See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/232570561

Influence Tactics and Objectives in Upward, Downward, and Lateral

Influence Attempts

Article in Journal of Applied Psychology · April 1990

DOI: 10.1037/0021-9010.75.2.132

CITATIONS READS

709 19,259

2 authors, including:

Gary Yukl
University at Albany, State University of New York
103 PUBLICATIONS 27,722 CITATIONS

SEE PROFILE

All content following this page was uploaded by Gary Yukl on 01 February 2014.

The user has requested enhancement of the downloaded file.

Journal of Applied Psychology Copyright 1990 by the American Psychological Association, Inc.
1990, Vol. 75. No. 2, 132-140 0021-9010/90/500.75

Influence Tactics and Objectives in Upward, Downward,

and Lateral Influence Attempts
Gary Yukl and Cecilia M. Falbe
Management Department
State University of New \fork at Albany

Two studies were conducted to replicate and extend previous exploratory research by Kipnis,
Schmidt, and Wilkinson (1980) on influence tactics and objectives in organizations. A new question-
naire was developed that included measures of important influence tactics and objectives omitted in
the earlier research. Whereas the earlier research used only agent self-reports of influence behavior,
the present research used both agent and target reports. Differences in downward, lateral, and up-
ward influence attempts were replicated more for data from agents than for data from targets. Direc-
tion of influence had a stronger effect on influence objectives than on influence tactics. Despite some
differences due to data source and direction of influence, the relative frequency of use for the 8
influence tactics was remarkably similar across conditions. Consultation and rational persuasion
were the tactics used most frequently, regardless of the direction of influence.

One of the most important determinants of managerial were based on examples described by students, and the range of
effectiveness is success in influencing subordinates, peers, and influence tactics measured by the questionnaire is too narrow.
superiors. Despite the obvious importance of this subject, there Their questionnaire does not include some influence tactics
has been very little empirical research on the influence behavior found to be relevant for leadership effectiveness. Second, their
of managers. Considerable research has been conducted on list of influence objectives needs to be extended to include oth-
sources of managerial power (see Podsakoff & Schriesheim, ers central to managerial work. Third, the research examined
1985), but only a few studies have examined issues such as the only self-perception of influence tactics and objectives. The pos-
types of influence tactics used by managers and the objectives of sibility of systematic biases in these self-reports has not been
their influence attempts (Kipnis, Schmidt, & Wilkinson, 1980; examined. For example, respondents may have exaggerated
Mowday, 1978; Schilit & Locke, 1982). their use of socially desirable influence tactics, such as rational
In the exploratory study by Kipnis et al. (1980), a question- persuasion, and understated their use of less acceptable tactics,
naire was developed to measure categories of influence behavior such as coercive pressure. Likewise, respondents may have been
(called influence tactics) and common reasons for making in- biased to select as their target someone toward whom they were
fluence attempts in organizations (called influence objectives). more likely to use socially desirable influence tactics.
The questionnaire was administered to a sample of night stu- The present research had the following two major objectives:
dents, and each respondent described how often he or she used (a) to determine if the major findings in the Kipnis et al. study
the influence tactics in influence attempts with a particular sub- could be replicated with differences in methodology, and (b) to
ordinate, peer, or superior. On the basis of the results of their extend the research to include additional types of influence be-
exploratory research, Kipnis et al. concluded that managers havior and objectives. This article describes preliminary re-
tend to use different tactics and to have somewhat different ob- search to develop new measures and two studies designed to
jectives depending on the direction of influence. This study was address the potential limitations in the earlier research by
an important first step but, like all exploratory research, the Kipnis et al.
findings need to be verified and the potential limitations exam-
ined. Up until now the conclusions have been reported in many Development of the Measures
textbooks as established facts about influence tactics in organi-
zations rather than as tentative findings from a single explor- Influence Tactics
atory study. Moreover, subsequent studies have used the influ-
ence questionnaire to examine other types of research questions Preliminary research was conducted to pretest and refine a
(Ansari & Kapoor, 1987; Erez& Rim, 1982) and as the basis for more comprehensive questionnaire for measuring influence
developing a typology of managers based on patterns of tactics tactics. All of the items in our questionnaire were new, but in-
(Kipnis & Schmidt, 1988). cluded among them were items representing six of the eight
There are a number of limitations and potential problems in scales in the Kipnis et al. study: assertiveness, rationality, ingra-
the Kipnis et al. study. First, the items in their questionnaire tiation, exchange, upward appeals, and coalitions. Two of their
scales ("sanctions" and "blocking") were not represented in our
questionnaire because of conceptual problems and infrequent
Correspondence concerning this article should be addressed to Gary use. Most of their sanctions items (e.g., giving a pay increase
Yukl, Management Department, School of Business, SUNY, Albany, or promotion, suspending or firing the target) are reactions to
New York 12222. something the target has already done rather than specific, pro-
132
 INFLUENCE TACTICS AND OBJECTIVES 133

active influence attempts. In their study, sanctions were used Table 1

infrequently with subordinates (the mean item score in their Scale Definitions of Influence Tactics
study was 1.3 on a scale of 1 to 5) and not at all with peers or
Scale Definition
superiors (the mean item score was 1.0). Their blocking tactic
was used infrequently with coworkers (mean item score was 1.7) Pressure Tactics The person uses demands, threats, or
and almost never with superiors and subordinates (mean item intimidation to convince you to comply
scores were 1.1). The three blocking items with highest factor with a request or to support a proposal.
loadings appear to be examples of pressure tactics (threaten to (Similar to assertiveness)
Upward Appeals The person seeks to persuade you that the
stop working with the person, engage in work slowdown, request is approved by higher management,
threaten to notify an outside agency). In their more recent re- or appeals to higher management for
search, Kipnis and his associates (Kipnis & Schmidt, 1988) no assistance in gaining your compliance with
longer include blocking as one of their influence tactics. Thus, the request. (Similar to upward appeal)
Exchange Tactics The person makes an explicit or implicit
there seemed little point to include either sanctions or blocking promise that you will receive rewards or
when the questionnaire space could be devoted to more relevant tangible benefits if you comply with a
influence tactics. request or support a proposal, or reminds
The new questionnaire included items representing two in- you of a prior favor to be reciprocated.
(Similar to exchange)
fluence tactics not found in the Kipnis et al. questionnaire but Coalition Tactics The person seeks the aid of others to persuade
likely to be important, namely, inspirational appeals and con- you to do something or uses the support of
sultation. The major source of behavior examples for items rep- others as an argument for you to agree also.
resenting these tactics was the literature on managerial leader- (Similar to coalitions)
Ingratiating Tactics The person seeks to get you in a good mood or
ship. Inspirational appeals to values and emotions are an im- to think favorably of him or her before
portant aspect of charismatic and transformational leadership asking you to do something. (Similar to
(Bass, 1985;Bennis&Nanus, 1985; Conger &Kanungo, 1988; ingratiation)
House, 1977; Tichy & Devanna, 1986). Inspirational appeals Rational Persuasion The person uses logical arguments and factual
evidence to persuade you that a proposal or
include (a) use of emotional, symbolic language to emphasize request is viable and likely to result in the
the importance of a new project or task, (b) appeals to the target attainment of task objectives. (Similar to
person's sense of justice, humanitarianism, or organizational rationality)
loyalty, and (c) appeals to the person's desire to excel, to beat Inspirational Appeals The person makes an emotional request or
competitors, or to accomplish an important, challenging task. proposal that arouses enthusiasm by
appealing to your values and ideals, or by
Research on inspirational behavior by leaders is still in the ex- increasing your confidence that you can do
ploratory stage, but there is strong indication that this influence it.
tactic may be an effective approach for gaining subordinate Consultation Tactics The person seeks your participation in
commitment to a leader's objectives and strategies. making a decision or planning how to
implement a proposed policy, strategy, or
Consultation is a form of leadership behavior that has been change.
studied extensively during the last 3 decades (Bass, 1985; Yukl,
1989). It is widely accepted that managers are sometimes able
to influence people to accept a decision by involving them in
the process of making it, or at least in the process of planning the grouping of items into scales was strongly influenced by
how to implement it (Bradford & Cohen, 1984; Cotton, Voll- their scale definitions and item content.
rath, Froggatt, Lengnick-Hall, & Jennings, 1988; Vroom & The revised questionnaire had scales measuring eight influ-
Jago, 1988; Vroom & Yetton, 1973). The mechanism for influ- ence tactics, the definitions of which are shown in Table 1. The
ence through participation is not very well understood, but correlations among scales (see Table 2) indicate that the eight
when a person is invited to help decide what to do and how to tactics are sufficiently independent to be regarded as distinct
do it, the person is likely to identify with the decision and try to forms of influence behavior.
make it successful. Although consultation is not always success- Each of the scales in the revised questionnaire had four or
ful, it is an important, widely used tactic for influencing com- five items. Parallel versions of the questionnaire were developed
mitment to a decision (Heller, 1971). for agents and targets. Both versions had the same items, but
As in the Kipnis et al. study, the response choices for each minor changes in wording were needed to make the question-
item in our questionnaire indicated the frequency of use for the naire suitable for respondents to describe someone else's influ-
type of influence behavior described by the item. The prelimi- ence tactics rather than their own. The only other difference
nary questionnaire was administered to a sample of 293 target between the two versions was a not applicable option in the tar-
respondents that included both night master's of business ad- get version that did not appear in the agent version. The follow-
ministration (MBA) students and managers in several compa- ing response choices were used in the target version of the re-
vised questionnaire:
nies. Respondents were asked to use the questionnaire to de-
scribe the influence behavior of a peer or their boss. The ques- NA Not applicable; the behavior is something this person
tionnaire was revised and shortened after a variety of analyses, cannot do in my organization.
including factor analysis, item analysis, Q-sorts, and classifica- 1 Never uses this tactic under any circumstances
tion of items into predetermined scales by judges. Because the 2 Seldom uses this tactic (only once or twice a year)
purpose was to replicate and extend the Kipnis et al. research, 3 Uses this tactic occasionally (several times a year)
134 GARY YUKL AND CECILIA M. FALBE

Table 2
Intercorrelations of Influence Tactics
Variable

1. Pressure tactics
2. Upward appeals .50
3. Exchange tactics .11 .18
.16 —
4. Coalition tactics .31 .15 —
5. Ingratiating tactics .19 .17 .44 .27 — —
6. Rational persuasion .03 .10 .20 .51 .25
7. Inspirational appeals .20 .11 .31 .37 .44 .49 —
8. Consultation -.24 -.16 .15 .37 .18 .52 .36 —

4 Uses this tactic moderately often (every few weeks) 1. Ask the person to do a new task or work on a new project
5 Uses this tactic very often (almost every week) or account.
2. Ask the person to do a task faster or better.
A sample item from each scale in the target version is listed as
3. Ask the person to change his/her policies, plans, or proce-
follows:
dures to accommodate your needs.
1. Confronts you and demands that you carry out a re-
4. Ask the person to provide advice or help in solving a
quested action promptly, (pressure)
problem.
2. Complains to someone in higher authority if you do not
5. Ask the person to give or loan you additional resources
carry out a requested action, (upward appeal)
such as funds, supplies, materials, or use of equipment, facili-
3. Indicates that he/she will do a favor for you in return for
ties, or personnel.
doing what he/she wants, (exchange)
6. Ask the person to give a formal approval or signoff on a
4. Gets other people to provide evidence to you supporting
proposal, product, report, or document.
a plan or proposal that he/she wants you to help implement,
7. Ask the person to support your proposals in a meeting
(coalition)
with other managers or clients.
5. Compliments you on past accomplishments before asking
8. Ask for information needed to do your work.
you to do another task, (ingratiation)
The questions about influence objectives appeared in a sepa-
6. Provides evidence that the actions he/she is proposing will
rate section of the questionnaire following the section on influ-
lead to the successful completion of a task or project, (rational
ence tactics. Respondents were asked to indicate how often their
persuasion)
influence attempts have each type of objective. Response
7. Describes a proposed task or project with enthusiasm and
choices in the revised version were as follows:
conviction that it is important and worthwhile, (inspirational
appeal) NA Not applicable
8. Tells you what he/she is trying to accomplish and asks if 1 Never
you know a good way to do it. (consultation) 2 Seldom (only once or twice a year)
3 Occasionally (several times a year)
Influence Objectives 4 Moderately often (every few weeks)
5 Very often (almost every week)
Another purpose of the preliminary research was to develop
a more comprehensive and relevant list of influence objectives. For both the influence tactics and objectives, a Not Applicable
Studies on the nature of managerial work (Kanter, 1982; response was receded as 1 before doing any analyses.
Kaplan, 1986; Kotter, 1982; McCall & Segrist, 1980; Mintz-
berg, 1973;Pavett&Lau, 1983) suggest that an important com- Study 1
ponent of this work is influencing other members of the organi-
zation. A manager's eifectiveness depends on success in influ- This study is essentially a replication of the Kipnis et al.
encing others to (a) modify their plans and schedules, (b) (1980) research on agent self-reports of influence attempts, us-
approve and support the manager's plans and proposals, (c) pro- ing a new questionnaire with a broader range of influence tac-
vide additional resources needed to accomplish major tasks, (d) tics and objectives. As in their research, the primary design in
accept and carry out new assignments, and (e) provide relevant this study was a comparison of influence tactics used in upward,
and timely information. Some of these influence objectives are downward, and lateral relations.
not represented as separate items in the Kipnis et al. question-
naire. Hypotheses
The preliminary questionnaire had 12 influence objectives.
This list was shortened and revised after examination of fre- The exploratory research by Kipnis et al. did not propose
quency scores, intercorrelations among objectives, and judges' formal hypotheses, and their findings on directional differences
ratings of relevance. The following eight objectives were se- were not interpreted. In any replication there is the implicit hy-
lected for the revised questionnaire, as compared with only five pothesis that the same relationships will be found, but it is use-
in the Kipnis et al. study: ful to consider whether there is a credible rationale for these
 INFLUENCE TACTICS AND OBJECTIVES 135

findings. Prior theory and research were examined to determine signment, or (c) increase efforts on a task for which success is
if sufficient basis existed for making a priori hypotheses. Al- in doubt. The first type of influence objective may occur in
though prior research suggests the feasibility of the eight influ- downward, lateral, or upward influence attempts, but the latter
ence tactics for influence attempts in all three directions, there two objectives are more likely to occur in downward influence
was a credible rationale for proposing six hypotheses about ex- attempts (Kipnis et al., 1980). Thus, inspirational appeals are
pected directional differences. most likely to be used in influence attempts with subordinates.
Hypothesis 1: Pressure tactics are used most often in down- Hypothesis 6: Consultation is used most often in downward
ward influence attempts and least often in upward influence at- influence attempts and least often in upward influence attempts.
tempts. Pressure tactics are based on coercive power, and there Consultation appears to be most appropriate when a manager
is growing evidence that coercive power is used more frequently has authority to make a decision that must be implemented or
with subordinates than with peers or superiors (e.g., Kim & supported by the target person (Vroom & Yetton, 1973). Be-
Yukl, 1989). The strongest directional difference found by cause of the nature of authority relationships in hierarchical
Kipnis et al. was for assertiveness, which is similar to pressure organizations, this situation is most likely to occur for a man-
tactics. Hypothesis 1 reflects their findings. ager in relation to subordinates and least likely to occur in rela-
Hypothesis 2: Upward appeals are used more often in down- tion to superiors.
ward and lateral influence attempts than in upward influence
With regard to influence objectives, theoretical support was
attempts. Upward appeals are an attempt to invoke the author-
found for five hypotheses about directional differences. In for-
ity and power of higher management by (a) telling the target you
mal organizations, work objectives and standards are usually
are acting on behalf of higher management, (b) threatening to
determined by a top-down process, and influence attempts con-
go over the target's head (also a pressure tactic), or (c) directly
cerning them tend to follow the chain of command. Most man-
asking superiors to help you influence the target (also a coalition
agers have the authority to assign work to subordinates and es-
tactic). Upward appeals are easier to use with subordinates and
tablish performance standards for them. However, it is rare for
peers and are more likely to be successful (Sayles, 1989). When
work assignments to be made in an upward direction, and it is
an agent claims to have the approval of higher management and
the target is the agent's boss, the target is more likely to question awkward for subordinates to request faster or better perfor-
the agent's credibility. Threats to go over the target's head and mance by their boss. In the case of interdependent units, a man-
overt attempts at upward bypassing are more likely to have un- ager may have authority to ask peers to carry out their part of a
favorable repercussions for future relationships if the target is joint activity, and the inputs provided by peers may be evaluated
one's own boss rather than a peer or subordinate. according to established standards of performance. Thus, the
Hypothesis 3: Exchange tactics are used more often in down- following two hypotheses appear reasonable and are consistent
ward and lateral influence attempts than in upward influence with results from the Kipnis et al. study.
attempts. The concept of exchange implies that there is some- Hypothesis 7: Requests to get someone to do a new task occur
thing of value to be traded. Managers usually have control over most often in downward influence attempts and least often in
resources desired by subordinates, but it is less common for upward influence attempts.
subordinates to control resources desired by a superior. Fur- Hypothesis 8: Requests for faster or better performance occur
thermore, it is awkward for subordinates to initiate an exchange most often in downward influence attempts and least often in
with a superior, because resources under subordinate control upward influence attempts.
are usually things they are expected to provide without addi- The target of influence attempts involving resource alloca-
tional rewards. With peers, the potential for exchange depends tion and approvals is usually the person who has authority to
on the extent of task interdependence between them and their make these decisions. In hierarchical organizations, the author-
control over complementary resources. The potential for using ity to allocate resources, authorize decisions, and give formal
exchange tactics is considerable when peers depend on each approvals is greatest for managers in relation to subordinates,
other for information, assistance, resources, approvals, political and least in relation to superiors. Although these objectives
support, and cooperation'to accomplish their task and personal were not included in the Kipnis et al. research, the following
objectives (Cohen & Bradford, 1989; Kaplan, 1984). hypotheses are reasonable.
Hypothesis 4: Coalition tactics are used more often in upward Hypothesis 9: Requests for resources occur most often in up-
and lateral influence attempts than in downward influence at- ward influence attempts and least often in downward influence
tempts. Managers usually have sufficient power and authority
attempts.
to influence subordinates without using coalitions. However, for
Hypothesis 10: Requests for approvals or signoffs occur most
(a) introducing innovations in an organization, (b) convincing
superiors to change an unpopular policy or plan, and (c) influ- often in upward influence attempts and least often in downward
encing a peer over whom one has no authority, coalitions may influence attempts.
be one of the most effective influence strategies (Izraeli, 1975; The process of coalition formation includes attempts to get
Kanter, 1982; Kotter, 1985; Mechanic, 1962; Pfeffer, 1981; others to support one's proposals in meetings where the propos-
Strauss, 1962). als will be accepted or rejected (Cohen & Bradford, 1989; Ste-
Hypothesis 5: Inspirational appeals are used more often in venson, Pearce, & Porter, 1985). Support is more likely to be
downward influence attempts than in lateral or upward influ- sought from people who have considerable political power.
ence attempts. Inspirational appeals appear to be most appro- Thus, the following hypothesis is proposed.
priate for influencing somebody to (a) support an innovative Hypothesis 11: Managers seek support for proposals more of-
proposal or change in strategy, (b) accept a difficult task or as- ten from superiors and peers than from subordinates.
136 GARY YUKL AND CECILIA M. FALBE

Table 3
Mean Frequency of Influence Tactics as Reported by Agents
Direction of influence attempt

Downward Lateral Upward

Influence tactic (N = 62) (TV =75) (N = 60) F(2, 194) Eta2

Pressure tactics 2.2, 1.9b 1.5C 17.4** 15%

(Assertiveness) (2.4.) (1.6b) (1.4.) (243.9**)
Upward appeals 2.1. 2.2a 1.6b 16.4** 14%
(Upward appeal) (1.8.) (1.7.) (1.4b) (37.2**)
Exchange tactics 1.6, 1.7. 1.4b 6.5** 6%
(Exchange) (2.0.) (2.0.) (1.7 b ) (13.5**)
Coalition tactics 2.2 2.2 2.3 0.2
(Coalitions) (2.2) (2.2) (2.3) 0.1
Ingratiating tactics 2.6a 2.6a 2.2b 6.1** 6%
(Ingratiation) (2.6a) (2.7.) (2.4b) (16.7**)
Rational persuasion 3.3 3.2 3.3 0.5
(Rationality) (3.5b) (3.4b) (3.7.) (7.2**)
Inspirational appeals 2.9, 2.8ab 2.5b 4.2* 4%
Consultation 3.6a 3.4ab 3.3b 3.7* 4%

MultivariateFtest
Pillais criterion 5.2**
Hotelling's trace criterion 5.4**
Wilks's lambda 5.3**

Note. Means with different subscripts differ significantly at the .05 level by the Duncan multiple range test.
Results from Kipnis, Schmidt, and Wilkinson (1980) are shown in parentheses.
*p<.05. **p<.01.

Method trary to Hypothesis 4, but consistent with the findings by Kipnis

et al., there were no significant directional differences for coali-
The sample consisted of 197 respondents, including evening MBA
students who worked in regular jobs during the day and managers who
tion tactics. Results for rational persuasion were also nonsig-
were attending management development courses. Respondents filled nificant, and they failed to replicate the Kipnis et al. finding
out the agent self-report version of the revised influence questionnaire that this tactic was used more in upward influence attempts.
in class. Respondents were asked to describe their own influence at- The eta squared values in Table 3 indicate the percentage of
tempts with an upward, lateral, or downward target. The direction of variance in each influence tactic accounted for by direction of
influence in each case was determined randomly by the researchers. influence. Kipnis et al. did not report effect magnitudes for their
People not in their current job for at least 6 months were instructed ANOVA. However, estimates of eta squared values based on their
not to answer the questionnaire. Respondents were assured that their means and standard deviations suggest that, except for assert-
responses would remain confidential. iveness, the effect magnitudes were smaller in their study than
in ours, despite their larger F values.
Results Turning to the two new scales not included in the Kipnis et
Scale reliabilities for the eight influence tactics, computed in al. research, significant differences were found in each case.
terms of Cronbach's alpha, were as follows: .67 for pressure tac- Consistent with Hypotheses 5 and 6, inspirational appeals and
tics, .67 for upward appeals, .61 for exchange, .70 for coalition consultation were used more frequently in downward influence
tactics, .63 for ingratiation, .70 for rational persuasion, .79 for attempts than in upward ones, with results for lateral influence
inspirational appeals, and .71 for consultation. The scale reli- attempts in between.
abilities are comparable with those found in the Kipnis et al. Results for influence objectives are shown in Table 4 and are
study. Table 3 shows the results for the multivariate and univari- compared with results for objectives in the Kipnis et al. study.
ate analyses of variance (MANOVAS and ANOVAS). Results from Consistent with the Kipnis et al. research and Hypotheses 7 and
the Kipnis et al. study for corresponding scales (expressed in 8, assigning work and requesting faster or better performance
terms of mean item scores) are also shown in Table 3 to facili- were most likely to occur in downward influence attempts and
tate comparison. least likely to occur in upward influence attempts. Requests for
Most of the results found by Kipnis et al. for influence tactics resources were significantly more likely to be made in upward
were successfully replicated. Consistent with Hypothesis 1, and lateral influence attempts than in downward attempts,
pressure tactics were used most frequently in downward influ- which is partially consistent with Hypothesis 9. This objective
ence attempts and least frequently in upward influence at- is closest to requests for benefits in the Kipnis et al. study, al-
tempts. Consistent with Hypotheses 2 and 3, upward appeals though the benefits in our study are job-related resources, not
and exchange tactics were used less often in upward influence things like a better job or a pay increase.
attempts than in downward or lateral influence attempts. Con- Other results for objectives were less consistent with the re-
 INFLUENCE TACTICS AND OBJECTIVES 137

Table 4
Mean Frequency of Influence Objectives as Reported by Agents
Direction of influence attempt
Downward Lateral Upward
Influence objective (AT =62) (A r =75) (AT =60) F(2, 194) Eta2

Assign task or project 3.6a 2.8b 1.9. 39.9** 29%

(Assign work) (4.1.) (2.5b) ( 1 -9C) (329.7**)
Request better performance 3.3a 2.5b 1.8C 28.9** 23%
(Improve performance) (3.9a) (2.9b) (2.4C) (135.2**)
Request changes 2.7a 2.4ab 2.1 b 4.4** 4%
(Seek changes) (3.5.) (3.3b) (3.5.) (6.7*)
Request advice or help 3.6 3.6 3.5 0.1
(Seek assistance) (3.2.) (2.8b) (2.1 C ) (64.4**)
Request resources 1.7b 2.4a 2.6a 11.9** 11%
(Request benefits) (1.3.) (1.7b) (2.4.) (90.4**)
Request approval/signoff 1.9. 2.7b 3.2. 17.7** 15%
Request proposal support 2.0b 2.5a 2.6a 4.1* 4%
Request information 4.0 3.7 3.5 2.9

Multivariate Ftest
Pillais criterion 7.7**
Hotelling's trace criterion 10.5**
Wilks's lambda 9.1**

Note. Means with different subscripts differ significantly at the .05 level by the Duncan multiple range test.
Results from Kipnis, Schmidt, and Wilkinson (1980) are shown in parentheses.
*/><.05. **/><.01.

suits found by Kipnis et al., although the discrepancies may be ranked first in terms of frequency, regardless of the direction of
due to subtle differences in the definition of objectives. In our influence, and the other new objectives had moderately high
study, requests for changes in plans and procedures occurred frequency scores.
most often in downward influence attempts and least often in
upward attempts. In the Kipnis et al. study, trying to get some- Study 2
one to change the way the work is done or to accept an innova-
tion occurred more often in upward and downward influence The major purpose of the second study was to determine if
attempts than in lateral ones. In our study, no significant differ- the results could be replicated with a data source other than
ences were found with respect to requests for advice or help in agent self-reports. Study 2 was carried out at the same time as
solving problems. In the Kipnis et al. study, the objective "assist Study 1, but with a different sample. The research design was
me on my job" occurred most often in downward influence similar, and the same hypotheses were tested. However, all data
attempts and least often in upward influence attempts. How- in Study 2 were obtained from targets of influence attempts
ever, their objective included getting the target person to "do rather than from agents.
some of your work."
With regard to the three influence objectives not included in Method
the Kipnis et al. study, requests for approvals were made most The sample consisted of 237 respondents, including evening MBA
often in upward influence attempts and least often in down- students with regular jobs during the day and managers in management
ward influence attempts, consistent with Hypothesis 10. At- development courses. Respondents filled out the target version of the
tempts to gain support for the agent's proposals occurred more questionnaire anonymously in class. All respondents were asked to de-
frequently with peers and superiors than with subordinates, scribe the influence behavior and objectives of an agent designated as a
consistent with Hypothesis 11. No significant directional superior, a peer, or a subordinate. The type of agent to be described by
differences were found with respect to frequency of requests for each respondent was randomly assigned by the researchers. If the agent
information. was a peer or subordinate, the respondent was asked to select a person
In summary, a large majority of the results found by Kipnis with whom he or she had interacted frequently over the past 6 months.
et al. for influence tactics and objectives were replicated, despite When the agent was a superior, respondents were asked to describe their
immediate supervisor if they had worked under that person for at least
the many differences in methodology between the studies and 6 months. People who were new on their jobs were instructed not to
the much lower power of our study (197 respondents vs. 754). answer the questionnaire.
The results also indicate that some relevant tactics and objec-
tives were missing in the Kipnis et al. research. Consultation
Results
and inspirational appeals were among the tactics people re-
ported using most frequently in their influence attempts. Scale reliabilities for the influence tactics, computed in terms
Among the influence objectives, requests for information of Cronbach alphas, were as follows: .65 for pressure tactics, .45
138 GARY YUKL AND CECILIA M. FALBE

Table 5
Mean Frequency of Influence Tactics as Reported by Targets
Direction of influence attempt
Downward Lateral Upward
Influence tactic (TV =87) (JV=71) (N = 79) F(2, 234) Eta2
Pressure tactics 2.0a 1.5b 1.5b 12.9** 10%
Upward appeals 1.7. 1.7. 1.5b 5.6** 5%
Exchange tactics 1.4 1.5 1.4 1.0
Coalition tactics 1.9 1.9 2.0 0.1
Ingratiating tactics 2.5 2.4 2.2 2.5
Rational persuasion 2.8 2.7 2.9 2.1
Inspirational appeals 2.7a 2.2b 2.4b 6.4** 5%
Consultation 3.2 3.1 3.1 0.3
MultivariateFtest
Pillais Criterion 5.0**
Hotellings Trace Criterion 5.0**
Wilks's Lambda 5.0**
Note. Means with different subscripts differ significantly at the .05 level by the Duncan multiple range test.
*p<.05. **p<.01.

for upward appeals, .65 for exchange tactics, .57 for coalition 1, there were no significant differences for exchange tactics, in-
tactics, .56 for ingratiation, .62 for rational persuasion, .70 for gratiation, or consultation.
inspirational appeals, and .75 for consultation. Reliabilities Table 6 shows results for objectives of influence attempts, as
were within acceptable limits for most of the scales but on the reported by targets. There were significant differences for most
low side for a few scales. influence objectives, and most of the hypotheses about objec-
Table 5 presents the mean scale scores for influence tactics tives received at least partial support. Assigning work and re-
used in downward, lateral, and upward influence attempts, as questing faster or better performance were more likely to occur
reported by targets, and the results for the MANOVAS and AN- in downward influence attempts than in lateral or upward in-
OVAS. In general, the data from target respondents yielded fewer fluence attempts, which is partially consistent with Hypotheses
significant differences than data from agent respondents. The 7 and 8. Requests for resources and approvals were more likely
results provide partial support for Hypotheses 1, 2, and 5, but to occur in upward influence attempts than in downward in-
Hypotheses 3, 4, and 6 were not supported. Significant differ- fluence attempts, which is partially consistent with Hypotheses
ences consistent with those in Study 1 were found for pressure 9 and 10.
tactics, upward appeals, and inspirational appeals. The nonsig- The other four influence objectives yielded less consistent re-
nificant differences for coalition tactics and rational persuasion sults for agents and targets. Directional differences involving
are consistent with the findings in Study 1. Contrary to Study requests for information and requests for advice were signifi-

Table 6
Mean Frequency of Influence Objectives as Reported by Targets
Direction of influence attempt
Downward Lateral Upward
Influence objective (AT =87) (N=7\) (N = 79) F(2, 234) Eta2
Assign task or project 3.7a 2.3b 2-2b 49.0* 30%
Request better performance 2.5, 1.5b 1.7b 20.0* 15%
Request change in plans 2.7a 2.1C 2.4b 6.7* 5%
Request advice or help 3.4b 3.3b 4.1, 12.8* 10%
Request resources 1.4,, 2.0a 2.3a 15.7* 12%
Request approval/signoff 2.1b 2.0b 3.1. 15.9* 12%
Request proposal support 2.0 1.9 2.2 1.2
Request information 3.7b 3.1C 4.1. 17.1** 13%
MultivariateFtest
Pillais criterion 13.1**
Hotelling's trace criterion 14.7**
Wilks's lambda 13.9**
Note. Means with different subscripts differ significantly at the .05 level by the Duncan multiple range test.
*/><.05. **p<.01.
 INFLUENCE TACTICS AND OBJECTIVES 139

Table 7
Rank Order of Tactic Frequencies
Downward Lateral Upward

Influence tactic Agents Targets Agents Targets Agents Targets

Consultation 1 1 1 1 2 1
Rational persuasion 2 2 2 2 1 2
Inspirational appeals 3 3 3 4 3 3
Ingratiating tactics 4 4 4 3 5 4
Coalition tactics 5 6 5 5 4 5
Pressure tactics 6 5 7 7 7 6
Upward appeals 7 7 6 6 6 7
Exchange tactics 8 8 8 8 8 8

e Kendall's coefficient of concordance, ff = .96, p < .001. Ranks were based on means carried out to
two decimal places.

cant for the target data but not for the agent data. Directional ther clarified by use of a more specific and comprehensive list
differences involving requested support for a proposal were sig- of objectives derived from descriptive accounts of managerial
nificant for the agent data but not for target data. Directional work. Most hypotheses about directional differences for influ-
differences involving requested changes in plans or procedures ence objectives were supported by data from both agents and
were significant for both data sources, but the results for the targets. The results are consistent with prevailing conceptions
target data actually supported the Kipnis et al. findings better about role relationships and the distribution of authority in or-
than the results for agent data. ganizations. The larger number of influence objectives in our
study fills in some gaps in the list proposed by Kipnis et al. and
Supplementary Analyses in Studies 1 and 2 provides a clearer picture of the variety and mix of influence
attempts made by managers.
The focus so far has been on results from the ANOVA compar- Our research only partially replicated the Kipnis et al. find-
ing downward, lateral, and upward influence attempts. An al- ings for differences in upward, downward, and lateral use of
ternative focus is to examine the relative frequency of the eight influence tactics. Contrary to their findings, no significant di-
influence tactics. Table 7 shows the rank order of scale means in rectional differences were found for rational persuasion, not
each condition for Studies 1 and 2. It is evident that substantial even for agents. The directional differences they found for ex-
agreement occurred between agents and targets on the relative change and ingratiation were replicated for agents, but the
use of the influence tactics. Kendall's coefficient of concordance effects were weak and could not be replicated for targets. Of the
among the rankings was .96 (Siegel, 1956). The four tactics used six influence tactics from their study, only pressure and upward
most frequently were consultation, rational persuasion, inspira- appeals had significant results for both agents and targets. The
tional appeals, and ingratiation. Exchange tactics were used
relatively weak effects due to direction of influence were over-
least often. Except for the distortion caused by the greater use
shadowed by the similarity in frequency rankings for the influ-
of pressure tactics in downward relations, the frequency rank-
ence tactics. The overall pattern of results suggests that the
ing of tactics was remarkably similar regardless of data source
Kipnis et al. conclusions for influence tactics are considerably
or direction of influence. Looking more closely at the Kipnis et
overstated. The big story is not directional differences but rather
al. results, we discovered a similar pattern of frequency rank-
the discovery that some tactics are used more than others, re-
ings. This pattern in their results may have been obscured by
gardless of whether the target is a subordinate, peer, or superior.
their computation of scale scores as the sum of the items in a
scale rather than as the mean item score; it is difficult to com- The present research also demonstrated that consultation
pare scale sums when the number of scale items varies from 2 and inspirational appeals are an important addition to the list
to 6.
of influence tactics identified by Kipnis et al. Agent and target
With regard to the frequency rankings for influence objec- respondents agreed that these two tactics were among the ones
tives, there was no consistent pattern across conditions. Re- used most frequently by managers, regardless of the direction
quests for information and requests for advice or help were most of influence. Consultation and inspirational appeals are rele-
common, but beyond this there was little similarity in rank or- vant and meaningful influence tactics that help to bridge the
der for objectives across data sources and directions of influ- gap between power research and research on leadership. These
ence. Likewise, in the Kipnis et al. study, frequency rankings tactics appear to be important for understanding the process by
for influence objectives varied considerably across conditions. which leaders influence follower commitment to new objec-
tives, strategies, and projects.
Discussion Although many results were consistent for agents and targets,
some inconsistent results also occurred. Social desirability bi-
The Kipnis et al. conclusion that managers have different rea- ases in the agent self-reports are one possible reason for these
sons for influencing subordinates, peers, and superiors was discrepancies. However, to account for the obtained pattern of
strongly supported, and the nature of these differences was fur- results, the biases would have to differ depending on the direc-
140 GARY YUKL AND CECILIA M. FALBE

tion of influence attempt, and there is no evidence of such an Kaplan, R. (1986). The warp and woof of the general manager's job
interaction. A more likely possibility is that the target data were (Technical Rep. No. 27). Greensboro, NC: Center for Creative Lead-
generally less accurate because of attributions and judgmental ership.
errors made when respondents retrospectively described the in- Kim, H., & Yukl, G. (1989). The importance of different types of power
for influencing subordinates, peers, and superiors. In W. A. Ward &
fluence behavior and intentions of another person. This expla-
E. G. Gomolka (Eds.), Managing for performance: Proceedings of the
nation is consistent with the finding that pairwise comparisons 26th annual meeting (pp. 4-6). Portland, ME: Eastern Academy of
for agent data usually provided stronger support for the hypoth- Management.
eses than the corresponding comparisons for target data. Also Kipnis, D., & Schmidt, S. (1988). Upward influence styles: Relationship
consistent with this explanation is the finding that scale reliabil- with performance evaluation, salary and stress. Administrative Sci-
ities were usually higher for agent data than for target data. ence Quarterly, 33, 528-542.
However, the reason for inconsistent findings across the two Kipnis, D., Schmidt, S. M., & Wilkinson, I. (1980). Intraorganizational
data sources is still only a matter of speculation. Additional re- influence tactics: Explorations in getting one's way. Journal of Applied
search is needed to resolve and explain the discrepancies, and Psychology, 65, 440-452.
this research may have to use another method of data collection, Kotter, J. (1982). The general managers. New York: Free Press.
Kotter, J. (1985). Power and influence: Beyond formal authority. New
such as diaries or observation.
York: Free Press.
Future research should examine the relative effectiveness of McCall, M., & Segrist, C. (1980). In pursuit of the manager's job: Build-
different influence tactics for different objectives, targets, and ing on Mintzberg (Technical Rep. No. 14). Greensboro, NC: Center
situations. The relevance of the eight tactics identified in our for Creative Leadership.
research should be further substantiated by direct evidence that Mechanic, D. (1962). Sources of lower participants in complex organi-
use of these tactics has important consequences for individuals zations. Administrative Science Quarterly, 7, 349-364.
and organizations. Finally, researchers should consider the se- Mintzberg, H. (1973). The nature of managerial work. New York:
quences and combinations of influence tactics used in different Harper & Row.
situations, not just the relative frequency of individual tactics. Mowday, R. (1978). The exercise of upward influence in organizations.
Administrative Science Quarterly, 23, 137-156.
References Pavett, C., & Lau, A. (1983). Managerial work: The influence of hierar-
chical level and functional specialty. Academy of Management Jour-
Ansari, M. A., & Kapoor, A. (1987). Organizational context and up- nal, 26, 170-177.
ward influence tactics. Organizational Behavior and Human Deci- Pfeffer, J. (1981). Power in organizations. Boston: Pitman.
sion Processes, 40, 39-49. Podsakoff, P., & Schriesheim, C. (1985). Field studies of French and
Bass, B. (1985). Leadership and performance beyond expectations. New Raven's bases of power: Critique, reanalysis, and suggestions for fu-
"York: Free Press. ture research. Psychological Bulletin, 97, 387-411.
Bennis, W., & Nanus, B. (1985). Leaders: The strategies for taking Sayles, L. (1989). Leadership: Managing in real organizations. New
charge. New \brk: Harper & Row. York: McGraw-Hill.
Bradford, D., & Cohen, A. (1984). Managing for excellence: The guide Schilit, W., & Locke, E. (1982). A study of upward influence in organi-
to developing high performance organizations. New York: Wiley. zations. Administrative Science Quarterly, 27, 304-316.
Cohen, A., & Bradford, D. (1989). Influence without authority: The use Siegel, S. (1956). Nonparametric statistics for the behavioral sciences.
of alliances, reciprocity and exchange to accomplish work. Organiza- New York: McGraw-Hill.
tional Dynamics, 17, 5-17. Stevenson, W, Pearce, J., & Porter, L. (1985). The concept of "coali-
Conger, J., & Kanungo, R. (1988). Charismatic leadership: The elusive tion" in organization theory and research. Academy of Management
factor in organizational effectiveness. San Francisco: Jossey-Bass. Review, W, 256-268.
Cotton, J., Vollrath, D., Froggatt, K., Lengnick-Hall, M., & Jennings, Strauss, G. (1962). Tactics of lateral relationship: The purchasing agent.
K. (1988). Employee participation: Diverse forms and different out- Administrative Science Quarterly, 7, 161 -186.
comes. Academy of Management Review, 12, 8-22. Tichy, N., & Devanna, M. (1986). The transformational leader. New
Erez, M., & Rim, Y. (1982). The relationship between goals, influence York: Wiley.
tactics, and personal and organizational variables. Human Relations, Vroom, V., & Jago, A. (1988). The new leadership: Managing participa-
55,871-878. tion in organizations. Englewood Cliffs, NJ: Prentice-Hall.
Heller, F. (1971). Managerial decision-making: A study of leadership Vroom, V., & Yetton, P. (1973). Leadership and decision-making: A
style and power sharing among senior managers. London: Tavistock. revised normative model. Pittsburgh, PA: University of Pittsburgh
House, R. (1977). A 1976 theory of charismatic leadership. In J. Hunt Press.
& L. Larson (Eds.), Leadership: The cutting edge (pp. 189-207). Car-
Yukl, G. (1989). Leadership in organizations (2nd ed.). Englewood
bondale, IL: Southern Illinois University Press.
Cliffs, NJ: Prentice-Hall.
Izraeli, D. (1975). The middle manager and the tactics of power expan-
sion: A case study. Sloan Management Review, 16, 57-70.
Kanter, R. (1982). The middle manager as innovator. Harvard Business
Review, 60, 95-105. Received May 17, 1989
Kaplan, R. E. (1984). Trade routes: The manager's network of relation- Revision received October 3, 1989
ships. Organizational Dynamics, ]2(4), 37-52. Accepted October 3, 1989

View publication stats