Organization Sets and the Structure

Выходные данные статьи:

Elesh, David. Organization Sets and the Structure of Competition for New Members, Sociology of Education, Vol. 46, Issue 4 (Autumn, 1973), Pp. 371-395.

Organization Sets and the Structure

of Competition for New Members[*]

David Elesh

University of Wisconsin

This paper gives empirical content to the concept of an organization set and tests two related bypasses: (1) organization sets defined in terms of competitive relations are more likely to name functionally similar than dissimilar organizations as competitors; (2) other things equal, the greater the similarity of functions between a focal organization and the members of its organization set, the greater the likelihood that it will compete with them. These hypotheses are tested on data concerning the competition of colleges and universities for new students and only the first is given support. The results are explained in terms of the nature of the market for new students.

Recent attempts to describe the context of an organization have centered on those entities that are "relevant or particularly relevant to goal setting or goal attainment" and that together constitute its task environment (Dill, 1958; Thompson, 1967). Evan has focused on the organizational elements of a task environment and termed them its organization set in an analogue to the concept of the role set (Evan, 1966; Caplow, 1964). Accordingly, the organization set consists of all those organizations with which a reference organization deals in pursuit of its goals. The number and nature of these organizations is a function of the reference organization's position in a hierarchy of organizations, otherwise defined, as the number and nature of the roles in a role-set are a function of a status. Although Evan used this concept to suggest a number of hypotheses about inter- and intra-organizational processes, as yet almost no one has reported any research on organization sets (an exception: Anderson, 1967). The major purposes of this paper are to give empirical content to this concept and to test two hypotheses about sets through an analysis of the structure of competition among colleges and universities for new students. Examination of this structure gives an indication of the boundaries of the task environment and thus defines the organization set with which a particular type of school compete.

Although competitive relations (and their converse, cooperative relations) among organizations do not exhaust the range of relations included within Evan's definition, they are probably those most central to the analysis of organizational processes.[1] Certainly, the literature on organizational environments is almost entirely focused on the effects and correlates of competitive or cooperative relations (e.g., Clark, 1956 Thompson and McEwen, 1958; Selznick, 1949; Levine and White, 1961 Elling and Halinsky, 1961; Litwak and Hylton, 1962; Perrow, 1961 Blau and Scott, 1962; Barth, 1963; Emery and Trist, 1965; Thompson, 1967; Yuchtman and Seashore, 1967; Aiken and Hage, 1968). Evan maintains this focus by framing eight of his eleven hypotheses in terms of competition or cooperation. Moreover, this restriction to competition or cooperative relations is quite consistent with Caplow's concept of sets as organizations of 'the same type continuously visible to each other (Caplow, 1964: 201). Caplow draws an analogy between reference groups and organization sets. Because interindividual comparisons define reference groups, this analogy amounts to a definition of sets in terms of competitive or cooperative relations (1964:201-03).

Organizations in a task environment will compete (or cooperate) for inputs and for clients, or customers for outputs; the market for any given input or output will differ from the market for any other. Consequently, an organization will compete in as many markets as it has inputs and outputs, and each of these markets can be described in terms of a subset of the organization set consisting of all those organizations with which it competes or cooperates. Thus, the organization set is actually a set of subsets, each one of the latter representing those organizations competing (or cooperating) with the reference organization in a specific market.

The available evidence indicates that all organizations exist in hierarchies based on the ability to compete for scarce resources. Colleges and universities are no exception. Students are one kind of scarce resource, and schools may be ordered in terms of their ability to attract them. In other markets, rankings using such indicators as endowments, number of Ph.D.s on the faculties, and average College Board Examination scores are frequently published to apprise relevant publics of current standings. But, as in all rankings of individuals, an institution's position in one hierarchy may be only loosely related to its position in another.

Position in a hierarchy is a function of several factors. Among 'the most important of these are the extent of its legal mandate, its sources of resources, its past success at obtaining resources, its current store of resources, its goals, its population, the efficiency of its internal structure, and its ability to communicate a desirable image to relevant publics.[2] In terms of the market for new students, the hierarchy may be observed simply by noting the degree to which schools of one type attract applicants who are also applying to schools of other types. (This, of course, may not reveal much about the kinds of students being attracted.)

The role of organizational images is particularly important since they tend to summarize the other determinants of rank and are the major (and, sometimes, sole) basis for the allocation of resources to organizations; that is, it is through these images that relevant publics are able to match their "wants" (to use the economists' term) with organizational characteristics to make desired choices. The operation of images in the competition of colleges for new students is well-documented (Clark, 1959, 1970; Hilsinger, 1963; Rossi et al., 1964; Jencks and Reisman, 1968; Feldman and Newcomb, 1969) and has been noted in other forms of organizational recruitment as well (Lesly, 1959-60). But these studies have often misplaced their focus on the degree to which individuals have well-developed conceptions of particular organizations. These studies have taken this focus under the apparent assumption that the impact of images on organizational recruitment depends upon their sophistication. When highly developed images are not found, the researcher often takes this lack of images as an indication that images are unimportant. Yet the assumption itself has gone untested. The utility of the concept of the organization set is its indication of the need to discover whether images have real consequences for the structure of interorganizational relations, regardless of their level of development.

It is easily argued that the importance of images does not depend upon their completeness, cogency, or accuracy. Indeed, it is frequently in the interest of both the organizations and the relevant publics that they not be. Organizations can be expected to distort their images to present themselves as favorably as possible, and the relevant publics can be expected to make do with less than full completeness, cogency, or accuracy because the cost of maximizing these factors is generally judged excessive. As Herbert Simon (1957: Chapters 11, 14, 15) has pointed out, optimization usually less accurately describes market behavior than the more modest criterion of satisfaction.

Insofar as relevant publics have specifiable wants, they will choose those organizations that appear to satisfy their wants. Their ability to choose these organizations depends upon their knowledge and evaluation of organizational characteristics – in other words, their images of them. Thus, on the assumption that characteristics of at least some organizations are visible to at least some of the relevant publics, we are led to the following hypothesis: Organization sets defined in terms of competitive relations are more likely to name functionally similar than dissimilar organizations as competitors. I shall also examine a corollary to a stronger hypothesis of Evan (1966: 183): "[Other things equal, the] greater the similarity of functions between a focal organization and the members of its [organizational] set, the greater the likelihood that it will compete with them." The corollary examined here is that the more refined the definition of functional similarity, the more likely it should be that functionally similar organizations will name one another as competitors.

For institutions of higher learning as considered here, functional similarities will be defined as similarities in the kinds of experiences obtained by students at different schools as a result of their having attended these schools. Indicators of these similarities can be found among the objective characteristics of these schools which are generally visible and understood to be related to the kinds of experiences obtainable at them. For example, the sex composition of a school's student body is well understood to be an indicator of the kind of social life an applicant can expect.

The selection of objective characteristics to indicate the schools' functional similarities was based upon the factors mentioned by students as influencing their choice of schools in the literature on factors affecting college selection. Hilsinger (1963), in a review of the literature from 1926 to 1962, found that a school's reputation, courses of study offered, size, cost, and proximity were consistently important bases for choice despite many differences in and methodological inadequacies among the studies. More recent studies (Rossi et al., 1964; Richards and Holland, 1965) have confirmed the continuing importance of these criteria and have added others: a school's admissions policy, religious affiliation, social life, and sexual composition.

To a large degree, of course, these factors are correlated. A school that offers a large number of courses of study is likely also to be large, coeducational, and nonsectarian. A school that has a reputation for high academic standards is likely to be small to moderate in size, expensive, located in particular areas of the country, difficult to get into, and nonsectarian. These correlations point to a characteristic – the auspices of a school – that is rarely measured in studies of college criteria, although it was found to be important in the few cases where it was examined (Hilsinger, 1963). The auspices of a college refer to whether it is a public, private nondenominational, or religiously affiliated school. Although religious affiliation is in the list above, the critically important issue of whether a school is private and nonsectarian or public, is not. Clearly this distinction is crucial to understanding the other factors mentioned. For example, if a student wishes to attend a low-cost school offering a wide range of courses, he will almost certainly apply to a public institution. In all probability, it also will be large and coeducational, will have an active social life, will be fairly close to home, and will be relatively easy to enter. If he wishes to attend an academically excellent, small to moderate size school, his possible choices are largely ^high-cost, private nonsectarian institutions with high standards for admission and located fairly far from home. Thus, schools' auspices must almost always be taken into account when their functional similarities in other terms are examined. With the exception of a school's cost, it is probably the most fundamental characteristic determining to which schools students will apply.

THE DATA

The data were drawn from a 1962 mail survey of admissions officers at 811 institutions of higher learning.[3] For the purposes of this analysis, all schools that lacked accreditation or which were not four-year liberal arts colleges or universities were eliminated; the remaining sample numbered 650 schools. No claims are made for the representativeness of the sample with respect to the universe of such schools, although it is representative in terms of the geographic location, auspices, and type of the institution.[4] The admissions officers at each of these schools were asked to report the three schools to which their applicants most often also made application.[5] (College applications customarily ask for the names of the other schools to which the student is applying; they use the data to assess their competitive situation.) From these reports, a list of schools was created in which a school might be mentioned as a "naming school" (a school which named its three competitors), a competitor, or both. Each of the schools was then described in terms of several variables drawn from the College Characteristics Data Bank of the Bureau of Applied Social Research and selected for their probable visibility as referents of function for the college applicant population.[6] The variables were the cost, location, type of auspices, sex composition, size, residential status, and admissions policy of the schools. All but two of these variables fall within the above-mentioned list of college criteria drawn from the literature. And the two that were not mentioned – sex composition and residential status-are indicators, although in different ways, of a characteristic that was cited: a school's social life. The reputation of a school is not directly measured, but some indication of the effects of this factor probably can be gained from the joint effects of 'auspices and admissions policy, and the joint effects of auspices and cost. The schools with the highest reputations are likely to be those that are private and nonsectarian, and have high admission standards; they are also likely to have high costs.

The analytic technique employed is quite simple: characteristics of the naming schools are cross-tabulated against characteristics of the competitors named. The type of auspices of the naming schools is used to stratify most of the basic relationships in the examination of the first hypothesis because whether a school is private, public, or church-related is fundamental to its position in the market and highly related to its other characteristics. For example, public schools usually cannot legally impose stringent admissions standards. Stratification of both the naming schools and the competitors is used in the test of the second hypothesis. In the few instances where this more elaborate procedure affected the conclusions drawn from the simpler one, it is noted in the text.

Ideally, one would want to assess the distribution of the competition on some attribute by comparing it with the chance distribution represented by the distribution of the universe of schools on that attribute. However, the requisite data are not available for the universe. Consequently, chance distributions are given that are equivalent to the distribution on an attribute of all competitors named; these distributions represent the distribution one would expect if knowledge of a naming school's value on some attribute provided no advantage for the prediction of a competitor's value on that attribute.

FINDINGS

Cost and Region

Cost[7] and region – to the extent that the latter represents travel and extra away-from-home living costs – are organizational attributes that cause students to self-select themselves in terms of their financial resources. Of all the factors affecting college selection, cost of attendance is clearly the most fundamental. Although this factor is not always ranked as the most important in studies of college choice, Rossi et al. (1964: 97) suggest this is due to the fact that students from low socio-economic backgrounds appear to consider only colleges that are low cost and close to home.[8] Thus, one would expect students with limited resources to apply 'to inexpensive schools in their own section of the country. Because the alternatives available to students increase with their resources, one might additionally expect that the range of schools applied to with respect to cost and location would also increase. But there are countervailing forces. The expensive schools include many private schools of high quality and national reputation; an applicant to these schools typically has made his decision on noneconomic grounds. Consequently, one may expect to find more competition among these institutions than a simple economic argument would imply.

Table 1 shows that schools clearly find the majority of their competition within their own cost category. The concentration of competition is slightly greater among the high- rather than low-cost schools, perhaps suggestive of the potency of the images of the schools in this category.

TABLE 1

Percent of Competition with Same Cost as Naming Schools*

Naming Schools

Competitors

/ Less than $1000 / $1000-$1999 / $2000 or more / Expected Distinction
Less than $1000 / 66 / 27 / 5 / 36
$1000-$1999 / 30 / 55 / 24 / 42
$2000 or more / 4 / 18 / 71 / 22
Total / 100 / 100 / 100 / 100
N= / (350) / (581) / (174) / (1105)

χ2 = 441.5, p<.001

* Table tabulates competitors named; table totals vary because of missing data.

Region also has noneconomic meanings such as climate, proximity to friends and relatives, and distance to established patterns of social life (e.g., the New Yorker who cannot bear to be far from the city). Although there is some desire to "go away" to college, most students do not go very far. In part this is due to the fact that, except for the few schools of established national reputation, students' images of schools are likely to be weaker as their distance from them increases.

Table 2 shows the relationship between the region of the naming schools and the region of their competitors, controlling for cost. In only two instances is the percentage of competition within the same region less than 89 percent, and both exceptions are among the high-cost schools. At this economic level, almost half of the competition of southern schools and better than one-third of the competition of mid-western schools are in the Northeast. This is partially explainable by the location of 58 percent of the expensive schools in the Northeast, but it is also due to the widespread knowledge of a number of these schools outside their region. Note that in none of the subtables is the expected distribution remotely similar to the observed distributions.

TABLE 2

Percent of Competition in Same Region as Naming Schools by Naming Schools' Cost and Region*

Naming School
Less than $1000 / $1000 - $1999 / $2000 or more
Compe-titors / NE / S / MW / W / Expected Dist'n / NE / S / MW / W / Expected Dist'n / NE / S / MW / W / Expected Dist'n
NE / 98 / 2 / 14 / 91 / 8 / 6 / 2 / 34 / 94 / 48 / 35 / (7) / 69
S / 2 / 97 / 1 / 2 / 45 / 8 / 89 / 1 / 21 / 3 / 49 / 10
MW / 2 / 96 / 2 / 27 / 2 / 2 / 92 / 6 / 36 / 1 / 3 / 65 / 11
W / 1 / 1 / 96 / 14 / 2 / 93 / 9 / 2 / (93) / 10
Total / 100 / 100 / 100 / 100 / 100| / 101 / 99 / 101 / 101 / 100 / 100 / 100 / 100 / (100) / 100
N= / (49) / (161) / (96) / (44) / (350) / (192) / (118) / (216) / (55) / (581) / (100) / (31) / (28) / (15) / (174)

χ2=940.9, p<.001 χ2= 1373.6, p<.001χ2=282.8, p<.001