Understanding University Choice: a Multi-Attribute Approach

Ronald Vaughn, Bradley University
Joseph Pitlik, Bradley University
Behram Hansotia, Bradley University
ABSTRACT - This article presents a detailed analysis of the components of a university's image. University choice criteria are studied as to their relative determinance (using a new modified determinant attribute approach). The underlying cognitive structure of university choice is also investigated via factor analysis.
[ to cite ]:
Ronald Vaughn, Joseph Pitlik, and Behram Hansotia (1978) ,"Understanding University Choice: a Multi-Attribute Approach", in NA - Advances in Consumer Research Volume 05, eds. Kent Hunt, Ann Abor, MI : Association for Consumer Research, Pages: 26-31.

Advances in Consumer Research Volume 5, 1978      Pages 26-31


Ronald Vaughn, Bradley University

Joseph Pitlik, Bradley University

Behram Hansotia, Bradley University

[Ronald Vaughn, Joseph Pitlik and Behram Hansotia are, respectively: Associate Professor in the Department of Marketing; Instructor in the Department of Marketing; and Associate Professor in the Department of Business Management and Administration. All three professors are at Bradley University.]


This article presents a detailed analysis of the components of a university's image. University choice criteria are studied as to their relative determinance (using a new modified determinant attribute approach). The underlying cognitive structure of university choice is also investigated via factor analysis.


The need for the application of marketing theory to college admissions work has become apparent in recent years as the market has changed from the sellers market of the 60's to a buyers market for the 70's. Available information at present seems to indicate that the market for prospective college students will remain a buyers market into the 1990's and perhaps beyond.

Several factors external to the university system have contributed to this shift. The pool of applicants available from high school classes has peaked with the graduation class of 1970 and is expected to decline steadily through 1983. Furthermore, the abolishment of the draft and the ratio of college graduates to job openings in most academic majors appear to be vital forces in the prospective students initial decision as to whether to attend college. In the present competitive environment the four year institution now faces several factors that were almost nonexistent in the 1950's. The early 60's were characterized by the rapid development of community colleges, and closely followed by the development of technical programs offered at two year institutions. These two factors taken together have further reduced the pool of applicants or market potential available. The private university has yet another environmental threat. The large state university, long dormant in aggressively recruiting students, has awakened in recent years and appears to becoming active in its search for students to meet enrollment objectives.

The implication is clear. To successfully compete in the educational marketplace of the present and future, universities, especially private universities, must begin refining their marketing strategies. These refinements will often be in the form of tactical shifts in the promotion and product areas, and a sound base of research is almost a prerequisite to guide these strategies and tactical changes.


Image research has had wide use in marketing generally, e.g., company and brand studies, and in numerous store imagery studies. In fact, since 1960, more than forty image-related articles have appeared in the marketing literature, more than half of these in the last five years. Further, several environmental trends seem to dictate that this type of research will intensify in the future due to the increasing sophistication of the shopper (including the educational shopper), increased shopper mobility, product proliferation, and other competitive pressures. Just as retailers are facing an intensification of inter-type competition which is causing retailers to combat retail sectors whose merchandise had never before competed with theirs, so to higher education's already declining market is being impinged upon by proprietary and industrial schools. Not only does IBM, General Electric and other corporations now offer bachelor's degrees, but there is increased competitive pressures from community colleges and other more traditional colleges. An article by Kotler and Levy (1969) point out that marketing is a pervasive societal activity and that student recruitment by colleges is even a marketing activity.

Despite the apparent need for imagery research on nonprofit educational institutions and despite the wealth of imagery research on the profit-oriented business sector there has been a paucity of research dealing with the marketing problems of a university and hardly any research specifically investigating the image of a university. Although Kotler and Levy (1969) noted much earlier than nonprofit organizations have been typically ignored by the student of marketing, most of the university-oriented marketing literature to-date has been quite general and largely conceptual (e.g., Kotler 1976, O'Brian 1973, Berry and George 1975, etc.), except for a very limited study by MacLachlan and Leister (1975) which used multidimensional scaling to develop perceptual maps of the institutional image and a more extensive study of a graduate school's image by Naidu (1969).

This paper attempts to provide insight on the image of a university. Specifically, evidence will be contributed on some of these issues. Can a university's image be measured along salient image dimensions? If so, which attributes are most important? Do determinant attributes exist? How is the information about a center structurally organized, i.e., rather than just describing a university as very good or bad, how is information about a university coded and grouped into cognitive categories? Studying these relatively new issues to educational choice behavior should provide insight as to how educational institutions are evaluated with respect to each other. Finally, understanding imagery development and the choice process in a university setting should serve as a prerequisite for developing better diagnostic and predictive models. While this paper has not been intended as a dissertation on the usage and development of multi-attribute models, a number of reviews have been done which would be quite helpful in later attempts to model the educational choice process, e.g., Cohen (1972), Hansen (1976), and Pessemier and Wilkie (1972). The conceptual article by Kotler (1976) would also be quite helpful in developing more explicit models of educational choice. Towards this objective, a variety of issues and directions for future research are provided in the final section of this paper.


To determine an initial set of criteria that represents the salient components of university preference a preliminary study was conducted. Freshman College of Business Administration students at Bradley University were asked to give the five most important criteria they used in evaluating a university. This questioning process produced an extensive list of criteria that was further modified through discussions with the admissions director and associated admissions personnel plus a review of the literature. A final set of 16 criteria, shown in Table 1 was utilized for further study. While not exhaustive, these criteria are thought to represent those criteria most commonly used by prospective students and their parents. A questionnaire was then designed to determine the relative importance of these criteria, to both parents and students. Respondents were asked to rate in importance each criteria on a seven point scale where a "7" was extremely important. Respondents were also asked to identify and then evaluate each of their top three choices along the same criteria on a scale from 1-7 where a "1" was a poor evaluation and a "7" was a good evaluation.

In January of 1975 402 questionnaires were mailed to all students and parents of students who were admitted to the College of Business Administration for the Fall of 1974. The 30% response rate is primarily attributed to the rather low (16%) response rate in the category of "Parents of Students Admitted, but not Enrolled." However, the factors of a lengthy questionnaire and no monetary incentive provided for the respondent also probably influenced the low response rate. Due to the limited sample, the results presented are somewhat tentative and additional research is needed.




One of the most complex processes the student and his parents go through is in making the final choice among colleges that have accepted him or her. In this final choice process the student is believed to rate each college's characteristics on the criteria they hold to be most important. Thus, basic to developing proper college marketing procedure is an understanding of the university choice process, which, in turn, must be grounded in a thorough analysis of what attributes of a college are evaluated. We will attempt to provide insight to the university evaluation process by several analysis stages. The first stage of analysis will determine the relative importance of sixteen university choice criteria. Second, the evaluation of several alternative universities on each attribute will be compared. Third, selection criteria will be evaluated for relative determinancy, and fourth, the underlying cognitive structure of university choice will be explored using factor analysis.

Relative Importance of University Selection Criteria

Sixteen choice variables were selected for importance analysis. Table 1 shows that the variable means ranged from a low of 4.060 for the influence of friends who are attending the school, to a high of 6.449 for the quality of education received. Furthermore, the typical respondent put a premium on the quality of faculty, reputation of the business programs, academic reputation of the university, amount of individual assistance that could be provided by the faculty, number and variety of courses offered, the number of students per class and so on. These findings appear somewhat consistent with those of Holland (1958 and 1959) in his studies of what aspects of a college made it "best." Using a sample of parents of National Merit Scholars they found the key aspects to be quality of faculty, scholastic standards, curriculum, reputation and facilities. However, in the actual choice of college "practical" and "financial" factors were of substantial influence with academic factors playing a moderate, secondary role. Knowledge as to which attributes are most important undoubtedly suggests that certain educational attributes must be particularly considered for attribute development, maintenance and promotional activities.

Institutional Evaluation

The marketing task of a college is to articulate a mission and purpose for the college that makes sense in view of its history, resources, opportunities, competition and other situational variables (Kotler, 1976). Further, a coordinated marketing effort must assist in developing a distinct posture for the university, i.e., an institutional position that will meet the educational needs and desires of its particular market segment as a means to satisfying institutional goals. With this in mind, one of the first steps in developing an institutional position is to measure and evaluate the university's current perceived position in the market in relation to competitors. While knowledge of which attributes are most important provides some direction for faculty and administrative efforts at developing and maintaining its position, periodic assessment of the university's current perceived position provides additional insight to institutional strengths and weaknesses and suggests areas of the educational product where improvement or change might be necessary. As Gould (1971) points out, constructive change by a college must be preceded by proper introspection: "higher education must be more truthful in its own evaluation of its current weaknesses in order to portray itself effectively to its many constituencies."

Stage two of the analysis was to evaluate Bradley University's image on each of sixteen criteria and compare it to the perceived image of several competing colleges (grouped into an other private and a state school classification). While it was expected due to the nature of the sample (people who had applied and been admitted to Bradley University), Bradley had a more positive evaluation than its competitors on most of the attributes. It should be noted that other publics of the university if sampled might have provided different average evaluations. Table 1 shows some attributes on which Bradley seems to be viewed as quite distinct while on others there seems to be an insignificant perceived difference. Bradley seems to be viewed as relatively different from its competitors (in a positive sense) on size of the student population, class size, college displayed a personal interest, the amount of individual assistance provided by the faculty, campus size, housing facilities and reputation of the business program. Bradley was viewed as quite different from its competitors in a negative sense on only the basic cost of attending. Only a slight variation among the evaluation seemed to exist on the influence of friends attending, academic reputation of the university, quality of the faculty, quality of the education received and availability of financial aid.

While this study of perceived evaluations by attributes is helpful in pinpointing relative strengths and weaknesses, additional information is needed. Although a university is perceived as quite different in a positive sense from its competitors (on a given attribute), a relative strength does not necessarily exist unless that attribute happens to be of considerable importance to the educational consumer. Likewise, the fact that a university is viewed more poorly is not necessarily cause for alarm unless the criterion happens to be of crucial importance. Thus, this analysis suggests that a university should first, discover the relative importance of the various university selection criteria; second, investigate the image of themselves and their competitors on these same criteria; and, third, analyze the importance and evaluation scores in combination. The analysis of determinant attributes presented in the next section illustrates a way of analyzing attribute importance and evaluation together.

Determinant Attributes

The third stage of the analysis was to assess the principal university decision variables and their relative determinance in university selection decisions. Determinant attribute analysis should help to understand those features which move the educational consumer to action, i.e., to make him prefer a particular university, actually attend there and/or recommend it to friends. Because of their crucial relationship to preference, a universities perceived position with respect to these determinant attributes must be closely investigated to check for strengths, weaknesses, opportunities, etc.

Attribute determinance was estimated by evaluating (1) the importance associated with a particular university selection criterion and (2) the degree to which competing universities were perceived to differ in terms of that selection criterion (Myers and Alpert, 19 8). While these two aspects of a determinant criterion are part of the original definition of "determinance," the operationalization of this concept represents a new approach. Using Hansen's (1977) "modified determinant attribute analysis" each individual's response on a given attribute was cross-tabulated based on their response to each of the two questions (i.e., importance and perceived variation). Table 2 presents a sample of such a cross-classification for a product attribute, where the table cell values are reported as percentages. Hansen (1977) presents a critical evaluation of this new operationalization compared to the more traditional operationalization of attribute determinance. Using the Myers and Alpert definition of determinance, the attribute having the highest proportion of respondents classified in cell 9 of Table 2 where both importance and variation are high, would be the currently most determinant analysis was based on the present or current perceptions of the product. Cell 7 of Table 2 is also of interest as it presents the proportion of the respondents who indicated the attribute was high in importance, but low in perceived inter-university variation. This cell 7 is referred to as "potential" determinance because if one of the competing universities could produce an educational product that is perceived or evaluated much higher than its competitors, then that university would have a distinct advantage. This is because the current high importance-low variation attributes may offer the greatest potential for improvement since an increase in a university's evaluation on that dimension would lead to a distinct perceived difference on a "potentially" crucial or determinant choice criteria. The other cell values in Table 2, e.g., 6, 5, etc. have similar implications in that strategies may be suggested to increase the university's evaluation, or to increase the perceived importance of an attribute, or both. The remainder of the analysis, however, concentrates on cell 7 and cell 9.





The results of the determinant attribute analysis are shown in Table 3. Several attributes stand out as currently relatively determinant: college displayed a personal interest (22.81%); size in terms of the college's student population (20.86%); campus size (19.83%); and, amount of individual assistance that could be provided by the faculty (19.30%). The potentially determinant university choice criteria include: academic reputation of the university (63.94%); quality of the faculty (63.82%); reputation of the business programs (55.46%); and, the number and variety of courses offered (46.22%). It must be noted, however, that the variables that came out as currently or potentially determinant in this sample are quite possibly not determinant or crucial to the selection process of other universities. Remember that the sample being studied in this case was students who had applied to and were admitted to the College of Business Administration at Bradley University (and the majority of these students actually enrolled). Thus, the nature of the sample suggests that the currently determinant criteria are more likely to be those variables that were crucial to the decision to attend Bradley. The fact that Bradley is evaluated much higher than its competitors in Table 1 lends additional credibility to this interpretation. On the other hand, the potential determinant criteria suggest attributes that could contribute more to Bradley being chosen if the perceived evaluation of Bradley could be improved to the point where it was viewed as quite different from its competitors. Table 1 also lends credibility to this interpretation as the evaluations of Bradley University on the potentially determinant criteria show relatively little difference from the evaluations of its competitors.

Cognitive Structure of the University

In stage four of the analysis a factoring procedure was used to determine the underlying constructs or dimensions which the respondents considered when selecting a university. The same 16 selection criteria listed in Table 1 were the subject of the factor analysis. PA2, the iterative procedure for determining communalities, was used from the SPSS package. In 23 iterations, five factors with eigenvalues greater than one were extracted. The fifth factor explained only 6.4% of the total variance and since after the varimax rotation the fifth factor had no significantly high factor loadings on it, we decided to drop the fifth factor. An "L" test as described in Van de Geer (1971) also confirmed our dropping the fifth factor, which had an eigenvalue of only 1.016.

The first four factors explained 56.3% of the total variance. Reported below is the matrix of factor loadings after the varimax rotation. The asterisks (*) adjacent to the higher factor loadings are placed to help identify the important attributes of each salient university selection dimension. Factor I, for example, is loaded heavily on variables 2, 5, 3, 4, and 1, which are: quality of faculty, amount of individual assistance provided by faculty, reputation of the College of Business Program, academic reputation of the university, and quality of education received. Apparently the first dimension is "university and College of Business Program quality." Interestingly, the respondents do not seem to look at only the quality of the Business program but also at the quality of the entire university. Furthermore, they seem to relate the amount of individual assistance to quality.



The second factor is loaded heaviest on variables 8, 13, and 7; which are size in terms of the college's student population, campus size, and size in terms of the number of students per class. Obviously, this is a "size" dimension.

The third factor is loaded on variables 10 and 15, which are basic cost of attending and availability of financial aid. This seems quite clearly to be a "total university cost" dimension. Thus, the respondents do not seem to look at just the cost of attending a university, but at the total cost obtained by offsetting tuition and housing costs by availability of financial aid.

The fourth and last factor has a high loading on only the eleventh variable, namely location. Hence, it can be labeled a "location" dimension. Note that of the 16 variables originally selected for the analysis (see Table 4), 5 did not have appreciably high loadings on any of the four factors. They are variables 6, 9, 12, 14, and 16. Of these, variablesl2 and 16, namely job placement service after graduation and influence of friends who are attending, are least correlated to any of the four factors discussed above. Variables 6, 14 and 9 on the other hand do have modest correlations with factors 1, 2 and 3 respectively. Implying that variable 6 which is number and variety of courses offered is moderately correlated with the "quality" dimension. Variable 14, which is housing facilities, is moderately correlated with the "size" dimension and variable 9, college displayed a personal interest, is correlated with the "total cost" dimension.

Table 5 summarizes our findings in this section.




To continue to improve our understanding of the university choice process, research is needed on many issues, including the following:

1. How do university choice criteria differ between the parent (an important decision influencer) and the student? Are there differences in attribute saliency or determinancy between the student and the parent? Further, are there such differences for students who apply and do not enroll versus those who do enroll, i.e., users versus nonusers. Also, are there differences in saliency or determinancy for students of differing quality, i.e., for students with very promising academic potential versus average academic potential (Naidu suggested in 1969 that due to a greater number of alternatives, financial aid is more important to attracting students of high quality)?

2. What image does the university have among its other important publics, e.g., alumni, donors, high school and junior college counselors, and the community at large? As Shapiro (1972) suggests an institution such as a college is concerned with two primary sets of transactions or ex- changes, one is with donors and the other is with clients and for a college highly dependent on external funding sources the donor group may be at least as important as the client group (students) to study and understand.

3. Do university selection criteria change (i.e., are there new criteria) after the prospective student enrolls and attends the college for some time? Perhaps more subtly do criteria saliences or evaluations change through time? A study by Naidu (1969) suggests that the "attractiveness" of graduate education varies from level to level, e.g., he notes that financial aid and faculty reputation are considered to be more important by the student at graduate levels, but at the under- graduate levels location of the campus and general reputation of the university seemed more important in evaluating the educational product. It may be particularly important to understand the differential needs of seniors so that they will graduate as satisfied customers. Also, a more thorough analysis of the needs or disconfirmed expectancies of sophomores and juniors may help reduce the problem of retaining students for four years.

4. Are the perceptions of university faculty, administrators and managers consistent with the perceptions of their publics? As the images held by collegiate personnel help to determine the university's projected image to other publics, this aspect of image may also be of real concern.

5. What are the key information sources which students and other publics use, and which are the basis for attitude formulation?

6. Can the educational market be meaningfully segmented? Just as in other businesses, a school needs to profile its perhaps rather small potential market as a guide to effective and efficient allocation of resources.

7. Once a university has assessed its current perceived position, what strategies and tactics should it consider? Kotler (1976) suggests that such strategies and tactics go beyond just the work performed by the admissions office, but should include such activities as institutional positioning, portfolio planning, applicant development work, applicant evaluation and notification work, recruitment effort evaluation, college improvement planning, and alumni loyalty work. Little empirical work has been done to determine the effectiveness of different strategies or ways of implementing them. Likewise, the differential sensitivity of different market segments to certain marketing strategies or tactics has had little focus by marketing people.

In summary, while there has been a considerable amount of research focused on nonprofit organizations by marketing people in the past 10 years (as illustrated by Rothschild's (1976) 43-page bibliography of works relating to marketing for public sector and nonprofit organizations), there has been a lack of focus on the educational sector, which is one of the largest and most important service industries in the U.S. It is hoped that the reported research study and suggestions, along with consideration of the benefits that may accrue directly to academicians, will prompt a more active interest in this area by marketing researchers.


Berry, Leonard L. and William R. George. "Marketing the University: Opportunity in an Era of Crisis," Atlanta Economic Review, 25 (4) (July- August 1975), 4-8.

Cohen, Joel B. "Toward an Integrated Use of Expectancy-Value Attitude Models," Paper presented at the ACR/AMA Workshop on Buyer/Consumer Information Processing, University of Chicago, 1972.

Gould, Samuel B. "Bridging the Interpretation Gap," College and University Journal, 10 (May 1971), 14- 16.

Hansen, Flemming. "Psychological Theories of Consumer Choice," Journal of Consumer Research, 3 (December 1976), 117-142.

Hansen, Robert A. "Determinant Attribute Analysis and the Search for Product Modification Ideas: Theoretical and Managerial Considerations," an unpublished working paper, Ohio State University, 1977.

Hie, Norman, et al. Statistical Package for the Social Sciences (SPSS). New York: McGraw-Hill, Inc., 1975.

Holland, John L. "Parental Expectations and Attitudes about College," College and University, 34 (Winter 1959), 70.

Holland, John L. "Student Explanations of College Choice and Their Relation to College Popularity, College Productivity, and Sex Differences," College and University, 33 (Spring 1958), 313-20.

Kotler, Philip and Sidney J. Levy. "Broadening the Concept of Marketing," Journal of Marketing, 33 (January 1969), 10-15.

Kotler, Philip. "Marketing Theory and Its Application in College Admissions," paper presented at the Colloquium on College Admissions, The Abbey, Lake Geneva, Wisconsin, May 1976.

MacLachlan, Douglas L. and Douglas V. Leister. "Institutional Positioning: A Metamarketing Application,'' Marketing's Contribution to the Firm and to the Society, ed. R. C. Curham. Chicago, American Marketing Association, 1975.

Myers, James H. and Mark I. Alpert. "Determinant Buying Attitudes: Meaning and Measurement," Journal of Marketing, 32 (October 1968), 13-20.

Naidu, Gurramkonda M. Systems Approach to the Marketing Aspects of Higher Education, an unpublished dissertation at Michigan State University, 1969.

O'Brian, Edward J. "Marketing Higher Education," College and University Journal, 12 (4) (1973), 22- 23.

Pessemier, Edgar A. and William L. Wilkie. "Multi- Attribute Choice Theory--A Review and Analysis," Institute Paper No. 372, Institute for Research in the Behavioral, Economic and Management Sciences, Purdue University, 1972.

Rothschild, Michael L. "An Incomplete Bibliography of Works Relating to Marketing for Public Sector and Nonprofit Organizations," an unpublished working paper, University of Wisconsin--Madison, January 1976.

Shapiro, Benson P. "Marketing in Nonprofit Organizations,'' an unpublished working paper, Marketing Science Institute, 1972.

Van de Geer, John P. Introduction to Multivariate Analysis for the Social Sciences. San Francisco: W. H. Freeman and Company, 1971.