leaders may have an orientation emphasizing either flexibility or stability. Similarly, managerial leaders may display either an internal or an external emphasis on the horizontal continuum. This results in four models of leadership style models:
(1) the open system model (an emphasis on flexibility and external focus);
(2) the rational model (an emphasis on stability and external focus);
(3) the internal process model (an emphasis on stability and internal focus); and
(4) the human relation model (an emphasis on flexibility and internal focus).
Within the framework of his theoretical model, Quinn (1984, 1988) identified eight leadership roles - innovator, broker, producer, director, coordinator, monitor, facilitator, and mentor roles - with two roles pertaining to each of the four quadrants. These eight roles are presented in a circular pattern in Figure 1, and characteristics of each role are described in Table I.
First, the open system model is comprised of two roles - the innovator role and the broker role. The primary goals of this open system model are growth and resource acquisition. As such, these two roles are considered to be adaptive and inventive. Second, the producer role and the director role, positioned in the rational goal model, characterize leaders who are task-oriented and directive. These roles are appropriate, because in the rational goal model an organization is viewed as a rational economic tool; this model emphasizes productivity, accomplishment, direction and goal clarity. The third model, the internal process model, emphasizes the primary goals of stability and control. Thus, the two roles in this system include the coordinator role and the monitor role. Leaders who emphasize these roles may be stability-oriented and conservative in their decision- making styles. Finally, the human relation model stresses participation, openness, commitment and morale, and consists of two roles: the facilitator role and the mentor role. Leaders who emphasize these two roles may be characterized as people-oriented/supportive leaders (Denison et al., 1995; Hooijberg, 1996; Quinn, 1984; Quinn and Rohrbaugh, 1983).
Two important conditions in understanding how these eight roles affect managers on a circular pattern are paradoxicality and multifunctionality. First, these eight roles are often referred to as paradoxical, in that the roles on the opposite continuum represent conflicting demands (e.g. the roles of a monitor and a coordinator conflict with the roles of an innovator and a broker). Nevertheless, effective managers are expected to perform these multiple leadership roles, although to a different degree, depending on the lifecycle stage of the manager and the organization (Denison et al., 1995; Hooijberg, 1996; Quinn, 1984, Quinn and Rohrbaugh, 1983). Importantly, the model has considerable face validity for retail organizations and their managers and leaders because of the many competing and often conflicting demands that managers in retailing face on a daily basis.
Denison et al. (1995) examined how 176 mid-level executives in the public utilities industry were rated by their subordinates (n = 670) and supervisors (n = 222). The findings indicated that highly effective mangers were perceived to exhibit a greater degree of behavioral complexity (i.e. the ability to perform highly cognitive tasks, the ability to do "more of everything", and the ability to respond to paradoxical and contradictory roles simultaneously) than less effective managers. The authors, however, concluded that a similar study of managers in a different sector (e.g. entrepreneurs or managers in highly dynamic industries) might reveal different findings than those displayed by the utility managers. Quinn and his colleagues reported that managers who balance competing leadership roles will tend to be more successful than those who display a specific leadership function (Hooijberg, 1996; Quinn, 1988; Quinn et al., 1991; Hart and Quinn, 1993). However, a few researchers have determined that certain leadership styles are more essential than others for effective leadership (Shamir et al., 1993; Sosik et al., 1997).
Hypotheses development
Leadership style clusters
By measuring the eight roles and employing a cluster analysis technique, Quinn et al. (1987) identified and profiled clusters of effective and ineffective managers as evaluated by their subordinates. The clusters that emerged among effective managers were labeled: conceptual producers (above the mean on six of the roles but falling to the mean on the coordinator and the monitor role), aggressive achievers (high scores on the bottom four roles but at or below the mean on the top four roles; particularly low on the facilitator role), peaceful team builders (above the mean on six of the roles and near the mean on the broker and producer roles), committed intensives (a highly varied profile - high on one of the roles in each quadrant), open adaptives (the highest scores on the top four roles), and masters (higher than the mean on all eight roles). While the masters cluster displayed well-rounded and high scores on all roles, the rest of the clusters in the effective groups had one or two roles that were below or at the mean scores. This suggests that effective managers may fall below the mean on certain roles as long as they can maintain enough balance among other roles.
The clusters that emerged among ineffective managers were labeled: chaotic adaptives (very low on the coordinator role), abrasive coordinators (particularly weak on roles dealing with people and with change), drowning workaholics (somewhat high on the producer role but low on practically all other roles), extreme unproductives (low scores on all roles), technical incompetents (somewhat high on external or internal roles but low on other roles), obsessive monitors (extremely high on the monitor role but low on all other roles), and disorganized externals (above the mean on the broker role but low on other roles). Those in the ineffective clusters tended to do poorly on nearly all roles or to exceed the mean on three or four roles.
Quinn's eight roles have proven to be effective in classifying managers into a few homogeneous groups based on their leadership styles. Therefore, H1 was established as follows:
H1. Different clusters of retail managers will emerge based on the extent to which they place varying degree of emphasis on eight leadership roles.
Personal values and leadership styles
A value is defined as "an enduring belief that a specific mode of conduct or end-state of existence is personally or socially preferable to an opposite or converse mode of conduct or end-state of existence (Rokeach, 1973, p. 5)". Personal values are associated with executive mindsets and figure prominently in shaping their leadership orientations (England, 1967; Fleishman and Peters, 1962; Hambrick and Brandon, 1988; Matsui et al., 1978; Geletkanycz, 1997). Furthermore, Quinn (1988) argued that leadership functions in the four quadrants are based on emphases on different types of values or motivations that leaders might hold. Thus, we expect that leadership styles would be associated with personal values as follows:
H2. Different leadership clusters will exhibit differing degrees of various personal values.
Organizational commitment, job characteristics and job satisfaction
Organizational commitment is defined as a psychological link between the employee and his/her organization and has been found to be related to various job-related characteristics (Allen and Meyer, 1990, 1996; Meyer et al., 1991). Earlier studies on employees' commitment to their organizations often attempted to identify factors that influence this commitment, i.e. antecedents to commitment. Job characteristics (e.g. Hunt et al., 1985) and job satisfaction (e.g. Dubinsky and Skinner, 1984; Good et al., 1988) have been found to be such antecedents in previous studies.
Job characteristics consist of four dimensions: variety, autonomy, task identity, and feedback (Becherer et al., 1982; Hackman and Lawler, 1971; Hertzberg, 1966). Variety refers to the degree to which employees are allowed to undertake various options in their work, while autonomy refers to employees' freedom to do what they want to do on the job. Task identity refers to the extent to which employees can identify with the project from inception to closure. Finally, feedback is the degree to which they receive feedback from their supervisors. These job characteristics have been found to influence employee's organizational commitment (Hunt et al., 1985) and job satisfaction (Becherer et al., 1982; Hackman and Lawler, 1971; Herzberg, 1966).
Lucas et al. (1990) claimed that job satisfaction consists of two components (intrinsic and extrinsic) that play different roles in job attitude and behavior. Intrinsic job satisfaction refers to one's satisfaction with the nature of the job itself, whereas extrinsic job satisfaction refers to one's satisfaction with such external factors as financial reward, benefits, and working conditions, etc.
Although studies seeking to define the relationship between leadership styles and career factors (i.e. job characteristics, satisfaction, and commitment) are scarce in the literature, we expect that there will be significant relationships as follows:
H3. Different leadership clusters will perceive job characteristics, job satisfaction and organizational commitment differently.
Career progression, organizational and demographic characteristics
Quinn et al. (1987) found that career (e.g. position or rank) and demographic (e.g. education, gender) characteristics were significant factors in differentiating leadership clusters of effective and ineffective managers. For instance, all effective managers classified as aggressive achievers were males in middle management. The majority of individuals in the conceptual producers cluster had graduate training. Most individuals in the masters cluster were in upper-middle or top management. Those labeled extreme unproductives tended to be in first-line or lower management positions. Therefore, there is sufficient evidence that career progression, and organizational and demographic characteristics are related to leadership styles, resulting in the following hypothesis:
H4. Different leadership clusters will exhibit differences in career progression, and organizational and demographic characteristics.
Methods
Data collection and sample characteristics
Questionnaire data were collected from 205 retail managers and executives from national retail chain store companies, including department stores, grocery retailers, discount stores, drug stores, and specialty retailers. A pre-survey postcard was mailed to inform the subjects, and two mailings of the survey were sent out. A total of 63 percent of the respondents were male, and 37 percent were female. Approximately 60 percent had a college degree or graduate degree. The majority of the respondents (88 percent) were white. A wide range of ages was represented, with a mean age of 42 years old (SD = 9.5). Almost all states (42 states and over 100 cities) were represented. The average years of management or supervisory experience in general (M = 18 years, SD = 9.1) and in retailing specifically (M = 17 years, SD = 9.2) indicate that respondents in this study have significant experience in retail management.
Department stores (40 percent) and discount mass merchandising retailers (31 percent) accounted for more than two-thirds of the representations. Store managers (46 percent) were most represented in this study, followed by human resource managers (26 percent), executives (17 percent), and entry-level managers (9 percent). A total of 85 percent of the respondents indicated that they had profit-and-loss responsibility. The annual base salary ranged widely, from $25,000 or less to over $130,000 (M = $80,537, SD = 86,140). Cash bonuses for 1999 also ranged widely, from $5,000 or less to $46,000 or more (M = $25,562, SD = 36,948). Approximately 75 percent of the respondents had stock options with values ranging from $10,000 or less (41 percent) to $51,000 or more (10 percent). Most respondents (94 percent) indicated that their overall compensation had increased over the last three years. Approximately two-thirds of the respondents indicated that they had received one or more promotions over the last three years (68 percent) or had received serious inquires from other firms, executive search firms or other sources soliciting them to seek alternative employment (66 percent). About 44 percent of the respondents indicated that they were on a fast track in their career progress as compared to their peers.
Instrument
Quinn's (1988, p. 128) self-assessment of leadership roles was used to assess retail managers' and executives' leadership functions. Table II presents questionnaire items by role, consisting of two statements for each role and Cronbach alpha coefficients.
Table III also summarizes the measurements for independent variables. The Rokeach Value Survey (RVS) items were pre-classified by faculty and graduate students panelists into four a priori categories:
(1) social affiliation;
(2) growth/development;
(3) achievement; and
(4) security.
Four dimensions of job characteristics (variety, autonomy, task identity, and feedback) were adopted from Hunt et al. (1985). The commitment scale (affective) was adopted from Allen and Meyer's (1990) organizational commitment scales. Affective commitment measures one's emotional attachment and is similar to other measures frequently used by other researchers (e.g. Good et al., 1996).
Six items of job satisfaction (three intrinsic and three extrinsic) were adapted from the Michigan Organizational Assessment Questionnaire (Cammann et al., 1983). Finally, variables related to career progression (i.e. retail management experience in years, base salary, cash bonuses, increases in compensation, promotions received, and number of recruitment solicitations over three years) were measured in open-ended question format. Career track was measured by asking the respondents to indicate on a scale (1 = definitely fast track; 5 = definitely slower track) the rate at which they thought their career had progressed over the last three years as compared to that of their peers at the same career stage. Finally, individual demographic (e.g. age, gender, education, ethnicity) and company organizational characteristics (retail format) were obtained using a categorical format.
Results
Cluster analysis on leadership functions
k-means clustering, a nonhierarchical clustering procedure, was used in this study (Hair et al., 1995). The parallel threshold method was used, in which several cluster seeds are simultaneously selected in the beginning, and objects are assigned within the threshold distance to the nearest seed. The Quick Cluster technique via SPSS, this study's analytical tool, overcomes the shortcoming of other nonhierarchical algorithms by selecting initial cluster centers with well-separated values, unlike algorithms that rely on random initial assignment (SPSS Users' Guide, 1988). This technique results in greater cluster solution stability, and cluster solution validity is further enhanced by the fact that nonhierarchical algorithms are not affected by outliers to the same extent as hierarchical algorithms (Hair et al., 1995).
Cluster solutions of two to four groups were initially generated. The solutions were evaluated on the basis of the extent to which the cluster scores differed significantly across the k-clusters (evaluated via F-ratios) and the extent to which the cluster scores differed significantly on a pairwise basis (evaluated via Scheffe's test). After careful examination, the three-cluster solution was chosen as optimal because it exhibited the highest percentage of pairwise differences and produced large F-scores that were significant at the p < 0.001 level. It also generated the largest multivariate F score. The three-segment solution minimized within-group error.
Table IV presents the three clusters on their mean scores for eight leadership functions. To validate the cluster solution, multivariate and univariate analyses were conducted among three groups. Owing to the disparity in sample sizes among the three groups, both the multivariate test for homogeneity of dispersion matrices (Box M test) and the univariate homogeneity of variance tests (Barlett-Box test) were run to assess the equivalence of covariance matrices across groups for leadership functions. Insignificant levels of Barlett-Box Fs and Box M Fs at the 0.05 level dispelled any concern that the variances in the three groups might not be equal.
A multivariate F-ratio of 36.0 (df = 16, 378, p < 0.001) and the effect size of 0.60 indicated that the three segments were significantly different overall. Therefore, H1 was accepted. Univariate analysis of each leadership function indicated that the three groups differed significantly in all of the leadership functions. Scheffe tests revealed that Cluster 2 differed most from Clusters 1 and 3.
It should be noted that the size of the significant F-ratios varied, indicating that the three clusters were differentiated more clearly on some factors than on others. For instance, the broker leadership role (F = 113.6, p < 0.001) had the largest F-ratio, followed by the monitor leadership role (F = 79.7, p < 0.001).
Loner/internal focused
Cluster 1 was labeled loner/internal focused and included 36 percent of the subjects (n = 71). This cluster tended to have lower mean scores on all functions, ranging from 3.43 to 5.59 on a seven-point scale. However, this tendency was most pronounced among the roles in the upper quadrants (i.e. the open system and the human relation model) in Figure 1. For instance, Cluster 1 had an extremely low mean score (M = 3.43) on the broker role, which was almost two standard deviations apart from the scores of other groups. Managers in this group reported not making decisions at higher levels or not maintaining a network of influential contacts. People in this group also had an extremely low mean score on the facilitator role (M = 4.70) and the innovator role (M = 4.97). For the roles in the lower quadrant, managers in this group fell slightly behind the total group mean. Therefore, it appears that managers in this group are internally focused and tend to make decisions without involving others.
Team builder/goal-oriented
Cluster 2 was labeled team builder/goal-oriented. It comprised 46 percent of the respondents (n = 91). Managers in this group tended to have high scores on all functions, with the mean scores ranging from 5.40 to 6.56. Of the eight functions, managers in this group had extremely high scores on the roles in the lower right quadrant (the rational model) - the producer role (M = 6.56) and the director role (M = 6.30). This indicates that managers in this group are highly committed to their job and goal/task oriented. Managers in this group also had a much higher score on the human relation model (the upper left quadrant), particularly in the area of mentoring (M = 6.01) than those in other clusters.
Conceptual producer/external-focused
Cluster 3, labeled conceptual producer/external-focused, comprised 18 percent of the respondents (n = 37). Managers in this cluster had high scores on six roles. However, this group was distinguished from other groups due to its extremely low scores on the lower left quadrant (the internal process model) - the role of monitor (M = 3.72) and the coordinator role (M = 4.72). Paradoxically, this group had the highest score on the broker role (M = 5.58) in the open system model. Therefore, managers in this group tended to be externally focused and unconcerned with internal processes.
Profiling leadership style clusters
Personal values
Table V presents results of multivariate analysis of variance (MANOVA) and univariate analysis of variance (ANOVA) among the three groups regarding personal values, job characteristics, job satisfaction, organizational commitment and various career experience variables. An examination of Barlett-Box Fs and Box M Fs indicated that variances among these factors were relatively equal among the three groups. Overall, the three groups differed in their personal values (Multivariate F = 10.2, p < 0.001) (see Table V). Further univariate analyses of variance revealed that the differences originated from all leadership functions. Scheffe tests indicate that Cluster 2 (team builder/goal-oriented) differed from the other two clusters. Cluster 2 had the highest mean scores on all four dimensions of values. No particular differences were found between the personal values held by those in Cluster 1 (loner/internal focused) and the personal values of those in Cluster 3 (conceptual producer/external-focused).
Job characteristics, job satisfaction, and organizational commitment
MANOVA indicates that the three clusters differed overall in their perception of job characteristics (F = 6.48, p < 0.001) (see Table V). A series of univariate analyses also indicated that the three groups differed in each of the four aspects of job characteristics. Scheffe tests indicated that Cluster 1 (loner/internal focused) tended to have lower scores on all four aspects of job characteristics than the other groups. Cluster 2 (team builder/goal-oriented) had a higher score on the task identity (M = 6.00) and feedback variables (M = 5.30) as compared to other clusters. The three groups also differed overall in their organizational commitment (Multivariate F = 5.85, p < 0.001). This difference was mainly due to the groups' affective commitment. Cluster 1 (loner/internal focused) displayed a lower level of commitment (M = 4.29) than the other clusters.
Career progression variables
The three clusters differed overall in their career progression variables (Multivariate F = 4.56, p < 0.001) (see Table V). Univariate analyses indicated that the three groups differed in various aspects of financial compensation, the number of recruitment solicitations, and their subjective evaluation of career progress. However, no differences were found in the number of years of retail management experience or in the number of promotions received within the last three years.
Organizational and demographic characteristics
Chi-square analyses were conducted to determine differences in categorical variables (Table VI). Group differences (χ2 = 24.2, p < 0.01) were found in the types of retail organization format in which the groups were currently employed. Interestingly, the majority of the managers working in discount mass merchandising were concentrated around Cluster 1 (loner/internal focused, 49 percent) and Cluster 2 (team builder/ goal-oriented, 44 percent). More than half of the department store managers (56 percent) were classified into Cluster 2 (team builder/goal-oriented), followed by 32 percent in Cluster 1 (loner/internal focused). A slightly greater percentage of managers in grocery/food or specialty retailing (43 percent, respectively) were classified into Cluster 3 (conceptual producer/external focused).
Job title (χ2 = 38.8, p < 0.001) had a significant effect on leadership style clusters. Upper level of managers (e.g. district/regional VPs or senior executives) tended to belong to Cluster 3 (56 percent), whereas human resource managers tended to belong to Cluster 2 (57 percent). Entry level/departmental managers were gathered around Cluster 2 (50 percent). The majority of store managers tended to be split between Cluster 1 (50 percent) and Cluster 2 (43 percent), and very few were classified as Cluster 3 (7 percent).
Ethnicity (χ2 = 5.99, p < 0.05) and gender (χ2 = 6.0, p < 0.05) had some effect on group differences. Caucasians/whites were spread between Cluster 1 (39 percent) and Cluster 2 (43 percent) while a greater percentage of non-whites (65 percent) occupied Cluster 2 (team builder/goal-oriented). A greater percentage of female managers (55 percent) were classified into Cluster 2 (team builder/goal-oriented) while male managers were more evenly distributed among the three clusters.
Of all the demographic characteristics, education (χ2 = 35.4, p < 0.001) had the most significant effect on group differences. A large percentage of those with a high school degree (71 percent) or some college or a college degree (63 percent) were classified into Cluster 2 (team builder/goal-oriented). A greater percentage of those with a college degree (46 percent) were classified into Cluster 1 (loner/internal focused), while a greater percentage of those with some graduate school or a graduate degree (48 percent) were classified into Cluster 3 (conceptual producer/external-focused). No significant differences were found in marital status and age.
Conclusion, discussion and implications
Exploring leadership style archetypes among retail managers and executives
Utilizing Quinn's (1984) model of competing leadership roles, our study attempted to identify archetypes of leadership styles that may exist among retail managers. Based on a self-assessment of their leadership functions, we identified three prototypes and labeled them:
Cluster 1 (loner/internal focused);
Cluster 2 (team builder/goal-oriented), and
Cluster 3 (conceptual producer/external focused).
The broker and the monitor roles, which are diametrically opposed in the model and therefore competing with each other, were the most decisive factors in differentiating between the groups, at least among the respondents in this study. This finding supports the concept of "competing" roles among the leadership function roles that are diametrically positioned in the model. Adopting Quinn's concept, Figure 2 Figure 3Figure 4 were constructed, representing the size and the shape of each profile.
Cluster 1 (loner/internal focused) had the smallest size, with much less emphasis on the upper right quadrant (the open system model). Managers in this group tended to engage in all functions to a lesser degree than others, but particularly in the innovator and broker roles. They were less likely to seek consensus from others and much less likely to change or maintain the unit's external legitimacy through external networking. Managers in this group tended to perceive their job to be more repetitive than those in other groups. This group can be compared, although cautiously, to the extreme unproductives identified by Quinnet al. (1987) in that people in the extreme unproductives cluster were described as scoring lowest on all functions.
The loner/internal focused managers tended to view their jobs as not very fulfilling, i.e. they experienced less skill variety, less autonomy, less task identity, and less feedback from superiors regarding their job performance. Not surprisingly, loner/Internal focused managers were less likely to be satisfied with their job and displayed much less organizational commitment than other groups. Managers in this group tended to have the least aggressive career progression characteristics in terms of financial compensation and the number of recruitment solicitations. They also perceived their careers as being on a slower track than those of their peers. Most in this group were more likely to be employed by discount mass merchandisers, followed by department store retailers and to hold a store manager title. Most of these managers had college degrees.
Cluster 2 (team builder/goal-oriented), on the other hand, was the largest profile, with a stronger emphasis on both of the lower quadrants (especially the lower right quadrant, the rational goal model). Managers in this group perceived themselves as engaging in all of the functions more frequently than those in other groups. However, they tended to place a greater emphasis on the producer and the director roles than on other roles. Managers in this group also tended to seek consensus among team members and to work with external constituencies. Thus, these managers tended to join the competing roles of directing and being goal-oriented with the roles of facilitating and mentoring to build a team-oriented atmosphere.
People in this group stood out in the intensity of their personal values in that they had a higher level of social affiliation, growth/development, achievement and security values than other groups. Managers in this group also perceived their job as task-oriented and as offering opportunities to receive feedback from their supervisors to a greater degree than those in other groups. These managers were more likely to be employed by department store retailers, followed by discount mass merchandisers. They were also more likely to be non-white or female and tended to have lower educational levels than those in other groups.
Cluster 3 (conceptual producer/external focused) was unique in that it had an extremely weak emphasis on the lower left quadrant (the internal process model) but had a relatively strong emphasis on the upper quadrants (especially on the broker and facilitator roles). Managers in this group were well balanced in all areas except for the monitor and coordinator roles. They were strong on facilitating internal group processes while networking with external constituencies. This group was somewhat similar to conceptual producers in Quinn et al. (1987) in that conceptual producers scored highly on broker and facilitator roles but lower on the monitor and coordinator roles.
Managers in this group were not too different from those in Cluster 1 in the areas of personal values and job characteristics but had a higher level of job satisfaction and organizational commitment. They were the most rewarded financially as indicated by their annual average compensation which was more than twice the two other groups. Also over the last three years their compensation had increased over twice the rate of those in the other two clusters. These managers received nearly twice the recruitment solicitations from other employers than those in the other two groups, and in general perceived themselves to be on a fast career track. People in this group were much less likely to be employed by traditional retail formats (e.g. discount stores, department stores) and more likely to work for grocery/food or specialty retailers. They were also more likely to hold a upper-level management title such as district or regional manager/vice president or senior executive in the home office. A slightly greater percentage of males were represented in this group, and a substantially greater percentage of individuals in this group had some graduate education or a graduate degree.
Theoretical implications
The findings of this exploratory study provide some interesting implications theoretically and shed some light on important questions to address in future studies. In general, it can be concluded that different leadership styles displayed by retail managers and leaders are influenced by a variety of factors, including personal values, job characteristics, organizational commitment and job satisfaction, career progression, and organizational and personal demographic characteristics. Several studies have been conducted to determine the relationships among such variables as organizational commitment and job satisfaction (i.e. Dubinsky and Skinner, 1984; Good et al., 1988). Our study offers seminal findings regarding how leadership roles might affect the theoretical framework of organizational commitment and job satisfaction, along with other variables such as personal values, job characteristics and career progression.
Quinn's model of leadership styles has proved effective in classifying retail managers into three distinct groups based on the degree to which they emphasize various leadership functions. It is important to emphasize that this study was based on a self-assessment of leadership functions. Therefore, cluster analysis based on assessments by others such as subordinates, peers, and/or supervisors might provide different results.
The findings regarding differences in personal values among the leadership styles suggest that personal values are significantly related to the leadership styles displayed by managers. This result supports Quinn's (1988) assumption that the four leadership styles reflect an emphasis on different types of values or motivations. It also supports previous studies that found significant relationships between personal values and leadership orientations (e.g. Kotey and Meredith, 1987).
Managerial implications in the context of the retailing sector
Perhaps the most enlightening implication from this study may be that retail managers and executives can self-assess their leadership styles using Quinn's instrument (see Table II) to determine how their leadership style compares to the three styles found in this study. This self-assessment can be used to identify leadership roles that such managers and executives need to emphasize or de-emphasize in order to achieve their desired leadership style. This instrument can also be used by those who lead retailers' human resources departments to train and assist current and/or new managers and executives.
Other meaningful implications exist for retail management in the correlation of leadership styles with job characteristics, job satisfaction and organizational commitment. Retail managers who view their work to be repetitive and to place little demand on their management skills tended to be in the loner/internal focused leadership group. The fact that many of these people tended to be store managers as opposed to other types of managers (e.g. human resource managers, entry/departmental managers, or upper level executives) in this study indicates the possible value of a serious educational training program designed to stimulate store managers. It was interesting to note that 50 percent of the entry/departmental managers were in Cluster 2 (team builder/goal-oriented). This observation may imply that many retail entry level managers begin with a more positive leadership style at first. However, as they move up the career ladder to store manager, they become more isolated and thus develop a loner/internal focused style. Store managers are the key driving force behind a retail firm's profit, yet many of these managers in the loner/internal focused group find their job to be less than fulfilling, while having the lowest satisfaction level and the least organizational commitment. A longitudinal study would reveal the process by which entry level managers develop their leadership styles as they move up the career ladder.
The finding that those in Cluster 3 (conceptual producer/external focused) appeared to have enjoyed a more rapid career progression than those in other groups suggests that these managers might be most effective and most rewarded in retail organizations, both financially and in terms of career development. If this holds true, the size of the leadership profile (i.e. the-larger-the-better concept) might not be the best approach in identifying the most effective and/or successful leaders. A balanced leadership profile with emphasis placed on a few roles might indeed be the most effective. That those in Cluster 2, whose scores were high on all functions, did not significantly differ from those in Cluster 1 in these career progression characteristics also supports this argument. In other words, the highest scores on all functions may not be perceived as most desirable.
The finding that the retail format was a significant factor affecting leadership styles indicates that not only the general type of industry but a different format within the industry can influence the leadership styles developed by management. Managers employed in grocery retailing, specialty retailing and/or newly-emerging entities such as non-store or warehouse retailers, tended to be more clustered around the conceptual producer/external focused group. These retail formats are characterized as less conventional, more aggressive and somewhat smaller in scale than large department store or discount retailers and, thus, may attract managers who are already more conceptually/externally-oriented, or allow managers to be more conceptually/externally-oriented.
It can also be concluded that leadership styles are related to certain personal demographic characteristics (e.g. education, gender, ethnicity). It is interesting to note that college graduates were more likely to be placed in the loner/internal focused cluster than those with a graduate diploma or with no college diploma whatsoever. That those with a graduate diploma were likely to be clustered around the conceptual producer role indicates that further training after a four-year college degree can serve as a significant tool in helping managers develop higher-level thinking and management skills.
Limitations and future studies
Our study is among the first in the retail management field to adopt Quinn's leadership model in classifying retail managers and leaders into a small number of clusters based on their leadership styles. Three clusters of leadership styles were successfully identified based on the managers' emphasis on various competing leadership functions. Managers in these leadership clusters also differed in their personal values, their perception of job characteristics, their job satisfaction, and their organizational commitment. The three groups of retail managers and leaders displayed differences in career progression as well as in organizational and demographic characteristics. Therefore, it can be concluded that leadership styles are a function of multi-dimensional, complex individual and organizational factors.
Quinn's instrument can be used effectively both for self-assessment and for assessment by others. Our study focused only on self-assessment. Therefore, further studies based on assessment by others should be conducted. It would also be enlightening to determine how consistent or inconsistent these perceptions might be. Future studies should also seek to determine the profile of ideal and/or effective managers/leaders as assessed by subordinates. Studies focusing on senior top executives may also result in different profiles and characteristics. Longitudinal studies assessing management trainees' styles and how their styles change over time as they progress through their career would provide insight into the developmental and socialization process by which these styles are developed.
Figure 1Quinn's (1988) leadership functions
Table IDefinitions of Quinn's (1984, 1988) model of leadership roles
Table IIThe measurement of the eight leadership roles
Table IIIA summary of the measurements
Table IVCluster analysis on leadership functions
Table VMultivariate and univariate analysis of variance among three clusters of leadership styles
Table VIChi-square analysis among three clusters of leadership styles
Figure 2Cluster 1: loner/internal focused
Figure 3Cluster 2: team builder/goal-oriented
Figure 4Cluster 3: conceptual producer/external focused
