Predictors of academic performance and pro-social behavior of business administration students.

ABSTRACT

Using data from a private business school in Germany, I analyze the importance of educational qualification, social skills and maturity as predictors of academic performance and of pro-social behavior of business administration students. I find that academic performance is

predicted by educational qualification and social skills. The latter are also weakly significant predictors of pro-social behavior. Experience and maturity are not significantly related to either academic performance or to pro-social behavior. Drawing on the empirical findings, I argue that business schools should do more to ensure that their programs appeal to the experience of students, and are practically relevant. I also argue that academic and social objectives of business administration programs should be integrated more strongly.

1. INTRODUCTION

"Few chief executives would have a hope in hell of getting into business school today" (Lucy Kellaway: The Logic of Fools. In: Financial Times, 9 December 2002, p. 6)

Over the past twenty to thirty years, business education has been one of the fastest growing industries in many Western economies. Many people, young high school graduates and experienced managers alike, flock into business administration programs. Their expectations are to develop skills and become acquainted with concepts and tools that will enable them to manage increasingly complex organizations in an ever-more turbulent environment. Yet, what are the qualities and skills fostered at business schools? If many of today's senior managers have little chance of entering these institutions, as Lucy Kellaway claims, what would this tell us about the nature of business school education, and about its relevance? Do practical experience and social skills, however defined, matter at all in business administration programs, or are these programs nothing but collections of abstract, formal subjects on the one hand and courses on idiosyncratic cases on the other, with no general lessons to be learned from either of them?

Broadly speaking, there are three ways to address these issues. The first approach is to analyze business administration programs, and the students enrolled in them, by comparing them with other programs, respectively with students of other subjects (for an example of this approach, see Barber et al., 2003). The benefit of this approach is that it provides sharper 'contours' of the nature of business administration programs by delineating them from other educational offers. However, this method does not provide significant insight into the issue of the relevance of the content provided in business administration programs.

A second approach is to relate the type and the level of education that individuals gain at business school to outcomes later on in their lives, such as job performance. If the measures ex-ante and ex-post are well specified, evidence of a significant relationship would support the relevance of business school education in a positive or negative sense. As the ultimate criterion for business school education is whether it prepares people for their roles later on in their careers, this method is, on a conceptual level, ideal. From an economic point of view, it follows Gary S. Becker's (1964) classical analysis of education as an investment into human capital, which is expected to yield returns in terms of higher productivity or other socially and individually beneficial outcomes. The problem with this approach is, of course, an empirical one: It is hard to establish standardized measures of 'outcomes' across a large number of business school graduates. Personal income is often used as a proxy but tells little about individuals' ability to be effective managers.

outcomes (performance, behavior, etc.) during the business administration program. By establishing the predictors of performance and behavior at business schools, one can obtain valuable insights into the nature of the tuition provided at these institutions. As an example, many business schools claim that they provide net only 'theoretical' tuition, but also foster 'soft', social skills among their students. If this claim were true, one would expect the social skills observed prior to entering business school to be a reasonable predictor of the pro-social behavior of business administration students during the program. If the relationship between social skills prior to entry and pro-social behavior during the program were insignificant, the claim that business schools value social skills among their students would look doubtful. Therefore, relating 'outcomes' during business administration programs to 'inputs' serves a twofold purpose. Firstly, it helps to establish what it takes to get into business school (thereby responding to Lucy Kellaway's article). Secondly, this method also provides insights into the nature of the 'outcomes': The individual qualities produced at business school. From there on, one can draw conclusions about the design and the content of business administration programs, both in terms of their academic curricula and in terms of their non-academic aspects such as the social atmosphere that they foster and the experiences that students gain during these programs.

The specific aims of this paper are threefold. First, I seek to identify the predictors of academic (grade) performance and of pro-social behavior of students of business administration. The term 'pro-social behavior', coined by O'Reilly and Chatman (1986), is defined here as the students' engagement and initiative in support of communal activities under the overall aegis of the students' union. Three classes of potential predictors of academic performance and pro-social behavior are analyzed: Educational achievement and qualification, social skills, and experience and maturity prior to entering business school. Some demographic characteristics are also taken into account as control factors. The second aim is to describe the business administration program, and the students enrolled in it, of the business school from which the data for the empirical part of the paper is drawn. The school concerned is a high profile, private organization in Germany, which for the past decade or so has consistently been among the most highly rated business schools in central Europe. Its general educational approach is a unique blend of an 'American-style' approach (e.g. significant use of case-based classes) and a traditional German curriculum with its heavy focus on commercial subjects such as accounting, commercial law, and so on. For educators from the Anglo-American community, this analysis may provide a refreshing insight into the structure and the content of business education in other parts of the world.

The third objective of the paper is to provide a critique of the education at business school that draws on the insights generated by the empirical section of the paper, and to derive recommendations regarding a re-orientation of business school education. Specifically, I shall argue that business school education, through curriculum design and active cooperation with students, should integrate academic tuition and the provision of social skills in a much better way. In addition, greater value should be placed on content that appeals to the experience of the students, and that is practically relevant. These recommendations follow from evidence regarding 'missing links' between observed individual-level input factors and outcomes during business administration programs.

The structure of the paper is as follows: In the next section I provide a literature review and develop hypotheses for testing. Thereafter I describe the data and the statistical methodologies used to test the hypotheses. In section 4 I give an overview of the results, which are discussed in greater detail in section 5. In that section I also develop the implications of the findings for business education in general and for achieving excellence in teaching management and business in particular. The final section of the paper contains a summary and an outline of directions for further research.

2. PREDICTORS OF ACADEMIC PERFORMANCE AND PRO-SOCIAL BEHAVIOR OF BUSINESS ADMINISTRATION STUDENTS: REVIEW AND HYPOTHESES

An extensive body of theoretical and empirical research focuses on the determinants of academic performance of students of business, economics and related subjects. In contrast, the literature on the determinants of students' pro-social behavior is much sparser and less systematic. In the following, I first look into potential predictors of academic performance before turning to the determinants of pro-social behavior. In each of the two sections, I investigate three classes of individual characteristics prior to entering business school that might potentially qualify as determinants of outcomes post-entry: First, measures of prior educational achievement and qualification such as high school grades and scores in standardized math tests; second, measures of social skills such as the quality of individuals' interaction in and contribution to teams; and third, biographic data that capture aspects of experience and maturity. I also briefly discuss the role of demographic background factors (e.g. gender) that need to be taken into account as control factors.

2.1 Predictors of Academic Performance of Business Administration Students

Much of the existing literature finds that measures of educational achievement during or at the end of secondary education are reasonably good, but nevertheless incomplete predictors of the performance of students in general during tertiary education. For example, Ransdell (2001) reports results according to which high school GPA accounts for up to 20% of the variance in first-year college GPA (see also Anderson, 2001, Linn, 1989 and Wolfe and Johnson, 1995 for similar findings). More specifically, with respect to the performance of students of business-related subjects (e.g. accounting, economics, finance, marketing), most studies (Anderson et al., 1994; Borde, 1998; Borg and Shapiro, 1996; Borg and Stranahan, 2002; Brasfield et al., 1993; Devadoss and Foltz, 1996; Filbeck and Smith, 1996; Gist et al., 1996; Johnson et al., 2002; Richter, 2003; Williams et al., 1992; Ziegert, 2000) confirm that schooling grades and standardized tests of educational achievement such as the SAT or the ACT are powerful predictors of the academic performance of students, when compared to other factors. A number of these studies (e.g. Williams et al., 1992) show that mathematics skills are particularly important for students of business administration and related subjects.

There are exceptions to this result, however. Bartlett et al. (1993), in a study of 39 students in an undergraduate accounting program in the UK, find that A-level scores are not significantly related to accounting exam results at the beginning of the program, after one and after three years of study. Their result may be distorted due to the relatively small sample size. However, in a study of 1286 university students in the USA, Laband and Piette (1995) do not find evidence of significant effects of the scores in the verbal and the math portion of the SAT taken prior to entry on the average GPA in upper-level economics courses. Tay (1994), focusing on 1314 students in Singapore, also fails to find a significant effect of high school GPA on performance in a first-year economics course. These results may partly be accounted for by the fact that in both studies exposure to, respectively performance in, previous economics courses are included among the independent variables, which should pick up some of the abilities measured by high school GPA or SAT scores. In line with the vast majority of studies, I expect that comprehensive measures of educational achievement and qualification prior to entering business school, in particular in mathematical subjects, predict the performance of business administration students post-entry reasonably well.

Hypothesis 1: Educational achievement and qualification of students prior to entering business school is a strong, and positive, predictor of the academic performance of students of business administration.

Very little evidence is available with respect to the relationship between measures of social skills--such as the ability to cooperate with others, the capacity to assume an appropriate role in a group and so on--on the grade performance of students of business administration. Several authors (e.g. Borg and Shapiro, 1996; Borg and Stranahan, 2002; Filbeck and Smith, 1996; Ziegert, 2000) find significant effects of personality type--as captured by measures such as the Myers-Briggs-Type Indicator, the 'Big Five' or similar personality concepts--on academic performance. Specifically, some authors (Gul and Chun Cheong Fong, 1993) report that introverted students tend to do better academically than extroverted students. Naser and Peel (1998) find that students who engage in solitary revision have better exam grades than students who prepare in small teams. Both studies focus on students in accounting courses. These results, although not directly related to the issue of social skills, appear to run counter to the consideration that such skills are--or should be--important requirements for managers, and thus for students of business administration who are preparing for such roles.

In contrast to this indirect evidence on the role of social skills, the hypothesis put forward in this paper is that these skills are important determinants of the academic performance of students of business administration, if meaningful measures of such social skills can be defined. The advantage of the data set used in this paper (see section 3.1) is that it contains manifold observations about candidates' behavior in groups in the context of an entry-level assessment center from several, independent observers. Equally important, these observers differ from the academics and supervisors who assess the students' academic performance during their studies. Moreover, the business school at which the data was collected emphasizes group interaction and social cohesion among the students not only as a "nice-to-have" adjunct to, but as an essential part of its academic program. For example, some--although, of course, not all--of its courses are assessed on the basis of team exercises, group presentations and jointly written papers. If the claim by the school concerned--and by many others--that their aim is to foster social skills through their programs is valid, one should expect 'academic performance' to be defined and measured in such a way that social skills are essential conditions for good performance. In consequence, measures of social skills prior to entry should then be significant predictors of academic performance post entry. One of the aims of this paper is to test this claim. Therefore, my hypothesis is as follows:

Hypothesis 2: Social skills of students prior to entering business school have a positive and significant effect on the academic performance of students during their business administration program.

Very similar arguments can be put forward with respect to the relationship between maturity and practical experience on the one hand and students' academic performance in a business administration program on the other hand. The typical claim by business schools is that the type of education they provide is highly relevant from a practical point of view. If this claim is valid, one would expect mature students with prior practical experience to have an advantage in grasping and applying the concepts taught during the program. The empirical evidence on the issue is very inconclusive. The few studies that address this question use students' age as a measure of their maturity. Of these studies, Didia and Hasnat (1998) and Tay (1994) find a positive effect of age on performance, while Filbeck and Smith (1996) and Laband and Piette (1995) report that the relationship is insignificant. Giving the benefit of doubt to academic institutions that look for maturity and experience among their candidates, my hypothesis is as follows:

Hypothesis 3: Experience and maturity have a positive and significant influence on the academic performance of students during their business administration program.

A large literature is concerned with the effects of socio-demographic characteristics such as ethnicity, parental status and family background (e.g. Jaggia and Kelly-Hawke, 1999) and even with the local environment of students during their studies (Reid, 1983) on their academic performance. These aspects are not the focal point of this study. One variable that is included as a control factor is gender. Without going into detail, suffice it to say that many but not all studies written during the 1970s and early 1980s (Allison, (1977; Siegfried, 1979; Ferber et al., 1983) found that males outperformed females in economics and business-related subjects. The majority of studies written since the mid-1980s show that gender differences are insignificant (Williams et al., 1992; Richter, 2003; for a critical discussion see Kleinfeld, 1999). In line with the more recent evidence, I do not expect gender to be a significant predictor of the academic performance of business administration students.

2.2 Predictors of Pro-Social Behavior of Business Administration Students

Whereas the evidence on the determinants of academic performance is substantive if somewhat inconclusive, a literature on the effects of prior education on social behavior of students of business administration is simply non-existent. Some general analogies can be drawn from the psychological and organizational behavior (OB) literature. For example, organizational commitment--a concept widely discussed in the OB literature--might lead to pro-social behavior in the sense of the willingness and ability to engage meaningfully in peer activities and to show initiative in support of common goals. Assuming this argument were true--and the empirical evidence on the issue is weak (see, for example, Meyer and Allen, 1991)--then it would be worthwhile for the purposes of this paper to look into the antecedents of organizational commitment. Steers (1977) finds that the relationship between education and organizational commitment is negative, presumably because more highly educated people have better access to alternative options; they may also view their organizations more critically. As the theoretical and empirical evidence on the relationship between education and organizational commitment is, at best, inconclusive, I refrain from drawing inferences regarding any relationship between educational achievement prior to entry and pro-social behavior of business administration students. The same goes for demographic characteristics and biographic variables such as experience or maturity as potential determinants of pro-social behavior.

More immediately plausible, however, is the argument that social skills prior to entry should have a positive and significant effect on pro-social behavior post entry, if these concepts can be measured in any meaningful way. For example, if it is possible to observe and assess individuals' behavior in teams at the outset of an academic program--as is frequently done in recruitment and admission processes--one would expect these assessments either to be reflected in the behavior of individuals post entry, or to be meaningless and hence wasteful. O'Reilly and Chatman (1986) argue that observations of behavioral dimensions of individuals such as commitment and the willingness to participate in group activities above and beyond the immediate task at hand have positive effects on pro-social behavior. In line with their finding, the hypothesis 4 to be tested in this paper is as follows:

Hypothesis 4: Social skills of business administration students prior to entry have a positive and significant effect on the pro-social behavior of students during their business administration program.

In summary, four hypotheses regarding the potential determinants of academic performance and pro-social behavior have been developed. The model displayed in figure (1) depicts the hypothesized relationships as outlined above. In the next section, I describe the data used to test these hypotheses and the methodological approach taken to do so.

[FIGURE 1 OMITTED]

3. DATA AND METHODOLOGY

3.1 Data Description

The quantitative information used in this paper is drawn from a rich data set containing information on the 1998 cohort of students at a high profile, private business school in Germany. The data was collected in 2002 from three sources: The admissions office, the student registry and the examinations office. The data set covers information from students' application materials, which include high school grades and self-assessment reports, the results of the entry-level assessment center and written tests and information on the grade performance of students during the program.

The diploma program in business administration at the school concerned consists of a tightly organized curriculum. During the first three terms in particular, students take mandatory courses in mathematics, micro- and macroeconomics, accounting and finance, law, statistics, general management, English, introductory IT and accounting and controlling. In addition, they focus on either information systems (Wirtschaftsinformatik, WI) or foreign languages (zweisprachige Studienrichtung, ZS) as an area of specialization. Whereas WI-students take courses in programming, application design and the like, ZS-students focus on Spanish or French as a second foreign language in addition to English.

The focus of this paper is on the academic performance and pro-social behavior of the students during the first three terms of the program, which lasts eight half-year terms (four years) in total. After the first three terms they are awarded a so-called 'pre-diploma', which provides them with an intermediate certificate. During the remaining five terms of the program fixed requirements are loosened, and henceforward the curriculum provides a combination of mandatory courses in strategy and organization, corporate accounting and controlling and economics, with electives from a range of options. Also, students are required to spend the third year of study abroad at foreign partner universities. Overall, the school and the diploma program it offers is known throughout central Europe for its rigor, international orientation and emphasis on fostering social cohesion and commitment among the students (see also the description on the importance of the 'social grades' below).

The following variables are used in order to test the hypotheses regarding the relationships among the theoretical constructs described in the previous section, Prior educational achievement is measured principally by the overall high school examination (the German Abitur) grade. The Abitur grade represents a weighted average of the results across a broad range of subjects (some of them mandatory, some of them elective) taken during the final two years of the students' high school education. While performance standards vary somewhat across the different states (Bundeslander) within Germany, the broad system of high school education is the same across Germany (Steedman et al., 1997). Richter (2003) finds the Abitur to be a valid measure of overall educational achievement and a good predictor of future academic performance. As an additional measure of educational achievement and qualification, the score of the students in the math part of the admission tests at the business school concerned is taken into account.

Social skills at entry-level are measured by the results from the assessment center that students take part in in the course of the admission procedure. The assessment center consists of three parts: First, a group presentation on specific topics that are assigned to groups of 3-4 candidates in advance, second, a group discussion following the presentation and third, an individual interview with each candidate. Each of the three exercises is observed by a panel of 2-4 observers (respectively, interviewers) that take notes independent of each other. The majority of the panelists are managers of companies with which the school cooperates. In advance of the assessment center, these 'practitioners' undergo a structured training program on HR assessment and selection provided by academics from the school. They use standardized observation sheets to award Likert-type scores on a range of variables such as 'team cooperation', 'social and intercultural competence' and 'communication skills'. In this paper, I use the average grade for each of the three parts of the assessment center across sub-categories and observers. In line with the actual operation of the assessment center in practice, I take the group presentation exercise as the most important measure of social skills (see section 3.3 on the orthogonalization procedure of the explanatory variables which presupposes an appropriate ordering of the variables).

Experience and maturity, the third set of explanatory variables investigated here, are measured by the age of the students, respectively the number of extended stays in foreign countries. The latter is a highly important variable due to the international orientation of the business administration program analyzed. I also experimented with the number of internships as an alternative measure of experience. However, I found that the number of internships correlates highly with the number of foreign stays, so that including both measures would have raised multicollinearity problems. Therefore, I decided to use only one of the two measures. In addition, the number of internships turned out to be a more problematic variable than the number of extended foreign stays as the criteria for what activity qualifies as an internship are not clearly defined on the application forms from which this information is gathered.

As a measure of academic performance I use the students' overall pre-diploma grade (Vordiplomnote), which represents the weighted average of the grades of the various subjects taken during the first three terms of the business administration program. As an alternative, a focus on the performance in different subjects and terms would have been possible. As the emphasis of this paper is on the determinants of academic performance in business school in general, I opted for the pre-diploma grade as a comprehensive measure.

As a measure of pro-social behavior, I use the so-called 'social grade', which students at the business school concerned receive at the end of the third term of the program, concomitant with their pre-diploma grade. This grade reflects the students' commitment and engagement in social activities (e.g. organization of events, initiatives, etc.) and in the students' union. It is awarded by the students' union committee. Students have the incentive to score well since the social grade is taken into account in the allocation of study places at foreign partner schools at which students study during the third year of their program; better social grades will result in admission at a more prestigious universities. Overall, the 'social grade' reflects a lateral assessment of the students' social behavior by peers, rather than a hierarchical assessment by supervisors. Due to the availability of a peer-based, continuous measure of pro-social behavior, the data set used in this paper provides social scientists with a rich basis for researching the determinants of behavioral phenomena in organizations.

In addition to the variables outlined above, controls for gender and area of specialization are included in the regression analysis outlined in section 3.3 below. Nationality and ethnicity are omitted as control factors since the vast majority of the student cohort analyzed here was born and raised in Germany. Regardless of regression specification, including these factors does not yield any additional explanatory power.

3.2 Descriptive Statistics

Of the total of around 200 students who started the program in 1998, complete information on the variables investigated here is available for 115 individuals that form the sample in this study. Descriptive statistics are provided in table 1. Correlation coefficients among the continuous variables and their significance levels are shown in table 2.

In terms of the demographic characteristics of the sample under investigation, 43% of the students are female, the rest male. The age range at the beginning of the program was between 18 and 24 years; the average age was just over 20 years. Students reported that they had done on average two to three extended foreign stays prior to beginning their studies. With an enrolment rate of 78%, foreign languages were the far more popular specialization than information systems.

With respect to the educational achievement of the candidates before beginning their studies, the average Abitur grade in the sample is around 2.0, the range extending from 1.0 to 3.2. With a highly significant correlation coefficient of 0.28, there is an apparent relationship between the Abitur grade and the math score of the students from the admission test. Similarly, the results from the three parts of the initial assessment center--group presentation, group discussion and single interview--display clear interrelations.

The academic performance of the students during their business administration program is measured by their pre-diploma grade, which, at an average of 2.46, is located very near the center of the overall performance distribution. Given that less than 15% of the students achieve a pre-diploma grade better than 2.0, the data show that the school concerned does not award 'excellent' grades too easily, and that performance expectations are high. The pre-diploma grade displays a high correlation, in particular with the Abitur variable but also with the math score from the admission test and the group presentation and group discussion scores from the assessment center.

The average 'social grade' is 2.36, with a relatively high variance as compared to the other pre-entry or post-entry performance variables contained in the data set. In allocating the social grades, the students' union made use of almost the entire grading scale, thereby yielding a differentiated overview of its assessment of the pro-social behavior of the school's student body. I applied t-tests in order to explore any potential relationships between binary variables (gender; area of specialization) and academic performance respectively pro-social behavior. The results are contained in table 3.

The t-tests show that academic performance is not significantly related to gender differences, and that the performance differentials between those students with an emphasis on information systems and those focusing on foreign languages are only very weakly significant. 'Social grades', in contrast, are more strongly related to gender and course of study. Females achieve significantly better social grades than males, and students with an emphasis on foreign languages achieve higher scores than those concentrating on information systems. While these relationships are interesting, it remains to be seen whether they represent any causal effects.

3.3 Methodological Approach

The aim of the empirical section of this paper is to investigate the effects of a single set of explanatory variables on two different outcome variables using regression analysis. It is clear from the correlations reported in table 2 that variables within two distinct groups--namely the educational achievement variables ABITUR and MATH on the one hand and the grades from the assessment center (GRPRES, GRDISC, and INTVIEW) on the other--are highly correlated. Including these data without further manipulation in a regression would result in serious multicollinearity problems. Therefore, I use a modified Gram-Schmidt procedure (Golub and Van Loan 1996, pp. 218-219) to orthogonalize the variables by estimating the following equations:

(1.1) [ABITUR.sub.orth] = [q.sub.01] + [q.sub.1]ABITUR

(1.2) [MATH.sub.orth] = [q.sub.02] + [q.sub.12]ABITUR + [q.sub.22]MATH

(2.1) [GRPRES.sub.orth] = [r.sub.01] + [r.sub.11]GRPRES

(2.2) [GRDISC.sub.orth] = [r.sub.02] + [r.sub.12]EGRPRES + [r.sub.22]GRDISC

(2.3) [INTVIEW.sub.orth] = [r.sub.03] + [r.sub.13]GRPRES + [r.sub.23]GRDISC + [r.sub.33]INTVIEW

Since in this procedure the order of variables determines the orthogonalization, I experimented with different orders achieving fairly similar results in each case. The resulting orthogonalized variables are not significantly correlated, not even across groups (e.g. [GRPRES.sub.orth] is not significantly correlated with [ABITUR.sub.orth], and so on). Apart from avoiding the problem of multicollinearity, the benefit of orthogonalizing the variables is that this procedure eases the interpretation of the results. The orthogonal variables represent 'pure' phenomena in that the effect of all the preceding variables (e.g. ABITUR in the case of the MATH variable) is removed. Hence, [MATH.sub.orth] represents those mathematical abilities of the students in the sample that have not already been captured by the ABITUR variable, and so on. The dependent variables investigated in this paper are the academic grade and the social grade. In order to investigate the determinants of these variables, I use Zellner's Seemingly Unrelated Regression (SUR). The educational production function that I estimate is

(3.0) [MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII]

where OUTCOME is specified as the pre-diploma grade (ACAD_GRADE) in one instance and as the social grade (SOCIAL_GRADE)-in the other. The first line of equation (3.0) represents the students' educational achievement prior to entering business school, the second line their performance in the assessment center and the third line their demographic and biographic characteristics. Complete variable definitions for the variables are contained in table 1; e denotes the error term. The quadratic terms in the model are included in order to avoid omitted variable problems. Ramsey regression specification error tests suggested that these problems existed in the regression without those terms. I also tested for heteroscedasticity using the Cook and Weisberg (1983) test and did not find evidence for it.

Using a single set of explanatory variables, SUR produces the same results as would be obtained with two separate Ordinary Least Square (OLS) regressions. However, SUR allows testing whether the coefficients in the two models are jointly zero. It also allows estimating the correlation between the residuals in the pre-diploma grade and the social grade equations and testing the hypothesis that this correlation is zero.

4. RESULTS

Table 4 shows the results of the SUR analysis, with the pre-diploma grade and the social grade as the two dependent variables in regressions (1) and (2) respectively. Both regression models display good fit. They explain 53 per cent and 25 per cent of the variance in the students' academic performance and pro-social behavior respectively. In the following, I report first on the determinants of academic performance (regression (1)) before turning to the predictors of pro-social behavior (regression (2)). A graphic representation of the overall findings is shown in figure 2.

[FIGURE 2 OMITTED]

4.1 Predictors of Academic Performance

As the results of regression (1) show, the most important predictor of the students' pre-diploma grade among the variables investigated is clearly their high school grade. More specifically, the fact that the coefficient on the square of the [ABITUR.sub.orth] variable is both highly significant and negative, suggests that academic performance is concave in high school performance. Above and beyond the abilities captured by the general high school grade (Abitur), the students' specific mathematical abilities are also significantly related to academic grade in business school, and have the expected sign. This finding confirms the expectation that mathematical aptitude is an important precondition for success in studying business administration. Overall, the results strongly support hypothesis 1 regarding previous educational achievement and qualification as a predictor of academic performance of students at business school.

The results of regression (1) also show that academic performance is positively and significantly influenced by the students' social abilities evidenced in the entry-level assessment center. In particular, group discussion abilities are an important predictor of academic performance, more so than group presentation skills whose impact on performance is of smaller magnitude and less strongly significant. In contrast, the effect of the observers' assessment of the students in the individual interview is with an unexpected sign, but insignificant. As the results from group discussion and presentation are clearly more important measures of students' social abilities than the results for a single interview setting, in the sense of the framework presented in figure 1, hypothesis 2 regarding the effects of social skills on academic performance is supported by these findings.

The coefficients on the variables reflecting demographic and biographic variables on academic performance are insignificant in all cases. This includes the variables AGE and FOREIGN, which are taken here as proxies for maturity and experience respectively. I also experimented with the number of internships as an explanatory variable instead of the number of foreign stays, and found an insignificant relationship, too. Therefore, hypothesis 3 regarding the positive impact of these characteristics on performance in business school is not supported by the data. A causal effect of the area of specialization, which was tentatively suggested by the (weakly) significant differences between business administration students with an emphasis on foreign languages and those concentrating on information systems (see table 3), disappears once other factors are taken into account.

4.2 Predictors of Pro-Social Behavior

In regression (2), only two types of variables are found to be significant predictors of the students' social grade: First, social skills, represented by group presentation abilities in the assessment center, is positively related to social grade. Specifically, the positive sign on the quadratic term of the GRPRES variable indicates that pro-social behavior in business school is convex in group presentation skills of the students. However, it should be noted that the significance of these effects is weak. Taking into account the fact that the coefficient on the GRDISC variable is below significance level, I find only tentative support for hypothesis (4).

The second variable that reaches relatively strong significance is the students' area of specialization. Ceteris paribus, business administration students with an emphasis on foreign languages achieve significantly higher social grades than students with an emphasis on information systems. In contrast, the gender differences shown in table 3 are found not to represent direct causal effects once other factors are taken into account.

Similar to regression (1), the data do not point to a significant effect of maturity and experience on pro-social behavior. Also, prior educational achievement and qualification do not impact pro-social behavior in business school significantly.

The correlation coefficient between the residuals of the academic grade and the social grade equations is -0.050. A Breusch-Pagan test of independence of the residuals does not yield significant results; hence the test hypothesis that this correlation is zero cannot be rejected. The weakly significant correlation of 0.19 between academic and social grade reported in table 2 disappears in the multivariate SUR model when the co-determinants of the two dependent variables are taken into account.

5. DISCUSSION AND IMPLICATIONS FOR BUSINESS EDUCATION

The results provided in the previous section hold a lot of good, as well as some not-so-good news for business school education and for those involved in providing it, including teachers, admission staff and curriculum developers. On the positive side, the findings support the argument that the business administration program analyzed here is intellectually rigorous. The grade performance of the students is very clearly influenced by how well they did during their final two years at high school, measured by the overall Abitur grade. In addition, their math skills are a significant predictor of performance at business school, even after controlling for the mathematical abilities already contained in the general high school grade. Therefore, mathematical abilities are an important requirement for grade performance at the business school concerned. Based on this relationship between educational qualification and achievement prior to entering business school and grade performance of 2nd year students, the business administration program analyzed here requires those intellectual abilities which are also important 'to make it' at high school. It also has a mathematical spike. (Of course, whether these findings are regarded as positive depends on one's personal point of view. Those who perceive high school education as irrelevant to managers will view the parallels between high school education and business school education uncovered by the data as evidence that the tuition provided at business schools is equally irrelevant. Those who take a more positive view of the value of general education will view the finding that business education builds on the abilities fostered in secondary education as welcome evidence for its intellectual quality).

The second positive finding is that social skills are beneficial to students of business administration, not only in terms of their pro-social behavior among peers but also in terms of their academic performance. In fact, they seem to be even more important for academic performance than for pro-social behavior. For example, the interactive abilities exhibited in group discussion exercises during the assessment center prior to admission are highly significant in predicting academic performance, whereas they are insignificant for the 'social grade' by which pro-social behavior is measured in this paper. If social skills are significant predictors of academic performance, then the academic program itself appears to involve activities and behaviors that draw on social skills. This might be the case, for example, in team projects, group-based exercises, and joint presentations and papers. The finding of a significant relationship between social skills ex-ante and academic performance during the program also is a positive signal to business school administrators to adopt, or keep, well-designed group exercises as part of their admission procedures. This recommendation does not extend, however, to individual interviews as selection instruments. Assessments based on individual interviews are neither significantly related to academic performance nor to pro-social behavior. This result is consistent with the prevailing view in the literature (e.g. Cascio 1991, pp. 271ff.) that, for testing purposes, individual interviews have, at best, low predictive validity. Many institutions--business schools and firms as employers alike--have an urge to assess candidates in individual interview settings, presumably in the hope of obtaining an immediate, 'eye to eye' impression of their candidates. Individual interviews can, of course, serve other purposes than to assess people (Watson 1989), but as a testing instrument they add little value. Since they are fairly expensive (Cook 1988, p. 246), they may well be value destroying.

The third positive aspect that emerges from the description of the business school concerned is that prosocial behavior at business school actually matters. The mere fact that pro-social behavior among peers is recognized by a special grade that has some (although limited) importance to students' further advancement, should be regarded as a positive sign that education at a business school is, or can be, more than gray and drab theory. Also, the finding that group presentation skills are a, however weak, predictor of pro-social behavior provides welcome evidence that business schools in selecting their candidates, have access to indicators that they should look for in order to identify those candidates who are likely to fit into the social environment. Of course, it would be helpful to have more specific measures of social skills that could be used to predict pro-social behavior in a better way. The weak significance of only one out of three parts of the assessment center (namely, the group presentation exercise) suggests that schools which place great value on pro-social behavior among their students should try to identify better characteristics in order to screen applicants for their social skills in a more reliable way. Since the primary beneficiaries of pro-social behavior are the students themselves, they--through their representatives--should be involved in developing and implementing such measures.

In addition to these positive findings, there are some results that shed doubt on particular aspects of the business education system at the school analyzed here, and possibly for business education in general. The data point to two specific areas of concern.

First, according to the results, experience and maturity are significant predictors neither of academic performance nor of pro-social behavior at business school. A potential reason for this finding might be the use of unsatisfactory proxies (age, number of foreign stays, number of internships) as measures of experience and maturity in this paper. A measure of the total time spent in foreign countries respectively in internships, together with reliable descriptions of the type of activities carried out, might alleviate these problems. However, the fact that the coefficients on the variables concerned are nowhere near significance, does not provide a lot of hope that experience, if measured in a better way, could be regarded as an important predictor of outcomes during business school. This result is disappointing in that it suggests that business school education does not value students' experience in foreign countries or practical experience in business situations very highly. If it did, students with significant experience would have a head-start compared with those who have not had the benefit of becoming familiar with foreign cultures or seeing 'management at work' in real business situations. If this were the case, there might well be some truth in Lucy Kellaway's claim quoted at the beginning of this paper that current chief executives would have little chance of getting into business school today. In addition, if business schools do not make good use of the experience that students bring to the table, it seems improbable that they will make good use of the experience, which students gain during their studies. Several internships and studying a full year abroad are exactly the requirements for students participating in the program at the business school concerned.

The recommendation to be derived from this finding is that business schools should develop curricula that make better use of students' practical experience. For example, students could be invited to report and discuss their insights from internships or stays abroad in the classroom, or to write papers that integrate academic concepts acquired during their coursework with their insights from practical experience. A more international composition of teaching staff, greater immersion of teachers in business situations and better teacher training might also help to make business education more practice-oriented. Overall, business school should ensure that they provide programs that combine the objectives of rigor and relevance in an ideal way.

The second point is more subtle. The empirical results suggest that the correlation between the two dependent variables in this paper--academic grade performance and pro-social behavior--is limited to the influence of a particular type of social skills, namely those evidenced in group presentation and group discussion exercises during the assessment center prior to admission. After the influences of these specific factors are accounted for, the correlation between academic performance and pro-social behavior disappears. The concern is that there is insufficient 'common ground' between the objectives of the academic curriculum (grade performance) and the objectives of the student body (pro-social behavior). While the fact that both objectives are explicitly stated and that there is some common ground between them should be applauded, a stronger integration between them would seem beneficial. Is it not the aim of business schools to foster pro-social behavior--beyond the ability to give group presentations--through their academic programs? Conversely, is it not the aim of the students to learn, in a thorough, academic sense, from their experiences in the context of their communal activities? As a practical suggestion, a closer integration of these objectives could be achieved by organizing some of the students' communal activities--such as the preparation and management of student symposia, the editing of the students' newspaper and so on--as projects that carry credit in the sense of regular coursework. Similarly, such projects and initiatives could be used more strongly as practical examples in the context of academic courses. These are just some examples of how a stronger integration could be achieved. From my personal point of view, excellence in business education is not achieved when academics succeed in trans-acting a 'set message' to their students regardless of their situation. Rather, the common aim of educators and students alike should be to interweave management education with the actual experience of the learners. Pro-social behavior in particular can only be learned by doing. The closer the integration between academic and social objectives, the greater the likelihood that what is taught at business schools will stay relevant for learners in the long run.

6. CONCLUSION

In this paper, I have analyzed the predictors of academic performance and of pro-social behavior of students of business administration in order to obtain insights into the nature of business school education. Using data from a high profile business school in Germany, I have looked at three classes of explanatory variables: Educational achievement and qualification prior to entering business school, social skills, and personal experience and maturity. The first two of these three categories of variables are found to be of significant importance for academic performance, whereas the third category is not. With respect to pro-social behavior, only social skills prior to entry aRe weakly significant predictors. I have used these results as a basis for a more general critique of business school education. My overall conclusion is that education at business school has come a long way not only in terms of academic rigor and quality, but also in terms of recognizing soft, social skills as important aspects of education. However, there is still scope for achieving a closer integration between the academic and the social objectives of business administration programs. In addition, business schools should do more to ensure that the tuition they provide builds on and appeals to the experience of students, and that it is practically relevant.

Clearly, further research is needed to substantiate these findings. One avenue for achieving this objective would be to pursue similar studies, but with greater sample sizes and even more differentiated measures of constructs such as maturity and experience. An alternative way forward would be to investigate the effects of business school education on a range of professional and behavioral outcomes. Taken together, the results of these two types of studies should produce a compelling analysis of the nature of education at business schools. The findings from this paper are a first, and important, step in that direction.

TABLE 1
VARIABLE DEFINITIONS AND SAMPLE DESCRIPTION (a,b)

Variable         Mean    S.D.    Range       Variable definition

A. Educational achievement prior to entry

  ABITUR         1.97    0.50    1.0-3.2     Overall high school
                                             (Abitur) grade of student

  MATH           61.73   14.59   17-97       Student's score in math
                                             test during admission
                                             procedure

B. Assessment center prior to admission

  GRPRES         2.25    0.41    1.50-3.41   Student's average score
                                             achieved in group
                                             presentation across all
                                             observers and observation
                                             categories

  GRDISC         2.35    0.48    1.3-3.5     Student's average score
                                             achieved in group
                                             discussoon across all
                                             observers and observation
                                             categories

  INTVIEW        2.14    0.60    1.00-3.66   Student's average score
                                             achieved in single
                                             interview across all
                                             observation categories

C. Demographic and biographic characteristics

  AGE            20.20   1.15    18-24       Age of the student at the
                                             beginning of the program
                                             (Oct. 1998)

  GENDER         0.43    0.49    [0;1]       Dummy: 1 if the student is
                                             female; 0 otherwise

  FOREIGN        2.40    1.62    0-8         Number of extended foreign
                                             stays prior to entering
                                             business school

  COS            0.76    0.42    [0;1]       Dummy: 1 if the student is
                                             taking foreign languages
                                             as his/her area of
                                             specialization; 0 if
                                             information systems

D. Outcomes in  business school

  ACAD_GRADE     2.46    0.45    1.3-3.4     Student's pre-diploma
                                             grade (end of 3rd term)

  SOCIAL_GRADE   2.36    0.82    1.3-5.0     Student's social grade
                                             (end of 3rd term)

(a) n = 115

(b) All variables except dummies and MATH represent grades on the basis
of the standard German grading scale, ranging from 1 (very good) to 5
(fail). MATH represents a score ranging from 0 (fail) to 100 (full
points)

TABLE 2
CORRELATION COEFFICIENTS AND SIGNIFICANCE LEVELS AMONG CONTINUOUS
VARIABLES

                                   A. Educational
                                   achievement/
                                   qualification
                                   prior to entry

                                   ABITUR     MATH

A. Prior educat.                   1.00       --
achievement         ABITUR
                    MATH           -0.28 **   1.00

B. Social skills    GRPRES         0.14       -0.16 ([dagger])
(assessment         GRDISC         0.02       -0.04
center)             INTVIEW        0.12       -0.03

C. Demogr./biogr.   AGE            0.22 *     -0.02
characteristics     FOREIGN        -0.07      0.00

D. Outcomes in      ACAD_GRADE     0.58 ***   -0.39 ***
business school     SOCIAL_GRADE   0.21 *     -0.11

                                   B. Social skills (assessment
                                   center)

                                   GRPRES     GRDISC    INTVIEW

A. Prior educat.                   --         --        --
achievement         ABITUR
                    MATH           --         --        --

B. Social skills    GRPRES         1.00       --        --
(assessment         GRDISC         0.39 ***   1.00      --
center)             INTVIEW        0.29 **    0.22 *    1.00

C. Demogr./biogr.   AGE            0.01       -0.06     0.12
characteristics     FOREIGN        -0.12      0.03      -0.12

D. Outcomes in      ACAD_GRADE     0.24 **    0.27 **   0.06
business school     SOCIAL_GRADE   0.21 *     0.21 *    0.13

                                   C. Demographic    D. Outcomes in
                                   and biographic    business school
                                   characteristics

                                   AGE    FOREIGN    ACAD_     SOCIAL_
                                                     GRADE     GRADE

A. Prior educat.                   --     --         --        --
achievement         ABITUR
                    MATH           --     --         --        --

B. Social skills    GRPRES         --     --         --        --
(assessment         GRDISC         --     --         --        --
center)             INTVIEW        --     --         --        --

C. Demogr./biogr.   AGE            1.00   --         --        --
characteristics     FOREIGN        0.06   1.00       --        --

D. Outcomes in      ACAD_GRADE     0.10   0.07       1.00      --
business school     SOCIAL_GRADE   0.00   -0.07      0.190 *   1.00

([dagger]) p <. 10

* p < .05

** p <. 01

*** p <. 001

TABLE 3
OUTCOMES BY GENDER, AREA OF SPECIALIZATION AND SIGNIFICANCE LEVELS

                                    Mean   S.D.   Significance of
                                                      t-value

ACAD_GRADE--males                   2.52   0.05   n.s.
ACAD_GRADE--females                 2.38   0.06

SOCIAL_GRADE--males                 2.52   0.11   p<0.05
SOCIAL_GRADE--females               2.15   0.08

ACAD_GRADE--information systems     2.60   0.06   p<0.10
ACAD_GRADE--foreign languages       2.41   0.05

SOCIAL_GRADE--information systems   2.92   0.22   p<0.0001
SOCIAL_GRADE--foreign languages     2.18   0.06

TABLE 4
RESULTS OF SUR REGRESSION: PREDICTORS OF ACADEMIC PERFORMANCE AND
PRO-SOCIAL BEHAVIOR (a)

                              Outcome (dejeendent variables)

Independent                      (1)            (2)
variables                     ACAD_GRADE   SOCIAL_GRADE

CONSTANT                        2.93 ***     3.83 **
                               (4.79)       (2.69)
[ABITUR.sub.orth]               0.26 ***     0.10
                               (8.47)       (1.40)
[([ABITUR.sub.orth]).sup.2]    -0.07 **     -0.06
                              (-3.11)      (-1.20)
[MATH.sub.orth]                -0.08 **     -0.02
                              (-2.80)      (-0.40)
[GRPRES.sub.orth]               0.07 *       0.13 ([dagger])
                               (2.05)       (1.66)
[([GRPRES.sub.orth]).sup.2]    -0.01         0.11 ([dagger])
                              (-0.60)       (1.63)
[GRDISC.sub.orth]               0.10 ***     0.11
                               (3.23)       (1.61)
[INTVIEW.sub.orth]             -0.04         0.01
                              (-1.33)       (0.19)
AGE                            -0.02        -0.04
                              (-0.68)      (-0.70)
GENDER                         -0.09        -0.15
                              (-0.30)      (-0.89)
FOREIGN                         0.02         0.00
                               (0.15)       (0.05)
COS                            -0.02        -0.57 **
                               (0.77)      (-3.05)

[R.sup.2]                       0.53         0.25
F                              10.74 ***     3.15 ***

(a) Variable coefficients are unbracketed, with associated t-values
shown in parantheses

([dagger]) p < .10

* p < .05

** p <. 01

*** p < .001

7. REFERENCES

Allison, Elisabeth K., Educational Production Function for an Introductory Economics Course, Harvard University Discussion Paper, 545, April 1977.

Anderson, Gordon, Benjamin, Dwayne, and Fuss, Melvyn A., "The Determinants of Success in University Introductory Economics Courses", Journal of Economic Education, 25 (2), 1994, 99-119.

Anderson, L.W., "Predicting Academic Performance of College Students in the United States and in Estonia", International Journal of Educational Research, 35 (4), 2001, 353-355.

Barber, Clifford S., Borin, Norm, Cerf, Douglas C., and Swartz, Teresa A., "Measuring the Effectiveness of Innovative Business Programs", Journal of the Academy of Business Education, 4 (1), 2003, 41-53.

Bartlett, Susan, Peel, Michael J., and Pendlebury, Maurice, "From Fresher to Finalist: A Three-Year Analysis of Student Performance on an Accounting Degree Programme", Accounting Education., 2 (2), 1993:111-122.

Becker, Gary S., Human Capital: A Theoretical and Empirical Analysis with Special Reference to Education, National Bureau of Economic Research, New York, 1964.

Borde, Stephen F., "Predictors of Student Academic Performance in the Introductory Marketing Course", Journal of Education for Business, 73 (5), 1998, 305-310.

Borg, Mary O., and Shapiro, Stephen L., "Personality Type and Student Performance in Principles of Economics", Journal of Economic Education, 27 (4), 1996, 3-25.

Borg, Mary O., and Stranahan, Harriet A., "Personality Type and Student Performance in Upper-Level Economics Courses: The Importance of Race and Gender", Journal of Economic Education, 33 (1), 2002, 3-14.

Brasfield, David W., Harrison, Dannie E., and McCoy, James P., "The Impact of High School Economics on the College Principles of Economics Course", Journal of Economic Education, 24 (2), 1993, 99-111.

Cascio, Wayne F., Applied Psychology in Personnel Management. 4th edition, Prentice Hall, Englewood Cliffs, N J, 1991.

Cook, Mark, Personnel Selection and Productivity. Wiley, London, 1988.

Cook, R.D., and Weisberg, S., "Diagnostics for Heteroscedasticity in Regression", Biometrika, 70, 1983, 1-10.

Devadoss, Stephen, and Foltz, John, "Evaluation of Factors Influencing Student Class Attendance and Performance", American Journal of Agricultural Economics, 78, 1996, 499-507.

Didia, Dal, and Hasnat, Baban, "The Determinants of Performance in the University Introductory Finance Course", Financial Practice & Education, 8 (1), 1998, 102-107.

Ferber, M.A., Birnbaum, B.G., and Green, C.A., "Gender Differences in Economic Knowledge: A Reevaluation of the Evidence", Journal of Economic Education, 14 (2), 1983, 24-37.

Filbeck, Greg, and Smith, Linda L., "Learning Styles, Teaching Strategies, and Predictors of Success for Students in Corporate Finance", Financial Practice & Education, 6 (1), 1996, 74-85.

Gist, Willie E., Goedde, Harold, and Ward, Bart H., "The Influence of Mathematical Skills and Other Factors on Minority Student Performance in Principles of Accounting", Issues in Accounting Education., 11 (1), 1996: 49-60.

Golub, G.H., and Van Loan, C.F., Matrix Computations, 3rd edition, Johns Hopkins University Press, Baltimore, 1996.

Gul, Ferdinand A., and Chun Cheong Fong, Steve, "Predicting Success for Introductory Accounting Students: Some Further Hong Kong Evidence", Accounting Education, 2 (1), 1993, 33-42.

Johnson, Dean L., Joyce, Patrick, and Sen, Swapan, "An Analysis of Student Effort and Performance in the Finance Principles Course", Journal of Applied Finance, 12 (2), 2002, 67-72.

Kleinfeld, Judith, "Student Performance: Males Versus Females", The Public Interest, Winter, 1999: 3-20.

Laband, David N., and Piette, Michael J., "Better Learning from Better Management: How to Improve the Principles of Economics Course", American Economic Association Papers and Proceedings, 85 (2), 1995, 335-338.

Linn, R.L., Educational Measurement. The American Council on Education / Macmillan Series in Higher Education, 1989.

Meyer, J.P., and Allen, N.J., "A Three-Component Conceptualization of Organizational Commitment", Human Resource Management Review, 1 (1), 1991,61-89.

Naser, Kamal, and Peel, Michael J., "An Exploratory Study of the Impact of Intervening Variables on Student Performance in a Principles of Accounting Course", Accounting Education, 7 (3), 1998: 209-223.

O'Reilly, C., and Chatman, J., "Organizational Commitment and Psychological Attachment: The Effects of Compliance, Identification, and Internalization on Prosocial Behavior", Journal of Applied Psycholoqy, 71 (3), 1986, 492-499.

Ransdell, Sarah, "Predicting College Success: The Importance of Ability and Non-Cognitive Variables", International Journal of Educational Research, 35 (4), 2001,357-364.

Richter, A., "Who Is Going to Take the Prize? Performance Predictors of Business Administration Students", European Business School Working Paper No. 4, 2003.

Siegfried, John J., and Strand, Stephen, "Sex and the Economics Student", Review of Economics and Statistics, 59, May, 1977, 247-249.

Siegfried, John J., "Male-Female Differences in Economic Education: A Survey", Journal of Economic Education, 10 (2), 1979, 1-11.

Steedman, Hilary, Green, A., Bertrand, Oliver, Richter, Ansgar, Rubin, Marcus and Weber, Klaus, Assessment, Qualifications and Standards: The UK Compared to France, Germany, Singapore and the US, London: Centre for Economic Performance Report, London School of Economics and Political Science, 1997.

Steers, Richard M., "Antecedents and Outcomes of Organizational Commitment", Administrative Science Quarterly, 22, 1977, 46-56.

Tay, Richard S., "Students' Performance in Economics: Does the Cultural Norm Hold Across Cultural and Institutional Settings?", Journal of Economic Education, 25 (4), 1994, 291-301.

Watson, T., "Recruitment and Selection", in: Sisson, K. (ed.), Personnel Management in Britain, Basil Blackwell, Oxford, 1989, 125-148.

Williams, Mary L., Waldauer, Charles, and Duggal, Vijaya G., "Gender Differences in Economic Knowledge: An Extension of the Analysis", Journal of Economic Education, 23 (3), 1992, 219-231.

Wolfe, R.N., and Johnson, S.D., "Personality as a Predictor of College Performance", Educational and Psychological Measurement, 55, 1995, 177-185.

Ziegert, Andrea L., "The Role of Personality Temperament and Student Learning in Principles of Economics: Further Evidence", Journal of Economic Education, 31 (4), 2002, 307-322.

Ansgar Richter, European Business School, Germany

Author Profile:

Dr. Ansgar Richter earned his Ph.D. at the London School of Economics and Political Science in 1999. He worked for several years as a management consultant with McKinsey & Company, Inc., before returning to academia in 2002. Currently he is an assistant professor of organization and human resource management at the European Business School, Schloss Reichartshausen, Germany.