Assistant Professor and Coordinator, Department of Cooperative Education, Northeastern University, Boston, Massachusetts

Statement of the Problem

Cumulative grade point average (GPA) has long been used by educators both to categorize students and to predict their future academic success. Although grades are a valid indicator of academic performance, once students are categorized in this manner there is a temptation to use those grades to predict nonacademic achievement. For example, recruiters hiring co-op students often use cumulative grade point average as a predictor of job performance. As a result, minimum grade point average requirements are imposed on positions, even when there is little justification or validation of the need for such a restriction. Moreover, many co-op programs reflect the same assumption about grades and job performance. This assumption might be explicit, as in programs where a minimum GPA is required to participate in co-op, or it might be implied, as in programs where honor students are given preferential treatment. In either case, average to below-average students who might experience great benefit from participation in co-op are effectively screened from the process, or underserved by the co-op community and hiring employers.

Research on the correlation of GPA to job performance has primarily focused on long-term success factors measured several years after graduation. While no significant relationship between undergraduate grades and long-term job success has been found, there is little evidence to support the assertion that the same is true for co-op, where job performance evaluation is more immediate and short-term. Thus, the aim of this study was to determine the relationship, if any, between cumulative grade point average and job performance of students while on co-op.

Review of the Literature

Since grades are a commonly accepted measure of cognitive development, and provide a convenient, easily defined method of differentiating one student from another, much of the research to determine the predictors of postgraduate job success has centered on undergraduate grades. Whether global success measures (Weinstein & Srinivasen, 1974; Williams & Harrell, 1964) or individual job related criteria (Howard, 1986; Muchinsky & Hoyt, 1973) are used, the results are the same: there is no significant relationship between overall job success and undergraduate grade point average. A recent study conducted by Ann Howard (1986) at AT&T found that of the eleven criteria used to determine successful job performance, only three, intellectual ability, work involvement and nonconformity, were related to grades. She concluded that although grades are predictive of these very specific criteria, it would be misguided to use them to predict overall job success. In fact, other college factors, such as choice of major and extra-curricular activities, are far more predictive than are grades of the administrative and interpersonal skills deemed necessary for successful job performance.

Nonetheless, many employers, whether hiring postgraduates or co-op students, use grade point average as a primary screening device in selecting candidates. This presents a problem for co-op practitioners who are naturally concerned with meeting employer requests, but whose primary goal is enhancing the educational experience of all students, regardless of academic ability.

The dilemma of being asked to refer only "the best and the brightest" to co-op employers is especially difficult because educators recognize that co­op has special benefits for the average and low achieving student. Lindenmeyer (1967) found that co-op experience has a positive effect on the academic performance of all students, but especially for those in the academic lower half of their class. Additional studies corroborate these findings, and show that in addition to having a favorable influence on overall academic performance, cooperative work experience also reduces attrition (Somers, 1986; Smith, 1965; Davie & Russell, 1974), leading to a higher percentage of students in the lower half of the class persisting to graduation. In fact, integrating school and work experience has even been used successfully to redirect high school students identified as potential dropouts (Greenberg & Lieberman, 1981).

Other well-documented benefits of co-op, such as increased self- confidence, greater sense of relevancy and meaning in studies, enhanced career awareness and increased sense of responsibility (Brown, 1976; Mosbacker, 1957; Wilson & Lyons, 1961) can be assumed to have special benefits for those lower achieving students who have perhaps not experienced the same degree of affirmation, independence and confidence in their own judgement as have their higher achieving counterparts.

Despite these indications, lower achievers are under-represented in most optional co-op programs. It has been suggested ( Gotlieb, 1986) that lower achievers are screened out of optional programs, either intentionally, by imposing a minimum grade point average on co-op participation, or unintentionally, by a failure to recruit and serve such students. It would appear that the assumption that grade point average is a predictor of job performance is an underlying factor in this phenomenon. Students who perceive that all the "good" co-op jobs go to those with honor grades, cannot be faulted for choosing not to participate in co-op programs. Co-op practitioners, who fear that low achievers will reflect negatively on the institution and the program, may understandably institute a minimum grade point average as a program requirement. Also, co-op employers, fearing that lower achievers will perform poorly, may impose inappropriate GPA restrictions on positions within their organizations. The cumulative effect of these actions is to effectively screen out of co-op programs those very students who might benefit the most from inclusion.

Given that GPA is a significant factor in the screening and selection of candidates, both for employment and for inclusion in cooperative education programs, it is clear that more information is needed about the effects of grades on student job performance. The aim of this study is to determine the correlation between a student's cumulative grade point average measured just before the start of a co-op term, with job performance measured at the end of the term.


Subjects: Participants in the study were selected from the undergraduate population at Northeastern University's College of Business Administration. To be included in the study, subjects had to have been on co-op during the six­month period of June-December, 1986, and have an employer-completed job performance evaluation on file for this term. Of 1,212 potential subjects, 416 met these criteria. For those who did not meet the criteria, failure to have an evaluation on file at the time of the data collection was the most common reason. To test for possible GPA differences between those students who submitted evaluations and those who did not, the mean grade point average was computed for each group. The results are listed in the table below, and show no significant difference between the two groups.

Table 1
Comparison of GPA for Students With and Without Evaluations

N x sd
With Evals 416 2.788 0.497
No Evals 796 2.880 0.514

Materials: Students' job performance was measured by a five-point "overall performance" rating scale included in the standard evaluation form used by Northeastern University. This form is sent out to all students on co-op about three weeks before the end of the term. They are asked to give it to their direct supervisors who are to complete it and return it to the university. In addition to the overall performance rating, the form invites brief comment on students' strengths and weaknesses, dependability, judgment, and interpersonal skills.

For computation, the overall performance rating categories on the evaluation form of" outstanding," "very good," "average," "marginal" and "unsatisfactory" were converted to a numerical scale from 5 (outstanding), to 1 (unsatisfactory). In converting those scores where supervisors ranked individuals between two categories, .5 was added to the lower rating.

Although every student is expected to have an evaluation on file for each co-op work term, many do not. There are two possible reasons for this, and they represent one limitation to this study. Students may fail to give evaluation forms to supervisors out of concern for the possible rating, in which case it might be expected that only positive evaluation forms would be returned. In addition, some supervisors, although given the form, neglect to send it to the university. Perhaps these supervisors are simply forgetful, but it is also possible that those who fail to send in evaluations do so out of a reluctance to give negative feedback. This would also contribute to the likelihood of only positive evaluations being returned. Since the GPA of those students whose evaluations were returned was not significantly different from those whose evaluations were not returned (see Table 1), it was determined that the phenomenon of missing evaluations did not threaten the validity of the study. Nonetheless, it does represent a limitation, in that a large number of students without evaluations were excluded from the study. Therefore, caution should be used in generalizing to the entire co-op population, information gathered in this study about the overall caliber of co-op evaluations.

Procedure: Co-op at Northeastem's College of Business Administration is an alternating program which begins in the sophomore year and continues throughout the five-year undergraduate curriculum. All students in good standing with the college are required to participate in co-op.

To control for differences in student's co-op situations, participants in the study were placed in one of three groups based on the following criteria:

Group I: (N = 290) included those students who worked for six months in university-sponsored co-op positions.
Group II: (N = 47) included students who worked for employers with established co-op programs, but only for three months.
Group III: (N = 79) included those students who worked for either three or six months in positions they found on their own through contacts with family or friends. These positions, called "own jobs," were approved by a co-op advisor and were subject to the same procedures and regulations which apply to standard co-op jobs. However, since many of the supervisors in these positions had a prior relationship with the student which might influence the evaluation process, this group was established to control for that effect.

The co-op period for the study was the six months from June-December 1986. Before the work period began, the cumulative grade point average for each potential subject was noted. At the end of the work period, each subject's job performance rating was recorded and converted to a numerical value, and product-moment correlations using job performance rating and GPA were calculated for each group.

Results: The results, as shown in Table 2 below, indicate a very low correlation between GP A and co-op job performance rating.

Table 2
Correlation of Job Performance and GPA

N r r2
Group I 290 .148 .022
Group II 47 .089 .008
Group III 79 .130 .017

The low correlation coefficients indicate that GPA is not a useful predictor of co-op job performance. Even using the highest coefficient (r = .148), it is evident that GPA accounts for less than three percent (r2 = .022) of the factors which influence overall job performance while on co-op.

Different measures of variability for grade point average and job performance rating are shown in Table 3. While GPA is normally distributed, evaluation ratings are skewed to the positive, with few low scores. Thus, although job performance evaluations range from 2.0 to 5.0, the mean is 4.375.

Table 3
Measures of Variability

N min max mean sd
GPA 416 4.567 4.000 2.788 .497
Evaluation 416 2.000 5.000 4.375 .606

There are several possible explanations for the clustering of evaluation ratings on the high end of the scale and the low variability among ratings. It may be that there are additional negative work performance evaluations which, for reasons previously mentioned, were never submitted. Or, it may be that the clustering reflects a limitation of the evaluation form itself. With a five­point scale, in which three of the five categories ("average," "marginal" and "unsatisfactory") have descriptors with somewhat negative associations, it is not surprising that ratings are skewed to the positive. Lastly, the clustering may be an accurate assessment of co-op performance reflecting the appropriateness of co-op placements.

Whatever the reason, this clustering may obscure the actual differences among students' work performance ratings, and therefore the correlation of those differences with grade point average. To explore these possible differences, GPA data for students in each rating category were calculated and the results are shown in Table 4, on the following page.

An analysis of variance indicates there is no significant difference in student GPA among rating categories.

Table 4
GPA Data by Rating Category

Performance Rating GPA
N min max x sd
5.0 ( outstanding) 177 1.57 4.000 2.888 .518
4.5 16 2.076 3.750 2.791 .641
4.0 (very good) 192 1.762 4.000 2.727 .459
3.5 7 1.755 3.524 2.489 .571
3.0 (average) 22 2.156 3.774 2.756 .490
2.5 1 2.900 2.900 2.900 .000
2.0 (marginal) 1 2.792 2.792 2.792 .000


The data indicate that GPA accounts for less than three percent of the factors which influence the co-op work performance of business undergraduates. While this is a statistically significant correlation, the magnitude of the relationship is too small to have any practical implications. Thus, the results indicate that GPA is·not a useful predictor of overall business co-op job performance.

Since subjects in this study were drawn exclusively from the business school, these findings cannot be generalized to the entire co-op population. It is possible that in some disciplines, particularly those in which technical expertise is critical to job success, the relationship between GP A and co-op work performance might be stronger. Replicating this study with engineering, computer science or nursing students would yield valuable information about the differences among disciplines in this regard.

Nonetheless, these findings have implications for business employers who broadly impose high GPA as a precondition of co-op employment, or who otherwise use academic achievement as a measure of overall potential. It would appear that these employers would do well to rely less on GPA as a predictor of job success and more on those selection criteria which are predictive of the characteristics related to overall occupational and managerial success, such as interpersonal skills, persuasiveness, precision, judgment, stability, and general

effectiveness (Howard, 1986; Muchinsky & Hoyt, 1973). While the identification of the criteria which would predict these characteristics is beyond the scope of this study, there are general indications that they might include factors such as extracurricular activities, relevance of job to major, and prior co-op work evaluations. Future research to determine the magnitude of the relationship between these factors and overall co-op job performance would be valuable to co-op employers and educators alike, and would further the understanding of the characteristics related to co-op job success.

The results of this study also have implications for co-op practitioners and administrators. Given the low correlation between GPA and co-op work performance, educators should be cautious about setting minimum GPA requirements for participation in co-op. The findings do not support the assertion that such action will enhance the reputation of the institution by improving the quality of co-op job performance.

Those interested in improving the overall effectiveness of their programs would perhaps do better to spend more time with students to ensure a good match between ability, interest and job responsibilities, and to devote more energy to job development in order to ensure a range of job opportunities appropriate to students of all academic levels. Appropriately placed students cannot only be expected to perform well, but also to have more frequent episodes of psychological success (Hall, 1976). In fact, these episodes of psychological success, and the corresponding increase in self-esteem, may account for the greater increase in academic achievement noted for lower achieving students who participate in co-op (Lindenmeyer, 1967). There appears to be little justification for denying admittance to these students who are likely to experience such profound advantages from participation in co-op.

In summary, there are several conclusions which might be drawn from this study. First, it would appear that the benefits which accrue to participants in co-op should not be denied to lower achieving students out of fear that their work performance will compromise the reputation of the institution. Second, it is reasonable to propose that in order to preserve the integrity and mission of cooperative education, educators should strive for appropriate placements, in which restrictions placed on positions are directly related to specific job criteria and not broadly imposed. And finally, these findings indicate that both co-op practitioners and employers should resist the temptation to use academic indicators as predictors of non-academic achievement.