Research Activity Essay

An Exploration of the

Path-Goal Theory of Leadership

in a Health Care Environment ANDREW D. SZILAGYI University of Houston HENRY P. SIMS, JR. Indiana University

Relationships between leader behavior and subordinate satisfaction and performance, unmoderated and moder- ated by a subordinate's role ambiguity, were studied at multiple occupational skill levels in a hospital. The re- sults supported path-goal theory's propositions concern- ing the relationship between leader initiating structure and subordinate satisfaction, but not leader initiating struc- ture and subordinate performance.

Leadership has been one of the most widely studied aspects of organiza- tion behavior. As a result, a number of theories have been advanced which have focused on (a) the traits of the leader ( 1 5 ) ; (b) the styles of leader behavior ( 1 , 25, 3 3 ) ; (c) the effects of subordinate behavior on subsequent leader behavior (5, 12, 2 6 ) ; and (d) the situational aspects of the leader- ship setting (6, 16, 35). The objective of this research was to determine whether one of the situational theories of leadership—the path-goal theory —could be supported with multiple occupational skill levels in a hospital environment.

THEORY AND RESEARCH

Theoretical Background

One of the current situational theories of leadership is the path-goal theory of leader effectiveness (4, 11, 16). A basic proposition of this theory is that one of the major functions of the leader is to enhance the psychologi- cal states of subordinates, which results in increased motivation to perform

Andrew D. Szilagyi (D.B.A.—Indiana University) is Assistant Professor of Organization Behavior and Management, University of Houston, Houston, Texas.

Henry P. Sims, Jr. (Ph.D.—Michigan State University) is Assistant Professor of Ad- ministrative and Behavioral Studies, Indiana University, Bloomington, Indiana.

622

1974 Volume 17, Number 4 623

and increased subordinate satisfaction with the job. From House's (16) re- search it has been inferred that this function of the leader consists of (a) recognizing and/or arousing subordinates' needs for outcomes over which the leader has some control; (b) increasing personal payoffs to subordinates for work-goal attainment; (c) making the path to these payoffs easier to travel by coaching and direction; (d) helping subordinates clarify expectan- cies; (e) reducing frustrating barriers; and (f) increasing the opportunities for personal satisfaction contingent on effective performance. Since the theoretical foundations are stated in terms of paths, needs, and goals, the theory is referred to as the Path-Goal Theory of Leadership ( 1 9 ) .

In developing path-goal theory. House (16) states that for high occupa- tional level employees (such as research personnel in an R&D organization), task clarifying behavior by the leader (as operationalized by the variable leader initiating structure) can function to clarify path-goal relationships of the subordinate because such high occupational level jobs are frequently ambiguously defined. House thus hypothesized that the more ambiguous the task, the more positive the relationship between leader initiating struc- ture and subordinate satisfaction and performance (16, p. 325). Also, for high occupational level employees the task performed can be hypothesized to be intrinsically satisfying in itself, and therefore the need for social sup- portive leader behavior (as operationalized by the variable leader considera- tion) will be less than for lower occupational level employees (16, p. 324).

For lower occupational levels such as blue collar employees, leader initiat- ing structure can be hypothesized to be perceived by the subordinate as an imposition of external control. Such lower level jobs are generally routine, and hence the path-goal relationships are usually quite clear. Leader initiat- ing structure contributes little toward further role clarification of such jobs and may prove dissatisfying for the subordinate. House does point out that while such structuring may prove dissatisfying to the subordinate, it may concomitantly increase productivity by preventing slow downs, idle periods, and so on. Leader consideration serves as a source of extrinsic social satis- faction and support for the employee, which makes the path easier to travel and thus more satisfying.

Additional factors, beyond occupational level, may serve to clarify task demands in organizations. House further explicates the theory and provides a possible link to a hospital environment by stating:

When task demands are self-evident due to a high degree of routinization, or where roles are clearly defined by such factors as mechanization, legal con- straints, contracts, professional ethics or group norms, initiating structure will not result in role clarification and will be unsatisfying to subordinates (16, p. 325).

The hospital includes a variety of jobs and tasks which can vary from the very routine to the most highly skilled and professional of occupations. As Georgopoulos states:

Because of the nature of its work—the hospital shows a great concem for favorable outcomes and for clarity of responsibility and accountability, with little tolerance for ambiguity or error (9, p. 18).

624 Academy of Management Journal December

In a hospital the clarity of role demands may originate, therefore, from sources other than specific routines or duties set down by department re- quirements, as in the case of lower occupational level food preparation em- ployees or janitorial services. The clarity of role demands for higher occu- pational level individuals such as nurses and therapists may originate from professional standards and practices established by educational training or professional societies (9, 10).

Research Support

House (16) applied path-goal theory to convenience data collected from previous research work. In a study of office employees of a heavy equipment manufacturing company, the variable role ambiguity was introduced as moderator of the relationship between leader initiating structure and sub- ordinate satisfaction. Role ambiguity was used as a variable to approximate unclear path-goal relations. As hypothesized, the results revealed a signif- icant negative correlation between leader initiating structure and role am- biguity, and significant positive correlations between leader initiating struc- ture and subordinate satisfaction variables. In a second analysis of previously collected data. House (16) sought to test hypotheses about the moderating effect of task autonomy on the relationship between leader initiating struc- ture and consideration and subordinate behavior. In particular, he hypothe- sized that job autonomy would have a positive moderating effect on the relationship between leader initiating structure and subordinate satisfaction and performance. The results were generally supportive of the stated hy- potheses.

In follow-up studies to House ( 1 6 ) , Dessler (3) and House and Rizzo (20) expanded on the initial formulations of path-goal theory. House and Rizzo (20) studied managerial and professional-technical employees in an equipment manufacturing company. Measures of organizational and leader- ship practices were studied in relationship with subordinate satisfaction, per- ceived effectiveness, anxiety, and propensity to leave. Role confiict and role ambiguity were used as mediating variables. Hypotheses similar to the orig- inal House study were tested. The results not only provided additional sup- port for the theory, but the authors concluded that even more emphasis should be placed on role ambiguity within the organization behavior model.

In line with previous research, Dessler (3) proposed that the variable task certainty (or predictability of the task environment) be studied as mod- erating the relationship between leader behavior and subordinate attitudes. In particular, he hypothesized that as task certainty increases, the relation- ship between leader initiating structure and subordinate satisfaction de- creases, while the relationship between leader consideration and subordinate satisfaction increases. The results of the study were generally supportive of the hypothesized relationships; however, the moderating effect of task cer- tainty proved to be curvilinear in nature.

These research findings indicate that path-goal theory, when operation- alized by such task related variables as role ambiguity, task autonomy, and task certainty, is a viable approach to the study of leadership.

1974 Volume 17, Number 4 625

HYPOTHESES

Given the discussion of the theoretical background and previous research on path-goal theory, the following hypotheses were formulated and tested:

Hypothesis 1: A negative relationship exists between leader initiating structure and subordinate role ambiguity.

Hypothesis 2: A negative relationship exists between leader initiating structure and subordinate satisfaction.

Hypothesis 3: A positive relationship exists between leader initiating structure and subordinate performance.

Hypothesis 4: A positive relationship exists between leader considera- tion and subordinate satisfaction.

Hypothesis 5: The relationship between leader initiating structure and subordinate satisfaction will be positively moderated by role ambiguity.

Hypothesis 6: The relationship between leader initiating structure and subordinate performance will be positively moderated by role ambiguity.

METHOD

Sample

Data were collected from paramedical and support personnel at a major midwestern university medical center. The services provided by the medical center included not only general patient care facilities, but also major divi- sions for research and medical education. This research focused on the gen- eral patient care function of the medical center, which included a 900 bed capacity facility employing approximately 1,600 individuals in various med- ical, support, and staff positions. Employees who were temporary, part-time, or with primary responsibilities in the area of research and/or medical edu- cation were not included in the sample. The medical center was chosen as the site for this research because it offered a particularly rich research en- vironment in terms of the variety of skills and professions encompassed in one organization.

In order to classify the respondents into homogeneous categories to test path-goal theory, the job classification system used by the medical center was adopted. The classification system, which categorized jobs on the basis of area of specialty, education level required, and the extent of post-degree or on the job formal training required, included five main subcategories: (a) administrative, consisting of such positions as department heads, associate and assistant heads, and program coordinators in departments such as hospi- tal administration, accounting, data processing, and personnel; (b) profes- sional, including registered nurses, medical technologists, occupational and physical therapists; (c) technical, including licensed practical nurses, labo- ratory and therapeutic technicians; (d) clerical; and (e) service, consisting of nurses aides, food preparation, and building services personnel. Personal interviews and discussions were conducted with various department heads.

626 Academy of Management Journal December

directors, and supervisors in order to verify the utility of the classification system and to make adjustments for changes, alterations, or variations made by individual departments.

The classification system, which will be identified in this study as occupa- tional skill level, resulted in the following sample breakdown: 53 adminis- trative employees (assistant and associate department heads and program coordinators); 249 professional employees (134 registered nurses, 93 thera- pists, and 22 medical technologists); 132 technical employees (72 licensed practical nurses and 60 laboratory technicians); 227 clerical employees; and 312 service employees (266 food and building service workers, and 46 nurses aides). The clerical group was dropped from the analysis since the majority of the respondents were part time, temporary, or secretarial pool employees.

Only nonsupervisory personnel were included in the professional, tech- nical, and service groups. This eliminated such functions as head nurse, chief therapist, and food supervisor. The administrative group consisted of employees who performed managerial or professional tasks and also had indirect supervisory responsibility. This group, however, did not include direct supervisors of the other occupational skill groups.

The data for this study were collected by means of a questionnaire which was completed during normal working hours in separate classroom and sem- inar facilities on the premises of the medical center. Ouestionnaires were collected from a total of 1,161 employees, of which 79.6 percent were fe- male. A total of 230 questionnaires were discarded due to incomplete re- sponses.

Measures

Leader Initiating Structure (18 items)—This variable described the de- gree to which the leader initiates psychological structure for subordinates, by doing such things as assigning tasks, specifying procedures, clarifying his expectations of subordinates, and scheduling the work to be done. This variable was taken from the Supervisory Behavior Description Ouestion- naire (SBDO) ( 8 ) .

Leader Consideration (28 items)—This variable was defined as the de- gree to which the leader creates a supportive environment of psychological support, friendliness, and helpfulness by doing such things as being friendly and approachable and looking out for the personal welfare of his group. This variable was also taken from the Supervisory Behavior Description Ouestionnaire (SBDO) ( 8 ) .

Role Ambiguity (6 items)—This variable was defined as the lack of clarity and predictability one perceives in his work related behavior. It re- fers to the degree to which an individual perceives that he is working under ambiguous conditions. A six item instrument developed by Rizzo, House, and Lirtzman (29) was used to measure the subject's perception of role ambiguity.

1974 Volume 17, Number 4 627

Job Satisfaction—This variable was defined as the degree to which an individual's desires, expectations, needs, or demands are fulfilled by his employment in an organization. The Job Description Index (JDI) con- structed during the course of the Cornell studies of job satisfaction (31) was the device used to measure the individual's job satisfaction. The JDI meas- ures satisfaction over five areas of a job: the work itself, supervision, pay, co-workers, and opportunities for promotion on the job. This paper focuses only on satisfaction with the work and the supervisor.

Performance Evaluation (10 items)—This variable was defined as the degree to which an individual carries out his job in adherence with certain specified standards or acceptable practices of the organization. Each individ- ual's job performance was measured by having his supervisor evaluate the subordinate's current performance with respect to the following aspects: quantity of work, quality of work, dependability, ability to get along with others, attendance and punctuality, knowledge of work, planning ability, initiative on the job, and overall performance. The supervisor was required to evaluate each of his subordinates for each of the criteria on a five-point scale ranging from extremely low performance to extremely high perform- ance. This instrument used items which resemble those developed by Porter and Lawler ( 2 8 ) .

Validity and Reliability—With the exception of the Job Description In- dex (whose validity and reliability are well supported by the literature), the various instruments were submitted to a factor analysis to determine con- struct validity. Detailed results of the factor analyses, reported by Szilagyi ( 3 4 ) , confirm the construct validity of the variables. The reliability of each scale was estimated through split-half analysis, corrected by the Spearman- Brown formula, and was as follows: 0.93 for Leader Initiating Structure; 0.89 for Leader Consideration; 0.76 for Role Ambiguity; and 0.94 for Per- formance Evaluation.

RESULTS

The Pearson Product-Moment correlations of the major variables which relate to the hypotheses in this study are presented in Tables 1,2, and 3. Detailed results of the individual hypotheses are discussed below.

TABLE 1

Pearson Product-Moment Correlations Between Leader Initiating Structure and Subordinate Satisfaction, Performance, and Role Ambiguity

Correlation Between Leader Initiating Structure and:

Satisfaction with work Satisfaction with supervisor Performance Role ambiguity

Adminis- trative

(N = 53)

.380**

.188 - . 2 9 5 * - . 4 0 6 * *

Occupational Skill Level

Frofes- sional

(N = 240)

- . 1 0 0 - . 1 2 0 - . 0 2 0 - . 1 1 1

Technical (N = 117)

- . 1 0 3 - . 0 5 6 - . 0 2 0

.000

Service (N = 231)

- . 1 4 4 * - . 3 3 2 * * * - . 1 5 4 * - . 0 7 1

.05 **p < .01 ***p < .001

628 Academy of Management Journal December

TABLE 2 Pearson Product-Moment Correlations Between Leader Consideration

and Subordinate Satisfaction

Correlation Between Leader Consideration and:

Satisfaction with work

Satisfaction with supervisor

Adminis- trative

(N = 53)

.337*

.800***

Occupational Skill Level

Profes- sional

{N = 240)

.320***

.826***

Technical (N = 117)

.400***

.762***

Service (N = 231)

.420***

.750***

*P < .05 ***p < .001

Hypothesis 1 The results, presented in Table 1, indicate a neutral to small negative

relationship between leader initiating structure and role ambiguity for the professional, technical, and service groups. The hypothesis was supported for the administrative group, which shows a strong negative relationship.

Hypothesis 2 An analysis of the results, also presented in Table 1, reveals negative re-

lationships between leader initiating structure and subordinate satisfaction with work and the supervisor for the professional, technical, and service groups. Only for the service group, however, was the hypothesis significantly supported. For the administrative group, the results indicate a significant positive relationship between leader initiating structure and satisfaction.

Hypothesis 3 As Table 1 indicates, this hypothesis was not supported in any of the four

groups studied. The relationship between leader initiating structure and subordinate performance for the administrative and service groups was shown to be both negative and statistically significant.

Hypothesis 4 The results presented in Table 2 indicate that there was a significant, posi-

tive relationship between leader consideration and subordinate satisfaction as hypothesized. This positive relationship is statistically significant and consistent across the four main skill classifications. The relationship between leader consideration and satisfaction with supervisor was extremely strong.

Hypothesis 5 A test of the hypothesized moderator effect of role ambiguity was pro-

vided by the use of subgroup analysis. The subgroup analysis was performed by dividing each occupational skill classification into two groups, above and below the role ambiguity mean for each total group, and calculating Pearson Product-Moment correlations between leader initiating structure and subordinate satisfaction. The results of the subgroup aneilysis are pre- sented in Table 3. For the technical and service groups, role ambiguity did not serve as a significant positive moderator between leader initiating struc-

1974 Volume 17, Number 4 629

1 I I

So

Ii 1.5

.-I 00 O O 00 Ov

; C4

I I I

I I I

I I I

• * Ov r )

2 g S I I I

w^ — Ov r) r) "oq o o \ \

ro c-i n m 00 00 ri N q

I I I

+

18 2

+

03 6

g 1

61 6

I I I

S ^ tI ^ •S •£ '% '% ^ <3 B U O O 5

1 1 E n CQ kri

•s -s >§ a a u

2 g 5 V V V

ei .4-)

00

Its Q

630 Academy of Management Journal December

TABLE 4

Role Ambiguity Means and Tests of Differences Between Means

Skill Level

Administrative (N = 53)

Professional (N = 240)

Technical ( N = 117)

Service (N = 231)

Role Ambiguity

Mean

21.15

18.28

17.62

16.71

Test of Significance Between (t-test—two-tailed)

Frofessional Technical

1.60 3.52**

.94

Mean

Service

4.44**

3.00*

1.52

*p < .01 **p < .001

ture and subordinate satisfaction. It is apparent, however, that role am- biguity does begin to be a factor for the professional group, as the variable becomes slightly significant as a positive moderator.

In the administrative group, strong moderating effects of role ambiguity were clearly evident, especially in the relationship between leader initiating structure and satisfaction with work. For those who perceive low role am- biguity, leader initiating structure had a slightly negative relationship with satisfaction with work. For the high role ambiguity group, however, the posi- tive relationship between leader initiating structure and satisfaction with work was very strong, with the difference in correlation coefficients between high and low role ambiguity groups being significant beyond the .01 level.

These results can be further developed through the analysis of Table 4, which presents role ambiguity mean scores for the four occupational skill groups. The table reveals two central points. First, even though there were significant differences across skill levels, in absolute terms the general level of perceived role ambiguity for the technical and service groups was low. The range of scores on the role ambiguity scale extends from a minimum score of six (low role ambiguity) to a maximum score of forty-two (high role ambiguity). The results from this supplementary analysis of the data suggest that the task demands for the majority of the respondents in the hospital sample were relatively clear and well defined. The results for the administrative group, however, were an exception to this general finding of low role ambiguity.

Second, it is evident that a continuum, in terms of role ambiguity, was established across the occupational skill groups. At one extreme was the service group, perceiving low role ambiguity because of the routine nature of the job. At the opposite extreme was the administrative group, perceiving the highest level of role ambiguity in the sample. Near the midpoint of the continuum were the professional and technical groups, whose roles may have been well-defined because of professional standards and practices. The exist- ence of this underlying continuum of role ambiguity is an important key to the interpretation of research results which differ from one occupational skill group to another.

1974 Volume 17, Number 4 631

Hypothesis 6

As with Hypothesis 5, subgroup analysis was used to test the hypothesized moderator effect of role ambiguity. The results, presented in Table 3, reveal that only in the case of the administrative group did role ambiguity signif- icantly moderate the relationship between leader initiating structure and subordinate performance in the hypothesized positive direction. Insignif- icant results were obtained for the three remaining groups. The results fur- ther show that in the administrative group a negative relationship exists with subordinate performance, which is not in accord with the theory.

DISCUSSION

The major objective of this research was to determine whether the path- goal theory of leadership could be supported with multiple occupational skill levels in a hospital environment. Overall, the results supported that part of the theory which dealt with the relationship between leader initiating structure and subordinate satisfaction. Crucial to the interpretation of this relationship was the existence of a role ambiguity continuum which revealed that role ambiguity increased as changes in occupational skill level reflected increasing managerial responsibilities.

At one extreme of the continuum was the service group, perceiving the lowest level of role ambiguity in the sample. This group included employees in jobs relating to food preparation and janitorial and building services. The findings of the research for this group, particularly the significant negative relationship between leader initiating structure and subordinate satisfaction (Hypothesis 2) and the lack of a moderating effect of role ambiguity (Hy- pothesis 5 ) , could be interpreted within the framework of path-goal theory as reflecting the impact of jobs which are routine in nature or in which the role demands are clear and unambiguous.

At the other extreme of the continuum was the administrative group, perceiving the highest level of role ambiguity in the sample. The results for this group—especially the significant positive relationship between leader initiating structure and subordinate satisfaction (Hypothesis 2 ) , the strong negative relationship between leader initiating structure and role ambiguity (Hypothesis 1), and the significant moderating effect of role ambiguity (Hypothesis 5)—were supportive of path-goal theory's interpretation for jobs which are nonroutine or in which the role demands are unclear and ambiguous. Since the tasks performed by this group were managerial in na- ture (e.g., assistant and associate department heads and directors), the find- ings for this group are consonant with path-goal theory.

The findings for the professional and technical groups were generally insignificant except for the slight moderating effect of role ambiguity for the professional group. The results, however, followed a pattern which was more similar to the service group than to the administrative group. Since the respondents in these two groups included nursing, therapeutic, and labora- tory professions of the medical environment, a possible explanation could be that the role demands of the job were well defined because of professional standards and practices.

632 Academy of Management Journal December

The results involving leader consideration (Hypothesis 4) confirmed the hypothesized positive relationship with subordinate satisfaction. The strength of the relationship for the service group was in accord with the theory; however, the strong relationship between leader consideration and subordinate satisfaction for the administrative group was not entirely in conformity with path-goal theory, which would predict a neutral relation- ship (16, p. 324). These results may be reflective of the nature of the health care environment which, as Georgopoulos and Mann (10) state, should emphasize broad member participation at all levels.

While supportive with regard to relationships with satisfaction, the results of this study did not support path-goal theory in its interpretation of the relationship with subordinate performance. The theory hypothesized that there should be a positive relationship between leader initiating structure and subordinate performance under conditions of both clear and unclear task demands. This study, however, revealed a consistent negative relation- ship between leader initiating structure and subordinate performance for all the occupational skill levels. The positive moderating effect of role am- biguity on the relationship between leader initiating structure and subordi- nate performance (Hypothesis 5) was supported for the administrative group; the relationship, however, was negative for both low and high role ambiguity groups of the subgroup analysis.

This lack of support for the theory may be explained by one or more of the following hypotheses:

1. A deficiency may be present in the present formulation of path-goal theory. It may be that the present path-goal formulation is sufficient vis-a-vis satisfaction, but requires further development vis-a-vis per- formance.

2. Path-goal theory has been formulated from research conducted pri- marily in industrial organizations. It may be that the performance as- pect of the theory is not sufficiently relevant in organizations where performance is organizationally defined primarily in terms of human- istic or service goals rather than profit goals ( 9 ) .

3. As House (18) has noted, different results may be possible depending on the particular scale used to measure leader initiating structure— the SBDO (used in this research) or the newer Form XII (LBDQ) scale (32).

4. Despite the previous work by Porter and Lawler ( 2 8 ) , the perform- ance evaluation measure used in this study may not have been suf- ficiently valid. As with most performance evaluation measurement techniques, this measure suffers from the lack of a common anchor point. In effect, the degree to which evaluations made by different supervisors on different jobs are directly comparable is seriously open to question. Results of the performance evaluation technique must therefore be interpreted with caution.

Beyond the aspects pointed out with respect to leader behavior and per- formance evaluation, a second area in need of research involves the use of

1974 Volume 17, Number 4 633

Other than surrogate variables (e.g., role ambiguity, task certainty) to meas- ure an individual's path-goal relationships. A research study which utilized measures of expectancies (i.e., effort-to-performance, performance-to-re- ward relationships) would not only provide a direct test of path-goal theory, but would also provide a link to expectancy theory.

In summary, this research was supportive of path-goal theory and its in- terpretation of the relationship between leader behavior and subordinate satisfaction. This research, however, was not supportive of path-goal theory's interpretation of the relationship between leader initiating structure and sub- ordinate performance.

REFERENCES

1. Bowers, D. G., and S. E. Seashore. "Predicting Organizational Effectiveness with a Four-Factor Theory of Leadership," Administrative Science Quarterly, Vol. 11 (1966) 238-263.

2. Comrey, A. A First Course in Factor Analysis (New York: Academic Press, 1973). 3. Dessler, G. "A Test of Path-Goal Theory of Leadership," Academy of Management

Proceedings, 31st Annual Meeting (1972), pp. 178-181. 4. Evans, M. G. "The Effects of Supervisory Behavior on the Path-Goal Relationship,"

Organizational Behavior and Human Performance, Vol. 5 (1970), 277-298. 5. Farris, G. F., and F. G. Lim. "Effects of Performance on Leadership, Cohesiveness,

Influence, Satisfaction, and Subsequent Performance," Journal of Applied Psychology Vol. 53 (1969), 490-497.

6. Fiedler, F. E. A Theory of Leadership Effectiveness (New York: McGraw-Hill, 1967). 7. Filley, A. C, and R. J. House. Managerial Process and Organizational Behavior (Glen-

view, 111.: Scott, Foresman, 1969). 8. Fleishman, E. A. "A Leader Behavior Description for Industry," in R. A. Stogdill and

A. E. Coons (Eds.), Leader Behavior: Its Description and Measurement (Columbus- Ohio State University, 1957).

9. Georgopoulos, B. S. "The Hospital as an Organization and Problem-Solving System," in B. S. Georgopoulos (Ed.), Organization Research on Health Institutions (Ann Arbor: Institute for Social Research, University of Michigan, 1972).

10. Georgopoulos, B. S., and F. C. Mann. The Community General Hospital (New York- Macmillan, 1962).

11. Georgopoulos, B. S., G. M. Mahoney, and N. W. Jones. "A Path-Goal Approach to Productivity," Journal of Applied Psychology, Vol. 41 (1957), 345-353.

12. Greene, C. N. "A Longitudinal Analysis of Relationships Among Leader Behavior and Subordinate Performance and Satisfaction," Academy of Management Proceed- ings, 33rd Annual Meeting (1973), pp. 433-440.

13. Harmon, H. H. Modern Factor Analysis, 2nd ed. (Chicago: University of Chicago Press, 1967).

14. Hays, W. L., and R. L. Winkler. Statistics—Probability, Inference, and Decision (New York: Holt, Rinehart and Winston, 1971).

15. Hollander, E. P., and J. W. Julian. "Contemporary Trends in the Analysis of Lead- ership Processes," Psychological Bulletin, Vol. 71 (1969), 387-397.

16. House, R. J. "A Path-Goal Theory of Leader Effectiveness," Administrative Science Quarterly, Vol. 16 (1971), 321-338.

17. House, R. J. "Strong Inferences in the Social Sciences," in J. H. Turner, A. C. Filley, and R. J. House, Studies in Managerial Process and Organizational Behavior (Glenview! 111.: Scott, Foresman, 1972).

18. House, R. J. "Leader Initiating Structure and Performance, Satisfaction and Moti- vation: A Review and a Theoretical Interpretation" (Unpublished paper University of Toronto, 1973).

19. House, R. J., and G. Dessler. 'The Path-Goal Theory of Leadership: Some Post Hoc and A Priori Tests" (Paper presented at the Second Leadership Symposium: Con- tingency Approaches to Leadership, Southern Illinois University, Carbondale, 111., 1973).

634 Academy of Management Journal December

20. House, R. J., and J. R. Rizzo. "Role Conflict and Ambiguity as Critical Variables in a Model of Organizational Behavior," Organizational Behavior and Human Performance, Vol. 7 (1972), 467-505.

21. House, R. J., A. C. Filley, and S. Kerr. "Relation of Leader Consideration and Initiating Structure to R & D Subordinate's Satisfaction," Administrative Science Quarterly, Vol. 16 (1971), 19-30.

22. Kahn, R. L., D. M. Wolfe, R. P. Quinn, and J. D. Snoek. Organizational Stress-Studies in Role Conflict and Role Ambiguity (New York: Wiley, 1964).

23. Korman, A. K. "Consideration, Initiating Structure, and Organizational Criteria: A Review," Personnel Psychology, Vol. 19 (1966), 349-361.

24. Korman, A. K. "On the Development of Contingency Theories of Leadership: Some Methodological Considerations and a Possible Alternative" (Unpublished paper, New York University, 1972).

25. Likert, R. New Patterns of Management (New York: McGraw-Hill, 1961). 26. Lowin, A., and J. R. Craig. "The Influence of Level of Performance on Managerial

Style: An Empirical Object Lesson in the Ambiguity of Correlational Data," Organi- zational Behavior and Human Performance, Vol. 3 (1968), 440-458.

27. Oaklander, H., and E. Fleishman. "Patterns of Leadership Related to Organizational Stress in Hospital Settings," Administrative Science Quarterly, Vol. 8 (1964), 520-532.

28. Porter, L. W., and E. E. Lawler, III. Managerial Attitudes and Performance (Home- wood, 111.: Irwin, 1968).

29. Rizzo, J. R., R. J. House, and S. E. Lirtzman. "Role Conflict and Ambiguity in Com- plex Organizations," Administrative Science Quarterly, Vol. 15 (1970), 150-163.

30. Saunders, D. R. "Moderator Variables in Prediction," Educational and Psychological Measurement, Vol. 16 (1965), 209-222.

31. Smith, P. C , L. M. Kendall, and C. L. Hulin. The Measurement of Satisfaction in Work and Retirement (Chicago: Rand McNally, 1969).

32. Stogdill, R. M. Manual for the Leader Behavior Description Questionnaire (Columbus: Bureau of Business Research, Ohio State University, 1965).

33. Stogdill, R. M., and A. Coons. Leader Behavior: Its Description and Measurement, Research Monograph 88 (Columbus: Bureau of Business Research, Ohio State Uni- versity, 1957).

34. Szilagyi, A. D. An Empirical Investigation of the Interaction Relationships of Super- visory Behavior and Selected Task Variables on Subordinate Satisfaction and Perform- ance (Doctoral dissertation, Indiana University, 1973).

35. Yukl, G. "Toward a Behavioral Theory of Leadership," Organizational Behavior and Human Performance, Vol. 6 (1971), 414-440.

36. Zedeck, S. "Problems with the Use of Moderator Variables," Psychological Bulletin, Vol. 76, No. 4 (1971), 295-310.

37. Zedeck, S., C. J. Cranny, C. A. Vale, and P. C. Smith. "Comparison of Joint Moderators in Three Prediction Techniques," Journal of Applied Psychology, Vol. 55, No. 2 (1971), 234-240.