Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
CONTENT, STRUCTURE,
AND GENERALIZABILITY
Lynn R. Offermann*
George Washington
University
Decision
Systems
Philip W. Wirtz
George Washington
University
Implicit theories of leadership (ILTs) have received increasing attention in recent years
as a means of understanding
leader attributions
and perceptions (e.g., Lord, Foti, &
DeVader, 1984) as well as a source of error in the measurement
of leader behavior
*
of Psychology,
George
Washington
44
LEADERSHIP
QUARTERLY
45
determining perceptions of effective leaders, and whether they differ from leaders,
would also be instructive.
Separate examination
of ILTs by stimulus targets such as leader, effective leader,
and supervisor, can answer two questions, one dealing with the issue of factor stability
across stimulus objects and the other with differences in levels of rating within a
particular stimulus object across different groups of perceivers. The first question
addresses the commonality
of characteristics used to describe leaders, effective leaders,
and supervisors. Although researchers have typically made the tacit assumption
that
factor structure does not vary significantly as a function of stimulus target (as evidenced
by a common willingness to transpose the terms leader and supervisor), there are reasons
to question this assumption
(Schneider,
1973). Correlations
among traits have been
shown to differ both as a function of stimulus person and relevance of trait (Koltuv,
1962). In the case of leadership stimuli, it is possible, for example, that raters see
characteristics
of sensitivity
and dedication
as descriptive
of both leaders and
supervisors, but positively related in leaders and negatively related in supervisors (or
vice versa).
The questions of factor structure differences across raters highlights the importance
of considering perceiver characteristics as well as stimulus characteristics in ILTs. Bieri
et al. (1966) and Goldstein and Blackman (1978) argue that there are systematic,
measurable differences in peoples implicit personality theories. Borman (1987) makes
a similar claim about implicit theories of subordinate
performance.
Peoples implicit
theories do not simply appear, fully formed, out of nowhere. Rather, they are generated
and refined over time as a result of peoples experiences
with actual leaders or
descriptions of leaders. Given that people have different exposure to and experiences
with leaders, there may be interesting and important differences in their implicit theories.
Because sex is a variable that may be reflective of differences in socialization history
and experiences
with leaders, it is possible that men and women structure their
perceptions of leaders differently. For example, Ayman (1993) suggests that respondent
sex may affect expectations
of leaders. Yet, previous research studies with ILTs have
typically collapsed male and female responses under untested assumptions of similarity.
We propose to test the assumption
of similar factor structures for male and female
perceivers.
In addition to understanding
the differences between perceptions of leaders, effective
leaders, and supervisors, information
on the content of ILTs may advance knowledge
in two ways. First, the content of ILTs may help us to better understand and ultimately
predict their effect on ratings of leader behavior. Second, and most importantly,
leadership researchers
may find that certain aspects of leadership are commonly
understood or inferred (as indicated by their presence in ILTs) that are not taken into
account in current theories and models of leadership. The study of implicit theories
can provide clues that will help in the development
of explicit theories to understand
the phenomenon
called leadership.
Sternberg (1985) has employed just such an approach to differentiate the concepts
of intelligence, creativity, and wisdom. His results indicate that people have well-defined
implicit theories of these three constructs and that they use them in their ratings of
both themselves and others. Furthermore,
Sternbergs results illustrate both the overlaps
and omissions between explicit theory and practice and laypersons conceptions.
For
46
LEADERSHIP
QUARTERLY
47
METHOD
AND RESULTS
The current research was carried out in five stages. Using undergraduate
samples, the
first stage consisted of generating a pool of items for our test instrument;
the second
attempted to identify the underlying factor structure of implicit theories of leadership;
the third focused on verifying the item content validity of the resulting factors; and
the fourth tested the research hypotheses. As a test of generalizability,
the fifth stage
involved the administration
of the instrument developed to a sample of working adults
for further validation.
Item Generation
Sample
Respondents
(n = 115) were provided with a sheet of paper with instructions
and
25 blank lines and were asked to list up to 25 traits or characteristics
of a leader. No
definition of the term leader was provided in the instructions; subjects with questions
on this issue were instructed to use whatever definition was meaningful to them. Using
the same procedure, 77 participants listed traits associated with the term supervisor.
Results
Working first with the responses to the term leader, a total of 455 unique items
were obtained. The frequency with which each item was nominated was then computed.
Those items that were clearly behaviors (e.g., gives orders), and those items that were
only nominated
once or twice (e.g., religious) were deleted. Several items which were
clearly synonyms (e.g., funny-humorous)
were combined in order to reduce the number
of items to a reasonable size. On this basis, a pool of 160 items was retained.
Comparison
of our resulting items with Lord, Foti, and DeVaders (1984) attributes
generated to basic-level leader categories showed that over half of their 59 items were
represented exactly in our 160. As would be expected, these items had high family
resemblance scores (thus represented multiple categories of basic-level leaders). Some
48
LEADERSHIP
QUARTERLY
of Lords attributes were represented by synonyms or multiple items on our scale (e.g.,
Lords verbal skills could be represented in our items by articulate, good
communicator, and eloquent). Items found by Lord but unrepresented in our item pool
were those with low family resemblance scores, indicating relevance to particular basiclevel leader domains rather than multiple leader categories (e.g., minority and patriotic).
This absence in our item pool would be expected given the items probable relevance
to certain types of leaders (i.e., political, minority, or military leaders) rather than leaders
in general.
Responses to the term supervisor were then compared to those of leader. Since
the superordinate term 1eadershould theoretically contain more items than the basiclevel term supervisor, our particular concern was to check on whether there were
items generated by the supervisor cue that were not reflected in the 160 items selected
for leader. The most typical supervisor responses were well reflected in the pool
of 160 items, either as elicited or by clear synonyms or by their opposite, as in patient
(generated to leader) and impatient (generated to supervisor?. There was one notable
exception-30 respondents generated the term leader to the supervisorcue, whereas
respondents to leader naturally did not repeat the cue as a response. In order to use
the same instrument for the different stimulus categories, which included leader, the
term leader was not used as an item in the final questionnaire, Thus, the 160 items
originally generated were retained. Based on our data from terms elicited by
supervisor, we believe that these 160 items included perceptions of supervisors as well
as leaders.
Factor Identification
The subjects for this phase were 763 undergraduate volunteers enrolled in
introductory psychology courses at the same two universities. Subjects received course
credit for their participation.
The 160 trait items were combined in random order into a single measure with lopoint response scales indicating the extent to which each trait was considered
characteristic of a stimulus person, where 1 was not at all characteristic and 10 was
extremely characteristic. Using this scale, subjects were asked to indicate how
characteristic each of the trait items was for one of three different stimulus persons:
either a leader, an effective leader, or a supervisor. Again, no explicit definitions of
the terms were provided. Respondents were also asked to indicate their sex. Subjects
were run in large groups of approximately 100 and were randomly assigned to one
of the stimulus person conditions. They were allowed to complete the ratings at their
own pace. Subjects who left more than 10 of the 160 items blank were omitted from
the analysis. Means were substituted for any blank items for the remaining subjects.
Complete scale data were obtained from 686 subjects (267 males, 379 females, 40 of
unknown sex).
49
Table 1
Factor Names, Sample Items, and Reliabilities
Leadership Theory Factors
Factor Names/
Number of Items
Sample Items
Sensitivity
Sympathetic,
(10)
Dedication
Tyranny
(10)
(10)
Charisma
Domineering,
(10)
Attractiveness
Dedicated,
Charismatic,
(5)
Attractive,
Reliability
sensitive, compassionate,
disciplined,
inspiring,
Male, masculine
Intelligence
(6)
Intelligent,
(4)
Strong,
pushy, manipulative
.90
involved,
dynamic
26
.78
tall
.88
clever, knowledgeable,
forceful,
.94
.90
classy, well-dressed,
(2)
understanding
hard-working
prepared,
power-hungry,
Masculinity
Strength
of Implicit
bold, powerful
wise
.85
.74
Results
In order to capture as much of the total variation in the original set of 160 items
as possible, subject ratings were submitted to principal components
factor analysis
rotated to a varimax solution. Initial factor analyses were performed separately for each
of the three stimulus objects. A parallel analysis was run combining the data from the
three targets into a single merged file. Since examination
of the three separate and single
combined solution revealed no differences in high-loading items across stimulus type,
results of the combined analysis are subsequently
reported below. Six interpretable
factors with eigenvalues greater than 1.00 were obtained and were identified as indicating
dimensions
measuring Sensitivity, Dedication,
Tyranny, Charisma, Intelligence,
and
Attractiveness.
Exploratory
analyses of the data conducted separately for male and
female respondents suggested the presence of an additional meaningful factor measuring
Strength and that the Attractiveness
factor should be split into two separate factors
(one consisting of items related to gender and the other consisting of items more clearly
reflecting
attractiveness).
In order not to prematurely
eliminate
any potential
dimensions of peoples implicit theories of leadership, the eight factors were tentatively
retained for further analysis.
The first four factors were composed of a large number of items (20 or more items
with factor loadings greater than .40). In these instances, the 10 items with the highest
loadings on the factor were considered to comprise the factor. For the smaller factors,
all items with loadings greater than .40 were used to name the factor. Using these criteria,
57 items representing eight factors were retained from the initial 160. Table 1 presents
the names, number of items, some representative
items, and the coefficient alphas for
each of the eight factors.
50
LEADERSHIP
QUARTERLY
The 44 subjects (21 males, 23 females) for this phase of the study were introductory
psychology students. All subjects were native English speakers and received course credit
for their participation.
Procedure
Descriptive definitions of the eight implicit leadership theory factors identified in the
previous section were derived by the consensus of six subject matter experts. Each of
the 57 trait items comprising the factors was put on an individual card. Subjects were
run in groups of four but completed the task individually. The task consisted of sorting
each of the 57 trait cards into the factor definition considered most appropriate.
No
definitions of the trait adjectives were provided. Subjects completed the task at their
own pace. After all of the items had been assigned to factors, subjects were given an
opportunity
to look back over their work and make any changes that seemed
appropriate. The subjects were then debriefed as to the purpose of the study.
Results
The data were analyzed using a modified version of the procedure employed by
Hinkin and Schriesheim (1986). Data were aggregated to indicate the relative frequency
with which each trait item was assigned to each of the factor definitions. An item was
considered to be a useful representation
of a particular dimension if 70% of the subjects
assigned it to the factor it was intended to represent. Forty-one of the 57 items exceeded
this criterion. Because the trait-sorting task involved multiple judgments across multiple
raters, a measure of the extent to which raters agreed that certain groups of traits
belonged within each of the eight dimensions was indicated (Tinsley & Weiss, 1975).
Fleisss (1971) modified version of Cohens Kappa was employed. This coefficient
indicates the probability,
beyond chance, that a second, independent
rater will assign
an item to the same category as the initial rater. Kappa was computed for each of the
eight dimensions.
In general, agreement among raters that trait items belonged in
the a priori dimensions was strong. Agreement was greatest for the Sensitivity factor
and lowest for the Strength factor. However, intra-category agreement was well beyond
chance for all eight factors. No significant differences between female and male raters
in intra-category
agreement were found. On this basis, 41 items comprising eight factors
were retained for hypothesis testing (see Figure 1). Table 2 presents the intercorrelations
of the eight revised factors in the sample from which they were derived. Whereas some
of the factors were found to be moderately intercorrelated,
the data from the sorting
task indicate that individuals can, and do, make clear conceptual distinctions among
these dimensions.
Tests of the Research Hypotheses
likelihood
Factor Structure of Implicit Leadership Theory Model Derived from Entire Sample
Figure 1.
Note:
LEADERSHIP
52
QUARTERLY
Table 2
Intercorrelations, Means, and Standard Deviations
of Implicit Leadership Theory Factors
Facfors
I. Sensitivity
2. Dedication
3. Tyranny
4. Charisma
5. attractiveness
6. Masculinity
7. Intelligence
8. Strength
(.93)
.47
-.38
.53
.I6
-23
.52
.I9
(.8U
c.911
.oQ
32
.48
-.07
.38
C%
.33
-.05
.58
.48
f.75)
.35
.31
.33
f.88)
-.04
.20
(.W
.42
(.6@
4.58
1.77
7.58
1.35
5.15
1.84
3.65
2.61
7.73
1.36
6.94
i.80
-.OI
.62
.21
-.07
.58
.3s
Mean
6.83
8.16
S.D.
1.69
1.22
reliibilities
Table 3
Summary of Models
Chfsqccare
Model
4f
MO: Null
2654.8
751
MI,:
3587.1
1502
3661.0
1571
4544.7
2253
4706.2
2391
.OoOl.
original 686 subjects. Given that the original factor structure was derived from 160 items,
it was deemed prudent to test for possible differences in factor structures using the
reduced number of variables (Joreskog, 1971; Joreskog & Sorbom, 1979, 1980). The
formal statistical test for invariant factor structures compares the chi-square value
associated with an unconstrained model to the chi-square value associated with a
nested (i.e., more constrained) model in which corresponding factor loadings are
required to be identical across the groupings. Under the null hypothesis of invariant
factor structures across groups, this difference in chi-square values would, itself, be
expected to follow a chi-square distribution with an expected value equal to the
difference in the degrees of freedom of the constrained versus unconstrained models.
As Tables 3 and 4 show, significant differences in factor structure across groups were
not found, either across sex or across stimuIus groups. The comparatively high factor
loadings reflected in Figure 1 (which reflects the factor structure of the constrained
stimulus condition model) supports the tenability of the proposed model.
Given the invariance of factor structures across stimulus condition, a test of the
hypothesis of no level differences by stimulus condition and rater sex was conducted.
53
Table 4
Summary of Model Comparisons
Model Comparison
MO versus
MO versus
MI, versus
ML, versus
Now
Chi-square
df
932.3
1889.9
73.9
161.5
751
1502
69
138
<.Ol
<.Ol
.33
.08
MI,
M20
M,b
Mzb
of the models.
Table 5
Mean Factor Ratings of Leaders, Effective Leaders, and Supervisors
Stimulus Condition
Factor Name
Leaders
Effective Leaders
Sensitivity**
Dedication**
Tyranny*
Charisma**
Attractiveness
Masculinity
Intelligence**
Strength**
7.05
8.35
4.78
7.96
5.11
3.58
7.97
7.51
7.26
8.31
4.15
7.79
5.03
3.50
7.95
7.07
Nom:
Supervisors
6.22
7.8 1
4.69
6.99
5.25
3.79
7.27
6.28
p<.Ol.
** p < .ooOl
54
LEADERSHIP
QUARTERLY
Sample
Individual volunteers were solicited in public waiting areas of a large airport. Only
those indicating
full-time employment
were asked to participate.
There were 260
participants:
177 males and 83 females. Their mean age was 39.3 years, and they had
been employed full-time an average of 16.2 years. Individuals were employed in a range
of categories, notably professional (39%), managerial (300/o), sales (8%), and technical
(6%).
Procedure
A confirmatory factor analysis testing the hypothesis that each of the 41 items would
be significantly related to the same factor for working adults as was previously found
for undergraduates
was performed. No significant differences between samples were
found.
DISCUSSION
The present findings show eight primary dimensions
of peoples implicit theories of
leadership: Sensitivity, Dedication,
Tyranny, Charisma, Attractiveness,
Masculinity,
Intelligence, and Strength. Of these eight, those traits deemed most characteristic of
leaders, effective leaders, and supervisors (viz., Dedication, Charisma, Intelligence, and
Sensitivity) are all typically seen as positive attributes. Thus, these data support the
view that people generally view leaders, effective leaders, and supervisors in a positive
fashion and hold them up to high standards. However, our data also show that people
acknowledge
the possibility that individuals
in leadership positions may use their
positions to dominate,
control, and manipulate,
as well as inspire, motivate, and
support.
It was suggested that examining ILTs separately by nature of leader stimulus targets
and perceiver could yield information
about both factor stability across targets and
perceiver categories, and information
about differences in the level of ratings within
stimulus target. In terms of factor stability, the factor structure of implicit theories about
leaders did not differ across stimulus targets of leaders, effective leaders, and supervisors,
indicating that people use similar dimensions
for their perceptions
of these three
leadership targets. Likewise, ILT factors were similar for men and women perceivers,
and for student and worker groups. Since worker groups did not participate in item
generation,
it is possible that worker ILTs contain additional
unique elements
unrepresented
in our student-generated
item sample. This would not, however, alter
our findings of similar worker endorsement
of the student-generated
items, but rather
indicate that future research might be undertaken
to determine whether increased
exposure to target stimuli (as in a workers possibly greater exposure to multiple
organizational
superiors) significantly affects the richness of ILTs. The present research
55
indicates
that factors
used in considering
leader targets can show considerable
generalizability.
This generalizability
provides support for a view of ILTs as collectively
held expectations that are widely shared.
However, differences in level of rating between raters using these same factors were
found. As predicted, respondents perceived the three stimulus targets as possessing these
attributes to different extents. The largest difference was between supervisors and the
other two stimuli (leaders and effective leaders) with supervisors viewed as possessing
less of those characteristics typically considered most favorable, including Sensitivity,
Dedication, Charisma, Intelligence, and Strength. These results may contribute to the
ongoing debate among psychologists (e.g., Zaleznik, 1977) by suggesting that people
do not differ in the categories they use to think about leaders and supervisors but that
they do differentiate between supervisors and leaders in terms of prototypicality
ratings.
Expectations appear generally lower for supervisors than for leaders, with the exception
of the characteristic
represented by the Tyranny factor. Leaders and supervisors are
both seen as more likely to exhibit tyrannical tendencies than are effective leaders. These
data do not support an overall differentiation
in the perception of leaders versus effective
leaders, with leaders and effective leaders rated similarly on six of the eight factors.
It is possible that the cue of 1eadernaturally
calls forth the image of an effective leader.
Only when the leader cue includes a modifier presenting other information (e.g., political
leader) does the addition of the effectiveness cue appear to augment ratings (Foti et
al., 1982).
One goal of the present research was to determine how well explicit theories of
leadership reflect the salient issues raised in peoples ILTs, or whether some aspects
of leadership as evidenced by ILTs are neglected by existing explicit leadership theories.
Happily, a number of our factors are represented in present theory and research. Our
Sensitivity factor is similar to longstanding
concerns with leader consideration
(e.g.,
Fleishman,
1973). Charisma has a long history of study, and is currently enjoying a
resurgence of interest both within and apart from its key role in transformational
leadership (e.g., Bass, 1985). Physical characteristics summarized in our Attractiveness
factor were heavily researched by early theorists, with typically low positive relationships
found between leadership and attractiveness (see Bass, 1981, for a review). Intelligence,
although often found to be only weakly related to leader performance (e.g., Stogdill,
1974), or moderated by such variables as interpersonal
stress (e.g., Fiedler & Garcia,
1987) is nonetheless
strongly related to leader expectations
in the present study.
Research in the area of masculinity/femininity
and leadership continues to generate
interest, particularly
in regard to leader emergence. A recent study showed that
regardless of their sex, group members with masculine characteristics emerged as leaders
significantly
more than those with feminine,
androgynous,
or undifferentiated
characteristics,
and emergent
leaders received significantly
higher ratings
on
interpersonal
attractiveness (Goktepe & Schneier, 1989).
However, the key role of dedication, our most strongly endorsed factor, is not as
thoroughly researched. The importance of the hard work, diligence, preparation,
and
commitment
by leaders that is so strongly expected by others needs to be further
integrated into models of leadership. This is not the same as the well-researched areas
of task-orientation
(e.g., Fiedler, 1967) or initiating structure (e.g., Fleishman,
1973)
which reflect more of a concern with valuing task over people or getting the work done
56
LEADERSHIP
QUARTERLY
through organizing or directing the work of others, respectively. For our participants,
dedication appears to be more centered within the leader, reflecting the leaders own
work ethic and motivation.
In this respect, our findings are consistent with Boyatzis
(1982) concept of efficiency orientation for effective leaders.
Strength and Tyranny also appear to be conceptually distinct and less well-integrated
into leadership theory. Tyranny, and perhaps Strength as well, may relate to feelings
of abuse of power, suggesting that a closer integration of models of power and leadership
is needed (Hollander & Offermann,
1990). Current work on the so-called dark side
of charismatic leadership (Hogan, Raskin, & Fazzini, 1990) may also be tapping the
same type of concerns we see in our Tyranny factor. Increased research attention is
needed to better understand this fearful view of leadership in the way it is commonly
understood by perceivers.
But perhaps the most important
research consideration
needs to be given to the
integration
of these different factors into a comprehensive
view of leadership rather
than as separate areas of study. Not only do we know less about some of these factors
than desirable, but we also have not considered them in conjunction
with all that we
do know about leadership. Lay perceivers may be the key to such integration, as these
various components of their implicit theories exist simultaneously.
These data support the utility of considering implicit theories of leadership for what
their content and structure can tell us, rather than treating such theories merely as
sources of rating error. The ultimate importance of ILTs may lie not only in how they
can bias our questionnaire
measures of leadership but also in the way in which they
ILTs are undoubtedly
reflected in the
structure the leader/follower
interaction.
expectations
that followers bring to the leader/follower
relationship.
Subsequent
research is needed to understand
how these expectations
affect the course of the
developing leader/follower
exchange. For example, one key area of future application
would be in the area of effective diversity management.
Different cultural groups may
have different conceptions of what leadership should be (Bass, 1990; Hofstede, 1993).
The English word leader does not even translate directly into languages such as
French, Spanish, or German, where available words such as le meneur, el jefe, or de
Fuhrer tend to connote only leadership that is directive (Graumann,
1986). The ILTs
of individuals from such backgrounds might differ dramatically from the United Statesfollowers
who expect
leader
based findings
presented
here. For example,
authoritarianism
may view leader sensitivity as a sign of weak leadership and evaluate
such behavior negatively. Understanding
ILTs might be useful in linking different
follower expectations with responses to leader behavior.
The present research has shown that differences in implicit theories across leadership
targets and across raters can be systematically examined. Using a different methodology
from previous research, the results are nonetheless consistent with Lords leadership
categorization
theory, and further serve to highlight areas where leadership theory,
research, and practice needs to integrate important areas of leader perceptions.
Acknowledgments:
Thanks are due to Elaine Belansky and Audrey Goldman for
collecting the working sample data, to Don Gallo for conducting the content validation,
and to Larry James, Bob Lord, and George Rebok for comments on an earlier draft
of the manuscript.
57
NOTE
1. Similar results might be expected with the use of the word manager, when that term
clearly connotes the immediate superior. Because the term manager may be used by individuals
to identify superiors at several different hierarchical levels above them, whereas supervisor more
clearly refers to the immediate superior, we chose to use the term supervisor in this study to
provide greater clarity about the target.
REFERENCES
Ayman, R. (1993). Leadership perception: The role of gender and culture. In M.M. Chemers
& R. Ayman (Eds.), Leadership theory and research: Perspectives and directions. New
York: Academic.
Bass, B.M. (1985). Leadership and performance beyond expectations. New York: Free Press.
Bass, B.M. (1981). Stogdillk handbook of leadership. New York: Free Press.
Bieri, J., Atkins, A.L., Briar, S., Leaman, R.H., Miller, H., & Tripodi, T. (1966). Clinical and
social judgement. New York: John Wiley & Sons.
Borman, W.C. (1987). Personal constructs,
performance
schemata, and folk theories of
subordinate effectiveness: Explorations in an army officer sample. Organizational Behavior
and Human Decision Processes, 40. 307-322.
Boyatzis, R.E. (1982). The competent manager. New York: John Wiley.
Eden, D., & Leviatan, U. (1975). Implicit leadership theories as a determinant
of the factor
structure underlying supervisory behavior scales. Journal of Applied Psychology, 60, 736741.
Fiedler, F.E. (1967). A theory of leadership effectiveness. New York: McGraw-Hill.
Fiedler, F.E., & Garcia, J.E. (1987). New approaches to leadership: Cognitive resources and
organizational performance. New York: Wiley.
Fleishman, E.A. (1973). Twenty years of consideration and initiating structure. In E.A. Fleishman
and J.G. Hunt (Eds.), Current developments in the study of leadership (pp. l-37).
Carbondale: Southern Illinois University Press.
Fleiss, J.L. (1971). Measuring nominal scale agreement among many raters. Psychological
Bulletin, 76, 378-382.
Foti, R.J., Fraser, S.L., & Lord, R.G. (1982). Effects of leadership labels and prototypes on
perceptions of political leaders. Journal of Applied Psychology, 67, 326-333.
Goldstein, K.M., & Blackman, S. (1978). Cognitive style. New York: John Wiley & Sons.
Gioia, D.A., & Sims, H.P., Jr. (1985). On avoiding the influence of implicit leadership theories
in leader behavior descriptions. Educational and Psychological Measurement, 45.217-232.
Goktepe, J.R., & Schneier, C.E. (1989). Role of sex, gender roles, and attraction in predicting
emergent leaders. Journal of Applied Psychology, 74, 165-167.
Graumann, C.F. (1986). Changing conceptions of leadership: An introduction. In C.F. Graumann
& F. Moscovici (Eds.), Changing conceptions of leadership. New York: Springer-Verlag.
Hinkin, T.R., & Schriesheim, C.A. (1986, August). Influence tactics used by subordinates:
A
theoretical and empirical analysis and refinement of the Kipnis, Schmidt, and Wilkinson
subscales. Paper presented at the Annual Conference of the Academy of Management,
Chicago, Illinois.
Hofstede, G. (1993). Cultural constraints in management theories. 7he Executive, 7, 81-94.
Hogan, R., Raskin, R., & Fazzini, D. (1990). The dark side of charisma. In K.E. Clark & M.B.
Clark (Eds.), Measures of leadership (pp. 343-354). West Orange, NJ: Leadership Library
of America.
LEADERSHIP QUARTERLY
58
Hollander, E.P., & Offermann, L.R. (1990). Power and leadership in organizations: Realtionships
in transition. American Psychologist, 4.5, 179-189.
Joreskog, K.G. (1971). Simultaneous factor analysis in several populations. Psychometrika,
36,
409-426.
K.G., & Sorbom, D. (1979). Advances in factor analysis and structural equation
Cambridge, MA: Abt Associates.
Joreskog, K.G., & Sorbom, D. (1980). LZSREL-IV users guide. Chicago: National Educational
Resources.
Klatzky, R.L. (1980). Human memory: Structure andprocesses.
San Francisco: Freeman.
Koltuv, B.B. (1962). Some characteristics
of intrajudge train intercorrelations.
Psychological
Monographs,
76(33, whole no. 552).
Lord, R.G., Foti, R.J., & DeVader, CL. (1984). A test of leadership categorization
theory:
Internal structure, information
processing, and leadership perceptions.
Organizational
Joreskog,
models.
Behavior
and Human
Performance,
34, 343-378.
Lord, R.G., Foti, R.J., & Phillips, J.S. (1982). A theory of leadership categorization.
In J.G.
Hunt, U. Sekaran, & C.A. Schriesheim (Eds.), Leadership: Beyond establishment views.
Carbondale: Southern Illinois University Press.
Mervis, C.B., Catlin, J., & Rosch, E. (1976). Relationships among goodness-of-example,
category
norms, and word frequency. Bulletin of the Psychonomic Society, 7, 283-284.
Mervis, C.B., & Rosch, E. (1981). Categorization
of natural objects. Annual Review of
Psychology,
32, 89-
115.
Phillips, J.S., & Lord, R.G. (1982). Notes on the practical and theoretical consequences of implicit
leadership theories for the future of leadership measurement. Journal of Management,
12,
31-41.
Sternberg,
Personality
theories
of intelligence,
creativity,
and wisdom.
Journal
of
49, 607-627.
Sternberg, R.J. (1988). The triarchic mind: A new theory of human intelligence. New York:
Viking.
Tinsley, H.E.A., & Weiss, D.J. (1975). Interrater
reliability and agreement of subjective
judgments. Journal of Counseling Psychology, 22, 358-376.
Zaleznik, A. (1977). Managers and leaders: Are they different? Harvard Business Review, 5.5,
67-78.