PERSONNEL PSYCHOLOGY

1974, 27, 555-568.

CONSIDERATION, INITIATING STRUCTURE, AND

ORGANIZATIONAL CRITERIA—AN UPDATE
OF KORMAN'S 1966 REVIEW'

STEVEN KERR2 AND CHESTER SCHRIESHEIM

The Ohio State University

THE Ohio State leadership studies have been an important part

of the organizational behavior-industrial psychology literature, and
literally hundreds of studies have examined effects of the variables
isolated by the Ohio State researchers upon satisfaction, morale,
and performance criteria. Particular attention has been paid to
the leader behavior dimensions Consideration and Initiating Struc-
ture, usually measured by the Leader Behavior Description Ques-
tionnaire (LBDQ), the Leadership Opinion Questionnaire (LOQ),
or the Supervisory Behavior Description Questionnaire (SBDQ).
Probably the most important critique of the Consideration-
Initiating Structure literature to date is the one by Korman (1966),
in which he discussed five major shortcomings of the existing .re-
search, and concluded that future efforts would have to differ in
several important respects for useful applications to result.
The purposes of this article are to (1) summarize the criticisms
and recommendations of the Korman (1966) paper; and (2) review
those published and unpublished studies concluded since then, to
determine to what extent his criticisms have been met, and his
recommendations adopted.

iThe comments and advice of Fred E. Fiedler, Robert J. House, and

Abraham K. Korman are gratefully acknowledged.
2 Correspondence should be addressed to Steven Kerr, College of Admin-
istrative Science, The Ohio State University, 1775 College Road Columbus
Ohio 43210.
Copyright © 1974, by PERSONNEL PSYCHOLOGY, I N C .

555
556 PERSONNEL PSYCHOLOGY
Background
The following descriptions are taken from Fleishman and Peters
(1962):
Consideration reflects the extent to which an individual is
likely to have job relationships characterized by mutual trust,
respect for subordinates' ideas, and consideration of their feelings.
Initiating Structure reflects the extent to which an individual
is likely to define and structure his role and those of his sub-
ordinates toward goal attainment. A high score on this dimension
characterizes individuals who play a more active role in directing
group activities through planning, communicating information,
scheduling, trying out new ideas, etc. (pp. 43-44).
Korman (1966) stated that
The Leadership Opinion Questionnaire is a Likert-type attitude
scale which attempts to measure how the supervisor thinks he
should behave in his leadership role.
The Leader Behavior Description Questionnaire typically mea-
sures subordinate perceptions of supervisory behavior (p. 350).
The Supervisory Behavior Description Questionnaire is similar to
the LBDQ, in that it attempts to obtain information about a super-
visor's actual behavior from his subordinates. However, a number
of items are different from these on the LBDQ.
Korman criticized the research he reviewed in the following
ways:
1. While recognizing that several of the Ohio State researchers
showed awareness that the effects of Consideration and Structure
depend upon various situational contingencies, Korman noted that
"in most cases the researchers have made little attempt to either
conceptualize situational variables which might be relevant and/or
measure them" (p. 355). In this regard he indicated a need for
a "systematic conceptualization of situational variance as it might
relate to leadership behavior and a research program designed to
test derivations from such a conceptualization so that direction
might be given to the fleld" (p. 355).
2. Most of the reviewed studies yielded generally insignificant
correlations between leader behavior measures and criteria. This
was most true of relationship involving Initiating Structure, as
measured by the LOQ. Korman reported that the results were "quite
inconsistent," stating that "no discernible pattern appears except
for the presence of low correlations" (p. 354). Studies using the
LBDQ produced "slightly more consistent" patterns, but here also
Korman found "a great deal of inconsistency" (p. 354).
 KERR. AND SCHRIESHEIM 557
3. Many of the studies employed a design which required that
predictor and criterion ratings be made by the same individual.
Thus an LOQ study might have the leader describe both his pre-
ferred leadership role and the degree to which his work group
was troubled by stress. Similarly, a study using the LBDQ might
gather information from subordinates concerning leader behavior,
and then obtain satisfaction scores from the same individuals.
Korman expressed concern over the possibility that raters might
distort their perceptions so as to obtain "balanced" cognitions.
4. Because of the great emphasis in most of the reviewed studies
placed upon correlational data, and because most of the research
sought only to establish concurrent validity, Korman felt that it
was impossible to address the question of causalit3^ He was "unable
to locate any studies in the literature where 'Consideration' and/or
'Initiating Structure' were experimentally varied in order to deter-
mine criterion outcomes" (p. 354), and was able to uncover only two
predictive field studies. Consequently, it is possible that perform-
ance might well cause certain levels of Consideration and Structure
rather than being caused by those variables, as is commonly
assumed.
5. Korman's final criticism was that no one had attacked the
problem of determijiing which leader behavior scores are related
to criterion variance, and which are not. Since curvilinear rather
than linear patterns of relationships have been found by many
researchers, it is critically important to determine the range within
which Consideration and Initiating Structure scores "make a differ-
ence" (p. 360).

Discussion
Although Korman's review appeared in 1966, only one of the
studies he cited was published after 1964. It, therefore, appeared
necessary for this paper to review those studies employing the
Consideration and Structure scales which have been completed since
that time. D a t a obtained through use of the LOQ were examined
separately from those obtained by the LBDQ and SBDQ. This ap-
proach has been shown to be necessary (Weissenberg and Kavanagh,
1972), and is consistent with the procedure used by Korman (1966).
Criticism 1
Since the time of Korman's review, efforts have been made by
many researchers to conceptualize situationai variables relating to
leadesrship behavior and to test the effects of such variables. Thus,
while much more work needs to be done in this area, the criticism'
558 PERSONNEL PSYCHOLOGY
t h a t researchers are ignoring the effects of situational variance is
no longer true.
Of t h e m a n y studies since 1964 which have sought to examine
Consideration-criteria and Structure-criteria relationships as mod-
erated by such situational variables, perhaps the most important
h a v e been those pertaining to t h e P a t h - G o a l Theory of Leadership
(House, 1971), and those completed by the research group a t
Southern Illinois University (see, for example, H u n t and Liebscher,
1973). These studies and more t h a n 200 others, including m a n y
unpublished manuscripts and theses, have been included in a com-
prehensive review of situational implications of t h e Consideration-
Initiating Structure literature (Kerr, Schriesheim, M u r p h y , and
Stogdill, 19'74). This review identified m a n y variables which mod-
erated the relationships between leader behavior predictors and
various criteria, including:
Subordinate considerations. Expertise, experience, competence,
job knowledge, hierarchical level of occupied position, expectations
concerning leader behavior, perceived organizational independence,
and various psychological aspects.
Supervisor considerations. Similarity of attitudes and behavior to
those of higher management, and upward influence.
Task considerations. Degree of t i m e urgency, amount of physical
danger, permissible error rate, presence of external stress, degree
of autonomy, importance and meaningfulness of work, and degree
of ambiguity.
Criticism 2
While not all of the studies reviewed by Kerr et al. (1974)
yielded statistically significant relationships, the overall effect of
these studies is to show that use of the Consideration and Initiating
Structure scales often does produce such relationships. This con-
tinues to be more true of studies employing the LBDQ and SBDQ
than for those which utilize the LOQ. In general, then, Korman's
1966 observation that use of the Ohio State instruments is normally
accompanied by insignificant correlations between leader behavioi
predictors and criteria is not valid today.
Probably the most important reason why this is so relates to
the increased efforts by researchers to conceptualize and measure
situational variables relating to leadership behavior, discussed above.
Many of the early studies included in Korman's review employed
an "A causes B" design, and by so doing may have blurred the
effects of leader behavior on criteria, by combining people (or
 KERR AND SCHRIESHEIM 559
situations) positively affected by such behavior with others whose
reactions were neutral or negative. As more has become known about
the conditions under which leader behavior variables are associated
with satisfaction, morale, and performance criteria, it has become
possible to employ more sophisticated research desigris, and to
concentrate attention upon areas where leader behavior has been
shown to make a difference. As knowledge about such situational
variables further increases, it is likely that even more significant
relationships will be obtained.
Another possible reason why recent studies have been more
likely to obtain significant correlations between leader behavior
variables and organizational criteria stems from the decreasing
use of "averaged" data, whereby analysis is by work group rather
than by individual. Several of the studies reviewed by Korman
(1966) employed such an approach (Fleishman, 1953, 1957a;
Fleishman, Harris, and Burtt, 1955; Halpin, 1957a; Halpin and
Winer, 1957) with the likely result that predictor-criteria relation-
ships were to some extent obscured.
The group-average approach was based upon the assumption
that "the behavior of the leader is in fact reasonably constant for
all staff members" (Seeman, 1957, p. 95). Thus, differences in sub-
ordinate descriptions of the same leader were attributed to measure-
ment error, and it was further assumed that averaging these descrip-
tions would minimize this error (Graen, Dansereau, and Minami,
1972); In addition to using averaged leader behavior descriptions,
the dependent variables used in these studies often consisted of
group or averaged productivity or satisfaction measures. Empirical
support for this approach seems based upon correlational data
which show that there is a fair degree of agreement in the way
subordinates describe their leaders (interrater correlations appear
to g;roup around .6; Fleishman, 1957a; Fleishman et al., 1955;
Halpin, 1957b; Evans, 1970) and upon analyses of variance which
show that "subordinates differ more in describing different leaders
than in describing the same leader" (Halpin, 1957a, p. 66). Users
of this approach, however, ignored some early studies which yielded
data which indicated that leaders may behave differently for dif-
ferent subordinates. For example, Seeman (1957) found that leader
Structure was significantly negatively correlated with years of job
experience {r .= - .39; p < .01) and years under the leader
(r = -.34:; p < .05), but not with status (r = .12). He concluded
that leaders adjust their Structuring behavior to fit the experience
of their subordinates.
.560 P E R S O N N E L PSYCHOLOGY

Recent experimental findings further suggest that leaders do

not behave in a consistent manner towards all subordinates. For
example, Lowin and Craig (1968) found that supervisors are more
considerate and less structuring for competent than for incompetent
subordinates. The leader may therefore exhibit different leadership
styles within the same work group. Averaging Consideration and
Structure scores in such circumstances would create a fiictitious
average or middle-range score which the leader may never display,
with resultant "findings" becoming little more than artifacts of
measurement.
The growing popularity of moderator variable research designs
appears to have reduced the frequency of this practice, as many of
the moderators are individual-based variables (e.g., authoritarian-
ism; Dessler, 1973) which require analysis on an individual basis.
However, some recent studies continue to employ group-average
designs (e.g.. Cummins, 1972).

Criticism 3
Korman (1966) observed that in many of the studies he cited
both predictor and criterion ratings were made by the same in-
dividuals, so that those significant correlations he did find might
be explained in terms of respondent efforts to achieve "balanced"
cognitions. While this criticism remains true of many of the more
recent studies, in general the problem seems less acute today than
in 1966. This is true for the following reasons:
1. Many of the recent research studies have included performance
data as dependent variables, and in almost no case was such data
obtained from the same individuals who furnished leader behavior
information. Consideration and Initiating Structure data are fur-
nished by the leader himself (using the LOQ) or by his subordinates
(via the LBDQ or SBDQ). By contrast, performance scores have
most often been obtained from higher management (e.g.. Hunt and
Hill, 1971; Hunt, Hill, and Reaser, 1971; Cummins, 1972), or from
a combination of higher management and peers (House and Wigdor,
1969; House and Kerr, 1973), or from observers (Greenwood and
McNamara, 1969). In a few other cases (e.g., Lowin, Hrapchak,
and Kavanagh, 1969; Dawson, Messe, and Phillips, 1972) objective
measures of performance were obtained.
2. In instances where subordinate satisfaction has been examined
as the dependent variable the problem detailed by Korman has
been more likely to occur. However, this has only been true for
 KERR AND SCHRIESHEIM 561
those studies which obtained leader behavior scores from sub-
ordinates. In studies employing the LOQ, which obtains attitudes
toward Consideration and Initiating Structure from the leader
himself, Korman's criticism has been largely irrelevant. However,
it should be kept in mind that LOQ studies have typically yielded
much less significant relationships between predictors and criteria
than have those using the LBDQ or SBDQ. The extent, then, to
which relationships obtained through use of the LBDQ or SBDQ
have been infiated by respondent attempts to achieve "balanced"
cognitions, is impossible to ascertain.
3. In a few cases (e.g.. Skinner, 1969) objective indicators of
dissatisfaction have been employed as dependent variables. To the
degree that actual rates of absence, lateness, grievances, or turnover
are examined, the problem of obtaining predictors and criteria from
the same informational source is clearly resolved.
4. As has been mentioned earlier, many of the more recent re-
search efforts have examined the effects of moderator variables
upon relationship between the independent and dependent variables.
The utilization of such a moderator design can serve as a partial
control over the problem described by Korman (1966), in the
following way:-:,..
It is recognized that correlation (and regression) coefficients
may be spuriously high due to method specificity. This limits
the interpretability of correlations between variables which were
measured in the same way. However, there is no a priori reason
to suggest that the numerical difference between correlations of
two variables measured in the same way at two levels of a third
variable (i.e., interactions) is infiuenced by method specificity.
(Cummins, 1972, p. 657).
Criticism 4
Knowledge is still far from complete concerning the nature of
causality between leader behavior "predictors" and satisfaction,
morale, and performance "criteria." Overall, research focusing on
causality has taken one of the following forms.
Longitudinal studies. It is often possible to infer causality from
evidence that one variable did not occur before another. While
such an approach is inadequate to clearly establish cause-effect
relationships, it reduces the plausability of rival causal hypotheses
(Campbell and Stanley, 1963). Only two studies were discovered
which attempted to determine causality in this manner, and these
are briefiy summarized in Table I.
 562 PERSONNEL PSYCHOLOGY

mail
13
bin !

PP
•i'g
8.
o

-s.s
I I&

ft gO
a go

00
to

b£ CO
•g S
3 a
s 4 ^ 13
'a cS c
a;
s o
win
win

b -a
sa.
H
Ii o o
 KERR AND SCHRIESHEIM 563
In one of these studies, Greene (1973) employed cross-lagged
panel correlations and dynamic correlational analysis as recom-
mended by Lawler (1968). Greene points out that both cross-lagged
and dynamic correlational techniques require identical measure-
ments of the variables to be taken with the same respondents in at
least two points in time. The cross-lagged panel correlational model
can yield suggestive results concerning both the strength and direc-
tion of causation between variables, but it cannot prove causation.
This is because the model cannot distinguish between whether an
increase in variable A caused an increase in variable B or whether
a decrease in variable B caused a decrease in variable A. Other
limitations of the model include an inability to take into account
reciprocal causation and effects of moderating and extraneous causal
variables (Howard and Krause, 1970; Rozelle and Campbell, 1969;
Sandell, 1971; Yee and Gage, 1968). However, as Lawler and
Suttle (1973) have noted, "if it fails to yield data which supports
causation, it is very likely that the kind of causation which is
hypothesized does not exist" (p. 486).
In an attempt to control for spurious correlations due to extra-
neous variables, Greene employed dynamic correlational analysis
(Vroom, 1966) along with the panel design. Based upon the results
obtained, he concluded that in his study (1) there was some evi-
dence of reciprocal causation between leader Initiating Structure
and subordinate performance, but the data more strongly supported
subordinate performance as the causal variable; (2) there was no
significant relationship between Structure and subordinate satisfac-
tion; (3) subordinate performance appeared to cause leader Con-
sideration; and (4) there was some evidence of reciprocal causation
between leader Consideration and subordinate satisfaction, but the
data more strongly supported Consideration as the cause of satis-
faction.
Field and laboratory experimental designs. Four attempts to de-
termine Consideration and Initiating Structure causal relationships
through use of an experimental design were uncovered by the litera-
ture review. These are also referenced in Table 1. The variety of
designs and purposes of these studies makes it impossible to speak
in terms of any "weight of evidence" concerning causality. It appears
that under some conditions subordinate performance causes sub-
sequent leader behaviors (Lowin and Craig, 1968, described in
Table 1) while upon other occasions such performance is caused by
leader behavior (Lowin et al., 1969; Hand and Slocum, 1972;
564 PERSONNEL PSYCHOLOGY
Dawson et al., 1972). T h e evidence concerning subordinate satisfac-
tion is even more_ambiguous.
Overall, it is obvious t h a t much more needs to be learned about
the nature of cause-effect relationships between leader behavior
variables and those concerning subordinate satisfaction, morale,
and performance. K o r m a n ' s fourth criticism, relating to the im-
possibility of resolving the question of causality, is very nearly as
valid t o d a y as it was in 1966.

Criticism 5
T h e question raised by K o r m a n concerning which leader behavior
scores are related to criterion variance and which are not has stil]
not been directly confronted. T h e ranges within which Consideration
and Initiating Structure scores are relevant are still unknown.
However, substantial normative d a t a have accumulated concerning
means and s t a n d a r d deviations of scores on both t h e early (Halpin,
1957b) and later (Stogdill, 1963) versions of t h e L B D Q , as well as
for t h e S B D Q (Fleishman, 1957a) and LOQ (Fleishman, 1957b,
1969). Such d a t a accumulation does not directly permit identifica-
tion of the range of scores which relate t o criterion variance, but
does constitute an important first step in t h a t direction. Knowledge
of such descriptive statistics of Consideration and Structure scores
enables the researcher t o determine, albeit imperfectly, whether his
sample is high or low in either dimension vis-a-vis similar samples.

Conclusions
Several of the shortcomings of the Consideration-Initiating Struc-
ture literature which were prevalent have been systematically
addressed, and a t least partially resolved. T h u s it appears t h a t
present-day researchers are designing their studies in such a way
as to recognize the infiuence of situationai moderating variables
(criticism 1). Largely as a result, there has begun to emerge a
research trail of studies showing often-significant relationships
between leader behaviors and subordinate morale, satisfaction, (and
to a lesser degree) performance variables (criticism 2 ) .
Furthermore, it is t r u e t h a t , for a variety of reasons, the problems
which accrued as a result of obtaining information about both
independent and dependent variables from the same source are
somewhat less troublesome t o d a y (criticism 3 ) . Also, t h e practice
followed by m a n y early researchers of using a group or average
method of d a t a analysis has been largely discontinued in recent
studies.
 KERR AND SCHRIESHEIM 565
On the other hand, we still know almost nothing about the con-
ditions under which leader behavior is the cause of subordinate
outcomes, and those where leader behavior is largely the result
(criticism 4). Static correlational analysis is not an adequate tech-
nique to shed light on this matter. Yet all but a handful of studies
have continued to employ such analysis.
Also, we know nothing of the ranges within which Consideration
and Structure "make a difference," and therefore have no protection
against inadvertently dividing samples into "high" and "low" cate-
gories which are artificial, and which will consequently distort
findings (criticism 5).
Finally, the Korman (1966) review did not concern itself with the
psychometric shortcomings of the Ohio State scales, restricting
itself to limitations of the research of those who used the scales.
This paper has retained that emphasis. However, it has been pointed
out elsewhere (Bish and Schriesheim, 1974; Schriesheim and Bish
1974; Schriesheim and Kerr, 1974; Schriesheim and Schriesheim,
1974) that the scales inadequately control for agreement response
tendencies, typically generate responses which may be contaminated
by what seem to be social desirability or leniency, include behavior
dimensions other than those of Consideration and Structure, and
provide response choices of unequal intervals.
Thus, while we have unquestionably progressed in our knowledge
and understanding of Consideration and Initiating Structure, it is
obvious that a great deal of additional research needs to be done
before definitive statements about the effects of these two behavior
dimensions will be possible. Refinement of the current versions of
the scales will also be needed to overcome the psychometric diffi-
culties outlined above.
REFERENCES
Bish, J. and Schriesheim, C. An exploratory dimensional analysis of form
XII of the Ohio State leadership scales. Academy of Management Proceed-
ings, 1974, in press.
Campbell, D. T. and Stanley, J. C. Experimental and quasi-experimental de-
signs for research on teaching. In N. L. Gage (Ed.), Handbook of research
on teaching. Chicago: Rand McNally, 1963.
Cummins, R. C. Leader-member relations as a moderator of the effects of
leader behavior and attitude. PERSONNEL PSYCHOLOGY, 1972, 25, 655-660.
Dawson, J. A., Messe, L. A., and Phillips, J. L. Effects of instructor-leader
behavior on student performance. Journal of Applied Psychology, 1972 56
369-376.
Dessler, G. A test of the path-goal theory of leadership. Unpublished doctoral
dissertation, Baruch College, City University of New York, 1973.
566 PERSONNEL PSYCHOLOGY

Evans, M. G. The effects of supervisory behavior on the path-goal relation-

ship. Org-anizational Behavior and Human Performance, 1970, 5, 277—298.
Fleishman, E. A. The measurement of leadership attitudes in industry. Jour-
nal of Applied Psyehology, 1953, 37, 153-158.
Fleishman, E. A. The leadership opinion questionnaire. In R. M. Stogdill and
A. E. Coons (Eds.), Leader behavior: Its description and measurement.
Columbus: The Bureau of Business Research, The Ohio State University,
1957, (b).
Fleishman, E. A. A leader behavior description for industry. In R. M. Stogdill
and A. E. Coons (Eds.), Leader behavior: Its description and measurement.
Columbus: The Bureau of Business Research, The Ohio State University,
1957, (a).
Fleishman, E. A., Manual for leadership opinion questionnaire (1969 revision).
Chicago: Science Research Associates, 1969.
Fleishman, E. A., Harris, E. F., and Burtt, H. E. Leadership and supervision
in industry. Columbus: Bureau of Educational Research, Ohio State Uni..
versity, 1955.
Fleishman, E. A. and Peters, D. R. Interpersonal values, leadership attitudes
and managerial success. PERSONNEL PSYCHOLOGY, 1962, 15, 127-143.
Graen, G., Dansereau, F., and Minami, T. Dysfunctional leadership styles.
Organizational Behavior and Human Performance, 1972, 7, 216-236.
Greene, C. N. A longitudinal analysis of relationships among leader behavior
and subordinate performance and satisfaction. Academy of Management
Proceedings, 1973, 433-440.
Greenwood, J. M. and McNamara, W. J. Leadership styles of structure and
consideration and managerial effectiveness. PEESONNBL PSYCHOLOGY, 1969,
22, 141-152.
Halpin, A. W. The leader behavior and effectiveness of aircraft commanders.
In R. M. Stogdill and A. E. Coons (Eds.), Leader behavior: Its description
and measurement. Columbus: The Bureau of Business Research, Ohio State
University, 1957 (a).
Halpin, A. W. Manual for the leader behavior description questionnaire.
Columbus: Bureau of Business Research, Ohio State University, 1957(b).
Halpin, A. W. and Winer, B. J. A factorial study of the leader behavior
descriptions. In R. M. Stogdill and A. E. Coons (Eds.), Leader behavior:
its description and measurement. Columbus: The Bureau of Business Re-
search, The Ohio State University, 1957.
Hand, H. and Slocum, J. A longitudinal study of the effect of a human rela-
tions training program on managerial effectiveness. Journal of Applied Psy-
chology, 1972, 56, 412-418.
House, R. J. A path goal theory of leader effectiveness. Administrative Science
Quarterly, 1971, 16, 3.21-338.
House, R. J. and Kerr, S. Organizational independence, leader behavior, and
managerial practices: a replicated study. Journal of Applied Psychology,
1973, 58, 173-180.
House, R. J. and Wigdor, L. A. Cosmopolitans and locals: Some differential
correlations between leader behavior, organizational practices, and em-
ployee satisfaction and performance. Academy of Management Proceedings,
1969, 135-137.
Howard, K. I. and Krause, M. S. Some comments on "techniques for estimat-
 KERR AND SCHRIESHEIM 567
ing ihe source and direction of influence in panel data." Psychological Bul-
letin, 1970, 74, 219-224.
Hunt, J. G. and HOI, J. W. Improving mental hospital effectiveness: A look
at managerial leadership. Technical report 71-4, Southern Illinois University,
1971.
Hunt, J. G., Hill, J. W., and Reaser, J. M. Consideration and Structure effects
in mental institutions: an examination of two managerial levels. Technical
report 71-1, Southern Illinois University, 1971, and in Proceedings of the
Midwestem Academy of Management, 1971.
Hunt, J. G- and Liebscher, V. K. C. Leadership preference, leadership be-
havior, and employee satisfaction. Organizational Behavior and Human
Performance, 1973, 9, 59-77.
Kerr, S., Schriesheim, C, Murphy, C. J., and Stogdill, R. M. Toward a con-
tingency theory of leadership based upon the consideration and initiating
structure literature. Organizational Behavior and Human Performance,
1974, 12, 62-82.
Korman, A. K. "Consideration," "initiating structure," and organizational
criteria—A review. PERSONNEL PSYCHOLOGY, 1966, 19, 349-361.
Lawler, E. E. A correlational-causal analysis of the relationship between
expectancy attitudes and job performance. Journal of Applied Psychology,
1968, 52, 46i2-468.
Lawler, E. E. and Suttle, J. L. Expectancy theory and job behavior. Organ-
izational Behavior and Human Performance, 1973, 9, 482-503.
Lowin, A. and Craig, J. R. The infiuence of level of performance on managerial
style: An experimental object-lesson in the ambiguity of correlational data.
Organizational Behavior and Human Performance, 1968, 3, 440-458.
Lowin, A., Hrapchak, W. J., and Kavanagh, M. J. Consideration and initiating
structure: An experimental investigation of leadership traits. Administrative
Science Quarterly, 1969, 14, 23S-253.
Rozelle, R. M. and Campbell, D. T. More plailsible rival hypotheses in the
cross-lagged panel correlation technique. Psychological Bulletin, 1969, 71,
74-80.
Sandell, R. G. Note on choosing between competing interpretations of cross-
lagged panel correlations. Psychological Bulletin, 1971, 75, 367-368.
Schriesheim, C. and Bish, J. The validity of form XII of the Ohio State
leadership scales. Unpublished paper. The Ohio State University, 1974.
Schriesheim, C. and Kerr, S. Psychometric properties of the Ohio State leader-
ship scales. Psychological Bulletin, 1974, in press.
Schri6!sheim, C. and Schriesheim, J. Development and empirical verification of
new response categories to increase the validity of multiple response alterna-
tive questionnaires. Educational and Psychological Measurement, 1974, in
press.
Seeman, M. A comparison of general and specific leader behavior descriptions.
In R. M. Stogdill & A. E. Coons (Eds.), Leader behavior: Its description
and measurement. Columbus: The Bureau of Business Research, The Ohio
State University, 1957.
Skinner, E. W. Relationships between leader behavior patterns and organiza-
tional-situational variables. PERSONNEL PSYCHOLOGY, 1969, 22, 489-494.
Stogdill, R. M. Manual for the leader behavior description questionnaire—
568 PERSONNEL PSYCHOLOGY
form XII. Columbus: Bureau of Business Research, Ohio State University,
1963.
Vroom, V. H. A comparison of static and dynamic correlational methods in
the study of organizations. Organizational Behavior and Human Perform-
ance, 1966, 1, 55-70.
Weissenberg, P. and Kavanagh, M. J. The independence of initiating structure
and consideration: A review of the evidence. PERSONNEL PSYCHOLOGY, 1972,
25, 119-130.
Yee, A. H. and Gage, N. L. Techniques for estimating the source and direction
of causal influence in panel data. Psychological Bulletin, 1968, 70, 115-126.