Stereotype threat and women’s performance in engineering

Stereotype threat and women’s performance in engineering

Bell, Amy E


Recent research has demonstrated that stereotype threat-the concern that others will judge one negatively due to a stereotype that exists about one’s group-interferes with women’s performance on standardized math and engineering exams. In the current research we find that when a shortened version of the Fundamental of Engineering Exam is described as a test that is diagnostic of ability (i.e., when stereotype threat is high) women perform worse than men on the test. When stereotype threat is reduced, however (by characterizing the test as non-diagnostic or as not producing gender differences), women do just as well as men. The implication of these results for improving the engineering education environment is discussed.


A. Persistent Under-representation of Women in Engineering

Of the bachelor’s degrees awarded in science and engineering in the U.S. in 1996, 47 percent were awarded to women (up from 38 percent in the mid-1980s). However, the percentages differ greatly by field. For instance, in 1996, women earned the following percentages of bachelor’s degrees by field: 73 percent psychology; 50 percent biological and agricultural sciences; 51 percent social sciences; 37 percent physical sciences; 33 percent earth, atmospheric, and ocean sciences; 46 percent mathematics; 27 percent computer science; and, 18 percent engineering [1]. The percentages of master’s and doctoral engineering degrees awarded to women in 1996 were 17 percent and 12 percent, respectively [1].

The U.S. engineering bachelor’s degree figures have steadily improved for women from 15.4 percent in 1990 to 18.6 percent in 1998 to 21 percent in 2001 [1-3]. However, there is a disparity across the various engineering disciplines. For instance, women earned only 14 percent of the electrical, computer and mechanical engineering bachelor’s degrees in 2001, but they earned 34 percent, 36 percent and 39 percent of the industrial, chemical and biomedical engineering bachelor’s degrees, respectively [3]. The overall figure of 21 percent is evidence that electrical, computer and mechanical engineering represent the majority (52 percent) of bachelor engineering degrees awarded [3].

Not only do women earn a disproportionately lower share of the awarded engineering degrees, they also have lower retention rates than men. Table 1 shows the nationwide retention rates for women and men at three points in the undergraduate engineering pipeline for a 1982-1993 cohort. This data illustrates that the retention rates of women are lower at every point in the undergraduate engineering pipeline.

Paradoxically, although the degree completion rate for women is significantly lower than the rate for men, the grade point averages (GPAs) of women and men were nearly identical (GPA = 2.98, standard deviation = 0.437 for women; GPA = 2.88, standard deviation = 0.561 formen) [4].

Many studies have investigated the reasons why women remain persistently under-represented in science, mathematics, and engineering. Some of the primary reasons that are given to explain this phenomenon include: women’s loss of confidence/lower self-confidence [5]; a chilly classroom/learning environment for women [6]; gender-role socialization [7]; and women’s innate inability (relative to men’s) in science and mathematics [8, 9].

An innovative approach to understanding the lower performance of women on difficult math tests offers a new perspective on the performance and retention of women in fields that require a significant amount of mathematics, like engineering. This new approach suggests that-in addition to lower/loss of self-confidence, chilly climate factors and socialization effects-the differences in women’s performance and retention may be explained, at least in part, by the influence of stereotypes and prejudice on their performance. This influence has been characterized as stereotype threat.

B. Stereotype Threat Defined

Research on stereotype threat examines the experience of being in a situation where one risks being judged negatively due to a commonly held devaluing stereotype that exists about one’s group [10-16]. The primary hypothesis of stereotype threat research is that when one is in a situation in which a negative stereotype exists about one’s group, then the concern with being judged or of self-fulfilling the stereotype interferes with one’s performance. This predicament of being in a situation in which one faces judgment based on societal stereotypes begins with a prejudice that is widely known-even among people who do not believe the stereotype. For instance, our society alleges women have inferior math abilities compared to men. So when a woman finds herself in a situation in which her math skills are being tested (e.g., a formal test or answering a question in math class), she experiences a pressure that may degrade her performance. This is a predicament that others, not stereotyped in this way, do not face. Stereotype threat hypothesizes that this predicament creates a pressure that leads to performance degradation.

C. Stereotype Threat Example

Consider the following illustration of stereotype threat and performance [16]. A study examined black and white men’s performance on a golf putting task. The participants were randomly separated into two groups. The first group performed the putting task after being instructed that the task was a measure of athletic ability; the second group performed the task after being instructed that it was a measure of sports intelligence. The black men outperformed the white men when the task was characterized as testing athletic ability. However, the white men outperformed the black men when the task was characterized as testing sports intelligence. How the test was represented to the participants affected their performance. The characterization of the test as a measure of athletic ability induced stereotype threat for the white men while the characterization of the test as a measure of sports intelligence induced stereotype threat for the black men. Ironically, this simple putting task tested neither athletic ability nor sports intelligence.

D. Stereotype Threat Elaborated

Stereotype threat is not tied to one particular group; it is likely that everyone experiences it in various situations throughout her or his life. The strength of stereotype threat varies with the situation; for instance, a woman may suffer from its effects in her math class, but not in her elementary education class. Stereotype threat is not an internalized belief in the stereotype or a fear that it may be true. Nor is stereotype threat a belief that others will be prejudiced against you. It is a threat that is “in the air [13].” Stereotype threat can be experienced even if one does not believe the stereotype or worry that the stereotype could be true about oneself. A person can have high self-confidence and still suffer from the effects of stereotype threat; indeed, past research indicates that stereotype threat effects are largest among the best students who are most identified with the subject matter [10, 14].

An important difference between the stereotype threat explanation for women’s under-performance on math tests and the explanation based on lower self-confidence is that stereotype threat only occurs in specific situations-situations in which the negative stereotype applies (i.e. when a woman is taking a difficult math test). In contrast, the self-confidence explanation is intrinsic to the person-it is felt in all situations. Furthermore, the stereotype threat explanation does not imply that there is something wrong with the women; however, a lack of self-confidence insinuates that the women need to “fix themselves” and improve their poor self-image.

The present research further evaluates the impact of stereotype threat on women’s performance on standardized engineering tests. Section II outlines the previous, germane research on stereotype threat and gender. Section III describes a new study on stereotype threat in engineering: the design, participants, procedure, scoring, and results are presented. Section IV discusses the implications of these new results.


A. Stereotype Threat and Women’s Math Performance

Previous research has demonstrated the dramatic impact of stereotype threat on women’s performance on math tests. From a stereotype threat perspective, a student’s concern about being stereotyped by others should be highest when two factors are at play: (i) the student is performing poorly (e.g., the questions are difficult); and, (ii) a stereotype might be applied to the student (e.g., the stereotype that women are not good at math). Based on previous research, it is in this situation that differences between men and women’s performance should emerge. In one study, women with strong math backgrounds performed worse than men with similar backgrounds on difficult tests-yet women performed equally well on easy tests [10]. Although stereotype threat is the purported cause of the performance differential, other interpretations remain. For example, one could argue that women’s lesser math capabilities only become evident on more advanced material.

Another study examined the performance of highly selected women and men on a difficult math test when the relevance of the stereotype was manipulated by how the test was characterized. In the “relevant” stereotype threat condition, participants were told that the test had shown gender differences in the past. Conversely, in the “irrelevant” stereotype threat condition, participants were told that the test had never shown gender differences. There are two important points to note. First, in the relevant stereotype condition, it was not stated that men had outperformed women in the past, just that there had been gender differences. Second, in the irrelevant stereotype condition, the validity of the stereotype was not attacked; instead, the particular math test was characterized as having shown no gender differences in the past. The results showed that men significantly outperformed women in the relevant stereotype condition and women and men performed equally well in the irrelevant stereotype condition [10]. Since the women’s performance improved and equaled the men’s when the difficult test was characterized as having shown no gender differences in the past, it suggests that there was something in the testing situation that was responsible for the difference. After all, if the women were less mathematically capable, then a different test characterization would have no impact on their performance. Thus, there is strong evidence that women’s underperformance on the difficult tests resulted from stereotype threat and not from innate inability.

In addition to explaining differences in performance, stereotype threat may also illuminate the lower retention rates of women in math-related fields. Researchers have argued that the stereotype threat that women experience in math-related domains may cause them to “disidentify” with the domain [13]. In other words, women drop out of math in order to avoid the evaluative threat that they sometimes feel in this domain. Indeed, one study indicated that women expressed less interest in pursuing academic majors and careers involving high levels of mathematics after watching stereotypic TV commercials [12].

B. Stereotype Threat and Women’s Engineering Performance

Similar to the prejudice that women are inferior to men in mathematical ability, there exists a societal stereotype that women are less capable than men in engineering ability. In a previous study, we showed that stereotype threat undermines women’s performance on engineering exams [15]. Eighteen women and thirty men, engineering sophomores and juniors at Virginia Tech, participated in the experiment. The test questions were a subset of questions from the general portion of the standardized Fundamentals of Engineering Exam (FEE) [17]. One-half of the participants were randomly assigned to take an easy version of the engineering test and the other half took a difficult version of the engineering test. We only selected participants for the studies who indicated that they had a relatively high grade point average (GPA) in engineering, and who stated that they were good in engineering and that it was important for them to be good in engineering. The test directions, called the “diagnostic” directions (see Figure 1), were the same for both the easy and difficult tests. The diagnostic directions increased the stereotype threat for women with its assertion of being able to distinguish the good engineers from the bad.

Our hypothesis was twofold. First, we anticipated no difference in the performance between women and men on the easy exam since the stereotype threat would be eliminated (or severely mitigated) by positive performance. Second, we expected a gender difference to emerge on the difficult exam in which the stereotype threat condition was “high” (difficult questions and diagnostic directions). Our results indicated that women and men performed equally on the easy engineering exam (57 percent score for both women and men), whereas men performed significantly better than women on the difficult engineering exam (21 percent score for women, 34 percent score for men) [15]. This difference remained statistically significant even when controlling for GPA and self-reported competence and importance of engineering.


A. Design

The previous research on stereotype threat’s impact on women’s performance on engineering tests does not conclusively rule out inherent gender differences on the difficult test [15]. For example, one might argue that women’s lesser engineering capabilities only become apparent on more advanced material. However, this seems unlikely for two reasons. The results of the previous study indicated no gender differences in GPA, easy test performance, self-reported engineering competence, and self-reported importance of engineering. Also, the difficult test questions were hard due to a lack of familiarity, not complexity; it is unlikely that only unfamiliar questions would reveal innate gender differences.

To further investigate whether stereotype threat-and not inherent gender differences-is the explanation for differences in performance, we designed a study to directly assess the effects of stereotype threat on women’s engineering performance. All participants took the difficult engineering test; however, we manipulated the relevance of stereotype threat with three sets of directions that characterized the test differently. In the “relevant” stereotype condition, the diagnostic directions represented the test as being able to discriminate between capable and incapable engineers; this induced a high stereotype threat condition for the women (similar to the previous study [15]). In the non-diagnostic directions condition we reduced stereotype threat by characterizing the test as being unable to judge a person’s engineering competence; this mitigated the stereotype threat for the women since the test is unable to judge their performance. Finally in the gender-fair directions condition we reduced stereotype threat by characterizing the test as one in which no-gender differences have been found. Figure 1 highlights the germane passages from these two sets of directions.

We developed the difficult engineering test from questions available as practice tests for the general portion of the FEE. We reduced the number of questions and shortened the duration, but maintained the proportions of questions for each of the various engineering areas. The composition (and ordering) of the eighteen questions were as follows: six math, two electric circuits, two statics, two chemistry, and one each for thermodynamics, dynamics, material science, computers, ethics, and engineering economics.

B. Participants

The participants were recruited to take part in the study through email and in-class announcements. We only selected participants who indicated that they had a relatively high grade point average (GPA) in engineering, and who stated that they were good in engineering and that it was important for them to be good in engineering (GPA scores were subsequently verified by official university records). We selected participants using these criteria because previous stereotype threat studies have indicated that stereotype threat effects are largest among the best students who are most identified with the subject matter [10, 14]. Nine women and 18 men took the test in the diagnostic condition, 11 women and 18 men took the test in the non-diagnostic condition, and 9 women and 18 men took the test in the gender-fair condition.

C. Procedure

The participants reported to each testing session in small, mixed gender groups; students were randomly assigned to the three test conditions. To begin, the students read the engineering test directions-either diagnostic, non-diagnostic, or gender-fair (these labels were not available to the students). They then took the difficult engineering test that was computerized (a paper copy was also distributed for working out the problems); they were given 30 minutes to complete it. Upon completion of the test the participants were reassured that no one was expected to do well since the advanced test material was beyond their current level. The participants were thanked for their time and paid $15.

D. Scoring

In scoring the answers we separated the math questions from the engineering questions. Our previous study demonstrated that the difficult math questions were relatively easy for our participants even with the diagnostic instructions (there were no significant differences between the women’s and men’s math scores).

In calculating a participant’s score for the engineering part of the exam we did not include questions that were directly related to the participant’s major. The engineering major (i.e., expertise area) for each question is sometimes ambiguous. For instance, industrial and systems engineering (ISE) students take a materials science (MSE) course in their junior year; however, MSE is not an ISE expertise area. The final categorizations were determined by the required courses in only the freshman and sophomore years of each engineering department. For example, the two electric circuit questions were not included in the score calculation for electrical and computer engineering majors. However, the computers and ethics questions were included in the score calculation for all engineering majors. We did this because we reasoned that the questions in a student’s major would not be difficult for her or him. We refer to this calculated engineering score as the “non-expert engineering score”. Moreover, we controlled for guessing in the non-expert engineering score by subtracting one-fourth of a point for each wrong answer. This normalized a chance performance to zero instead of twenty-five percent; it also implied that the scores were not inflated by random guessing.

E. Hypothesis

We expect that in the diagnostic instructions condition in which stereotype threat is high, women will under-perform compared to men on the engineering test. However, in the non-diagnostic instructions and gender-fair instructions conditions, we anticipate that the women will perform equally with the men.

F. Results

We tested for differences in the data using focused comparisons that tested the specific hypotheses we predicted and further examined our findings by testing for differences between specific means.1 We compared means using the F statistic which is computed as the ratio of the variability due to differences between the mean scores of the conditions to the variability due to differences between the data and mean score within each condition. For each computed F statistic, a corresponding p-value is also calculated; it describes the level of statistical significance (p is a function of F; as F increases, p decreases). Small values of p, typically p

A focused contrast of our prediction that women who took the test with diagnostic instructions (high stereotype threat) would perform worse than people in the other five conditions (women/ non-diagnostic, women/gender-fair, men/diagnostic, men/ non-diagnostic, men/gender-fair) was significant. The F statistic testing this comparison was 4.42; the probability that such a large F value occurred by chance is less than 5 percent (p

A comparison of women and men’s performance in each of the three conditions was also conducted. Women (mean score = 9 percent) performed worse than men (mean score = 30 percent) in the diagnostic instructions condition (F = 4.92, p 0.40) and women performed as well as men in the non-diagnostic instructions condition (mean score = 21 percent for both women and men, F = 0, p > 0.90). It is interesting that the performance of both men and women tended to drop in the non-diagnostic instructions condition compared to the gender-fair condition. Perhaps the participants took the test less seriously and thus did not perform as well when the test was described as non-diagnostic.

In addition, the effect of the different conditions on each gender’s performance was examined. Women in the diagnostic instructions condition performed worse than women in the gender-fair instructions condition (mean score = 9 percent and 28 percent respectively, F = 3.93, p = 0.05) and women in the diagnostic instructions condition tended to perform worse than women in the non-diagnostic instructions condition, but this difference did not reach statistical significance (mean score = 9 percent and 21 percent respectively, F = 1.71, p = 0.20). Among men there was a tendency for men to score lower in the non-diagnostic condition than in the other conditions, but none of these differences reached statistical significance (mean score = 30 percent in diagnostic condition, 21 percent in non-diagnostic condition, and 36 percent in the gender-fair condition; all F values 0.05).

An analysis of performance on the math questions revealed that, consistent with our previous research [15], these questions were indeed relatively easy for all of the participants and there were no significant differences between women and men’s scores in any of the conditions. The mean math scores for women (W) and men (M) were: 56 percent (W) and 46 percent (M) in the diagnostic condition; 50 percent (W) and 57 percent (M) in the non-diagnostic condition; and, 50 percent (W) and 49 percent (M) in the gender-fair condition.


The results of this study extend previous research and further support the idea that stereotype threat undermines women’s performance on engineering exams [15].

As part of the stereotype that women are less capable than men in engineering ability, there is a commonly held belief (particularly among engineering educators) that students have to be “tough” in order to succeed in engineering and that the women students tend to not be as tough as the men students. One can almost imagine their interpretation of the results of this research: “If the women’s performance is going to depend on the phrasing of one or two sentences in the test directions, then they are too sensitive and not tough enough to be good engineers.” Is this true-are the women too weak? Consider the following. A group of white men with strong math backgrounds took a difficult math test [14]. One-half of the participants took the test after it had been suggested that, in general, Asians were better at math than whites (the stereotype threat condition). The remaining participants took the test when there was no mention of Asian-white math ability differences (the control group). The white men who took the test under the stereotype threat condition significantly under-performed the white men in the control group. Can it be concluded that the under-performing white men are too sensitive and not tough enough; after all, it was only one mention in one test. Unlike the women in math and engineering, white men are not burdened by a history of stereotypes regarding their lesser abilities. Of course, the answer is that the men are no more weak or sensitive than the women. Human nature-and not a lack of “toughness”-prompts us to be affected by the prejudices that exist about our group.

The current study strongly suggests that stereotype threat may be a more important impediment to women’s success and persistence in engineering than has been realized. Furthermore, a recent, comprehensive study collected and analyzed data from students, faculty and administrators at 53 universities; one of the primary conclusions of the study was that the classroom and department climates were significant factors in the women students’ persistence in engineering [19]. However, there may be hope even despite the rather depressing statistics reviewed earlier. If stereotype threat is undermining women’s success in engineering then there are tractable ways to reduce stereotype threat, improve the climate, and increase women’s success.

In creating classrooms and learning environments low in stereotype threat and warm in climate, it is not enough to just avoid negative behaviors-positive efforts must also be made. When stereotype threat is the remaining barrier for a student, Steele has suggested several “wise” strategies that a teacher can employ to mitigate the impact of threatening stereotypes [13]. The recommended wise strategies include the following:

* expressing optimism about the student’s potential to achieve and succeed;

* assigning challenging-not trivial-work at a challenging, not overwhelming, pace;

* affirming the student’s belongingness based on her or his intellectual potential; and,

* valuing multiple perspectives and approaches to the academic content.

Hall and Sandier have provided a comprehensive list of action items for administrators, faculty and students on warming up the climate for women students [6]. Among the negative behaviors for teachers to avoid are: making seemingly helpful comments that imply women are not as competent as men; disparaging women in general, women’s intellectual abilities, or women’s professional potential; and, using sexist humor as a classroom device. Some recommended positive behaviors for teachers to adopt are: use terminology that includes both men and women; call directly on women students as well as men students in class; and, ask women and men qualitatively similar questions (critical and factual questions).

If engineering educators can create environments in which stereotype threat is low then more women should succeed and persist. Perhaps in such an environment, in which many of the obstacles to women’s success are removed, each woman would be able to achieve the same level of success as her male colleagues. The current research suggests that creating such environments in our engineering programs may be an important way to reduce the underrepresentation of women in our field.


The authors would like to thank Mr. Satyabrata Rout for his assistance in collecting the data in this experiment. This work was supported by a grant from the Alfred P. Sloan Foundation.

1 Our study was designed to replicate the results of previous research on the effect of stereotype threat on women’s math performance [10, 11]; we wanted to determine whether similar variables influenced women’s engineering performance. Thus, we tested our hypotheses using focused comparisons in analysis of variance; this analysis provides a more sensitive test of our results than a standard ANOVA (a two-way ANOVA in this case), Rosenthal and Rosnow provide a detailed description of these procedures and a complete explanation of their proper use [18].


[1] National Science Foundation, “Women, Minorities, and Persons with Disabilities in Science and Engineering: 2000,” Arlington, VA, 2000 (NSF00-327).

[2] Engineering Workforce Commission of the AAES, “Engineering and Technology Enrollments,” American Association of Engineering Societies (AAES), Washington, D.C., 1998 and 1999.

[3] “Databytes,” American Society of Engineering Education Prism, September 2001.

[4] Adelman, C., “Women and Men of the Engineering Path: A Model for Analyses of Undergraduate Careers,” U.S. Department of Education and the National Institute for Science Education, Washington, D.C., 1998.

[5] Seymour, E., and N.M. Hewitt, Talking About Leaving: Why Undergraduates Leave the Sciences, Westview Press, 1997.

[6] Hall, R.M., and B.R. Sandler, “The Classroom Climate: A Chilly One for Women?,” Project on the Status and Education of Women, Association of American Colleges, Washington, D.C., 1982.

[7] Meece, J.L., J.S. Eccles, C.M. Kaczala, S.B. Goff, and R. Futterman, “Sex Differences in Math Achievement: Towards a Model of Academic Choice,” Psychological Bulletin, Vol. 91, 1982, pp. 324-348.

[8] Benbow, C.P., and J.C. Stanley, “Sex Differences in Mathematical Reasoning Ability: More Facts,” Science, Vol. 222, 1983, pp. 1029-1031.

[9] Benbow, C.P., and J.C. Stanley, “Sex Differences in Mathematical Ability: Fact or Artifact?,” Science, Vol. 210, 1980, pp. 1262-1264.

[10] Spencer, S.J., C.M. Steele, and D.M. Quinn, “Stereotype Threat and Women’s Math Performance,” Journal of Experimental and Social Psychology, Vol. 1999, 35, pp. 4-28.

[11] Quinn, D.M., and S.J. Spencer, “How Stereotype Threat Interferes with Women’s Math Performance,” Journal of Social Issues, Vol. 57, 2001, pp. 55-71.

[12] Davies, P.G., S.J. Spencer, D.M. Quinn, and R. Gerhardstein, “All Consuming Images: How Demeaning Commercials that Elicit Stereotype Threat can Restrain Women Academically and Professionally,” Personality and Social Psychology Bulletin, Vol. 28, 2002, pp. 1615-1628.

[13] Steele, C.M., “A Threat in the Air: How Stereotypes Shape Intellectual Identity and Performance,” American Psychologist, Vol. 52, No. 6, June 1997, pp. 613-629.

[14] Aronson, J., M.J. Lustina, C. Good, and K. Keough, “When White Men Can’t Do Math: Necessary and Sufficient Factors in Stereotype Threat,” Journal of Experimental Social Psychology, Vol. 35, 1999, pp. 29-46.

[15] Bell, A.E. and S.J. Spencer, “The Effect of Stereotype Threat on Women’s Performance on the Fundamentals of Engineering Exam,” Proceedings of the American Society of Engineering Education Annual Conference, Montreal, Canada, June 2002.

[16] Stone, J., C. I. Lynch, M. Sjomeling, and J.M. Darley, “Stereotype Threat Effects on Black and White Athletic Performance,” Journal of Personality and Social Psychology, Vol. 77, 1999, pp. 1213-1227.

[17] National Society of Professional Engineers, , accessed July 7, 2003.

[18] Rosenthal, R., and R.L. Rosnow, Contrast Analysis: Focused Comparisons in the Analysis of Variance, Cambridge University Press, Cambridge, England, 1985.

[19] Goodman, I.F., C.M. Cunningham, C. Lachapelle, M. Thompson, K. Bittinger, R.T. Brennan, and M. Delci, “The Women’s Experiences in College Engineering (WECE) Project,” Goodman Research Group Inc., April 2002.


Department of Electrical and Computer Engineering

Virginia Tech


Department of Psychology

University of Waterloo


Department of Psychology

University of Waterloo


Department of Psychology

University of Waterloo


Amy E. Bell is an Assistant Professor in the Department of Electrical and Computer Engineering at Virginia Tech. She received her Ph.D. in electrical engineering from the University of Michigan. Bell conducts research in wavelet image compression, embedded systems, and bioinformatics. She is the recipient of a 1999 National Science Foundation CAREER award, a 2002 National Science Foundation Information Technology Research award, and two awards for teaching excellence.

Address: 340 Whittemore Hall, Electrical and Computer Engineering, Virginia Tech, Blacksburg, VA, 24061-0111; telephone: 540-231-2940; fax: 540-231-8292; e-mail:

Steve Spencer is an Associate Professor in the Department of Psychology at the University of Waterloo. He has conducted numerous studies on the effect of stereotype threat on women’s math performance.

Address: 200 University Avenue West, Psychology Department, University of Waterloo, Waterloo, ON, N2L 3G1; e-mail:

Emma Iserman is a graduate student at the University of Waterloo.

Address: 200 University Avenue West, Psychology Department, University of Waterloo, Waterloo, ON, N2L 3G1; e-mail:

Christine E.R. Logel is a graduate student at the University of Waterloo.

Address: 200 University Avenue West, Psychology Department, University of Waterloo, Waterloo, ON, N2L 3G1; e-mail:

Copyright American Society for Engineering Education Oct 2003

Provided by ProQuest Information and Learning Company. All rights Reserved