Uploaded by naome shabangu

1-s2.0-S1048984309001507-main

advertisement
The Leadership Quarterly 20 (2009) 837–848
Contents lists available at ScienceDirect
The Leadership Quarterly
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / l e a q u a
Situational leadership theory: A test of three versions
Geir Thompson a,⁎, Robert P. Vecchio b
a
b
Handelshøyskolen BI, 0442, Oslo, Norway
Department of Management, Mendoza College of Business, University of Notre Dame, Notre Dame, IN 46556, United States
a r t i c l e
i n f o
Keywords:
Situational leadership theory
a b s t r a c t
Three versions of the leadership dynamics derived from Hersey and Blanchard's Situational
Leadership Theory were identified: (1) the original, 1972, statement of the theory; (2) the revised,
2007, theory; and (3) an alternative statement of the theory's essential principle of differential
follower response to “autonomy afforded by the leader” in conjunction with “follower
developmental level” (as indexed by employee job experience). Survey data collected from 357
banking employees and 80 supervisors, sampled from 10 Norwegian financial institutions, were
analyzed for predicted interactions. Results indicated that the 2007 revised theory was a poorer
predictor of subordinate performance and attitudes than the original version. The third,
alternative, version (which predicted an autonomy × job experience interaction) offered
promise for further exploration of the theory's essential principle that employee outcomes are
associated with prescribed leader behaviors in combination with follower developmental
level, although this version also did not add substantially to accounting for criterion variance.
© 2009 Elsevier Inc. All rights reserved.
1. Introduction
Although it is among the most widely-known theories in the domain of managerial leadership, Situational Leadership Theory
(SLT) remains among the less well-substantiated models. While popular management textbooks routinely include SLT in their
leadership chapter, they seldom critique the absence of empirical support for the theory. In contrast, major textbooks devoted to
leadership (e.g., Northouse, 2007; Yukl, 2006) typically note the lack of a strong evidence-base for SLT. Yet, outside of the
community of scholars who actively study leadership, SLT is less critically viewed (e.g., in management training programs and in
school-teacher training settings). The continued popularity of SLT principles is further attested by the recent appearance of a
revised version of SLT (Blanchard, 2007).
As originally proposed (Hersey & Blanchard, 1972; 1982), SLT predicts that an optimal style of supervision (definable by specific
combinations of leader relationship-focus and leader task-focus) can be prescribed for given levels of subordinate maturity (definable
as the combination of subordinate commitment and competence). The precise form of this relationship between supervisor behavior
and subordinate attribute can be stated as follows: for subordinates of low-level maturity, superiors should show relatively lower
consideration and higher task structuring; but as subordinates gain in maturity, task structuring should decrease, while
considerateness should rise for mid-level subordinate maturity and then subsequently subside as subordinates achieve high-level
maturity. For expositional purposes, four combinations of subordinate maturity and leader style are identified: (a) the subordinate of
very low maturity who should benefit from a “telling” style of supervision, (b) the subordinate of moderately low maturity who should
benefit from a “selling” style of supervision, (c) the subordinate of moderately high maturity who should benefit from a “participating”
style of supervision, and (d) the subordinate of very high maturity who should benefit from a “delegating” style of supervision. Stated
in its most essential terms, SLT forecasts that follower readiness for self-direction (“readiness for self-direction” being an alternative
and less pejorative label for the notion of follower maturity) is a critical contextual feature in dictating an optimal leader style.
⁎ Corresponding author. Tel.: +47 46410299.
E-mail address: geir.thompson@bi.no (G. Thompson).
1048-9843/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.leaqua.2009.06.014
838
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
1.1. Strengths and criticisms of SLT
As noted by Northouse (2007, pp. 96–97), SLT possesses several distinct strengths (vis-à-vis other theories of leadership
dynamics). Specifically, the theory has stood the test of time in the marketplace of leadership training programs, as it is wellknown and commonly used for training leaders. Also, SLT is easily understood, intuitively appealing, and seemingly applicable to a
wide range of leadership settings. Furthermore, SLT is distinctly prescriptive in nature (whereas other leadership approaches are
comparatively more descriptive) and offers guidelines for interpersonal relations. Related to the prescriptive guidance of SLT is the
inherent recognition that there is no single universal style of leadership that is best for all circumstances. Finally, SLT emphasizes
the value of (a) understanding subordinates in terms of differential readiness for taking greater responsibility and (b) developing
the skill-set of followers.
A number of criticisms of SLT have also been identified in the academic literature (Northouse, 2007, pp. 97–100). A major
criticism lies in the ambiguity surrounding the conceptual definition of follower development level. Originally, the dimension of
follower developmental level was defined as the (presumably equally-weighted) combination of follower commitment and
competence. However, the result of any type of weighting to create a developmental dimension is then converted (for purposes of
explicating the model's dynamics) from a continuous dimension to a set of discrete categories of follower level of development.
Precisely how these discrete categories should be created for purposes of testing the theory is not offered by the theory's advocates
and, as a partial consequence, any formal assessment of the theory requires creating a scheme for operationalizing developmental
level when subjecting the theory's prescriptive principles to empirical testing.
1.2. Prior research on SLT
While SLT has garnered substantial recognition within the area of management training (Blanchard & Nelson, 1997), the theory
has also come under much criticism in academic circles for its lack of clear empirical support (Northouse, 2007; Yukl, 2006). To
date, there have been four empirical efforts to fully test the theory's predicted three-way interaction (i.e., that the variables of
leader supportiveness, leader directiveness, and subordinate readiness/maturity will interact in determining such relevant
outcomes as (a) subordinate performance and (b) subordinate attitudes toward the leader, e.g., satisfaction with the leader and the
quality of leader–member relations in terms of LMX). The first published study (Vecchio, 1987) obtained data on several hundred
high school teachers in 14 institutions. Teacher descriptions of superior behaviors were combined with superiors' appraisals of
teacher maturity to predict teacher performance assessments and teacher attitudes. The results indicated that SLT's predictions
held for low-readiness subordinates (i.e., better outcomes were identified for low-readiness teachers who described their
superiors as being more structuring and less considerate). These results suggest that new recruits may benefit from greater levels
of supervisor directiveness, but that SLT may be of little relevance in predicting the performance and attitudes of other employees.
A second study (Norris & Vecchio, 1992) obtained evidence from a sample of nurses and their supervisors. Again, evidence was
more supportive of the theory for lower levels of employee readiness/maturity. In a constructive replication and extension of these
earlier studies, a third study of university employees (both faculty and staff) and their supervisors (Fernandez & Vecchio, 1997)
tested the original version of SLT, which is a within-jobs perspective, as well as a modified version of SLT that was based on an
across-jobs perspective. For the within-jobs comparison, the results suggested that the theory may be more correct in the middle
range of employee readiness/maturity. The across-jobs comparison (which defined employee readiness/maturity as being related
to job level such that the norms, or expectations, associated with employee job level could be substituted for follower maturity as
the key contextual construct) also failed to yield supportive evidence of SLT. However, further analyses did support a prediction
that closeness of monitoring by supervisors would be judged more favorably by low job level employees. A fourth study of SLT in
a military setting (Vecchio, Bullis, & Brazil, 2006) examined the performance and attitudes of 860 members of 86 squads at a
US service academy after four weeks of intensive field training. Directional results suggested that subordinate performance
and attitudes were improved in the low and moderate readiness/maturity levels when the prescribed leader behaviors were
manifest.
1.3. A second version of SLT
Judged in their totality, these empirical assessments of SLT indicate that the theory has minimal (often only directional) support
in the low maturity level condition. Hence, one cannot fully endorse the theory as originally stated. Based partly on feedback from
managers and early critiques of the theory (and perhaps the results of other efforts to test empirically various aspects of SLT, cf.
Hambleton & Gumpert, 1982; Blank, Weitzel, & Green, 1990; Goodson, McGee, & Cashman, 1989), SLT has undergone several
revisions (e.g., Blanchard, 1988; Blanchard, Zigarmi, & Zigarmi, 1985; Blanchard, Zigarmi, & Nelson, 1993) and has emerged most
recently as a restated set of prescriptive principles (Blanchard, 2007). In this newer version of SLT (which is labeled SLT-II), the
interaction between leader behavior and follower developmental level has been modified.
Drawing from theory, research, and practice involving employee empowerment (cf. Blanchard, et al., 1993; Lawler, 1995;
Malone, 1997; Siebert, Silver, & Randolph, 2004), Blanchard (2007) modified the prescriptive curve of SLT-I such that four alternate
definitions of follower developmental level and their four corresponding alternate optimal styles of leadership were specified.
Stated more precisely, the four follower developmental levels (from “developing” to “developed”) are: (1) the enthusiastic
beginner, who is characterized as low on competence but high on commitment, and who benefits from a directive style of
leadership (directive defined as low supportive behavior in conjunction with high directive behavior), (2) the disillusioned learner,
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
839
who is characterized by being low on competence to having some competence in combination with low commitment, and who
benefits from a coaching style of leadership (coaching defined as high supportive behavior in conjunction with high directive
behavior), (3) the capable but cautious performer, who is moderate to high on competence but has variable commitment,
who benefits from a supportive style of leadership (supportive defined as high supportive behavior in conjunction with low
directive behavior), and (4) the self-reliant achiever, who is high on both competence and commitment, and who benefits from a
delegating style of leadership (delegating defined as low supportive behavior in conjunction with low directive behavior).
In comparison to SLT-I, SLT-II still rests on the underlying dimensions that previously were used to define follower readiness/
maturity (i.e., follower competence and commitment). However, the two underlying dimensions that defined readiness/maturity
are no longer treated in an additive fashion (where the attributes of competence and commitment were, arguably, previously
viewed as somewhat compensatory). Instead, follower attributes are now used to create a clearly more discrete typology of
follower style. SLT-II also differs most directly from SLT-I in its relabeling of the most desirable follower developmental end-state
(i.e., where fully-developed followers are no longer in need of supervision, and therefore are allowed a high degree of autonomy)
as one that involves being “developed” rather than high on readiness/maturity. Although SLT-II is still tied to a prescriptive curve
(Blanchard, 2007, p. 89) that seems to suggest that delegation is a combination of low supportiveness and low directiveness, the
deliberate specification of “delegation” indicates that the optimal leader style may be something beyond the combination of low
levels of two types of leader behavior, and instead involves the deliberate relocation of decision-making power. In SLT-II,
considerably more attention is devoted to describing the value associated with attaining the ideal end-state where followers have
achieved a high degree of self-directedness, or self-leadership (Chapter 6). While SLT-II clearly appears to have a more defensible
view for specifying delegation as the optimal style of leadership for fully (or highly) developed followers, the revised theory now
allows for some combinations of the ranges on follower attributes to remain unspecified. That is to say, there are potentially
instances where combinations of competence and commitment will not have specific theory-based predictions for an optimal style
of leadership. Of course, until SLT-II is subjected to an empirical test, we do not know what proportion of cases will actually fall into
a non-predictive category.
1.4. An alternate statement of SLT's essential thesis
Beyond the unique specific predictions offered by SLT-I and SLT-II for the interaction of leader style and follower attributes, one
can derive a more general statement of the SLT approach's essential interaction hypothesis. A more general statement of the SLT
thesis avoids problems that emerge from creating over-specified typologies and attempting to narrowly define optimal styles of
leadership. A more general alternate statement (which we presently label as SLT-III) predicts that there will be differential follower
response to “autonomy afforded by the leader” in conjunction with “follower developmental level.” The measurement of leader
autonomy can be assessed directly from a follower's description of how a leader behaves toward him/her on the relevant
behavioral dimension. This measurement approach is in accord with Lueder's (1985a,b) and Hersey's (1985) observations that
descriptions of leader behaviors taken from the same leaders are suspect and that follower reports of leader behaviors should be
preferred. Furthermore, follower developmental level can also be assessed from the alternate perspective of employee job
experience. This is in accord with prior evidence that employees with greater seniority report less need for supervisory direction
(de Vries, Roe, & Tailleu, 1998). In short, SLT-III predicts that the outcomes of individual follower performance and attitudes will
reflect a “leader autonomy” by “follower experience” interaction. Support for SLT-III, rather than SLT-I and SLT-II, would suggest
that earlier statements of SLT are perhaps needlessly over-specified.
1.5. This study's purpose
In summary, the present study sought to test the validity of the SLT approach by contrasting the predictive utility of different
versions of the theory. Along with attempting to replicate the earlier findings based on SLT-I, the present study sought to provide
the first empirical test of the validity of SLT-II. In addition, the present study sought to test SLT's most essential prediction that
leader autonomy and follower experience interact in predicting follower performance and attitudinal responses. It should be noted
that the three versions are partially nested. That is to say, if one version receives support, another may receive some support as well
(i.e., the versions are not mutually exclusive). Nonetheless, it should be possible to assess the relative superiority of each version in
terms of its predictive utility. Formally stated, the present study tests three hypotheses:
Hypothesis 1. There will be a three-way interaction of leader considerateness, leader structuring, and follower developmental
level in accounting for the follower outcomes of job performance and attitudes (as forecast by the original version of SLT);
Hypothesis 2. There will be evidence of superior follower job performance and attitudes for followers whose developmental level
(as specified by the newer, 2007, version of SLT) matches the prescribed combination of leader considerateness and structuring,
relative to comparable followers of the same developmental level who do not match on the prescribed combination of leader style;
and
Hypothesis 3. Leader autonomy and follower experience will interact in determining follower job performance and attitudes such
that followers of higher experience will respond positively to greater leader autonomy (as predicted by the present re-statement of
SLT's essential thesis concerning the presumed value of affording greater autonomy to more experienced followers).
840
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
2. Method
2.1. Sample and settings
Data were collected from supervisors and subordinates in 10 Norwegian financial institutions who were contacted directly by
the lead author after obtaining approval from top management at each firm. From a total of 477 individuals who were invited to
participate in the study, 357 subordinates and 80 supervisors provided usable data (a 91.6% response rate). This high response rate
was partially due to support for the project from top-level administrators at each firm. Results of the surveys were also shared with
the firms. With the exception of the Vecchio, et al. (2006) study (which utilized 86 leaders), this is the largest number of leaders to
be examined in an empirical test of SLT principles. Supervisors were mostly males (55%) with an average age and education of 45
and 15.5 years, respectively, while subordinates were predominately female (56%) with an average age and education of 44 and
14.5 years, respectively. In addition to providing descriptive demographic information on age, gender, and education, respondents
also reported how long (to the nearest month) they had been employed in their current job, as well as how long they had been
supervised by the same individual. Financial institutions were selected for study because much of the prior research on SLT has
focused on service-oriented organizations (e.g., education, health-care, armed services), while profit-oriented firms have been
relatively neglected. Therefore, the present data sites have the potential to increase our knowledge of whether SLT principles are
demonstrably valid in a for-profit setting.
2.2. Instruments
Supervisors provided assessments of their subordinates on a five-item scale developed by Liden & Graen (1980). Pilot testing of
the instruments with a focus group of six supervisors indicated that the instruments had relevance for a financial institution
setting. As all instruments were originally developed in the English language, they were put through a translation–backtranslation
conversion process to ensure equivalence of item meaning (Brislin, Lonner, & Thorndike, 1973; Cavusgil & Das, 1997). More
specifically, each supervisor responded on a 7-point Likert scale for each item (item stems: Dependability, Planning, Know-how
and Judgment, Overall Present Performance, and Expected Future Performance; anchors, 1 = Unsatisfactory, 7 = Outstanding).
Responses to these five items were then summed to provide a measure of Performance. Each supervisor also provided assessments
of subordinate developmental level by completing a modified 10-item Employee Readiness Scale, that was previously developed
by Fernandez & Vecchio (1997) from an earlier (Blanchard, 1988) Employee Readiness Questionnaire. This scale combined items
that assessed both employee competence and commitment (sample items for competence: “Past job experience,” “Knowledge of
the subject area;” sample items for commitment: “Willingness to take responsibility,” “Positive work attitude;” anchors, 1 = Low,
7 = High).
Subordinates, in turn, provided assessments of their supervisor, along with descriptions of their own responses to work
experiences. The assessments provided by the subordinates included impressions of supervisor considerateness and structuring.
Leader consideration was measured on a three-item scale composed of items taken from the LBDQ-XII instrument (Stogdill & Coons,
1957). Sample item: “My supervisor's relations with me can be described as friendly and approachable,” anchors, 1 = Never,
2 = Seldom, 3 = Occasionally, 4 = Often, 5 = Always. Leader structuring was measured with five items taken from the LBDQ-XII,
using the same 5-point response scale for each item. Sample items: “My supervisor schedules for me the work to be done,” and “My
supervisor lets me know what is expected of me.” Leader–member exchange was measured with the 7-item version of the LMX
measure (Scandura & Graen, 1984). Sample item: “My supervisor has enough confidence in me that he/she would defend my
decisions if I were not present to do so,” anchors, 1 = Strongly Disagree, 2 = Disagree, 3 = Undecided, 4 = Agree, 5 = Strongly
Agree. LMX was included in the survey to provide an index of employee assessment of his/her working relationship with the
leader. Subordinate satisfaction with the supervisor was measured with the supervisor-satisfaction subscale of the JDI (Smith,
Kendall, & Hulin, 1969; Balzer, et al., 1997; Hanisch, 1992). Sample items: “Around when needed,” and “Knows job well,”
anchors = Yes, ?, and No. The JDI-satisfaction with leader subscale was intended to measure employee satisfaction with the leader,
per se. Organizational commitment was assessed with a 15-item Organizational Commitment Questionnaire developed by
Mowday, Steers, & Porter (1979). Although three facets of commitment were originally proposed, Mowday et al., (1982) maintain
that the questionnaire measures a single global construct. The present study also found (via factor analysis) that the scale was unidimensional. Sample item: “I am willing to put in a great deal of effort beyond that normally expected in order to help the
organization be successful,” anchors, 1 = Strongly Disagree, 2 = Disagree Somewhat, 3 = Slightly Disagree, 4 = Neither Agree nor
Disagree, 5 = Slightly Agree, 6 = Agree Somewhat, 7 = Strongly Agree. Organizational commitment was assessed from the
subordinate's perspective in order to provide an index of the validity of the leader's appraisal of follower developmental level.
Finally, each subordinate provided responses to a five-item measure of autonomy afforded by the leader, composed of Ohio State
Leadership items (Stogdill, 1963). Items: “My supervisor gives me the freedom to decide how to do my job,” “My supervisor trusts
me to use my judgment,” “My supervisor values my advice,” “My supervisor puts my suggestions into operation,” and “My
supervisor treats me as an equal,” anchors, 1 = Never, 2 = Seldom, 3 = Occasionally, 4 = Often, 5 = Always.
The six scales that were completed by the subordinates were subjected to confirmatory factor analyses (CFA) in order to
determine whether a six-factor solution better represented the data than a single, or general, affect factor. The CFA results for a sixfactor solution were: chi-squared = 2761.01, normed fit index (NFI) = .96, Tucker-Lewis index (TLI) = .97, comparative fit index
(CFI) = .98, and root mean square error of approximation = .06. In contrast, the CFA results for a one-factor solution were: chisquared = 4590.76, NFI = .93, TLI = .94, CFI = .95, rmsea = .08. A statistical test of the reduction in chi-square for the six- versus
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
841
one-factor solution yielded a significant result (Δ chi-square = 4590.76−2761.01 = 1829.75; with df difference = 1327−1311 =
16; p b .01). Hence, the subordinates' responses are better described as reflecting six factors.
For the three sets of responses provided by the leaders (follower performance, competence, and commitment), CFA was used to
contrast a three-factor solution with a single-factor solution. For the three-factor model, chi-square = 809.21, NFI = .96, TLI = .95,
CFI = .97 and rmsea = .15. For the single-factor model: chi-square = 949.04, NFI = .96, TLI = .95, CFI = .96 and rmsea = .16. The
resulting change in chi-square was significant (949.04−809.21 = 139.83, df difference = 91−87 = 4, p b .01). As each rmsea was
relatively high, we should also conclude that there was not a close fit overall, and that the models could be improved (Browne &
Cudeck, 1993). Therefore, the leaders' responses are better described as reflecting three factors. As the original version of SLT posits
that competence and commitment are compensatory attributes, the dimensions of competence and commitment were,
nonetheless, summed in order to create a single index of Employee Readiness/Maturity.
3. Results
Descriptive statistics and intercorrelations for the variables are provided in Table 1. Noteworthy in this display is evidence that
employee performance was significantly correlated with LMX (r = .27, p b .01) and leader consideration (r = .26, p b .01), but not
correlated with leader structuring (r = −.01, n.s.). Performance ratings were also significantly correlated with autonomy (r = .27,
p b .01). Employee demographics were not generally related to the other variables, with the notable exceptions that older
subordinates reported greater organizational commitment and autonomy (rs = .20 and .14, respectively, both p b .01) and female
subordinates reported being in their positions somewhat longer than male subordinates (r = .14, p b .01). Analyses controlling for
the demographic variables of education, experience, age, and gender (which had been identified by Vecchio and Boatwright, 2002,
as being of potential importance in explaining follower preferences for leader behavioral styles) did not alter the following results.
Follower self-reports of organizational commitment (which were obtained to provide corroboration of the Employee
Readiness/Maturity Index) were most notably correlated with Employee Readiness/Maturity and age (rs = .17 and .20, p b .01,
respectively), but less well correlated with actual job experience (r = .10, p b .10).
Ratings of follower competence and commitment were both correlated with ratings of performance (rs = .72 and .75, both
p b .01, respectively), and with each other (r = .86, p b .01). The present correlation of rated job performance with Employee
Readiness/Maturity (r = .87, p b .01) replicated earlier published findings of a substantial association. It should be noted that a high
correlation between the developmental level variable and the job performance criterion may make it more difficult to identify
support for SLT predictions, such that the obtainment of supportive results would be all the more noteworthy.
3.1. SLT-I
The original version of Situational Leadership Theory (SLT-I) predicts a three-way interaction among leader consideration,
leader structuring, and follower developmental level. To test this interaction, hierarchical regression was employed for the three
predictors, as well as for two-way interactions and a three-way interaction. Of special interest is the incremental contribution that
is made to criterion variance by the three-way interaction after controlling for main effects and lower-order interactions. Following
the procedures outlined by Aiken & West (1991), all predictors were mean-centered before creating two- and three-way
multiplicative interaction terms. Table 2 summarizes the results of these analyses for the dependent variables of employee
performance, LMX, and JDI-satisfaction with supervisor. As the results indicate in step 3 (where the three-way interaction term is
entered after all main effects and lower-order interactions), there was no evidence of a gain in criterion variance for any of the
dependent variables due to the inclusion of the three-way interaction.
Of course, it can be argued that hierarchical regression may not be the optimal approach for testing SLT-I as key regression
assumptions (e.g., heteroscedascity of errors) may not be reasonable (cf. Vecchio, 1987). As a further test of the theory's
predictions, cases were identified as representing low, moderate, or high levels of developmental level. Then, “matches” were
identified within each level by identifying proper combinations (i.e., low, moderate, or high) of leader consideration and
Table 1
Means, standard deviations, reliabilities, and intercorrelations.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Performance
Leader–member exchange
Consideration
Structuring
Satisfaction with super.
Organizational commitment
Readiness/Maturity
Age
Sex (0 = male, 1 = female)
Months with supervisor
Job experience (months)
Autonomy
M
SD
1
2
3
4
5
6
7
8
9
10
11
12
27.42
27.40
11.96
18.09
44.98
84.56
70.71
44.27
.56
42.56
110.23
20.57
4.51
3.88
1.99
3.43
7.79
11.39
10.84
9.91
.50
43.97
118.21
2.00
(.89)
.27⁎⁎
.26⁎⁎
−.01
.10+
.12⁎
.87⁎⁎
−.07
−.05
−.05
−.08
.27⁎⁎
(.87)
.69⁎⁎
.37⁎⁎
.62⁎⁎
.33⁎⁎
.31⁎⁎
−.02
−.09+
−.02
−.01
.58⁎⁎
(.73)
.35⁎⁎
.56⁎⁎
.22⁎⁎
.28⁎⁎
−.10⁎
−.05
−.11⁎
.03
.62⁎⁎
(.84)
.43⁎⁎
.26⁎⁎
−.01
−.03
.05
−.05
−.02
.17⁎⁎
(.79)
.29⁎⁎
.10+
−.01
−.05
−.11⁎
−.05
.41⁎⁎
(.88)
.17⁎⁎
.20⁎⁎
.01
.02
.10+
.32⁎⁎
(.92)
.02
−.09+
−.01
−.05
.29⁎⁎
(−)
.03
.23⁎⁎
.50⁎⁎
.14⁎⁎
(−)
.05
.11⁎
−.08
(−)
.34⁎⁎
−.03
(−)
.05
(.74)
N = 357; Cronbach alphas on primary diagonal;
+
p b .10; ⁎p b .05; ⁎⁎p b .01.
842
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
Table 2
Summary of hierarchical regression analyses: Test of SLT-I predicted three-way interaction.
Dependent variables
Performance
Predictors
Equation 1:
Consideration (C),
Structuring (S),
Maturity (M)
Equation 2:
C × S,
C × M,
S×M
Equation 3:
C × S × M interaction
+
Beta
.02
−.02
.86⁎⁎
LMX
R2
.75⁎⁎
−.01
−.02
.05+
.75⁎⁎
−.03⁎
.75⁎⁎
ΔR2
.75⁎⁎
Beta
.60⁎⁎
.15⁎⁎
.15⁎⁎
JDI
R2
.52⁎⁎
ΔR2
Beta
R2
ΔR2
.52⁎⁎
.48⁎⁎
.26⁎⁎
−.03
.38⁎⁎
.38⁎⁎
.40⁎⁎
.02⁎⁎
.41⁎⁎
.01+
.00
−.07⁎
.01
.06
.53⁎⁎
.01
−.07
−.14⁎⁎
.05
.00
.02
.53⁎⁎
.00
.08
+
p b .10; ⁎p b .05; ⁎⁎p b .01.
structuring. More specifically, three levels of development were created by trichotomizing leader assessments of follower levels of
development (as indexed by the Employee Readiness Questionnaire) into three groups at the 33rd and 67th percentiles.
Trichotomizing was utilized (rather than quartering, as would be suggested by graphic representations of SLT) in order to have an
adequate sample size for conducting the necessary statistical tests across the cells, as well as to avoid having empty cells. The cutoff values used for this scale were 67 and 77. For the leader behavior dimensions, consideration was dichotomized at the sample
median value of 12, and structuring was trichotomized at the values of 17 and 20. Dichotomizing at the median on structuring and
trichotomizing on consideration appear reasonable in light of the graphic portrayals of the theory commonly used in texts and
training programs (Vecchio, 1987). “Matches” (which are predicted to be superior combinations of developmental level,
consideration, and structuring) were then contrasted with “mismatches” across all developmental levels. The results of this
omnibus test of SLT-I are presented in Table 3. As indicated by the mean values, the means of the matched cases significantly
exceeded the means of the mismatched cases for the subordinate attitudinal dependent variables.2 Plus, the estimated effect sizes
(eta-squared) range from 1 to 4% of the variance in the criteria.
This overall, or omnibus, comparison of matches and mismatches, however, may mask unique pattern differences within each
development level. Therefore, further comparisons were made within each of the developmental levels. As shown in Table 4, the
observed pattern of means suggests that the significant omnibus results were largely driven by strong differences in the low and
moderate readiness/maturity groups. In the high developmental group, the results of tests for mean differences were nonsignificant. However, in 8 of the 9 mean comparisons, the mean differences were in the predicted direction (i.e., matches exceeded
mismatches). Also, estimates of effect sizes were greater in the low and moderate readiness/maturity groups. Despite the
directional support with the partitioned tests, it is difficult to argue that there is clear support for Hypothesis 1.
3.2. SLT-II
To test Hypothesis 2, it was necessary to identify the dimensional space specified for SLT-II's predictions by quartizing the
dimension of developmental level. As there is not a simple, direct point-to-point correspondence between the four different
developmental levels and the underlying dimensions of competence and commitment in SLT-II, it is critical to define each of the
four developmental levels in accord with the terms used by Blanchard, (1988; 2007), where developmental level 1 is the
combination of low competence and high commitment, developmental level 2 reflects some competence in combination with low
commitment, developmental level 3 is defined by moderate to high competence with variable commitment, and developmental
level 4 is the combination of high competence and high commitment. This system of classification differs substantially from the
earlier, SLT-I, system of defining categories of developmental level in that some cases are not potentially classifiable when they do
not fall into the specified subsets of the two-dimensional space mapped by the detailed combinations of competence and
commitment.3 Because of this lack of a straight-forward linear conversion of competence and commitment into developmental
level, a hierarchical regression approach is less attractive as an analytic technique (although it would be feasible to use categorical
dummy-coding). Instead, omnibus and partitioned tests provide for more direct comparison of results across conditions.
For the present data, development level 1 was defined as cases that were in the bottom third on competence (low) and in the
top third on commitment (high); developmental level 2 was defined as cases that were in the bottom half on competence (some
degree of competence) and in the bottom third on commitment (low); developmental level 3 was defined as cases that were above
the top third on competence (moderate to high) and below the top third on commitment (variable); and developmental level 4
2
To better convey the outcomes for the dependent variable, the means reported in the tables for these statistical tests are the average item-values (i.e., an
examination of mean item-values provides for easier comparison and interpretation via the anchor-labeled response metric).
3
Cases were omitted from the tests of SLT-II when (1) follower competence was between the values of 33 and 38 and follower commitment was greater than
39, or (2) follower competence was less than 33 and follower commitment was between 34 and 39. These cases, in essence, fell into portions of the potential twodimensional space for which SLT-II is silent.
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
843
Table 3
Results of omnibus test: Comparisons of “SLT-I specified” matched.
Group
M
Dependent variable = Performance
Match
5.61
Mismatch
5.43
Dependent variable = Leader–member exchange
Match
4.12
Mismatch
3.82
Dependent variable = JDI-satisfaction with supervisor
Match
2.62
Mismatch
2.44
SD
n
t
Estimated effect size
.68
.98
111
236
− 1.75+
.01
.40
.59
112
236
− 4.88⁎⁎
.02
.33
.46
112
236
−3.76⁎⁎
.04
Cases with mismatched cases. +p b .10; ⁎p b .05; ⁎⁎p b .01.
was defined as cases that were in the top third on both competence (high) and commitment (high). For a case that also had the
specified combination of leader consideration and leader structuring within its developmental level (based on splits on the sample
medians), the case was labeled a “match.” Other cases within the same developmental level were labeled “mismatches.” Following
this procedure, 225 of 357 cases were identifiable as falling within SLT-II's specified space (i.e., 132 cases did not occur in the
specified space, and no prediction can be offered by this revised model for such cases). Of the 225 identifiable or captured cases, 38
were matches and 187 were mismatches.
The results of an omnibus comparison for matches and mismatches (i.e., across all cases in the four developmental levels of
SLT-II) are presented in Table 5. As the test statistic values indicate, the mean values for the matches and mismatches did not
differ as had been predicted across the four dependent measures. In fact, the means for all of the mismatched cases were higher
than the means for the matched cases.
Partitioned tests were then conducted within each of the four developmental levels. That is to say, the matches and mismatches
within each developmental level were contrasted on the four dependent variables. As indicated in Table 6, the mean results were
only significant and in the predicted direction for the dependent measures of LMX and satisfaction with supervision in
Table 4
Results of partitioned tests: Comparisons within maturity.
Group
M
Low-readiness/Maturity group
Dependent variable = Performance
Match
4.81
Mismatch
4.50
Dependent variable = Leader–member exchange
Match
3.96
Mismatch
3.63
Dependent variable = JDI-satisfaction with supervisor
Match
2.60
Mismatch
2.40
Moderate readiness/Maturity group
Dependent variable = Performance
Match
5.65
Mismatch
5.53
Dependent variable = Leader–member exchange
Match
4.19
Mismatch
3.75
Dependent variable = JDI-satisfaction with supervisor
Match
2.66
Mismatch
2.36
High readiness/Maturity group
Dependent variable = Performance
Match
6.15
Mismatch
6.29
Dependent variable = Leader–member exchange
Match
4.13
Mismatch
4.06
Dependent variable = JDI-satisfaction with supervisor
Match
2.60
Mismatch
2.54
Levels specified by SLT-I.
+
p b .10; ⁎p b .05; ⁎⁎p b .01.
SD
n
t
Estimated effect size
.60
.81
27
86
− 1.83+
.03
.36
.55
26
85
− 2.93⁎⁎
.07
.35
.50
27
84
− 1.89+
.03
.56
.46
48
66
−1.19
.01
.39
.56
50
67
− 4.73⁎⁎
.16
.30
.48
49
67
− 3.81⁎⁎
.11
.35
.40
36
84
.42
.58
36
84
−.66
.00
.35
.40
36
85
−.76
.01
1.86+
.03
844
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
Table 5
Results of omnibus test: Comparisons of “SLT-II specified” matched cases with mismatched cases.
Group
Dependent variable = Performance
Match
Mismatch
Dependent variable = Leader–member exchange
Match
Mismatch
Dependent variable = JDI-satisfaction with supervisor
Match
Mismatch
+
M
SD
n
t
Estimated effect size
5.21
5.53
1.03
1.04
38
185
1.76+
.01
3.86
3.96
.53
.56
38
184
.95
.00
2.45
2.53
.45
.41
38
185
.99
.00
p b .10; ⁎p b .05; ⁎⁎p b .01.
developmental level 2. In all other comparisons (i.e., in 14 of the 16 mean comparisons), the means were either not significantly
different or, if they were, they differed in a direction opposite to that predicted by SLT-II. In general, these results are not supportive
of Hypothesis 2 (i.e., of SLT-II).
Table 6
Results of partitioned tests: Comparisons within maturity levels specified by SLT-II.
Group
Development level 1
Dependent variable = Performance
Match
Mismatch
Dependent variable = Leader–member exchange
Match
Mismatch
Dependent variable = JDI-satisfaction with supervisor
Match
Mismatch
Development level 2
Dependent variable = Performance
Match
Mismatch
Dependent variable = Leader–member exchange
Match
Mismatch
Dependent variable = JDI-satisfaction with supervisor
Match
Mismatch
Development level 3
Dependent variable = Performance
Match
Mismatch
Dependent variable = Leader–member exchange
Match
Mismatch
Dependent variable = JDI-satisfaction with supervisor
Match
Mismatch
Development level 4
Dependent variable = Performance
Match
Mismatch
Dependent variable = Leader–member exchange
Match
Mismatch
Dependent variable = JDI-satisfaction with supervisor
Match
Mismatch
+
p b .10; ⁎p b .05; ⁎⁎p b .01.
M
SD
n
t
Estimated effect size
5.47
5.70
1.27
.99
3
2
.22
.02
4.10
4.33
.30
.44
3
3
.78
.13
2.69
2.59
.14
.47
3
3
−.33
.03
4.44
4.38
.73
.80
19
66
−.28
.00
4.12
3.59
.37
.50
19
64
− 4.29⁎⁎
.19
2.74
2.38
.21
.46
19
64
− 3.33⁎⁎
.12
5.40
5.90
.40
.41
3
35
2.02⁎
.10
4.10
3.92
.08
.48
3
36
−.63
.01
2.31
2.46
.31
.47
3
36
.50
.01
6.22
6.29
.36
.39
13
82
.66
.01
3.38
4.25
.53
.45
13
81
6.36⁎⁎
.31
2.00
2.67
.42
.29
13
82
7.19⁎⁎
.36
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
845
Table 7
Summary of hierarchical regression analyses: Test of SLT-III predicted interaction for job experience and autonomy.
Dependent variables
Performance
Predictors
Equation 1:
Age
Gender,
Months with supervisor,
Job Level,
Education
Equation 2:
Experience (E),
Autonomy (A)
Equation 3:
E × A Interaction
+
Beta
LMX
R2
−.07
−.01
−.04
.13⁎
.09
.04+
−.02
.28⁎⁎
.01
ΔR2
Beta
JDI
R2
ΔR2
Beta
R2
ΔR2
.01
.01
.04+
−.02
−.07
.03
−.02
.04
.01
.01
.05
−.03
−.07
−.04
−.02
.11⁎⁎
.08⁎⁎
−.03
.58⁎⁎
.33⁎⁎
.32⁎⁎
−.04
.41⁎⁎
.17⁎⁎
.16⁎⁎
.11⁎⁎
.00
.09⁎
.34⁎⁎
.01⁎
.09
.18⁎⁎
.01
p b .10; ⁎p b .05; ⁎⁎p b .01.
3.3. SLT-III
Hypothesis 3 forecasted that followers who had more job experience (as indexed by months in their current position) would
respond more positively to leaders who provided greater autonomy. That is, the gradient of the relationships between autonomy
and the dependent variables should be steeper and more positive for more highly experienced followers. To formally test this
interaction prediction, the variables of job experience and autonomy were mean-centered prior to creating a multiplicative term to
represent the interaction in a hierarchical regression analysis. Because employee age, gender, time with the supervisor, job level,
and education were related to (a) job attitudes, (b) job experience, and (c) autonomy, these individual differences variables were
included as control variables in an initial step of the regression analysis.4 It should be noted, however, that the use of these control
variables (which are related to experience) reduces the shared variance in the relationships between job experience, autonomy,
and LMX.
Table 7 summarizes the results of these analyses. As the results indicate, the main effects for experience and autonomy added
significantly to accounting for variance in the dependent measures. However, evidence of a significant interaction for experience
and autonomy was only found for the dependent measure of leader–member exchange (ΔR2 = .001, p b .05). To graph the form of
this interaction in order to determine whether the interaction was of a predicted form, cases were selected that were in the highest
and lowest quartile on job experience (as the variable of job experience was not normally distributed). Fig. 1 portrays the resulting
bivariate associations of autonomy with the dependent variables of leader–member exchange for each level (high/low) of
employee job experience, respectively. Clearly, the gradients differ such that high job experience employees had a more positive
response to higher autonomy. This pattern for the gradients is supportive of the prediction offered by Hypothesis 3.
4. Discussion
The findings of the present study, in general, did not provide clear support for SLT, in any of its versions. The more statistically
sophisticated hierarchical regression analysis approach to testing SLT yielded no evidence of a three-way interaction (in terms of
both statistical significance and effect sizes). Moreover, although patterns of mean differences were (in the instance of omnibus
tests for SLT-1) suggestive of support for the theory, the partitioned tests did not clearly support SLT's prediction that high
readiness/maturity subordinates would benefit from a combination of low consideration and low structuring on the part of the
leader. When significant differences were identified in omnibus and partitioned tests, the effect sizes were generally meager. Also,
the results for tests of SLT-II (based necessarily on fewer cases because of SLT-II's silence for a portion of the conceptual space)
yielded even weaker results (in terms of the patterns of mean differences). The third version of SLT, that posited an interaction for
autonomy and experience, yielded a somewhat more supportive pattern of results. As the third version is a simplified and less
elegant re-statement of SLT, one could argue that the interaction is not especially close to SLT as it has been variously stated, and
that other principles of leadership (e.g., from LMX-theory) may as easily account for this result. For example, as job experience
increases, a subordinate will gain in job expertise, and both autonomy and LMX will likely increase at the same rate. Job experience
is correlated with expertise (and is likely a predictor of delegation). High LMX relationships, in turn, typically involve greater
delegation and empowerment. Arguably, the present results can also be said to replicate this alternative dynamic.
4.1. Practical implications
Without compelling empirical evidence of the validity of SLT's principles, it is difficult to endorse the use of the model in
leadership training programs. In comparison to other approaches to understanding leadership dynamics, SLT may offer an
4
Analyses of all leader and follower variables by “organizational membership” did not reveal evidence of significant differences across the ten organizations.
846
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
Fig. 1. Graph of experience by autonomy interaction for the dependent variable of leader–member exchange.
attractive framework for discussing interpersonal relations. Yet, it does not have sufficient empirical grounding in its original, 1972,
version or its more recent, 2007, revised statement to warrant advocating a close adherence to its prescriptive guidelines. In the
absence of more substantial research findings (i.e., beyond directional mean differences), those who instruct others within
leadership training programs should, as a matter of professional honesty, advise their trainees that SLT still lacks a strong empirical
grounding, and that its alluring character should not substitute for the absence of empirical substantiation. At best, there is some
evidence to support the original theory's prescriptions for dealing with newer employees with greater directiveness, and then
substituting directiveness with supportiveness as employee seniority increases.
The third version of SLT, which proposed that more experienced employees benefit from greater autonomy may be a point
worth emphasizing in leadership training programs. However, this principle, while in accord with SLT, is also in accord with other
views of leadership. The possibility that a principle can be identified within several theoretical streams is arguably a positive for the
principle, and implies that an important social dynamic is being identified. Yet, in light of the present and prior findings, leadership
training programs probably should not completely abandon discussing SLT. Perhaps SLT should be taught as a starting point in
discussing both leader–subordinate social interaction and the notion of modifying leader behavior to match subordinate attributes.
Clearly, the portion of SLT that has received the greatest support (i.e., low consideration and high structuring for newer recruits)
should be taught, and the broader notion of coupling greater autonomy with subordinate experience should be noted as well.
However, the theory's advocacy of low–low consideration/structuring leadership for high maturity subordinates should be
identified as being in error.
4.2. Strengths and limitations
Despite its inherent intuitive appeal, it has proven exceedingly difficult to verify the principles of SLT, even in its various
manifestations. Assuming for the moment that the principles are valid, obstacles to verifying SLT would seem to lie in the areas of
context, instrumentation, and/or research design. Concerning context, there has been substantial effort to test SLT in a sufficient
variety of organizational settings. Plus, a review of the studies that have tested the full model, as well as studies that have focused
on specific aspects of the model, suggests that organizational context is not likely a problem. Similarly, level of leadership does not
appear to be an issue, as past research has followed the model's recommended focus on settings where leader–subordinate
interaction is relatively more direct and continuous. High-level leaders, who one might label as more strategic or visionary in terms
of their potential impact, have not been studied relative to more immediate forms of direct supervision. So again, the settings of
prior research on SLT have been appropriate (i.e., researchers have been searching for the proposed dynamics in locations where
the principles are more likely to be manifest).
Instrumentation remains, comparatively, more of a concern as much of the published research has followed an approach that
collects follower and leader descriptions of each other. It is possible that these sets of respondents do not describe one another in
accord with the proposed principles, yet actually behave in accord with the proposed principles. This suggests that independent
observers of leader–subordinate social interaction may be better able to identify the suggested social dynamics. The different
components of the model also vary in terms of how well-developed their respective research streams have been in providing
construct valid instruments. The study of leader behaviors is fairly well-established as a field, such that reasonable confidence can
be placed in available instruments that focus on the dimensions of leader supportiveness and directiveness (Judge, Piccolo, & Ilies,
2004). The construct of follower developmental level, however, is far less well-established. Moreover, the definition of
developmental level has shifted over time from “ability and willingness” to “competence and commitment,” and from “maturity”
to “readiness.” While these labels appear highly similar in meaning, their precise operational definition remains somewhat elusive.
Also, the manner in which underlying attributes of a follower should be combined to define a precise developmental level has,
arguably, become somewhat less clear as the model has undergone revision (Graeff, 1983; 1997). This problem is further evident in
the present test of SLT-II where a number of cases fell outside of the model's definable construct space, thereby resulting in no
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
847
specific prediction for this set of followers (i.e., 37% of the respondents in the present study). Additionally, the number of
“matched” cases versus “mismatched” cases is a cause for further concern, as this suggests that cases that are presumptively ideal
are not commonplace. Of course, the low frequency of “matched” cases is not an indictment of SLT as it may merely be the reality
that few subordinates are being dealt with most appropriately by their supervisors. Nonetheless, the low frequency of “matches”
makes it difficult to conduct tests of the theory (Cairns, Hollenback, Preziosi, & Snow, 1998). In order to better address the construct
of follower developmental level, future research may need to focus on jobs which are more homogenous in character, such that
only a few tasks or only one task is involved. This would allow for greater certainty in appraising and comparing followers on
personal job competency. The assessment of follower commitment would seem, however, to rely necessarily on more subjective
reports (from either job incumbents, supervisors, or independent observers). Alternatively, developmental level can continue to be
indexed by job seniority. This provides a comparatively more objective metric for inferring developmental level that likely reflects
differences in both follower job competency and personal commitment.
Research design is a further likely factor that may be limiting the obtainment of evidence in support of SLT. To date, research on
SLT has been cross-sectional in nature. Such an approach seeks to capture a “snap-shot” of the proposed dynamics in that leader–
subordinate dyads should map into the various neighborhoods of the construct space if the theory's principles are valid. It is
possible, however, that the model's proposed dynamics occur quite rapidly, with followers gaining job competence fairly quickly.
This is a plausible scenario as employers typically seek to bring their new hires “up to speed” rather quickly. Therefore, the
commitment factor may largely be responsible for how people are appraised on performance and their resultant attitudes. Other
social forces also may impact follower performance and attitudes and thereby minimize the possibility of detecting the model's
proposed dynamics. These issues suggest that a longitudinal study of new hires may offer insights on whether SLT is valid. This
developmental or growth thesis of subordinate change has not received empirical study in the organizational literature, perhaps
because the theory is silent with respect to when change is likely to occur. These remain among the major open questions with
respect to SLT's validity (i.e., whether the performance and attitudes of new hires can be shown to follow the model's
developmental curve over time in response to leader–subordinate social interaction, and precisely when, and how, change will
occur).
Finally, there is the possibility that settings, instrumentation, and research design are not the true obstacles to validating SLT.
Instead, it may be that the theory is itself fundamentally misstated. Simply because intuitively appealing storylines can be offered
does not necessarily mean that they occur with much frequency. Moreover, the present results, which offer by contrast some
modest support for a more general statement of SLT's basic proposition (i.e., SLT-III), suggest that an overly precise, and perhaps
over-specified, statement of social dynamics, while attractive or satisfying at a purely conceptual level, may be more difficult to
document empirically than a conceptualization of SLT that is comparatively more generic in character. Further tests and extensions
of a more general statement of SLT's most basic interaction principle may ultimately enable the offering of prescriptive
recommendations to managerial leaders that are both valid and not too narrowly specified.
References
Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.
Balzer, W. K., Kihm, J. A., Smith, P. C., Irwin, J. L., Bachiochi, P. D., Robie, C., et al. (1997). Job Descriptive Index (JDI). Users' manual for the JDI Bowling Green, OH:
Bowling Green State University.
Blanchard, K. (1988). Situational leadership II. Blanchard training and development.
Blanchard, K. H. (2007). Leading at a higher level. Upper Saddle River, NJ: Prentice-Hall.
Blanchard, K. H., & Nelson, R. (1997). Recognition and reward. Executive Excellence, 14, 15.
Blanchard, K. H., Zigarmi, D., & Nelson, R. B. (1993). Situational leadership after 25 years: A retrospective. Journal of Leadership Studies, 1, 22–36.
Blanchard, K. H., Zigarmi, P., & Zigarmi, D. (1985). Leadership and the one minute manager. New York, NY: William Morrow.
Blank, W., Weitzel, R., & Green, S. G. (1990). A test of situational leadership theory. Personnel Psychology, 43, 579–597.
Brislin, R., Lonner, W., & Thorndike, R. (1973). Cross-cultural research methods. New York, NY: Wiley.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural models (pp. 136–162). Newbury Park,
CA: Sage.
Cairns, T. D., Hollenback, J., Preziosi, R. C., & Snow, W. A. (1998). Technical note: A study of Hersey and Blanchard's Situational Leadership Theory. Leadership and
Organizational Development Journal, 19, 113–116.
Cavusgil, S. T., & Das, A. (1997). Methodological issues in empirical cross-cultural research: A survey of the management literature and a framework. Management
International Review, 37, 71–104.
de Vries, R. E., Roe, R. A., & Tailleu, T. C. B. (1998). Need for supervision: Its impact on leadership effectiveness. Journal of Applied Behavioral Science, 34, 486–501.
Fernandez, C. F., & Vecchio, R. P. (1997). Situational leadership theory revisited: A test of an across-jobs perspective. Leadership Quarterly, 8, 67–84.
Goodson, J. R., McGee, G. W., & Cashman, J. F. (1989). Situational leadership theory: A test of leader prescriptions. Group and Organizational Studies, 14, 446–461.
Graeff, C. L. (1983). The situational leadership theory: A critical review. Academy of Management Review, 8, 285–291.
Graeff, C. L. (1997). Evolution of situational leadership theory—A critical review. Leadership Quarterly, 8, 153–170.
Hambleton, R. K., & Gumpert, R. (1982). The validity of Hersey and Blanchard's theory of leader effectiveness. Group and Organizational Studies, 7, 225–242.
Hanisch, K. A. (1992). The job descriptive index revisited: Questions about the question mark. Journal of Applied Psychology, 77, 377–382.
Hersey, P. (1985). A letter to the author of “Don't be misled by LEAD”. Journal of Applied Behavioral Science, 21, 152–153.
Hersey, P., & Blanchard, K. (1972). Management of organizational behavior, 2nd ed. Englewood Cliffs, NJ: Prentice-Hall.
Hersey, P., & Blanchard, K. (1982). Management of organizational behavior, 4th ed. Englewood Cliffs, NJ: Prentice Hall.
Judge, T., Piccolo, R. F., & Ilies, R. (2004). The forgotten ones? The validity of consideration and initiating structure in leadership research. Journal of Applied
Psychology, 89, 36–51.
Lawler, E. E. (1995). Creating high performance organizations: Practices and results of employee involvement and total quality management. San Francisco, CA: Jossey-Bass.
Liden, R. C., & Graen, R. (1980). Generalizability of the vertical dyad linkage model of leadership. Academy of Management Journal, 23, 451–465.
Lueder, D. C. (1985a). Don't be mislead by LEAD. Journal of Applied Behavioral Science, 21, 143–151.
Lueder, D. C. (1985b). A rejoinder to Dr. Hersey. Journal of Applied Behavioral Science, 21, 154.
Malone, T. W. (1997). Is empowerment just a fad? Control, decision making, and IT. Sloan Management Review, 38, 23–35.
Mowday, R. T., Porter, L., & Steers, R. (1982). Employee–organization linkages. New York, NY: Academic Press.
848
G. Thompson, R.P. Vecchio / The Leadership Quarterly 20 (2009) 837–848
Mowday, R. T., Steers, R., & Porter, L. (1979). The measurement of organizational commitment. Journal of Vocational Behavior, 14, 224–247.
Norris, W. R., & Vecchio, R. P. (1992). Situational leadership theory: A replication. Group and Organizational Management, 17, 331–342.
Northouse, P. G. (2007). Leadership: Theory and practice. Thousand Oaks, CA: Sage Publications.
Scandura, T. A., & Graen, G. B. (1984). Moderating effects of initial leader–member exchange status on the effects of a leadership intervention. Journal of Applied
Psychology, 69, 428–436.
Siebert, S. E., Silver, S. R., & Randolph, W. A. (2004). Taking empowerment to the next level: A multiple-level method of empowerment, performance, and
satisfaction. Academy of Management Journal, 47, 332–349.
Smith, P. C., Kendall, L. M., & Hulin, C. L. (1969). The measurement of satisfaction in work and retirement. Chicago, IL: Rand-McNally.
Stogdill, R., & Coons, A. (1957). Leader behavior: Its description and measurement (Research Monograph No. 88). Columbus, OH: Ohio State University, Bureau of
Business Research.
Stogdill, R. M. (1963). Manual for the Leader Behavior Description Questionnaire—Form XII. Columbus, OH: Ohio State University.
Vecchio, R. P. (1987). Situational leadership theory: An examination of a prescriptive theory. Journal of Applied Psychology, 72, 444–451.
Vecchio, R. P., & Boatwright, K. J. (2002). Preferences for idealized style and supervision. Leadership Quarterly, 13, 327–342.
Vecchio, R. P., Bullis, R. G., & Brazil, D. M. (2006). The utility of situational leadership theory: A replication in a military setting. Small Group Research, 37, 407–424.
Yukl, G. (2006). Leadership in organizations. Upper Saddle River, NJ: Prentice-Hall.
Download