Cognitive Structure and Information Search Patterns of Prospective Graduate Business Students

ABSTRACT - A study examining two key aspects of the decision processes used by prospective graduate students is reported. Cognitive structure and the salience of its underlying dimensions to various key segments of the graduate student market are analyzed. Information search efforts and overall patterns of source usage are examined within the same segments. Tentative implications for the marketing effort of graduate business schools are drawn.


Michael J. Houston (1980) ,"Cognitive Structure and Information Search Patterns of Prospective Graduate Business Students", in NA - Advances in Consumer Research Volume 07, eds. Jerry C. Olson, Ann Abor, MI : Association for Consumer Research, Pages: 552-557.

Advances in Consumer Research Volume 7, 1980     Pages 552-557


Michael J. Houston, University of Wisconsin-Madison


A study examining two key aspects of the decision processes used by prospective graduate students is reported. Cognitive structure and the salience of its underlying dimensions to various key segments of the graduate student market are analyzed. Information search efforts and overall patterns of source usage are examined within the same segments. Tentative implications for the marketing effort of graduate business schools are drawn.


As a result of significant shifts in the demographic nature of the population, institutions of higher education are approaching a period of declining enrollments. Aggregate demand for higher education on the part of traditional consumers of higher education, while at peak levels in recent years, is expected to decline significantly in the 1980's but return to higher levels in the 1990's. While in the long run it appears that the higher education industry will maintain a healthy state of demand from society, the next decade is a crucial one for many individual colleges and universities and educational programs within them. Their ability to maintain satisfactory enrollment levels during this period will determine the strength with which they enter the period of increased demand in the subsequent decade and their ability to reap the attendant benefits. To some extent it will even determine, especially for individual programs within an institution, their ultimate survival.

Two implications are evident from the above. First, much like firms who compete for increased market shares in an industry where primary demand for its products is declining, more vigorous competition between different institutions and between different departments within an institution can be expected in the traditional demand segment of the student market. Second, attempts to stimulate demand for the university's services from non-traditional segments will become more prevalent.

Each of these implications suggests a stronger marketing orientation on the part of universities and individual departments within them. The recognition of the need for a marketing orientation on the part of educational institutions is certainly not unique to this paper. The recent emergence of marketing applications in the nonprofit sector (Kotler 1976) has included the higher education industry (Berry and Allen 1977; Hugstad 1975; Krachenberg 1972). The next decade presents perhaps the greatest challenge to date to university administrators in effectively implementing marketing strategies.

The probability of success in any marketing endeavor is enhanced considerably when sound research on its target market serves as input to the formulation of marketing strategy. The marketing of universities is, by no means, an exception. Understanding consumer behavior as it relates to university enrollment decisions can provide valuable input to the formulation of marketing strategies directed at traditional student markets. Furthermore, the extent to which a university or department seeks enrollment from nontraditional segments calls for a segmentation approach whereby behavioral characteristics unique to these segments and any resulting marketing consequences are ascertained.

The emerging process orientation to the study of consumer behavior (see, for example, Bettman 1979) offers a useful perspective to assume in examining consumer behavior with respect to higher education and in developing marketing strategy. In such an approach the focus is on processes leading to a choice rather than the choice itself. Interest lies in cognitive structure and decision criteria used, information search and sources utilized, and integrative processes by which utility estimates are made, to name a few. While an explanation of the ultimate choice made by a student is certainly useful input, knowledge of the processes leading up to that choice offers the administrator an excellent basis for determining program components and the means by which such information can be most effectively communicated to prospective students.

In spite of the increased attention to applied marketing issues in the higher education arena, little is known about university choice processes employed by either traditional or nontraditional students. The general purpose of this paper is to report a study that examined the cognitive structure of students engaged in the choice of a graduate business program and certain aspects of their information search process. Particular attention is given to differences across certain market segments.


The study reported here was undertaken as part of an overall project designed to upgrade the marketing of the Masters-level programs in business at a major state university. Interest focused on two major categories of dependent variables: cognitive structure of students engaged in the choice of a graduate school and external information search patterns used by these students. Each category of dependent variables was analyzed with respect to several independent variables, each representing a potential basis for segmenting the market.

Dependent Variables

Cognitive Structure.  The cognitive structure of an individual in a particular choice situation contains a variety of components. Included are the nature and scope of the criteria used to evaluate choice objects and any existing beliefs regarding the choice objects. Cognitive structure can be further characterized in terms of its size and complexity. The focus in this study is on the nature and scope of salient criteria used by graduate business students in choosing a university.

Previous research on university choice decisions by undergraduates (Vaughn, Pitlik, and Hansotia 1978) suggests that prospective students employ a rather complex multi-attribute framework for evaluating schools. In their research Vaughn et al. (1978) identified four broad dimensions of cognitive structure from attribute importance scores on several specific evaluative criteria. The four components of cognitive structure included university/business program quality, size, cost, and location. In perhaps the only study of graduate business school choice, Punj and Staelin (1978), while not studying cognitive structure or decision making processes, included similar variables in a predictive model of ultimate decision outcome.

Accounting for differences in graduate and undergraduates, a similar approach to that of Vaughn et al. (1978) was taken in this study. Attribute importance was examined for a series of specific evaluative criteria (see Table 2). The attribute importance by subjects matrix was then reduced via factor analysis to identify the underlying dimensions of cognitive structure.

Information Search.  An aspect of the decision process for university choice that has been virtually ignored in research to date is the external information search process. In this study certain features of the information search process used by prospective graduate business students were examined. These included measures of aggregate search effort, i.e., number of schools seriously considered and number of information sources consulted. Measures of patterns of search within the external information environment were also examined. These included measures of the extent of usage of each of several sources of information. To determine the existence of any dominant overall patterns of information search the source usage by subjects matrix was reduced via factor analysis.

Segmentation Variables

To ascertain any differences in the dependent variables between members of certain segments of the potential student market measures of several variables relating to student quality and nontraditional students were included. These variables and the rationale for their inclusion are discussed below. The extent that marketing consequences are associated with any observed differences between segments formed from these variables can reveal the need for specialized programs directed at a particular segment.

Student quality.  While many schools are looking towards nontraditional segments to attract new students, virtually every graduate business school will maintain an interest in attracting the highest quality students. With the numbers of such students declining, competition for these students will become even keener. Therefore, to determine if higher-quality students exhibited any differences in cognitive structure or information search, two measures of student quality were included: undergraduate grade-point average and percentile of performance level on the Graduate Management Admission Test (GMAT).

Sex.  While men and women have been attending universities together for many years now, the influx of women into the historically male-dominated domain of graduate business schools is a relatively recent phenomenon. As a result, many schools may be depending on the nontraditional female segment to offset enrollment declines in the male segment. Thus, sex is included as a key segmentation variable in this study.

Undergraduate Major.  As the job outlook for undergraduates in such areas as the social sciences, humanities, and education grows dimmer, more and more of these students are turning towards graduate business schools to further their education and enhance their job opportunities. Many of these students are of a nontraditional nature. They include students who at one time may have been negative towards a business profession, are seeking management skills to apply in the nonprofit sector, or in previous years would have pursued a career in the now oversupplied area of law. Therefore, undergraduate major is included as a segmentation variable to examine differences between these nontraditional students and the more traditional ones from such undergraduate programs as engineering and business.

Work Experience.  A characteristic of a student recognized by most administrators and faculty as an extremely desirable one is previous work experience. The student who can bring prior work experience to the classroom is generally considered a high-quality student. In fact, some schools require such experience as a condition for admission, while others use it as one of a set of criteria. For many schools such a student also takes on a nontraditional nature and is represented in small proportions within their graduate student bodies. These schools may be developing specialized efforts targeted at such students to increase enrollment. Therefore, as an index representing a combination of student quality and nontraditional students, prior work experience was included as a segmentation variable.


Subjects and Data Collection

The subjects for the study included all students accepted into a Masters program in business for the Fall 1977 semester at a major state university. In July of 1977 self-administered questionnaires were mailed to each of 371 accepted students. Approximately 10 days later a follow-up mailing to the entire sample occurred. A total of 240 questionnaires were returned for a response rate of 64.7%. The greatest fear regarding non-response bias was that a disproportionately large number of students deciding to attend the university would return the questionnaire, leaving those deciding to attend elsewhere underrepresented. These fears were not realized as official figures for the semester revealed statistically equivalent probabilities of enrollment for respondents and nonrespondents.

Measurement Procedures

Dependent Variables.  A set of 22 specific attributes, generated from the literature, intuition, and exploratory interviews with students already enrolled in a Masters program at the university, were included. The importance of each attribute was measured on a three-point scale ranging from "not at all important" to "very important." To determine the cognitive structure underlying the decision process attribute importance scores were factor analyzed using principal components analysis. Factor scores were computed to determine the salience to each individual of each dimension of cognitive structure.

Information search procedures were measured in several ways. First, to determine the number of schools considered, i.e., evoked set size, each subject was asked to list in open-end fashion the schools that he or she seriously considered. Each subject was then presented with a set of 13 external information sources (generated from the same procedures used for attributes) and asked to indicate on a three-point scale ranging from "not at all" to "quite a lot" how much each source was used to obtain comparative information on the schools being considered.

Aggregate search effort was measured by the number of sources for which the subject indicated "quite a lot." Patterns of source usage were determined through a factor analysis of individual source usage scores. The extent to which a subject exhibited each overall pattern of search was determined through factor scores for the extracted factors.

Segmentation Variables.  The segmentation variables were obtained through self-reports on questions on the survey. For each variable the following segments were formed:

1.  Undergraduate grade-point--Subjects were placed in one of five categories to represent grade-point: 3.0 or less, 3.01-3.25, 3.26-3.50, 3.51-3.75, and 3.76-4.00.

2.  GMAT percentile--Subjects were first split into two groups: those falling into the upper one-third of scores of all individuals taking the GMAT and those performing below that level. The higher group was then further subdivided into three groups. Subjects were thus placed in one of four segments: 67th percentile or less, 68-79, 80-89, and 90 or above.

3.  Undergraduate major--Based on the distribution of responses to a question concerning undergraduate major, the following combinations of majors represented different segments: marketing/management (MKTMGT), accounting/finance (ACTFIN), technical areas, e.g., math, engineering (TECH), humanities/ education (HUMED), social sciences (SOCSCI), and natural sciences (NATSCI).

4  Work experience--Subjects indicating they had at least one year of fulltime work experience subsequent to their undergraduate degree represented one segment, while subjects with less than one year of work experience represented a second segment.

The levels of each segmentation variable and the sample sizes associated with each level are summarized in Table 1.



Data Analysis

Means were computed for importance scores for each attribute, usage scores for each individual source, evoked set size, and number of sources consulted. Separate analyses for each segmentation variable were conducted. Comparisons between different segments were made using one-way analysis of variance (ANOVA). Dependent variables in the segmentation analyses included:

1.  Salience of each underlying dimension of cognitive structure, as measured by factor scores and sums of raw scores on variables loading on an extracted dimension;

2.  Aggregate search effort, as measured by evoked set size and number of sources used;

3.  Extensiveness of use of overall patterns of source usage, as measured by factor scores and sums of raw scores on variables representing each overall pattern.


Cognitive Structure

In Table 2 the results of the analysis of attribute importance are presented. Attributes are listed in descending order of mean importance for the overall sample. Factor loadings derived from a varimax rotation are presented along with percentage of variance explained and eigenvalues for each factor.



The specific attributes valued highly by prospective students seem to relate to two major aspects of a business school. First, with importance attached to job potential and the reputation of the university, business school, and department, reputation/prestige appears to be a dominant consideration in the decision process. Second, program-related issues, i.e., the opportunity to specialize, program flexibility, number of electives, are a major consideration.

Such an interpretation is confirmed when the dimensions of cognitive structure are inferred from the factor analysis results. Using standard criteria to arrive at a factor matrix (i.e., eigenvalues exceeding one), five meaningful factors were extracted to represent major components of cognitive structure. The minimum cutoff used for inclusion of a variable in the domain of a factor was a factor loading of .30. In the one instance where a variable loaded above .30 on two different factors, it was retained in the factor on which it loaded the heaviest rather than eliminated because of the wide discrepancy between the loadings (.61 vs. .34). The interpretation of the five factors suggests the following components of cognitive structure:

1.  Factor I--Cost (tuition, financial aid, costs of living)

2.  Factor II--Size (large business school, large university)

3.  Factor III--Nature of Program (flexibility, electives allowed, required courses, length for business undergraduates)

4.  Factor IV--Reputation/Prestige (business school, university, entrance requirements)

5.  Factor V--Nature and Reputation of Major Area (general reputation, research reputation, quantitative and behavioral emphases)

Segmentation Analyses of Cognitive Structure

As indicated, differences between segments in terms of the salience of each dimension of cognitive structure were analyzed using factor scores and sums of raw scores based on the attributes included in the domain of a factor. Factor scores and raw sums generated identical findings. Thus, in Table 3, which summarizes the findings, means are presented in terms of raw sums for ease of interpretation. However the F-values and their corresponding significance levels are those obtained from the analysis of factor scores.

Discussion of Cognitive Structure Findings

The complexity of cognitive structure in the choice process for graduate business schools is, not surprisingly, greater than that observed by Vaughn et al. (1978) in their study of undergraduates. Five components make up the structure of graduate students, while Vaughn et al. observed only four. Furthermore, the nature of the structures are different to some extent. While there is overlap with respect to the dimensions relating to overall prestige, cost, and size, graduate students also are sensitive to matters more internal to the university and the business school. Specifically, they deal with issues relating to program format and the reputation and nature of the major department. Location is not a consideration as it is with undergraduates. The marketing implications of these findings are clear. Graduate programs must be very precise in the information they communicate to prospective students. The nature of the university as well as specific information concerning program structure and the nature of specific departments within the business school must be communicated to students. Potential markets do not appear to be geographically restricted.

Differences in the salience of some of the components are evident in the segmentation analysis. Cost is a greater consideration to students with higher grade points, perhaps because they are more expectant of financial aid. Women are also more concerned with cost, perhaps because many of them are spouses concerned with household budget matters. Program structure is a stronger component for business majors, probably because many of them are interested in one-year programs. Students from the natural sciences are least concerned with the program dimension. Finally, for rather elusive reasons the prestige component is stronger for women than men. This may be due to perceived difficulties in obtaining a desirable position, thus requiring the weight of a prestigious university and school in breaking down sex-related barriers.



Information Search

Aggregate Measures of Search.  Both measures of overall search effort suggest a surprisingly limited amount of effort (see Table 4). The overall sample reveals an average of about three schools seriously considered. This may be misleading, however, if students employ a two-stage process whereby they eliminate several schools from consideration using one or a few attributes to arrive at a small set that they "seriously consider." The overall sample reports an average of about two information sources that are "used a lot."

Segmentation analyses on the aggregate measures of search reveal only one difference among the various market segments. Individuals with prior work experience consider a significantly smaller number of schools than those without it. This may be due to geographical restrictions associated with an employer or less of a willingness to move a family, which this group is more likely to have. It may be also that these students restrict their search to schools who require experience for admittance.

Patterns of Information Search.  In Table 5 the results of the analyses of source usage are presented. Sources are listed in descending order of mean extent of usage. Factor loadings associated with each source are presented along with eigenvalues and percentage of variance explained. Results show program catalogs as the dominant source of information and four distinct overall patterns of search:

1.  Print media specific to programs (catalogs, directories)

2.  Expert and peer-related personal sources (fellow students/family, Masters alumni, current Masters students)

3.  Sources at undergraduate school (faculty, articles, bulletins)

4.  Direct contact with Masters schools (visits, phone calls)



Segmentation Analyses of Search Patterns

Again, analyses of factor scores and factor sums generated identical results and the results presented in Table 6 provide mean factor sums and ANOVA results from factor scores.

Discussion of Information Search Findings

The findings on information search effort and patterns reveal a variety of distinct search patterns occurring within a rather limited evoked set of schools. The effort and patterns of search are quite consistent within most of the segmentation variables. Behavioral consequences appear to be associated with the prior experience variable. Individuals with working experience exhibit less overall effort in their search processes. They also exhibit less usage of search patterns involving personal sources of information and information from their undergraduate institutions. Neither finding is surprising. The interpersonal realm of such individuals will typically include fewer individuals who are relevant sources of information than will that of the individual entering graduate school directly out of his or her undergraduate program. Of course, the working individuals will also have less opportunity to use their undergraduate institution, given the greater distance from that institution compared to currently enrolled students. Finally, the more extensive usage of undergraduate sources by undergraduate business majors is not surprising, given their enrollment in the area of a university with the most information about graduate business programs.



These results suggest a rather difficult marketing task for graduate schools of business seeking to maintain or increase enrollment. They must gain access to an evoked set that tends to be limited. Creating awareness is the first step but the critical one appears to be the effective dissemination of evidence supporting not only the university's and school's reputation but also the quality of its major areas. Graduate students appear to recognize the variations in quality that often exist across departments within a school. These tasks are especially difficult when the market segment with prior working experience is considered. These prospective students consider fewer schools and engage in more limited search patterns.




A study examining two key aspects of the decision process for students seeking a graduate program in business has been reported. The underlying cognitive structure and the salience of its components across various segments of prospective students represented one major aspect. Information search processes and their differences across the segments was the second key concern. The findings revealed five components of cognitive structure: cost, size, program, prestige, and major area. Certain variations in the salience of these components were evident across segments based on undergraduate grade-point and major. Four distinct information search patterns were revealed with variations in the usage of the patterns associated with the work experience of the student. These findings and conclusions are, of course, limited to the university and students involved in the study. To the extent that this university and its applicants are similar to other major state schools, the results may be generalizable.


Berry, L. and Allen, B. (1977), "Marketing's Crucial Role for Institutions of Higher Education," Atlantic Economic Review, 25 (July-August), 4-8.

Bettman, J. (1979), An Information Processing Theory of Consumer Choice, Reading: Addison-Wesley.

Hugstad, P. (1975), "The Marketing Concept in Higher Education: A Caveat," Liberal Education, 61 (December), 504-12.

Kotler, P. (1976), Marketing for Nonprofit Organizations, Englewood Cliffs: Prentice-Hall.

Krachenberg, A. (1972), "Bringing the Concept of Marketing to Higher Education," Journal of Higher Education, 43 (May), 369-80.

Punj, G. and Staelin, R. (1978), "The Choice Process for Graduate Business Schools," Journal of Marketing Research, 15 (November), 588-98.

Vaughn, R., Pitlik, J., and Hansotia, B. (1978), "Understanding University Choice: A Multi-Attribute Approach,'' in Advances in Consumer Research, V, H. K. Hunt, ed., Ann Arbor: Association for Consumer Research, 26-31.



Michael J. Houston, University of Wisconsin-Madison


NA - Advances in Consumer Research Volume 07 | 1980

Share Proceeding

Featured papers

See More


When Sharing Isn’t Caring: The Influence of Seeking the Best on Sharing Favorable Word of Mouth about Unsatisfactory Purchases

Nicholas J. Olson, Texas A&M University, USA
Rohini Ahluwalia, University of Minnesota, USA

Read More


Cultivating Collaboration and Value Cocreation in Consumption Journeys

Melissa Archpru Akaka, University of Denver
Hope Schau, University of Arizona, USA

Read More


Gossip: How The Relationship With the Source Shapes the Retransmission of Personal Content

Gaia Giambastiani, Bocconi University, Italy
Andrea Ordanini, Bocconi University, Italy
Joseph Nunes, University of Southern California, USA

Read More

Engage with Us

Becoming an Association for Consumer Research member is simple. Membership in ACR is relatively inexpensive, but brings significant benefits to its members.