Skip to main content

Full text of "The political economy of organizations and the role of administration : coping with demands from multiple interest groups"

See other formats





Digitized by the Internet Archive 

in 2011 with funding from 

University of Illinois Urbana-Champaign 

Faculty Working Papers 



David A. Whetten 


College of Commerce and Business Administration 

University of Illinois at Urbana-Champaign 

College of Commerce and Business Administration 
University of Illinois at Urbana-Champaign 

February 22, 1977 


David A. Whetten 





David A. Whetten 
Assistant Professor 

C rganizational Behavior Group 
College of Commerce 
University of Illinois 

February, 1977 

Comments from Howard , Jean Bartunek , Manuel London and Michael Moch on 
an earlier version of this paper, have been most helpful. 

Draft: Do not cite or quote. Comments welcomed. 

A bstract 
This study focuses on the difficulty institutional leaders have 
satisfying multiple interest groups. It highlights the conflict between 
three Df these interests and discusses the implications of these conflicts 
for decisions made by the institutional leaders regarding how the organization 
is to be structured and administered. In addition, the effects which the 
conflict between two of these interests (staff's evaluation of effectiveness 
and the number of clients processed) has on the role of the agency director 
are examined. It is shown that leaders facing conflicting expectations 
experienced significantly more role related difficulties than those leaders 
facing complementary expectations. The implications of these results for 
public policy are discussed briefly. 

The politii al economy theory of organizations proposed by Zald (Zald, 
1970; Wamsley ai.d Zald, 1973a, 1973b) posits that organizations operate in an 
environment of multiple interest groups which place conflicting demands on 
the organ izatio:. 's administration. While this theory lias spawned considerable 
discussion (Benson, 1975; Aldrich and Pfeffer, 1976; Pfeffer, 1976a; and 
Jacobs, 1974), it has generated very little empirical research on the nature 
of these constraints and the effects of these conflicts on organizational 
administration. The purpose of this paper is to report data collected on 
the claims of multiple interest groups on the administration of a sample of 
employment and training public agencies and to discuss the implications of 
this research for the further development of the political economy model of 
organizations aid the administration of public programs. 

The Political Economy Model of Organizations 

Wamsley and Zald (1973a) proposed that "organizations exist in an 
immediate environment of users and suppliers, of interested and disinterested 
'others '... [that] may be competitive, hostile, overseeing, etc.; [together 
they form] a policy subsystem which shapes the conditions of existence for 
an agency." (p. 64- ) These interests within a policy subsystem affect the 
distribution of both political and economic resources required by public 
agencies. On the political side the members of the policy subsystem control 
legitimacy and power and therefore affect the propriety of an agency's 
existence. These "relevant others" also influence the arrangement of the 
division of labor between agencies and the allocation of economic resources 
to them. The central theme of this theory is that the administration of 
public agencies is a highly politicized enterprise. Political struggles 
between the focal agency and other interest groups within its policy sub- 
system result from efforts by external actors to influence the agency's 


niche and related goals and the agency's attempts to manipulate its relevant 
others in order to enhance its legitimacy and increase its share of the 
distribution of economic resources. (Wansley and Zald, 1973a) 

The political economy model is compatible with the resource control 
theory of organizational effectiveness proposed by Yuchtman and Seashore 
(1967). The basic assumption underlying the resource control model is that 
critically needed resources are in short supply in the environment and con- 
sequently an organization's ability to survive and prosper is a function of 
its ability to outmanuever other organizations in the acquisition of these 
scarce resources. An effective organization, therefore, is characterized as 
having a strong bargaining position which enables it to dominate it;; 
environment . This power over the environment is based on the organization's 
ability to control such contingencies as: input acquisition, output 
disposal, capital acquisition, acquisition of production f act ars , and the 
acquisition of a labor force. (Jacobs, 1974) 

These theories are compatible not only because they emphasize the impor- 
tance of controlling environmental resDurces but also because both explicitly 
recognize that the environment is compDsed of a heterogeneous mix of interests. 
In rejecting the goal approach to organizational effectiveness, Yuchtman 
and Seashore (1967) pointed out that administrators must serve a multitude 
of different parties with conflicting criteria of organizational effectiveness. 
This criticism of the goal approach has been echoed by Perrow (1970) and 
Hannon and Freeman (1976). Rather than having single identifiable purposes, 
organizations consist of coalitions which are vying for control over the 
power to make policy decisions regarding the distribution of resources 
(Cyert and March, 1963; Thompson, 1967). Following this line of reasoning, 
it is more appropriate to speak of the goals which individuals or groups, 

both inside and outside the organization, have for the organization's 
means than to refer to the organization's goals, per se. (Mohr, 1971) 

To date, the political economy model has stimulated a few research 
studies and theoretical articles which examine such things as the factors 
influencing whether organizations will engage in cooperative or conflictual 
relations (Hall, et al., 1974) and the bases for the distribution of 
power and resources within (Pfeffer and Salancik, 1974; Salancik and 
Pfeffer, 1974; and Pfeffer, 1976a) and between organizations (Jacobs, 
1974; Benson, 1975). These articles share the common orientation that 
conflicting interest groups work to affect the outcomes of an organization's 
decision making process. However, this literature has almost totally 
ignored the effect which operating in a highly political context has on 
the roles of an organization's senior administrators. This is curious, 
since as Selznick (1957) pointed out it is at the institutional level of 
management that the conflicts between internal and external interest 
groups converge. 

The importance of Selznick 's (1957) insight that senior administrators 
are responsible for coping with both the constraints of the organization's 
environment and the needs of lower participants has not been fully assimi- 
lated into the administration literature because generally the two respon- 
sibilities are treated as being separate and independent activities. For 
example , models designed to increase worker involvement and motivation 
(cf, Lawler, 1973) do not consider environmental conditions as constraints 
on the programs which top management can initiate. Similarly, one of the 
short commings of the resource control model of effectiveness is that it 
does not adequately acknowledge the relationship between the eff ieicncy of 
an organization's utilization of its resources and its needs to acquire 


resources . An organization which efficiently utilizes its resources can 
tolerate greater environmental dependence tfan one which is lot as frugal. 

It is interesting to note, however, that the importance of adminis- 
trators being able to cope with multiple corstraints is rece ving an 
increased amount of attention in the organizational effectiv -ness literature 
(cf, Pennings and Goodman, 1976; Mott , 1972; Reimann, 1975; tnd Katz 
and Kahn, 1968). For instance, Pennings and Goodman (1976: L5) propose 
that an organization is effective if "...relevant constraint; can be 
satisfied and to the degree that organizational results approximate or 
exceed a set of referrents for multiple outcomes." 

A few studies have focused on an administrator's need to satisfy both 
internal and external constraints. For instance, a s :udy oi 97 companies 
by Freidlander and Pickle (1968) demonstrated the difficulty of this task 
since they reported generally low and many negative correlations between 
measures of how well the organization was satisfying aach of 7 different 
interests. In a study of general hospitals, Pfeffer (1973) found that the 
chief administrator and members of the hospital's board of directors were 
chosen for their ability to satisfy either the organisation's needs for 
internal coordination or its needs for controlling sources of financial 
support in the environment. 

While these studies describe several of the multiple constraints which 
institutional leaders face, neither of them examine the effects of role 
conflict on the administrators themselves or the nature of the work they 
perform. One of the only studies which has examined both the sources of 
institutional leaders' role conflict and its consequences was a study of 
school principals by Gross, et al. , (1958). In this research the authors 
examined the conflicting expectations these administrators faced from 18 


different internal and external interest groups in making decisions on: 

(1) the hiring and promotion of personnel, (2) allocation of :he principal's 

after-school time, (3) teacher salary recommendations, and (4) budgetary 

recommendations. The 105 principals reported substantial rol ; conflict for 

all four issues with the highest amounts for budget and salar/ recommendations. 

This study also reported that those principals experiencing rnle conflict 

had lower job and career satisfaction and worried more than those who did not 

have to cope with conflicting expectations. This finding has been confirmed 

by most of the subsequent research on boundry spanning positions and role 

conflict (Kahn et al. , 1964; Organ and Green, 1972; Leifer, 1976; Keller and 

Holland, 1975; and Keller et al. , 1976). While these studies are important to our 

understanding of the consequences of the role conflicts experienced by 

institutional leaders their major limitation is that they have typically 

examined only the psychological effects of role conflict. Consequently 

we feel that a more complete understanding of the consequences of coping 

with conflicting demands can be obtained by examining the effects of role 

conflict on the nature of administrators' day-to-day work activities. 

This information will help fill a gap in the previous boundary spanning 

research which has generally failed to specify the link between role 

conflict and negative psychological outcomes such as frustration and low 

job satisfaction. 

The Study 

This study collected data on the nature and effects of conflicting 
demands faced by the directors of a sample of 69 manpower programs in New 
York State (Whetten, 1974). These agencies primarily serve .is links 
between the poor, under-priveleged and unemployed members of a community 
and the social service, vocational and educational training and employment 

opportunities provided by other public and private organizations. 

Relevant interest groups for these organizations were t le organization's 
staff, the local community leaders and the manpower agency's state and 
regional administration. These groups were chosen because they control 
political and economic resources essential for the effective operation 
of the organization. In a social service agency, it is not difficult for 
disgruntled staff members to covertly restrict output by such means as 
misplacing clients' files (Blau, 1963). They can also reduce the quality 
of service provided by referring the client to the wrong agency or speci- 
fying the wrong type of training. Manpower organizations function primarily 
as people processing organizations (Hazenfeld, 1972) utilizing a mediating 
technology (Thompson, 1967). Consequently, they ere highly dependent on 
the leaders of other local organizations which provide training, education 
and employment opportunities for the manpower program's clients or who 
work directly with different groups of disadvantaged in the community, 
because they can seriously affect the flow of referrals into and out of 
the manpower agencies. In this context the importance of legitimacy as a 
political resource (Benson, 1975) is underscored since the evaluation by 
the heads of local organizations to some extent de termines the availability 
of client resources in the task environment. Coni rol over the financial 
resources required to operate the local manpower c rganizations is held 
by the central administration of federal and state manpower programs. At 
the time the data were collected funding for the local manpower agencies 
was received through categorical (program) channels, e.g., the Neighborhood 
Youth Corp in a local community received its operating funds from Washington 
via the regional Manpower Administration office. As a result, the 
top level administrators of these programs were in a position to literally 

shut down a local agency if they did not comply with program guidelines. 

Information regarding the potential demands which these interest 
groups might exert on the heads of manpower organizations was collected 
by asking the organizations' staff and community leaders to evaluate the 
effectiveness of the agencies with which they were associated. We felt 
that this rating would reflect the political demands of these parties 
because each has an interest in the effectiveness of these agencies. They 
are therefore likely to use their control over valued resources as leverage 
to increase the agency's effectiveness as they define it from their 
particular perspective. Consequently, if these interest groups report 
inconsistent evaluations of an organization this suggests that its leaders 
are facing conflicting pressures as they make decisions regarding the 
goals and policies of their program. 

Data from the staff was gathered by asking all of the profess Lonal 
staff members to respond to the following question: "Overall how effective 
do you feel this organization is in fulfilling its mission as a manpower 
organization?" The response was coded on a 7 point scale from "not very 
effective" to "very effective." The rating by local community leaders 
came from a panel which consisted of the heads of seven key organizations. 
These were the Superintendent of Schools, Mayor, County Executive, or 
Chairman of the County Legislature, the Executive Director of the Chamber 
of Commerce, the Commissioner of the County Social Services Department, 
the Director of the Community Action Agency (CAP), and the Secretariate of 
the local Comprehensive Area Manpower Planning Council (CAMPS). These 
leaders were selected for two reasons: (l) because of their familiarity 
with the operations of these organizations and (2) because these leaders 
represented segments of the community which had a definite interest in the 


effectiveness of these agencies. The interests and concerns of clients 
were represented by the Social Services Department, the CAP agency, the 
school system and the elected officials. The relationships with community 
organisations providing services utilized by the clients was represented by 
the Chamber of Commerce, the CAP agency, the CAMPS council, the public 
schools and elected officials. We therefore felt that these seven people 
were in key positions in each community to receive feedback J rom the mem- 
bers of the manpower organization's task environment and as s uch acted as 
important opinion leaders in the community. These leaders responded to the 
following question: "Considering all that you know about each of these 
organizations how would you rate their effectiveness in meeting the man- 
power needs of this area?" Their response was coded on a fiv point scale 
from "ineffective" to "very effective." 

Obtaining an evaluation from the state and/or regional 1 laders of 
these programs was more difficult. These leaders indicated 1 hat they did 
not have any type of formal rating (or ranking) of the local agencies and 
they were not willing to perform such an evaluation foe our study. This 
reluctance was likely due to the uncertainty faced by the entire manpower 
system at this time. During the summer of 1973, Congress was preparing to 
pass a manpower revenue sharing bill which would dismantle the centralized 
categorical administrative structure and place the au hority to fund 
manpower programs in the hands of local community leaders. However, m 
our conversations with these administrators it became apparent that one 
of the primary evaluation tools they utilized in making funding allocations 
was' the monthly reports submitted by each agency. This statistical report 
reflected the level of an agency's activity In such areas as: recelvjng 
referrals, processing clients using various throughput process (e.g., 

counselling, testing) and outputting clients via placement in employment 
or completion of training. Our supposition that these program administrators 
were relying on statistical reports to evaluate an agency's effectiveness 
is consistent with the observations of others regarding the process by 
which this typo of public program is evaluated (Wilcox, 1969) and is 
borne out by this recent statement by a senior official in the Department 
of Labor, "...CETA success will henceforth be measured in terms of costs per 
unsubsidized placement" (Brandwein, 1976). The statistic which appeared 
to be the most appropriate surrogate measure for the potential demands of 
this interest group was the number of outputs from each organization. An 
average of the figures for twelve months was calculated to control for 
seasonal fluctations in the supply of inputs and demand for outputs caused 
by such things as seasonal employment opportunities and needs , and fluct- 
uations in the availability of funding. 


Table 1 shows the relationship between the evaluations by the staff 
and community leaders and the output statistics. This indicates that there 
is a strong negative relationship between the staff's evaluation and an 
organization's output which suggests a strong conflict between the interests 
of internal members and external program administrators. Since the other 
two correlations are not statistically significant it appears that they are 
independent evaluation criteria. This suggests that while they do not 
represent antagonistic demands an administrator would still find it 
difficult at times to pursue both interests simultaneously. 

Insert Table 1 About Here 

From Table 1 we learn that it is difficult for an administrator to do 


whatever is necessary to produce a high volume of output and simultaneously 
meet the staff's criteria for an effective organization. However, this 
information tells us very little about the nature of these antagonistic 
demands. To fully understand the consequences of this conflict of interests 
for administering these agencies we must be able to identify the staff's 
criteria for an effective organization as well as the organizational factors 
which contribute to high output and then determine which of these are in 
conflict. Conflicts between goals represent serious dilemma:; for admin- 
istrators when they involve antagonistic demands on the utilization of 
organizational means (e.g., to accomplish one objective the organization 
should be centralized; however, decentralization is necessary to roach 
another goal), otherwise, they are mere ideological disagreements. To 
determine whether the basic structural design and administrative procedures 
required to meet these three criteria of effectiveness are antagonistic 
the three dependent variables were correlated with several key organizational 
characteristics. These included: (1) the amount of coordination and 
control (formalization, centralization and communication) (2) staff 
training (years of education) (3) the manner in which work is processed 
internally during the throughput stage (variety of tasks, time spent 
processing each unit, task interdependence, number of services offered), 
(4) relationships with the environment (staff's rating of cooperation 
between manpower organizations in the community, size of the organization's 
set, concentration of the set, number of linkages with the seven community 
leaders used as raters) and (5) scale of operations (number of staff and 
occupational complexity). The specific operationalizations of these 
measures are included in Appendix 1. 

The organizational characteristics associated with the three measures 


of effectiveness are shown in Table 2. Focusing on the community loaders' 
rating first, it appears that the manpower organization's visibility in 
the community is the best predictor of this rating. Organizations which had 
a large organization set that was widely dispersed across the eight cate- 
gories of community organizations and included linkages with the organiza- 
tions administered by the community leaders were rated as being mos. 
effective. Visibility may have also been enhanced by the cooperation within 
the manpower system of organizations since the absence of negative publicity 
regarding the "in-fighting" between manpower organizations may resuJ t in 
the community leaders evaluating the entire group of organizations as 
being effective. The staff's level of education may be viewed as an indi- 
cator of their competence and expertise. Since employees ir this system of 
organizations are frequently drawn from the ranks of the unc erpriviledged 
and uneducated, the presence of articulate, well-educated representatives 
should bolster the confidence of outsiders in the ability oj the organization 
to competently service any clients referred to it and to mal e sound 
judgments regarding which of their clients should be referred to other 
local organizations. The correlation between formalization of rules and 
the community leaders ' rating is somewhat surprising since this aspect of 
an organization is generally not highly visible to external observers. 
However, formalization of rules and policies may be acting as a surrogate 
measure for an overall bureaucratic approach to administering the agency. 
If this is the case then from the point of view of interacting organizations 
this lack of flexibility may prove frustrating and result in a poor 
rating. This suggests the need to make a distinction between mutual 
formalization of a relationship, which Aldrich (1976) has shown to be 
positively correlated with the number of referrals exchanged between 


agencies , and unilateral formalization of policies which creates problems 
for the interacting agencies . 

Insert Table 2 About Here 

The finding that community leaders based their judgment of the man- 
power organization's effectiveness primarily on the agency's visibility 
is consistent with Perrow's statement that: "Competence is hard to judge 
so we rely upon familiarity." (1972: 11) This principle seems especially 
appropriate for the public sector where there is a lack of hard criteria 
(e.g., stock prices and profits) for judging the effectiveness of ai 
organization. It is especially difficult for an outsider to evaluate the 
quality of the primary characteristics of these agencies, i.e., the services 
they provide; hence it is only natural for an external observer to focus 
on the organization's secondary characteristics, such as its visibility 
in the community. 

As a test of this proposition we examined the community leaders 
rating of how familiar they were with each of the agencies they were 
evaluating. The alternative responses were: (1) I don't know anything 
about this organization; (2) I am somewhat familiar with this organization; 
(3) I am very familiar with this organization. The community leaders' 
average effectiveness rating. for each of these three levels of familiarity 
were 2.2, 2.8, and 3.5 (p = .01). Consequently, it appears that visibility 
is a very important determinant of a social agency's legitimacy within its 
immediate environment. 

The principal alternative explanation for these results is that these 
characteristics of manpower organizations enable them to successfully 
process many clients and this information is picked up by the community 


leaders via their contacts with their staff or clients whose interests 
they represent. However, if this were the case one would expect a high 
correlation between the comir'.unity leaders' ratings and the agencies' level 
of output which is not consistent with the results of Table 1. 

The most striking feature of Table 2 is that the staff's rating and 
the measure of output are correlated with basically the same set of or- 
ganizational characteristics but the signs of these correlations are 
opposite. Overall, it appears that an organically structured and admin- 
istered organization is rated as most effective by the staff while a 
mechanistic (Burns and Stalker, 1961) organization has the greatest output. 
The staff rating of effectiveness are higher in an organization which is 
decentralized, provides the mechanisms for a high level of internal 
communication (staff meetings), is staffed by highly educated people who 
serve in a small number of occupational titles and who perform a wide 
variety of tasks that involve spending a considerable amount if time 
dealing with each client. These characteristics of the throi ghput process 
are at least partially due to the fact that the organization nas instituted 
many supportive services, which are auxilliary to the main f auctions of 
the organization. The organizational members have also esta>lished a good 
working relationship with other manpower organizations. In contrast, the 
organization which has a high level of output is characterized by a high 
level of centralization of authority and formalization of rules and pro- 
cedures, a large number of staff who are not highly trained but are assigned 
to mar y occupational categories and who perform tasks with low variety 
which do not entail extensive contacts with clients, and, possibly because few 
auxilliary services are provided, require little coordination between 
members. The organization has also established a large network of 


relationships with other organizations in the community which presumably , 
aids in obtaining inputs and disposing of the organization's outputs. 

It is interesting to speculate regarding the source of this co iflict 
between these interests. One of the more plausible explanations is that 
the organization's staff have been socialized into a professional ethic 
which places greater value on the quality of service provided rather than 
the quantity of clients processed. It follows that staff members with this 
orientation would prefer to work in an organization that is organically 
structured and administered since it would be more compatible with their 
values. This explanation is supported by an extensive body of literature 
dealing with the conflict betweei the goals of quality of production versus 
quantity of production in industrial administration (cf . , Seashore, 1965; 
Freidlander and Pickle, 1968; Mcrr, 1973). 

An alternative explanation stems from the work on cognitive dissonance 
reduction (Festinger, 1957; Bern, 1967) This theory suggests that If the staff 
feel a strong pressure to evaluate their agency positively, then having 
done that they would look around for a justification of their rating. 
If they are aware that their organization is not processing a large number 
of clients (relative to others in the sample of organizations) then they 
might conclude that since the organization is not a high producer it must 
be focusing on the quality of outputs. While it :.s not possible to reject 
either of these explanations with the data in this study, the first 
alternative appears to be a more parsimonious explanation. 

Regardless of which explanation is correct, ":he result', are very 
striking. There is a very strong conflict between the inte 'ests of the 
staff members and the interests of central administrators i i documenting 
the effect of their categorical program by citing the numbe ' of placements 


they ha\ e accomplished. While the policy and administrative implications of 
this finding are numerous, we will delay our discussion of them until aft«:r 
the next section. 

Consequences of Role Conflicts 

The data in Table 2 tell us a great deal about the sources of the 
conflict faced by the heads of these organizations and they isolate the 
specif ic organizational characteristics which are affected by the incon- 
sistency in the three objectives. However, to better understand the 
consequences of these conflicts on the role of agency director, we need also 
to examine their affect on the day-to-day activities of these leaders. 
To do this we divided our organizations into two groups depending on whether 
their level of output and staff's evaluation were compatible or incompatible. 
We choose these two measures since they have a substantial negative 
intercorrelation which affects how many of the organization's mean:, are to 
be utilized. We therefore felt that the interaction between them was 
most likely to have a significant impact on the administrator's roLe. 
Each of these measures was divided at its mean and then the organisations 
were classified into the resulting four cells with the High/Low and 
Low/High cells comprising the conflicting condition and the High/High and 
Low/Low cells comprising the complementary condition. This grouping 
resulted in a 34 to 33 split of the 67 organizations. 

To determine the effects of intersender role conflict we asked each 
of the agency heads to describe the difficulty of their roles as adminis- 
trators in terms of the level of uncertainty they faced, the routineness of 
their work, and the amount of control they had over their personal environ- 
ments. Uncertainty was operationalized in terms of the predictability of 
their tasks, the time span of feedback and the extent to which it was 


necessiry to plan ahead. ;1 he first two measures have been frequently- 
used ii studies of uncertainty (cf, Lawrence and Lorsch, 1967; Downey, 
et al. , 1975) and the third attempts to measure the extent to which an 
administrator must look far ahead to consider the ramifications of daily 
decisions. We predicted that administrators in the conflict condition 
would have to be more concerned about the long-range consequences of their 
actions than their counterparts in the complementary condition. Similarly, 
because of the increased complexity of the decision making process in the 
conflict condition we predicted that these administrators would report 
a longer time span to receive feedback regarding the consequences of their 
decisions and less certainty in their ability to predict the eventual 
outcome of these decisions. Task routinization was measured by the number 
of unique and troublesome problems the administrator encountered each day 
(Perrow, 1967; Lynch, 1974). Because of the numerous dilemmas suggested 
in Table 2 we expected that administrators in the conflict group would 
report more problems in their work than those in the complementary group. 
The final measure of role difficulty was the amount of control the adminis- 
trators felt they had over the constraints within and outside the organi- 
zation which affected their job performance. Worker automony has been 
examined previously by Hackman and Oldham (1976), Turner and Lawrence 
(1965) and Haj.e and Aiken (1967). Consistent with our other hypotheses we 
expected that the administrators facing inconsistent role expectations would 
express less control over their environment than administrators facing 
compatible demands. The specific operationalizations of these six survey 
items are included in Appendix 2. 

Table 3 points out very dramatically the effects of role conflicl for 
thase institutional leaders. Not only are the means for the six role 


difficulty measures all significantly different from one another, but in 
addition, our hypotheses regarding the direction of the differences are all 
confirmed. It appears that serving as the head of an organisation where 
there is eithei a high staff evaluation and a low output of clients or a 
low staff evaluation and a high output of clients is a much more difficult 
task than the case in which you have similar levels on both dimensions. 

Table 3 About Here 

By collapsing our 2x2 table into two cells (conflict and complementary) 
we are masking the interaction effects between organization's output and staff 
evaluation for the six role characteristics. Even though the independent 
variables are moderately correlated there may be a significant interaction 
between them. To test for this possibility, we performed ar analysis of 
variance test on the entire 2x2 table (with unequal cell : izes). Table 4 
shows the cell means and the analysis of variance results for the six role 
characteristics. While it is disappointing that the interaction terms for 
only three of the analyses were significant, there is a strong pattern in 
the ranking of the cell means in Table 4 which supports our grouping the 
"interaction" cells together. This pattern is highlighted in Table 5, which 
shows that when the cell means are rank ordered from highest value to lowest 
value the means for the conflict cells (1 and 4) are higher than the means 
for the complementary cells (2, 3) for five out of the six variables. 

Tables 4 and 5 About Here 

In looking at Table 5, there are two individual cell means that 
should be highlighted. First, it is interesting to note that the High 
output/Low employee evaluation cell (#2) had the highest mean for five out 


of six variables. This suggests a possible dysfunctional consequence of 
public j olicies which emphasize or reinforce a high volume of outputs. 
Others lave expressed concern that this orientation is not in the best 
interests of the clients being served by our public institutions (Jones, 
1976) but this study has focused on a different implication of being 
overly numbers conscious, namely the effect it has on the role of agency 
administrator. Part of the reason why this particular type of conflict 
seems to be so problematic for an administrator is that it is very difficult 
to resolve. The alternatives are to raise the evaluation of the staff, 
which is presumably related to their professional orientation, or to 
decrease the volume of production. The staff's evaluation can be changed 
by lowering the professionalism of the staff through the hiring selection 
process. Unfortunately, this goes against the current trend of increasing the 
professionalism of social agency employees and would therefore likely 
decrease the agency's legitimacy with community leaders and clients. On 
the other hand, if the agency head chooses to reduce output there is the 
possibility of incurring negative sanctions from the program's central 
administration which controls the primary source of financial support for 
the local agency. 

The second particularly striking result in Table 5 is that both of 
the conflict cells have higher means than the Low/Low cell (Hi). This 
is a very interesting result since it appears that the condition of low 
output and Low evaluation is less problematic for administrators than 
either of the two conflict conditions. This suggests that when there is 
agreement between various interests that the organization is doinj r , poorly, 
at least in the public sector where there is a relatively low risk of" 
actually going out of business, the day-to-day task:; of administration arv 


less difficult to manage than when the administrator is faced with incon- 
sistent evaluations of effectiveness. In discussing this counterintuitive 
result with colleagues, it was suggested that maybe the agency heads in the 
Low/Low condition did not link the effects of negative evaluations from their 
staff and program administrators with the characteristics of their role because 
they were not expressively involved in their position as agency head. This 
idea was consistent with our observations that there was a group of manpower 
agency heads who had a long history of moving from one public program to 
another as openings with opportunities for advancement became available. These 
people typically were natives of the community and utilized their connections 
with "old chums" for upgrading their employment. These leaders typically spent 
less time than normal administering their agency and seemed to be less involved 
in the "cause" of providing training and employment to the underpriviledged. 
Under these conditions disagreements with the staff of the unfavorable evalu- 
ations by the regional administrators might be dismissed as unimportant if 
these leaders could fall back on the support of their primary reference gpoup-- 
their associates who were the heads of other local organizations. As a test 
of this explanation, we compared the community leaders' ratings of the agen- 
cies in each of the four cells and ir deed we found that those in the Low/Low 
cell were rated highest (F=2.78, p=.(>5). While this provides an interesting 
insight into the political nature of administering federally sponsored social 
service programs at the grass roots J.evel, our analysis shouLd be treated with 
some caution due to the small number of cases in this cell. 


Summary and Conclusions 

The political economy model of organizations posits that public 
organizations exist in a highly politicized environment characterized by 
conflicts between various interest groups within a policy sub-system. 
This study has shown that the environments of manpower organizations 
clearly vary on this dimension. While some organizations appear to be 
confronted by conflicting demands , others ara working within a highly 
complementary set of constraints. However, our major purpose was not to 
tesx the validity of the political economy model but instead to increase 
its utility by showing that the nature of the context within which an 
agency is operating has a significant impact on the role of agency head. 
In this regard we found that leaders facing conflicting demands described 
their job as being much more difficult than their counterparts in the 

complementary condition. 

* t " One of the strengths of the research design used in th. s study was 

the fact that it identified the source of conflicts faced by these leaders. 
Typically studies of boundary spanning roles simply compare the amount of 
role conflict reported by boundary spanners vs non boundary spanners and 
then show that this is related to the subjects' feelings towards their 
roles. By examining the role of boundary spanning in greater depth we 
are better able to understand the dynamics that produce the expressions of 
frustration and low job satisfaction among boundary spanners facing inter- 
sender role conflict which have been frequently reported in the literature. 
While our study would have been strengthened by self report measures of 
role conflict and job satisfaction our results show that administrator's 
in the role conflict condition are clearly describing their role "as if" 


they are facing conflict. This is likely due to the fact that the demands 
made by the interest groups we examined are highly visible and very 
pressing in these organizations. 

This research has also added to our understanding of the role of 
institutional leader--a position which is receiving an increased amount 
of attention (cf, Pfeffer, 1976b). The nature of the pressures from 
internal and external interests which are focused on this role make it a 
unique boundary spanning position which deserves further investigation. 
Our examination of the amount, and consequences, of role conflict faced 
by heads of social service agencies will hopefully stimulate more research 
in this area. A particularly attractive topic for future research is the 
strategies used by institutional leaders for reducing role conflict. Fcr 
this purpose alternatives suggested by Parsons (1951), Merton (1957) and 
Kahn, et al (1964) wouLd provide a sound basis for investigation. 






Staff's Evaluation 



Community Leaders' 



— — 


Agency's Output 



N = 69 
*** = .001 




Potential Demand 


Organization 1 


Rating by Panel 

Rating of 

Ou put 

of Community 

Organization' s 

(Training & 




(3 -H 

1 o o 








» ant 


Centralizat ion 
Communicat ion 



Level of 












•H ^ 

10 4-1 

Time Spent 

10 D 

tu a. 

tj .c 

n N1 

Each Client 



k Pr< 

Task Inter- 


U 4-> 

o ^ 

Number of 
















Between Man- 

power Organi- 





Ci 4J 
•H C 


to S 

Set Size 



c S 

o £ o 




Set Con- 


Linkage With 




Only statistically significant correlations are reported. 

p < .10 = r of .20 

p < .05 = r of .25 

p < .01 = r of .32 

N = 69 





Role Characteristic 

Compatible Demand 

Conflicting Demand 


A Unpredictability of 
Causal Relations 

1.71 (low) 

2.82 (high) 



B. Necessity of Long 
Range Planning 

2.47 (short) 

3.88 (long) 



C. Feedback Time 

2.27 (short) 

3.55 (long) 


D. Number of 




1.82 (few) 

2.70 (many) 


E. Lack of Control 

Over External 

~k ~k~k 


2.88 (low) 

4.52 (high) 


F. Lack of Control 

Over Internal 



1.91 (low) 

2.85 (high) 


«p < .05 
*** p < .01 

N = 69 



Role Characteristics 

Cell Means 

ANOVA F Statistic 

A. Unpredictability of 

1 = 3.17 

A = .77 

Causal Relations 

2 = 2.27 

B = 9.09** 

3 = 3.91 

AxB =1.54 

4 = 1.39 

B. Necessity of Long 

1 = 2.67 

A = 4.30** 

Range Planning 

2 = 4.09 

B = .03 

3 = 3.46 

AxB = 4.06** 

4 = 2.43 

C . Feedback Time 

1 = 1.83 

A = .04 

2 = 3.54 

B = 1.07 

3 = 3.54 

AxB = 6.76** 

4 = 2.36 

D. Number of 

1 = 2.00 

/ = 1.76 


2 = 2.77 

B = .12 


3 = 2.54 

4 = ] .77 

AxB =1.96 

E . Lack of Control 

1 = 3.67 

A = 5.04** 

Over External 

2 = 4.68 

B = .61 


3 - 4.18 

4 = 2.71 

AxB = 3.01* 

F . Lack of Control 

1 = 1.50 

A = .15 

Over Internal 

2 = 3.14 

B = 1.20 


3 = 2.27 

4 = 2.00 

AxB = .23 

A = Output 

B = Staff's Evaluation 

Cells: l=Low A Low B (N=6) 

2=Low A High B (N=22) 
3=High A Low B (N=ll) 
4=High A High B (N=28) 

* P < .10 

** p < .05 

*** p < .01 

N =■ 69 





Rank Order of the 
Four Cell Mean:. 

Role Characteristic Variables 

































Cells: 1 = Low A Low B 

2 = Low A High B 

3 = High A Low B 

4 = High A High A 


Organizational Characteristics 

Formalization : An index composed of the organization's standardized scores 
regarding the presence or absence of: 1 . An organization chart, 2. Written 
contracts of employment, and 3. Written records of job performance. The 
latter two variables were scored on the basis of whether they were available 
for administrators and supervisors only, or for all non-clerical personnel. 

Centralization : The staff members' average response regarding how often they 
participated in making the following decisions: 1. To promote any of the 
non-clerical staff, 2. To hire new staff members, 3. To adopt new policies, and 
4. To adopt new programs. Responses were coded on a 5 point scale from "Never" 
to "Always." 

Communicat ions : The numb ;r of regularly scheduled meetings within an organi- 
zation per month. This v triable was logged because of its highly skewed 

Professional Training : Tie average number of years of education of the 
staff members. 

Task Variety : Staff's av irage response to: "Would you describe your job as 
having little variety in Lt, or a lot of variety?" Responses were scored on 
a 7 point scale from "Litzle variety" to "A lot of variety." 

Time Spent Processing Each Client : Staff's average response to: "How much 
time does it usually take you to perform one unit of work, e.g., provide 
services to one client, process or investigate one claim, or solve one 
problem?" Response alternatives were: 1. One hour or less, 2. Half a 
day or less, 3. One working day or less, 4. 2-3 days or less, 5. One week 
or less, 6. 2-3 weeks or less, and 7. Several weeks. 

Task Interdependence : Staff's average response to the following question: 
"Work can flow through a work unit in several different ways. Three of 
these ways are depicted in the following illustrations. Place a check 
under the diagram which best depicts the flow of work through your work 
unit." Alternative responses were three diagrams with the following captions: 
1. Work does not flow between unit members, 2. Work flows between members, 
but in only one direction, and 3. Work flows between members in a reciprocal 


Number of Services Offered : Staff members were asked to identify the 
services offered by their organization from the following list: Outreach, 
Intake and assessment or diagnosis, Orientation of clients or program parti- 
cipants, Basic education, Work skill training, On-the-job training with the 
organization, On the job counseling or supervision (at sites in other 
organizations), Counseling, Supportive services (e.g., day care centers, 
transportation), Job development (solicitation of job openings), Sending 
referrals for job placement, Sending referrals to other organizations to 
receive personal services, Follow up on referrals, and Research and planning. 
This variable was logged. 

Occupational Complexity : On the basis of their job titles and a description 
of the tasks they performed, staff members were placed into eight categories: 
Administration, Basic Education Instructors, Work Skills Instructors, Guidance 
and Counseling, Interviewers, Job Placement and Development, Social Workers 
and Community Organizers, and Staff positions (e.g., publicity, evaluation, 
training, research) . The number of different categories represented was used 
as the organization's score. 

Organizational Size : The number of staff in an agency. Part-time members 
were counted on the basis of the fraction of full-time which they worked 
in the organization. 

Cooperation Between Manpower Organizations : The staff's average response to: 
"Overall how much cooperation is there among the manpower organizations in 
this area?" Responses were coded on a 7 point scale from "Very little cooper- 
ation to "High degree of cooperation." 

Organization Set Size : This is the total number of organizations in the local 
community with which the focal organization interacted. This number was obtained 
by compiling a list of all known public and non-profit organizations in the 
local community over a two year period. We then asked the director of each 
manpower organization to identify those organizations on the list with which 
his organization had any of the following types of relationships: 1. Send 
referrals to, 2. Receive referrals from, or 3. "Other" type of relationship, 
e.g., funding, shared office space and/or personnel. For a more complete 
description of this and the organization set size variable, see Whetten and 
Aldrich (1976). 

Organization Set Concentration : All of the organizations in the organization 
set were classified into nine categories referred to as community sectors. 
These included Education training and employment organizations, Public safety 
organizations, Medical and Health care organizations. The concentration of an 
organization set was calculated using the "H" measure of concentration which 
is the sum of the squared percentages of organizations in each of the eight 
community sectors. 

Linkage with Community Leader's Organization : Each of the community leader's 
organizations could be identified in the focal organization's set. This variable 
is therefore the number of the 7 community leader's organizations with which 
each manpower organization had interacted. 


Agency Head's Role Characteristics 

Predictability of Cause and Effect Relationships : "In some organizations 
things are fairly predictable — if you do this, that will happen. In others 
you are often not sure whether something will work or not. That percentage 
of the time would you say that you are not sure whether something you do 
will work out?" Response alternatives were 9 increments of percentages 
from "10% or less" to "90% or more." 

Necessity to Plan Ahead : "How far ahead do you find it necessary to plan 
your work?" Response alternatives were: 1. About one day, 2. Less than 
one week, 3. Less than 2-3 weeks, 4. Less than one month, and 5. Several 

Feedback Time : "In terms of the major tasks you are assigned, on the 
average how long is it before you know whether your work effort is successful?" 
Response alternatives were: 1. Within a day, 2. Within a week, 3. Within 2-3 
weeks, 4. Within a month, and 5. Several months. 

Number of Exceptions Encountered : "During the course of your work how often 
do you come across specific but important problems that you don't know how 
to solve, and you have to take some time to think them through by yourself 
or with others before you take action?" Response alternatives were: 
1. Daily, 2. 2-3 times a week, 3. Once a week, 4. 2-3 times a month, and 
5. Once a month or less often. The scoring on this variable was reversed. 

Lack of Control Over External Factors : "How much control do you personally 
feel you have over factors external to the organization which affect your 
ability to perforin your job?" Responses were coded on a 7 point scale from 
"A great deal of control" to "Very little control." 

Lack of Control Over Internal Factors : "How much control do you personally 
feel you have over factors within the organization which affect your ability 
to perform your job?" Responses were coded on a 7 point scaLe from "A great 
deal of control" to "Very little control." 


"""Keller, et al. (1976), Seiber (1974), Aldrich and Herker (19/7), and 
Pettigrew (1972), have noted that boundary spanning positions have several 
positive features which may outweigh the negative consequences of role 
conflict. These include high task variety, considerable automony and 
extensive opportunities to accrue power within the organization. However, 
since our research interest focuses on the consequences of role conflict 
experienced by a particular boundary spanning position (institutional 
leaders) rather than on an evaluation of the overall costs and benefits 
associated with boundary spanning roles, this body of literature will not 
be discussed extensively in thi; paper. 

During the year preceding the study we interviewed the heads of 
each of the manpower agencies to determine who in the local community was 
most closely associated with their activities and these seven organizations 
were among the most frequently mentioned. 

Under this plan local public officials could elect to fully fund, 
modify or totally eliminate the current categorical programs in the'r 
community. This manpower revenue sharing bill was passed in the form of 
The Comprehensive Education and Training Act (CETA). 


Aldrich, Howard. "Resource Dependence and Interorganizational Relations: 

1976 Relations Between Local Employment Service Offices and 
Social Services Sector Organizations," Administration and 
Society , February. 

Aldrich, Howard and Diane Herker. "Boundary Spanning Roles and Organization 

1977 Structure." To appear in Academy of Management Review . 

Aldrich, Howard, and Jeffery Pfeffer. "Environments of Organizations." In 
1976 A. Inkeles (ed.), Annual Review of Sociology . Palo Alto: 
Annual Reviews, Inc. 

Bern, D. J. "Self-Perception: An Alternative Interpretation of Cognitive 

1967 Dissonance Phenomenon," Psychological Review ; 74, 183-200. 

Benson, Kenneth, Joseph T. Kunce, Charles A. Thompson, and David L. Allen. 

1973 Coordinating Human Services, A Sociological Study of Inter- 
organizational Network . Columbia, Mo.: Regional Rehabilitation 
Research Institute, Monograph Series No. 6. 

Benson, Kenneth M. "The Interorganizational Network as a Political Economy," 

1975 Administrative Science Quarterly , 20: 229-249. 

Blau, Peter. The Dynamics of Bureaucracy . Chicago: University of Chicago 
1963 Press. 

Brandwein, Seymour. "Manpower and Vocational Education Weekly," Washington, D.C. : 

1976 Education News Service , Vol. 7, No. 41, October 14, page 5. 

Cyert, Richard M. , and James G. March. Behavioral Theory of the Firm . 
1963 Englewood Cliffs, N.J.: Prentice-Hall. 

Downey, H. Kirk, Don H. Hellriegel, and John W. Slocum, Jr. "Environmental 

1975 Uncertainty: The Construct and Its Application." Administrative 
Science Quarterly , 20: 613-629. 

Festinger, Leon. A Theory of Cognitive Dissonance . Evanston, 111. : Row, 

1957 Peterson. 

Freidlander and Pickle, "Components of Effectiveness in Small Organizations." 

1968 Administrative Science Quarterly , 13: 278-304. 

Gross, Neal, Ward S. Mason, and Alexander W. McEachern. Explorations in Role 

1958 Analysis: Studies of the School Superintendency Role. 
New York: John Wiley. 

Hackman, Richard, and Greg Oldham. "Motivation Through the Design of Work : Test 

1976 of a Theory." Organizational Behavior and Human Performance, 
16: 250-279. 

Hage, Jerald, and Michael Aiken. "Relationship of Centralization to Other 

1967 Structural Properties." Administrative Science Quarterly , 
12: 75-92. 

Hall, Richard H. , John P. Clark, Peggy Giodano, Paul V. Johnson, and Martha 
1974 Van Roekel, "Interorganizational Relationships." Paper 
presented at the Vlllth World Congress of Sociology, 
International Sociological Association, Toronto, Canada. 

Hannan, Michael T. , and John Freeman. "Obstacles to the Comparative Study of 
1976 Organizational Effectiveness." Paper presented at the 

Symposium on Organizational Effectiveness, Carnegie-Mellon 
Univei'sity, Pittsburgh, Pennsylvania. 

Hazenfeld, Yekeskel. "People Processing Organizations: An Exchange Approach." 

1972 American Sociological Review , 37: 256-263. 

Jacobs, David. "Dependency and Vulnerability: An Exchange Approach to the 

1974 Control of Organizations." Administrative Science Quarterly , 
19: 45-49. 

Kahn, Robert L. , Donald M. Wolfe, Roger P. Quinn, J. Diedrick Snoek. 

1964 Organizational Stress: Studies in Role Conflict and Ambiguity , 
New York: John Wiley. 

Katz, Donald, and Robert L. Kahn. The Social Psychology of Organizations . 

1966 New York: John Wiley. 

Keller, R. T. , A. D. Szilagyi , Jr., and W. E. Holland. "Boundary Spanning 
1976 Activity and Employee Reactions: An Empirical Study." 
Human Relations , 29: 699-710. 

Keller, R. T. , and W. E. Holland. "Boundary Spanning Roles in a Research and 

1975 Development Organization: An Empirical Imestigation." 
Academy of Management Journal , 18: 388-393. 

Lawler, Edward Motivation in Work Organizations . Monterey, Calif.: Brooks 

1973 Cole. 

Lawrence, Paul R. , and Jay W. Lorsch. Organization and Environment . Boston: 

1967 Graduate School of Business Administration, Harvard University. 

Leifer, R. "Boundary Spanning and Boundary Spanners: relationships With 

1976 Stress, Satisfaction and Task Routineness. ' Paper presented 
at the Midwest Academy of Management meetings, St. Louis, Mo. 

Lynch, Beverly P. "An Empirical Assessment of Perrow's Technology Construct." 

1974 Administrative Science Quarterly , 19: 338-357. 

Merton, Robert. Social Theory and Social Structure . New York: The Free Press. 

Mohr, Lawrence B. "The Concept of Organizational Goal," American Political 

1973 Science Review , 67: 470-481. 

Mott , Paul The Characteristics of Effective Organizations . New York: 
1972 Harper and Row. 

Organ, Dennis, and C. N. Green. "The Boundary Relevance of Two Project 

1972 Managers' Jobs." Research and Development Management , 3: 7-11. 

Parsons, Talcott. The Social System . Glencoe , 111.: The Free Press. 

Pennings, Johannes M. , and Paul S. Goodman, "Toward a Framework of Organizational 
1976 Effectiveness." Paper presented at the Symposium on Organi- 
zational Effectiveness, Carnegie Mellon University, Pittsburgh, 

Perrow, Charles "A Framework for the Comparative Analysis of Organizations." 
1967 American Sociological Review , 32: 194-208. 

Perrow, Charles Organizational Analysis: A Sociological View . Monterey, 
1970 Calif.: Brooks Cole. 

Perrow, Charles Complex Organizations: A Critical Essay . Glenview, 111. : 
1972 Scott Foresman. 

Pettigrew, Andrew. "Information Control as a Power Resource," Sociology, 
1972 6: 186-204. ~~ ~~~ 

Pfeffer, Jeffrey, "Power and Resource Allocation in Organizations," to appear 

1976a in Barry M. Staw and Gerald R. Salancik (eds.), New Directions 
in Organizational Behavior . Chicago: St. Clair Press. 

Pfeffer, Jeffrey "Beyond Management and the Worker: The Instil utional function 
1976b of Management." Academy of Management Review , 1: 36-46. 

Pfeffer, Jeffrey, and Gerald Salancik. "Organizational Decision-Making as a 

1974 Political Process: The Case of a University Budget." 
Administrative Science Quarterly , 19: 139-154. 

Reimann, Bernard C. "Organizational Effectiveness and Management's Public Values: 

1975 A Canonical Analysis." Academy of Management Journal , 18: 

Salancik, Gerald R. and Jeffrey Pfeffer. "The Bases and Use of Power in Organi- 
1974 zational Decision-Making: The Case of a University." 
Administrative Science Quarterly , 19: 453-473. 

Seashore, Stanley. "Criteria of Organizational Effectiveness," Michigan Business 
1965 Review , 17: 26-30. 

Seiber, S. "Toward a Theory of Role Ac< umulaticn," American Sociolo; ical 
1974 Review, 39: 567-578. ' 

Selznick, Phillip. Leadership in Administration . New York: Row, Peterson. 

Thompson, James. Organizations in Action . New York: McGraw-Hill. 

Turner, A. N. , and Paul Lawrence. Industrial Jobs and the Worker . Boston: 
196 5 University School of Business Administration. 

Wamsley, Gary, and Mayer N. Zald. "The Political Economy of Public Organi- 
1973a zations." Public Administration Review , 33: 62-73. 

Wamsley, Gary, and Mayer N. Zald. The Political Economy of Public Organizations . 
1973b New York: Praeger. 

Whetten, David A. Predicting Organization- Set Dimensions: An Interorganizational 
1974 Study of the Effectiveness of Manpower Programs in New York 

State. Unpublished doctoral Dissertation, Cornell University. 

Wilcox, Clair. Towards Social Welfare . Homewood, 111.: Irwin. 

Yuchtman, Ephraim, and Stanley E. Seashore. "A System Resource Approach to 

1967 Organizational Effectiveness." American Sociological Review, 
32: 891-903. 

Zald, Mayer N. "Political Economy: A Framework for Comparative Analysis." 

In Mayer N. Zald (ed.), Power in Organization s. Nashville, 
Tennessee: Vanderbilt University Press.