The Isles Leadership: The Voyage of the Behaviorists

The Leadership Box ã David Boje, December 7, 2000  




BEHAVIOR reseach is about TRANSACTIONS, not much on transformation

Lewin, 1938 Iowa University
Flanagan Critical Incidents for Behavior
Fleishman et al 1953 Ohio State
Katz et al 1950 Michigan University
Merton 1957 Role Theory
Tannenbaum & Schmidt 1958
Yukl 1971 Participative Leadership
Mintzberg 1973 Managerial Roles


Yukl 1989
Howell & Costley 20001, New Mexico State

Problems with Behavior & Process Leadership


Once upon a time,

most of the Leader Academy left the Isle of Traits and set sale for the Isle of Behaviors by the 1940s, just in time for WWII. They no longer believed in Myers or Briggs, and suspected X and Y like other trait research was some kind of fraud. The military wanted to know if leaders could be trained, and if so, what behaviors made them most effective. The Academy of Leader Professors wanting to get tenure, fame in time of world crisis, and fortune decided that some new theory of leadership must be found or all their jobs would be as extinct as dinosaurs. Working with the Army and with universities, two biggest bureaucracies in the world, it was mostly about transactional behavior, being autocratic or democratic to increase the transaction rate or quality. The game of life in organizations or ogranizations in society was never to be transformed.  

Their quest is this: to find universal leader behavior styles that correlate with effectiveness and are optimal transactions in ALL situations

Some desperate explorers led by Squire Fleishman and Sir Katz set off for the Isle of Behavior in separate ships. Their they discovered that Scribe Lewin had already established a behavioral settlement and an Iowa University since 1938. On the Isle of Behaviors, leader (transactional) behaviors became observable and their study turned objective and measurable.

Battle of the Universities. Each wanted to make its mark. Each wanted to study what do leaders do, use some statistical methods, and found their own state university.  Soon the Ohio State and Michigan University competed for the education of the peasants.  Fleishman became King of Ohio State and Katz was made King of Michigan University. Lewin was already King at Iowa. Each mustered their armies and prepared to battle for leader behavior territory.  As they became more warlike they built ships and did battle with the Motherland of Traits.

For decades the Ohio State and Michigan University kingdoms ruled the world of leadership.  But the taxes collected were too much and in 1957 Merton the Magician set off for new Isle of Behavior territory.  Roles Territory was established, a predacessor to theatre, to scripts.  Yukl led a revolution and established a new religion called "Participation" (kind of a rediscovery of what Scribe Lewin found in 1938).

Sir Mintzberg, knighted by the Canadians, resettled in the Isle of Behavior.  He decided to go and look to see if leaders did any planning, organizing, controlling, or leading.  He actually observed and charted what transactions that leaders do.  The world was shocked to discover, that leaders had a hectic, frantic, and fragmented transaction life, and did little of the behaviors thought to take place.  Some leaders were only figureheads, but he did confirm Sir Merton's view, but noting all the roles that leaders do.

While the Isle of Behavior was awash with two-factor studies of behavior and observations of roles here and everywhere, that great explored, Prince Yukl decided that process was more important than some list of universal behaviors.  And by 2001, Prince Howell and Knight Costley joined the search for process. They still liked to isolate and measure behaviors, but wanted to do this in the study of processes.  They made great maps of the world of leadership, charting each territory.  Leaders were reduced from traits or greatness to just psychoalgebraic behavioral equations, to "styles," or just transactions.

But alas most of the Leader Behavior Academy had already set sail for the Isle of Situation. It seemed obvious that Traits and Behaviors to be effective depended upon the Situation.  If there were universal behaviors, they are not optimal in ALL situations. Therefore a great expedition set forth to the Isle of Situation in the 1960s, with new waves of migration each decade since. This is where the arts of transformation were rekindled. Still a few die hearts remained on the Isle of Behaviors. 


There are two main theories of Leadership Behavior, Transaction and Transformation. This is what we call the "X" dimension of behavior leadership theory.

Figure 1: The XYZ In the Box Model of Leadership

The XYZ model has three dimensions. It is the X dimension that focuses on the Behavioral School of leadership.

X Dimension runs from Transactional to transformational leadership, as studied by Burns (1978) and Bass (1985). The debate rages, is this one dimension or two? This is a classic dualism in leadership studies.  Burns looked at modal thinking (the means over ends reasoning) in the early stages of development and held these leaders to be "transactional" in their behaviors. Transactional leadership "requires a shrewd eye for opportunity, a good hand at bargaining, persuading, reciprocating" (Burns, 1978:169). A "transformational leader," on the other hand, "recognizes and exploits an existing need or demand of a potential follower... (and) looks for potential motives in followers, seeks to satisfy higher needs, and engages the full person of the follower" (p. 4). Everntually transformational leaders were thought to engage in behaviors that changed the game, even changed the world.

In the Theatre approach,  leader behaviors are viewed as plots (grasping together characters, behaviors, and events). This is explained in the Myers-Briggs Study Guide.

Lead Plot s

along the X Behavior dimension (from transaction to transformation).

Y Dimension - From the Will to Server to the Nietzschean Will to Power. Again, is this one dimension or two? The Will to Power is specifically excluded from transaction and transformational leader theory by both Burns and Bass. I therefore treat it as a second dimension of leadership.  It is quite silly study leadership as just a well to serve; many leaders pursue power, some are able to do good things with it, others are swallowed by power. Nietzsche wrote about Will-to-Power (WTP) and Thus Spoke Zarathustra (TSZ) as having something to do with the will to initiate and implement a goal as well as the more macro construct of Darwin's theory of natural section, the power to transform the inherited advantages from generation to generation (WTP #362). And WTP is also a Will to Truth (TSZ, pp. 28, 113). The WTP is a will to overcome the small people, "they are the superman's greatest danger" (TSZ, p. 287). And the superleader is not satisfied with the happiness of the greatest number of workers or consumers (TSZ, p. 287). The Super leaders sees the abyss with the eyes of an eagle and grasps the abyss of poverty and misery with the talons of an eagle (TSZ, p. 288).

Z Dimension  - Participation is from monophonic (single voice) to (polyphonic) involvement in leadership. Some leaders cultivate one voice, their own, and other leaders are more pluralistic, able to create polyphonic and more participative leadership. 

First - there was one voice -In bureaucratic theater, there is mostly monologue. In bureaucratic leadership, for example, there is mostly monologue; other voices are there on the stage but forbidden to speak, or they can only be whispered, their words unhearable, drowned out by the one official narrator who is authorized to take center-stage and speak and speak some more.  As Kirkeby (2000: 232) argues it is the right of power to narrate events, to declare them romantic, tragic, comedic, or ironic, and then of course make them all into a romantic narratives that fits the bureaucratic pension for monophonic (single voiced) influence.   For any other voice to speak would be an act of bureaucratic espionage; certainly for the secretary to speak would be unthinkable rebellion.

Second - there were two voices - In the Quest two or more players take the stage, but it is rarely more than dialog. In dialogue the "I" and the "Other" take the stage and we hear voices, but little reflection. It is no longer the monologue of the I declaring the Other as villain. The Other gets to speak and be heard by the 'I."

Third - there were three voices - To me, this voice that Kirkeby describes is the same one discovered long ago by Adam Smith. Smith looked at global capitalism and say that without ethics events might well follow a logic of the market place that would not lead to ethical relations among buyer and seller, employer and employed, monopolist and entrepreneur. It is the internal spectator, the voice that speaks to us while observing the First and Second (the I and the Other) rehearse there dialogue on the stage in our mind's eye. And in this model, even two actors on the stage visualize the dialogue of the Triad in their own head, but as well in the head of the other.

Fourth - then there were four voices - This is a very special voice, one we sense is about to speak but does not, one that is on the stage but stays in the shadows. In the Fourth, "the event is never over and done with" (Kirkeby, 2000: 237). And with the about to speak voice of the Fourth, we are intuitively aware of the simulation and almost can here the polyphony of voices, a mob about to take storm the stage. We may hear a groan, a murmur, a mumbling sound, but we can never quite make out the words. We can sense somehow the bureaucratic machine, the quest journey, and even chaos itself are just mythic metaphors some people have speculated and articulated about the web of human events (web is yet another one, as it theater a metaphor). We sense the gap, and we know with one more step we will certainly fall into chaos. See Boje (2000c) for more on the multiple voices of leadership.

Table 1 gives you a summary of the relation of the XYZ dimensions to SEPTET (theatre dimensions), and to SEAM (Socio-Economic Approach to Management). For applications see ENRON is METATHEATRE Web Site.

  Table 1

Out of the box

In the box




Box  Dimensions


Plot s

X Behavior dimension (transaction/transformation)



Y Power dimension (will to serve/will to power)



Z Participation dimension (1 voice/many voices)

Inside Box



Traits (Myers & Briggs)




Situation of Box



Situation (time & place)


Spectacle s


We turn now to the History of authors in the Behavior School of Leadership:

Lewin et al -1938 3 STYLES MODEL of Iowa University

What transactional behavior styles matter? This is the study they say started the quest for Behavioral Leader Styles. The Buffalo Leaders is the autocratic one who Tells and Sells, and the result is a submissive group of employees.  The Democratic style is to Share and Consult, and the result is the cohesive team, the flight of the Geese; it's result is a cohesive group.  The Laissez-faire style is to give no direction at all, and the result can be frustration, disorganization, and low quality. 

Lewin, Lippitt, & White


AUTOCRATIC (Directive):
Centralize authority
Dictate work methods
Make unilateral decisions
Limit employee participation

DEMOCRATIC (Participative):

Involves employees in decision making
Delegates authority (now its called 'empowerment')
Encourages participation in deciding work methods and goals
Uses feedback as opportunity to coach employees
Participation sometimes results in higher satisfaction 
Greater decision acceptance sometimes



Seeks input and hears concerns and issues of employees but makes final decision.


Allows employees to have a "say" in what is decided; operates in group mode with leader as one more member giving input


Gives employees complete freedom to ask decisions and complete their work as they see fit.
Provides materials and answers questions
Lewin, Lippitt, & White studies of Boys Clubs found quantity of work in groups led by democratic and autocratic leaders equal, but quality and group-satisfaction higher in democratic groups. 
Subsequent studies (Bass, 1981; Miller & Monge; Wagner & Gooding, 1987; Yukl, 1989) have had mixed results
democratic leader styles sometimes had higher quantity & quality results than autocratic styles, and other times lower or equal to. 
studies using surveys found positive effects of democratic participation with effectiveness, whereas lab/field experiments and studies using independent raters had weak results. 
  1. Results depended upon method used
  2. Democratic participation sometimes leads to effectiveness, but often does not
  3. Impossible for correlation studies to determine the direction of causality (Yukl, 1989: 86)
  4. Research not specific about what the leader is doing with regards to participation (mostly perceptual measures of members feeling sense of influence)
  5. Field studies were confounded by "Hawthorne effect" the special attention of a participation program and special treatment, and not the leader behavior explained the results
  6. Too many studies focused on "satisfaction" of members as the outcome measure, rather than looking at performance criteria
  7. Few studies looked at situation variables


Flanagan 1951

Critical incident methods used by Trait researchers are also applied to leader behavior research. Subordinates, peers, and boss are asked to describe effective and ineffective leader behavior. 
Incidents are categorized by similar behavior profiles (by researchers, panel, or participants).
Planning, coordinating, & organizing operations
Supervising subordinates (directing, instructing, monitoring performance)
Establishing and maintaining good relations with subordinates
Establishing and maintaining good relations with superiors, peers, and outsiders
Assuming responsibility for observing organizational policies, carrying out required duties, and making necessary decisions
  1. Categories of behaviors found, differ from one study to the next
  2. Categories appear arbitrary and classification method is subjective
  3. Assumes that people (followers, peers, boss) know what behaviors are relevant to leader effectiveness
  4. Collective memory can be biased toward reporting incidents consistent with stereotypes or implicit theory of what is an effective leader (instead of reporting eye witness narratives)
  5. Incident reports recount behaviors for very specific job requirements and types of leader studied, making comparison to different leaders and contexts difficult

Adapted from Yukl (1989: 88)

Fleishman et al 1945 OHIO STATE 2 Factor Model

To determine the leadership styles, the Leader Behavior Description Questionnaire (LBDQ-XII) was administered to the world (Fleishman, 1953; Hemhill & Coons, 1957; Halpin & Winer, 1957; Fleishman & Harris, 1962). There is also a Leader Opinion Questionnaire (LOQ) completed by the manager. According to Starbuck (1996) "Initiating structure embodied the essential properties of the leadership concepts of 1910, and consideration embodied the concepts of the 1930s." 

The main point -> Initiating Structure and Consideration are just about transaction behaviors, and not about changing the game of life at work. Two two statistically independent dimensions behavior-factors or dimensions were said to be universal. A leader could be low, medium or high on the 2 dimensions at the same time:

This is the One Voice Leader who Directs with Transactional and Task oriented style. At best great for routine and repetitive task, at worst, the micro-manager.
  1. Lets work-unit members know what is expected of them
  2. Schedules the work to be done
  3. Encourages the use of uniform work procedures
  4. Assigns work-unit members to particular tasks
  5. Plans tasks for work-unit members
  6. Makes his or her attitudes clear to the work unit
  7. Clarifies work roles
  8. Asks for results
This is the Poly Voice Leader who is people-oriented and Participative, and a bit Transformational.
  1. Treats all work-unit members as his or her equal
  2. Is friendly and approachable
  3. Does little things to make work pleasant
  4. Puts suggestions made by the work unit into operation
  5. Looks out for personal welfare of work unit members
  6. Supportive socioemotional work atmosphere
  7. Maintains high morale in work-unit
  8. Collaborative work atmosphere
  1. These are both about transactions, not much transformation behavior here.
  2. There is no Will to Power/ Will to Serve assessment here. Both Initiating Structure and Consideration are but a some portion of the Leadership Box we study here.
  3. The model is U.S. and culture bound; fortunately most nations have other ways of being leader. 
  4. After 50 years of research, the LBDQ is thought to be unreliable yet as a measure of leader behavior; yet it is still used widely.
  5. The 2-factors are so widely taught in Business Schools that they have been born into managers' thinking
  6. LBDQ method relies upon recall and perceptions of respondents, not on behavior study (Campbell, 1977)
  7. Surveys like LBDQ have all kinds of bias to validity and reliability (e.g. recency, desirability, halo, primacy, stereotyping)
  8. First tested in the US, the LBDQ does not adequately depict behaviors in non-US cultures
  9. Cause vs. Effect? - "During much of the 1970’s, critics claimed research on leadership had little practical use. Many studies involved groups of twenty or so college students where one group was treated kindly and the other group was treated harshly, and both groups were given some simple tasks to accomplish such as assembling component parts" (Steve Hallam).
    researchers also jumped on the anti-leadership research bandwagon claiming that researchers were only
    “seeing” leadership because they expected to see it. If they observed a business turnaround and there was no other obvious cause, the turn around was attributed to leadership. 
  10. LBDQ was doomed to failure as a method, but the universal appeal of 2-factors keeps it alive to this very day. There must be other factors.
  11. One Best Style - The universalists immediately formed a `one-best´ leadership style fits all: the leader who can exhibit high consideration and high initiating structure (some research does support  this by Fleishman & Harris, 1962; House & Filey, 1971; Hoy & Brown, 1988: 27 in study of teachers). 
  12. The universal theory of leadership is an antihero approach, where impersonal factors and forces perform the executive function. 
  13. No Heroes here - It is a universal theory fit for the non-risk taker who initiates structure with consideration in a bureaucratic cage. Where else would anyone call this leadership? 
  1. Look at situational factors such as time pressure, ambiguity, external threat, etc. that suggests a more imitating structure would be warranted over consideration
  2. Challenging and interesting work and structured tasks may eliminate the need for imitating structure
  3. With clear goals and tasks, more initiating structure leads to feeling micro-managed, lowering work-unit satisfaction
  4. Workers low in growth needs, education, and knowledge may favor higher initiating structure
  5. Look for other factors (See Michigan State)


University of Michigan called the Survey of Organizations (Katz, Maccoby, & Morse, 1950). This behavioral survey, like the Ohio State LBDQ, is once again reliant upon the perceptions of only one source (subordinates) completing the questionnaire. It began in 1947 when Rensis Likert and his group of social researchers at University of Michigan launched series of leader studies in range of settings.

Main Point: -> Once again, 2 dimensions, this time production or employee centered, and these are about managing transactions, not about changing or transforming the rules of the game.


Production Centered: Employee Centered:
Emphasize technical or task aspects of the job
Concerned mainly with accomplishing group's goal
Regard group members as means to an end
Emphasize interpersonal relations
Take personal interest in needs of employees
Accept individual differences among members
High producing groups were led by leaders with an employee-centered style.

ROLE THEORY - Merton 1957

Merton's (1957) argument is that each social position of a leader is associated with a set of role-specific forms of behavior that together comprise a 'role set'. For example, "a medical student must act as a student not only in relation to other students, but also in relation to teachers, doctors, nurses, social workers, patients, medical technicians, and so on" (source). Leaders negotiate conflicting expectations of their stakeholder network: employees, bosses, superiors, peers, competitors, community, suppliers, etc. 

Latent Function - Merton contended that a behavioral pattern can have consequences (a) beneficial for a dominant economic or political structure; (b) unintended by the actors; and (c) not recognized by the beneficiaries as owing to that behavior (source). The question is does a pattern of leader behaviors have a function that can explain their presence? 

Implications - SCRIPTS - I would like to dust off Merton, and put him in the contemporary theatre school of leadership. The leader has a set of reole behaviors, that in theatre terms, constitute their "script." This behavior script is negotiated with the stakeholders with which the leader interacts. In addition, the leader directs, edits, and constructs the behavioral scripts of others. If the scripts are about transactions, and improving but, not changing the status quo, then it is a transactional script. If it is about transforming the rules of the game, or the rules by which the organiztion lives in its ecological and community enviornment, then it is a transformational script.

Merton, R. K. 1957. 'The Role Set: Problems in Sociological Theory'. British Journal of Sociology, 8.



TANNENBAUM & SCHMIDT 1958 Continuum of Leader Behavior (Autocratic to Democratic

Use of authority by manager <----------------------> Freedom for subordinates
Manager makes decision and accounces it

Manager "sells" decision

Manager presents ideas and invites questions Manager presents tentative decision subject to change Manager presents problem, get suggestions, makes decision manager defines limits; asks group to make decision

Manager permits subordinate to function within limits defined by superior 


Tannenbaum, R., Schmidt WH (1958), How to choose a leadership pattern. Harvard
Business Review 36/2, 1958: 95-101. 

Note: The Autocratic to Democratic continuum model of Tannenbaum and Schmidt (1958) builds upon the early work of Lewin et al (1938), both autocratic and democratic styles are apparent, but laissez-faire is absent.

Main Point -> Again this is all about transactions, be it tell, sell, consult, or share. Share and consult transactions are nicer and more social than tell and sell, but it is still about transactions.


Yukl References under construction.

Mintzberg Managerial Roles

Mintzberg's (1973, 1975) classification of managerial roles. Mintzberg did observation studies of five chief executives, and found that they did not divide their time into planning, organizing, influence, lead, and control. Rather the manager played ten fragmented roles in a high interruption environment..\ Half of these managers' activities lasted less than nine minutes and are very transactional. 







Out of the Box Interpretation  - Leadership is theatre, and the leader is suspended in a web of ten scripted roles. Some leaders use these roles with more persuasive power than others to influence spectators and other actors. 

The point - it is not transformational leadership until the leader changes the script of the organization.


Mintzberg web site and references
Mintzberg, H. 1975. The manager’s job: Folklore and fact. Harvard Business Review 53: 49-61.
Mintzberg, H. 1973. The nature of managerial work. New York: Harper & Row.



Howell and Costley (2001)  

Jon Howell in an interview with Sun-News (January 29, 2001: 7) states "A leaders' behavior must match the situation, and the news of his or her followers." And his summary is very appropriate here:


Leader effectiveness is determined by what people do, not by some inherent personal characteristic... I'm not saying personal characteristics don't help; they certainly do. But leaders have to adapt their behavioral styles to fit the situations in which they find themselves" (p. 7). 

Howell says the good news is most people can learn leader behaviors and learn to recognize situations in which certain behaviors are most important.  Howell and Costley (2001) argue for the match of leader behavior, leader traits and characteristics, follower characteristics, and the situation at hand.  And there are seven leader types, fit for various behavioral processes and situations in my read of their leader theory:

  1. Supportive Leaders (those considerate, people oriented leaders).
  2. Directive Leaders (fit for repetitive or work spread between sites and for cultures such as Mexico that prefer status well defined.
  3. Participative Leaders such as Dwight Eisenhower who could tame the Primadonna generals and politicians of WWII, and by careful listening to many voices craft an alliance. 
  4. Reward and punishment leaders (transactional).
  5. Charismatic (heroic) leaders
  6. Boundary spanning (network) leaders
  7. Leaders who build and forge social exchange (also networkers). 

And now we have left the obsession with one best style of leadership. There is no universal style. There is as I have suggested, a dimension of behaviors running between Transactional and Transformational. The behavior school to this point, is fixated on the transactional. To find transformational we must sail to the Isle of Situation.



Do you have as much sense as a goose? When geese fly in the "V" formation, the whole flock adds considerably more to its flying range than if each bird flew alone. Whenever a goose falls out of formation, it suddenly feels the drag and resistance of trying to fly alone and quickly gets back into formation to take advantage of the power of the formation. When the lead goose gets tired, it rotates back in the wing, and another goose flies point. The back geese honk from behind to encourage those up front to keep up their speed. Finally, when a goose gets sick and falls out, two geese fall out of formation with it until it is either able to fly or it is dead. They then launch on their own, or with another formation, to catch up with the group. 


(Afsaneh & Nahavandi, 2000: 155), In The Art and Science of Leadership. NJ: prentice Hall). 

The Flight of the Buffalo.  In recent years, authors such as Manz and Sims have proposed their Superleader model, and Belasco & Stayer (1993) The Flight of the Buffalo.  The choice given to leaders is to get beyond the 1950s behavioral models of consideration/ initiating structure and become geese, flying with incredible team work, self-leadership, and networking.  In the next table, the two approaches are compared, pointing out their fit to particular cultures and situations (For more see Flight of the Buffalo).

TABLE: Buffalo or Goose Leadership?


e.g. Mexico and France where Eiffel Tower obedience and respect for authority count

Task oriented
Initiating Structure
Task-motivated leaders who just do not delegate
These are not team players
Yet there is a fit to repetitive work processes in cultures that support the authoritarian manner, and where commitment is not a concern. 

e.g. US and Australia where egalitarian power distributions and individualistic styles count

e.g. Japan whose cooperative culture is imitated by the world; but US and others could not make the team cooperative style work and is moving on.



The belief in self-managing teams and in self-leadership.  The leader's new role is to be the facilitator, provide the training, and trust the team to find its way.  


Laissez Faire (a bit)
These are supportive and participative, relations motivated leaders.
They delegate and seek to become one of the team.
There is a fit here to more complex, multi-faceted tasks, and places where commitment counts, and the corporate and societal culture supports participative ways. 
Self-managing teams with facilitative leaders who help the team define its boundaries, find resources, and get multi-skilled.

In the above Flight of the Buffalo, the Super leader comes closest to transformational leader. X and Y are just transactional buffaloes.

How do leaders behave in teams?

In the next table, the horizontal dimension (is does not team -- yes to teams).  The vertical (delegates vs. does not delegate).  










Dumps the tasks onto the group who has no clear goals or structure of its own to rely upon. The result is frustratation, poor quality, and resistance.


Delegates to a team with goals and structure; sets expectations; provides training and help with managing conflicts; shares information. 









Tracks group progress and micromanages to point of idiocy; withholds info from the group to sustain power; makes the team submissive to will



Uses self-managing team to free self up for more strategic adventures; Trusts team has the skills and talent to handle the work; Prefers superleader role of facilitator and boundary spanner







See tables above for specific problems. The basic critique is that most of the behavior school of leadership studies situations and roles that are transactional. They study different scripts of transacitonal leaders, comparing one to the other. This could be because in their dislike for trait school, they did not want to get back into the great man/woman studies. So tranformational leaders were not study and instead an abundance of managers doing transactions were studied.

Another problem is that with the transaction school growing up in WWII, few trusted the charismatic and transformative leaders. No one wanted another Hitler. Everyone wanted to keep leaders in the box, and limit them to transactions that could be controlled. As people forgot about WWII, they flirted more with charismatic and transformational leaders. For more on this see Transformational and James McGregor Burns.

For more on this we need to go to the Isle of Situation.

go to Index of Choices in this page (top)     Go back to Choos In or OUt of Leader Box