Trust and Managerial Problem Solving Revisited

R. WAYNE BOSS

The centrality of interpersonal trust for effective managerial problem

solving was illustrated by inducing either a high-trust or a low-trust
mental set in experimental groups. The results indicated that subjects
operating in a high-trust environment were significantly more effective
in problem solving than those working in a low-trust environment.

Trust is essential to organizational effectiveness, as was indicated

in a of the literature (Golembiewski & McConkie, 1975)
recent review
and in other empirical studies (Likert, 1976; Likert & Willits, 1940). In
1972 Zand reported a classic study of the relationships between inter-
personal trust and influence, information, and control. Trust, as defined
by Zand, is making oneself vulnerable to others whose behavior one
cannot control (Zand, 1972, p. 231). After inducing a high- and a low-
trust mental set in different groups, Zand found that groups exhibit-
ing a high degree of trust were significantly more effective in problem
solving.
This article reports the results of an extension of Zand’s study. The
instructions and general approach were similar. However, this study
differs from Zand’s in its use of a larger and more varied population of
subjects, a different case study, videotaping, and the employment of
subjects to observe the problem-solving sessions of other teams.
Nevertheless, the results substantiated Zand’s finding that interper-
sonal trust is central to healthy and productive work groups.

This article is based on a paper presented at the 1977 National Academy of

Management Meetings, Orlando, Florida, August 22-24,1977.
Group & Organization Studies, September 1978, 3(3), 331-342.
Copyright 0 1978 by International Authors, B V. 331

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

 SUBJECTS
The 242 participants in this study included 139 master’s and doc-
toral students in business and public administration and 103 business
and public executives. Most of the graduate students were also working
full time in administrative positions, and each had taken at least one
graduate class in organization theory and administrative behavior. The
students took part in this experiment early in the semester of a course in
organization development; the executives participated as members of
several management-training seminars. The experiments were con-
ducted in groups often to sixteen. Participants were divided into teams
comprised of from three to five members with at least one high-trust
and one low-trust team per group. Over 90 percent of the sixty-five
teams participating contained four members. Zand’s subject popula-
tion, in contrast, consisted of only sixty-four business executives who
were combined into sixteen four-person teams during a series of
management-training seminars.
In both studies the participants were randomly placed on high-
trust or low-trust teams. The order in which the high-trust and low-trust
teams participated was also randomly assigned. Since the factual in-
formation given to each participant was identical, the only differences
were the specific roles each was to play and the separate instructions for

high- and low-trust teams. No team members were aware of the differ-
ent instructions prior to the role-play situations.

CASE STUDY

In contrast to the lengthy case study used by Zand, this experiment

employed a variation of the Dashman Company (Ritchie & Thompson,
1976, pp. 53-54), a one-page case about a manufacturing company
engaged in equipment production for the armed forces. Like the case
used in the original study, however, it contained a potentially serious
conflict among executives of the organization.
Post, a newly hired vice president of purchasing, has been charged by the
company president with assuring that a multi-plant organization will not
run out of certain potentially scarce raw materials in a period immedi-

ately preceding World War 11. Post ignores the advice of an administrative
assistant and takes a course of action that initially appears to be accepted
by the purchasing agents at the plants. The purchasing agents, however,
fail to change their procedures. The dilemma is that Post’s efforts appear
to have failed, and he may have aroused strong resistance among the
plant purchasing agents. If the problem is not resolved, the results could
seriously interfere with the organization’s effectiveness and Post’s use-
fulness to the company. Post calls a meeting of his administrative assis-
tants to address the problem.

All participants in the experiment who took a role in the meeting

332

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

received a more detailed briefing with additional background about
their particular positions.

INDUCTION OF HIGH- AND LOW-TRUST MENTAL SET

As in the Zand study, the mental set for the various participants was
shaped by a simple paragraph containing an explanation of the organi-
zational climate. The instructions for the high-trust teams were as
follows (Zand, 1972):
You have learned from your experience in the company that you can trust
the other members of your management team. You and the others can
express your differences and your feelings of encouragement or of disap-
pointment. You and the others also share all relevant information and
freely explore ideas and feelings that may be in or out of your defined
responsibility. The result has been a high level of give-and-take and
mutual confidence in each other’s support and ability. (p. 234)
The low-trust groups, on the other hand, were given a diametrically
opposite statement.
The reward structure was altered only for Post. In the high-trust
teams, he received the following information:
This morning Mr. Manson, the president, told you that if the problem was
not cleared up in eight weeks, you will be dismissed from the company.
Since your relationship to the others in your management team has been
highly collaborative, they are to be held as much responsible for the error
as you are. Therefore, it is likely that the president and the board of
directors will go outside for a successor and possibly other assistants to
the vice president as well. This would mean that you would lose your
$65,000 a year job.
In the low-trust teams, Post was also told that he would be dismissed in
eight weeks if the problem was not resolved. However, he was led to
believe that one of his assistants would be selected as his successor.

VIDEOTAPING

The participants had twenty-five minutes to review the case and

the instructions, after which they participated in a fifteen-minute
videotaped meeting. The specific task was to solve the Dashman Com-
pany’s problem. Each participant was asked not to discuss the material
until the videotaping had started so that all of the data would be on tape
for later processing. No discussion of the case took place in any team
prior to the taping.
Normally, those not participating in the problem-solving session
were absent during the taping of the other teams. In some cir-

cumstances, however, participants were permitted to observe other

groups conducting role-play sessions. The results were subsequently
analyzed to determine if the response patterns for the thirty-one partic-

333

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

ipants who observed the videotaping were different from those of the
participants who did not. Analysis of the data showed no difference
between groups.

EVALUATION OF THE PROBLEM-SOLVING SESSIONS

Immediately following the taping of the problem-solving activi-

ties, evaluation forms were distributed to members of the entire group.
The videotape session for each team was then systematically reviewed.
At the conclusion of each fifteen-minute segment, the tape was stopped
and all the participants evaluated the effectiveness of each item. Thus,
in addition to observing their own teams on videotape, the participants
acted as nonparticipant observers, rating as many as three other teams.
The 242 participants provided 562 usable ratings of teams other than
their own.
Zand’s subjects, who were not videotaped, rated their own team’s
effectiveness immediately after the problem-solving session, and ob-
servers in Zand’s study did not act as subjects but only observed each
group directly.
After all the evaluations were completed, the data were analyzed
and the results shared with the total group. A general discussion was
held on possible reasons for the results. At this point, the concept of
trust was first introduced.
The participants had understood that the objectives of this activity
were to give them experience in problem solving, to provide them with
an opportunity to learn more about themselves through the use of

videotape, and to gain experience in process analysis. Both the execu-

tives and the students confirmed in their post-experiment interviews
that they, as participants and as observers, were unaware of the differ-
ent instructions and the different mental sets induced in the teams.

RESULTS

Team effectiveness was measured by a questionnaire that asked

the participants to evaluate each team according to whether there was
much or little of each of the following items: (a) trust; (b) openness
about feelings; (c) clarification of the team’s basic problems and goals;
(d) search for alternative courses of action; (e) mutual influence on
outcomes; (f) satisfaction with the meeting; (g) motivation to imple-
ment decisions; and (h) closeness as a management team as a result of
the meeting. Each person was also asked to respond to the following
question as it related to his own team: &dquo;As a result of this meeting,
would you give little or much serious consideration to a position with
another company?&dquo;
The responses for the participants and observers are shown in
Table 1. Participant data are broken down separately for students and

334

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

 C
E
_7
O
1.1

L
s

5u
t9
7
9
¿
1.1

ia
c
0

vi
a

335

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

administrators. Observation scores include those of subjects who did
not participate in the problem-solving session but observed the teams
on videotape. Participant evaluations of their own teams are reported in
the participant column, and participant evaluations of the other teams
in their groups are shown in the observation column.
In each case, the experiment produced results almost identical to
those reported by Zand. Both participants and observers perceived the
high-trust teams to be significantly more effective than the low-trust
teams on the factors listed in Table 1. All chi squares were consistently
significant above the 0.001 level.
Thirty of the participants, comprising seven teams, were asked:
&dquo;Which team member was most responsible for your decision to give
little or much consideration to a position with another company?&dquo; Of
the twenty-three subjects who were not assigned the role of Post,
twenty-one said that Post’s behavior was the determining factor in
their decisions.
In order to determine the generality of these results, additional
experiments were conducted to determine the degree to which the
results shown in Table 1 were a product of either the specific set of
instructions or the case content. Each experiment utilized the same
design as that explained above in terms of team size, use of participants
as observers, and the use of videotaping.

Two experiments were designed to determine if the results shown

in Table 1 were a result of a low- or high-trust mental set or if the
respondents were merely following instructions regarding the degree
to which they communicated effectively. The first experiment in-
cluded twenty-eight participants who responded to an abbreviated set
of instructions that dealt only with the issue of trust. The high-trust
instructions were: &dquo;You have learned from your experience that you
can trust the other members of this management team.&dquo; The low-trust
instructions were: &dquo;You have learned from your experience that you
cannot trust the other members of this management team.&dquo; The results
from these abbreviated instructions were almost identical to those from
the more extensive instructions given earlier.
The second test for possible contamination included inducing
positive and negative mental sets regarding the quality of commu-
nication existing among thirty-two participants. Sixteen participants
were given the following abbreviated instructions taken from the orig-
inal design:
You have learned from your experience in the company that you and the
other members of your management team can express your differences
and your feelings of encouragment or of disappointment. You and the
others also share all relevant information and freely explore ideas and
feelings that may be in or out of your defined responsibility.
The remaining sixteen participants were given diametrically op-
posite instructions, also taken from the original design. In contrast to
the results from the abbreviated trust instructions, there were no sta-

336

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

tistically significant differences between results from those given the
positive and negative sets of instructions regarding the quality of
communication.
In order to determine the degree to which the results reported in
Table 1 may have been task specific, the impact of two additional
variations of the case content were examined. Both contained the high-
and low-trust instructions used in the original design. The first case
involved a potentially explosive problem-solving session between
three employees and their immediate supervisor. Results from the
thirty participants who experienced either the high- or low-trust in-
structions were almost identical to those shown in Table 1; a sta-
tistically significant difference existed between the high- and low-trust
groups.
The second variation of the case content also contained a difficult
and potentially explosive problem-solving situation involving mem-
bers of a natural team. Seventeen of the thirty-three participants met as
peers without their immediate supervisor. The remaining sixteen
people were asked to resolve the problem with the supervisor present.
Results from the natural team with the supervisor present were almost
identical to those reported in Table 1, while the results from the natural
teams with the leader absent evidenced no significant differences
between the high- and low-trust groups.

DISCUSSION

Analysis of the results shown in Table 1, from the four additional

experiments, and from the post-experiment interview data suggest a
number of significant findings. In the first place, the high or low trust
mental set proved to be the central variable in determining the degree
of problem-solving effectiveness, even though the instructions in-
cluded additional information regarding other aspects of the team
members’ working relationship. As previously mentioned, the partici-
pants also received information regarding the degree to which they
shared their differences and feelings, explored new ideas, and engaged
in a high or low level of give-and-take during the problem-solving
session. However, once the issue of trust had been raised, the subjects
concurred that the primary perception each had had regarding other
team members was the degree to which he trusted them and that the
level of interpersonal trust was the overriding variable that determined
the degree to which the team worked effectively. Many subjects could
not even remember the additional instructions they had received. Fur-
ther, the responses from the twenty-eight subjects given the abbrevi-
ated trust instructions corroborate the fact that the concept of trust is the
central variable contributing to group effectiveness.
Second, the subjects reported that the instructions effectively in-
duced either a low-trust or high-trust mental set among team members.
In terms of actual behavior, few differences emerged between the
students and administrators, particularly in low-trust situations in

337

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

which each team member tended to behave in dysfunctional ways. The
respondents further reported that once the low-trust mental set was
established, it was extremely difficult to change that perception.
The attempt to be open in such a degenerating situation had nega-
tive consequences, as the following example illustrates. In one low-
trust problem-solving session, the member playing Post shared the
information that he alone had received regarding his early-morning
meeting with the president and his possible termination. The other
team members smiled at each other and then proceeded to sabotage his
efforts even further. Post reacted with anger and frustration. After the
activity he reported the following:
If I had my way I would have fired the entire group. What a bunch of
turkeys. I was trying to be honest with them but they wouldn’t cooperate.
Everything I suggested they shot down; and they wouldn’t give me any
ideas on how to solve the problem. It was like pulling teeth to even get
them to speak. By the middle of the meeting I was fed up with the entire
group. I wanted to fire them all and get some new blood that I could
work with.
The other members of this team said they thought the information
about Post’s conversation with the president was fabricated and sus-
pected an attempt to manipulate therri. One team member said,
&dquo;Frankly, I was looking forward to your being fired. I was sick of
working with you-and we had only been together ten minutes.&dquo; It
thus appears that once the trust level degenerates beyond a certain
level, even honest attempts to correct the situation are suspect.
Third, the fact that trust was the overriding variable was not ini-
tially apparent to the subjects. When participants were asked to explain
the reasons for the obvious differences in team effectiveness, they
offered a number of possible explanations. Some thought that the indi-
viduals involved in the low-trust teams were wholly responsible and
that the dysfunctional behavior resulted from personality clashes.
Others were convinced that the problems resulted from the Dashman
Company case itself but could not explain why. Some guessed that the
instructions were different but were unable to identify what those
differences were. Others concluded that the low-trust members had
been told to make errors and intentionally sabotage the exercise. Of
those teams that observed other teams during the videotaping sessions,
some felt strongly that the order in which the teams participated deter-
mined the outcome: when low-trust teams preceded high-trust teams,
some concluded that the high-trust people learned from the low-trust
teams and therefore did not make the same mistakes; when high-trust
teams went first, the low-trust behavior that subsequently evolved was
thought to result from the lack of alternatives available to the team,
since they were reluctant to duplicate the previous teams’ solutions or
approaches.
When told of the different instructions, the group members reacted
with amazement and relief. They were amazed that they had not per-
ceived what seemed to them, after the fact, to be obvious; they were

338

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

relieved to learn that the differing results could be explained by the
instructions, rather than by their personality characteristics.
Fourth, the post-interview data supported the notion that the
leader’s behavior is more critical than that of the other team members in
determining the level of trust developed in the group (Argyris, 1973;
Boss, in press). When asked which team member most influenced their
team’s effectiveness, the overwhelming majority of those assigned to
subordinate positions maintained that Post’s behavior was the major
factor and helped to validate either their low-trust or high-trust feelings
about the team. Further, these interview data are consistent with the
responses to the question: &dquo;Which team member was most responsible
for your decision to give little or much consideration to a position with
another company?&dquo; Twenty-one of twenty-three subjects assigned to
subordinate positions reported their relationship with Post to be their
primary reason. Thus, subjects from both groups reported that their
desire to take a job with another company resulted from their lack of
trust of and resulting degenerative relationship with the leader. These
findings are consistent with those reported in Likert and Willits’ early
study (1940) of insurance salesmen, in which they found that the major
difference between low- and high-producing groups was the degree to
which the salesmen reported that they trusted their immediate super-
visor. Also, results from Likert’s (1976) recent study of school teachers
showed interpersonal trust to be directly related to supportive leader
behavior, open communications between superiors and subordinates,
work-group loyalty, and organizational effectiveness.
Fifth, the degree to which the induced mental set was validated
depended on the degree to which the perceived behavior of the leader
was congruent with the subjects’ initial expectations. In those situ-
ations in which the leader’s behavior was highly inconsistent with the
instructions, the subjects reported that they believed the behavior to be
the more reliable measure. For example, in the few low-trust groups
that were ranked relatively high on the evaluation instruments, the
participants reported that their leaders’ behavior was not consistent
with their instructions; based on that behavior, they felt more favorable
toward both the leaders and their fellow team members. Similarly,
those high-trust team members who reported relatively low scores
maintained that the leaders’ behavior was the major factor in determin-
ing the dysfunctional nature of the meeting. In most cases, the subjects
reported that the leaders’ initial comments determined the degree to
which they believed the low- or high-trust instructions. For example,
note the initial comments of Post in a low-trust situation:
You guys have a problem, and you have fifteen minutes to come up with a
solution. Mr. Manson talked to me this morning. We haven’t received any
contracts from our plants throughout the country. They were supposed to
send them to me if they were over $5,000, and we haven’t received any of
them. That’s a big problem. They’re doing contract purchases without
putting them through us, and your problem is to come up with a solution
to get these folks to cooperate in the next fifteen minutes.

339

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

 Contrast these comments to the following introductory comments
of a vice president in a high-trust situation:
I’d like to thank you for being so prompt in showing up this morning. You
know me well enough to know that I wouldn’t call this meeting unless it
was for pretty serious business. As you know, I have been recently in-

stituting some new purchasing procedures for our ten divisional plants in
the Midwest. What we hope to do is coordinate all our purchasing proce-
dures. As you know,I sent out a letter some weeks ago and received replies
that they planned to cooperate and so far they have not. We have three
weeks until the end of the season, and we have a problem. That’s why I’ve
called you together this morning. I’d like to get your input, and let’s solve
this problem the best we can. Does anyone have any ideas?
When asked why they responded the way they did in the problem-
solving sessions, many low-trust leaders reported that they perceived
that they had no choice. They clearly understood the implications of the
environment and felt a need to be directive, to protect themselves, and
to keep a tight control of the meeting. The high-trust leaders, on the
other hand, reported that they felt comfortable with their groups and
were quite willing to consider alternative points of view and to take

advantage of the subordinates’ expertise and abilities. Thus, although

subordinates used the leaders’ behavior as the chief indication of the
accuracy of the induced mental set, the leaders seemed to translate
their perceptions into actual behavior without behavioral input from
the subordinates. In the low-trust teams, the leaders’ behavior was seen
to be primarily responsible for the ineffective problem-solving be-
havior manifest during the experimental sessions.
Sixth, the impact of the low- or high-trust instructions appeared to
be most evident when the leader was present. Results from one of the
additional experiments evidenced few differences between the high-
and low-trust groups when the leader was absent, while the problem-
solving sessions that included the leader yielded results similar to
those reported in Table 1.
Seventh, the results reported in Table 1 also do not appear to be
task specific, as the task related to case content. Three different cases
were used with the same low- and high-trust instructions, and each
produced the same results. Under conditions of high trust, the natural
teams functioned effectively in the problem-solving environment,
while low-trust teams were perceived to be ineffective.
Eighth, the videotaping provided an opportunity whereby partici-
pants were able to view the group dynamics as they evolved without
the distraction and anxiety associated with taking part in the role-
playing activities. Thus, the tapes provided immediate, objective feed-
back, and the participants could see clearly how their actions contrib-
uted to the level of trust and the quality of the solutions.
Ninth, the activity provided an effective method of illustrating the
centrality of interpersonal trust for effective problem solving. Surpris-
ingly, both administrators and students were equally naive as to the
340

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

important role interpersonal trust plays in determining the way people
work together, even though the graduate students had been exposed to
substantial literature dealing with concepts relating to organizational
effectiveness. As one administrator put it:
I never knew that a lack of trust was our problem (back at the worksite)
until that exercise. I knew that things weren’t going well, but I never
really could quite understand why we couldn’t work well together. After
that experience, things fell into place.
Although the design provides an effective vehicle for illustrating
the centrality of trust in organizational problem solving, extreme cau-
tion should be exercised in the use of such an experiment. The use of
the design raises questions about the ethics of placing subjects in a
situation in which failure is highly probable. For example, the person
who is assigned the role of Post in the low-trust group is particularly
vulnerable, not because of any personal attributes or characteristics,
but because of the mental set induced by the instructions. If such an
experiment were used in a natural team or family group setting in
which efforts were oriented toward building an effective problem-
solving team, the consequences could be catastrophic, given the de-
gree to which some subjects take role-play situations seriously.
Further, if this design were used in an organization-behavior class
made up of undergraduates whose egos are relatively fragile and whose
self-concepts are not yet clear-or in a situation in which the psycho-
logical health of the participants is uncertain-the resulting behavior
could have serious consequences and might result in psychological
damage to those participants.
Subject vulnerability was decreased in these experiments by the
short time span of the experiment; the voluntary nature of the participa-
tion ; the stranger-group environment, which was oriented toward ex-
periential learning; an explicit norm condoning participant mistakes; a
group commitment to learn from these mistakes; and the maturity level
of the subjects. However, even under these circumstances a careful
diagnosis of the subjects’ needs and maturity level was undertaken
prior to the decision to proceed.
Another important consideration should be the investment of ade-
quate time for debriefing and discussion once the experiment has
concluded. This is particularly critical for those in the low-trust teams.
It is important that they understand the reasons why the process
evolved as it did and that certain behaviors were functions of the
different instructions, rather than the personalities involved. Under
circumstances where appropriate explanation and discussion have not
taken place, people have harbored feelings of failure, anger, and re-
sentment for having been manipulated. However, in those situations in
which adequate processing time was allowed, the participants have
seen the experiment as a positive learning experience about the impor-
tance of interpersonal trust and its relationship to organizational effec-
tiveness.
341

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016

 The results of this experiment add additional support to the notion
that trust is central to team effectiveness and managerial problem
solving. Further, they are consistent with Zand’s findings that the
degree to which people trust each other determines the degree to
which they will share relevant information, allow others to control their
behavior, and permit others to exercise influence over their decisions.
People are often hired because of their technical ability and are termi-
nated because of their inability to work effectively with others; and
implicit in that effective working relationship is the issue of interper-
sonal trust.
Even though there are numerous differences between the mechan-
ics of Zand’s experiment and this study, the results validate those
reported by Zand in terms of the influence of a person’s high- or
low-trust mental set as he enters a problem-solving situation. Under
conditions of high trust, problem solving tends to be creative and
productive. Under conditions of low trust, problem solving tends to be
degenerative and ineffective.
REFERENCES

Argyris, C. The CEO’sbehavior: Key to organization development. Harvard

Business Review, 1973,(2), 55-64.
51
Boss, R. W. The effect of leader absence on a confrontation-team building
design. Journal of Applied Behavioral Science, 1978,14(4).
Golembiewski, R. T., & McConkie, M. L. The centrality of interpersonal trust
in group processes. In C. L. Cooper (Ed.), Theories of group processes.
New York: John Wiley, 1975.
Likert, R., & Willits, J. M. Morale and agency management. Hartford, CT: Life
Insurance Agency Management Association, 1940.
Likert, R. Personal communication, December 7, 1976.
Ritchie, J. B., & Thompson, P. Organization and people. New York: West
Publishing, 1976.
Zand, D. Trust and managerial problem solving. Administrative Science Quar-
terly, June 1972, 17 (2), 229-239.

R. Wayne Boss is an associate professor in the Graduate

School of Public Affairs at the University of Colorado,
Boulder Campus. He received a doctorate in public ad-
ministration from the University of Georgia with an
emphasis in organization development. He has had ex-
tensive consulting experience in both the public and
private sectors and is currently writing a book on the
application of organization development to the public
sector. He is presently the chairperson of the Peer Re-
view Panel for Region VI of the International Associa-
tion of Applied Social Scientists.

342

Downloaded from gom.sagepub.com at Hunan University on October 27, 2016