SAFS Logo

The Use and Abuse of Systemic Discrimination

September 2000

With intentional discrimination now being rare, the concept of systemic or substantive discrimination is repeatedly invoked in the continuing debate on equity in Canadian universities. Three recent developments illustrate assumptions made about such discrimination, which is said to occur when institutional policies and practices unintentionally discriminate against women and certain ethnic groups. The first is the Natural Sciences and Engineering Research Council’s (NSERC) restriction of its University Faculty Awards to women. The second is a study of the University of Toronto (U of T) faculty’s ethnic composition, showing that it is much less diverse than the student body, and that it is unlikely to change much in the foreseeable future.

The third - and most troubling - is allegations of racism in U of T’s Physics Department, made by an Ontario Human Rights Commission (OHRC) investigator in his report on of a grievance by geophysicist Dr. Kin-Yip Chun.

The report claims that Chun’s failure to obtain a tenure-track appointment at the University in four separate competitions between 1987 and 1992 reflects an appointments procedure which "effectively screens out racial minorit[ies]." The investigator dismissed an internal U of T investigation finding no evidence of racial discrimination because it did not consider the implications of a "prevailing substantive theory of discrimination."

But there is no such "theory." In the name of combatting racism and sexism, activists have instead adopted a simplistic doctrine called the proportionality fallacy: a presumption that, if women and ethnic groups are not represented in the work force roughly in proportion to their population ratios, then some form of discrimination must be the cause. This doctrine obviates the need to examine historical, sociological, psychological and other factors that might account for the patterns of representation.

In universities, the doctrine opens the door for unsustainable policies such as preferential treatment of special interest groups, and almost always detracts from the principle of merit in education and scholarship. Moreover, because it absolves its proponents from the responsibility of actually identifying a discrimination mechanism, it legitimates shoddy arguments and may even support witch-hunts for racism and sexism in every aspect of human behaviour.

Unintentional discrimination may, of course, occur in educational institutions and workplaces. It can be rectified when the source of bias is identified; two examples are: the provision of facilities enabling the disabled to attend colleges and universities, and of left-handed writing tablet chairs in lecture halls. Thus the possibility of unintentional bias cannot be dismissed, but the emphasis must be upon identifying the actual bias mechanism, so that appropriate adjustments can be made.

This rational understanding of systemic discrimination, however, seems not to have gained acceptance because the proportionality fallacy is now deeply entrenched; as an example, aptitude tests used in screening students or for employment are often arbitrarily restructured to eliminate any deviation from proportionality. Thus, in the USA, universities frequently adjust scores on scholastic aptitude tests by race solely on the presumption that such differences must be the result of some unidentified bias factor.

The doctrine originated in attempts to eradicate the deeply-rooted effects of racial segregation in the USA. The landmark 1964 Civil Rights Act targeted only intentional employment discrimination; but the pace of change was glacial, and a US Senate report on the 1972 Equal Opportunity Employment Act viewed discrimination as being much more complex than the consequence of ill-will. It noted that "Experts... describe the problem in terms of ‘systems’ and ‘effects’ [such as] the mechanics of seniority [and] the perpetuation of ... discriminatory practices through various institutional devices, and testing and validation requirements." These remarks reflect a precedent-setting 1971 US Supreme Court ruling, which disallowed a high-school equivalency education requirement for promotion in a previously segregated Southern company, because blacks disproportionately failed this test. The Court viewed the company’s financial assistance for employees’ education as evidence of lack of intent to discriminate, but nevertheless prohibited the requirement because of its consequences for the pattern of employment. Furthermore, the Court placed upon the employer the burden of demonstrating that any employment test was clearly related to a specific job performance.

The 1972 Act granted enforcement powers to the US Government’s Equal Employment Opportunities Commission (EEOC), which has since aggressively promoted the notion that disparate impact, as opposed to disparate treatment, is a test for discrimination. Hence, if the numbers of "visible minority" employees in a job category fall below 80 percent of what the EEOC deems to be the relevant applicant pool, an employer must show that it has not discriminated, and that any screening procedure is clearly relevant to the specific skills required for the job. Failure to provide evidence deemed satisfactory by the Commission can result in the imposition of severe penalties.

Canada has adopted the philosophy underlying US policy. In 1985, our Supreme Court ruled intent and motive irrelevant in establishing discrimination so that, as in the US, investigations focus on effects. Statistical evidence pointing to "under-representation" of a designated group within a workplace, and anecdotal evidence by individuals have both been accepted by human rights tribunals as indicating systemic discrimination. Furthermore, these tribunals use the balance of probabilities test in adjudication, requiring only that an accuser establish a prima facie case. In practice this means that the accuser has only to make a plausible case, and the accused is burdened with providing a persuasive justification for the conduct. Such an approach is ripe for abuse by grudge-bearers or ideological axe-grinders.

The essence of the proportionality fallacy was identified thirty years ago by Oxford University psychologist Corinne Hutt in her analysis of female and male employment patterns. Concluding even then that these could be explained largely by sexual differences in psychology, and thus by women making entirely reasonable choices having nothing to do with discrimination, she noted that, in the absence of evidence of a mechanism, use of the proportionality criterion reduces the concept of discrimination to a meaningless tautology.

The current paucity of women faculty in the physical sciences and engineering is an excellent example; the evidence points overwhelmingly to a lack of qualified candidates for reasons having nothing to do with any kind of discrimination. US statistics show that, in 1998, women earned only 13 percent of doctoral degrees in physics and engineering, much less than the 50 percent advocated by those insisting on defining equity in terms of outcome rather than opportunity. This is consistent with psychological evidence; thus, amongst mathematically precocious youth, females are much less likely than males to choose mathematics, the physical sciences or engineering s a career, opting instead for fields such as the humanities and biology.

This is not to say that decades ago there was no discrimination against women in academia. A recent detailed study of Canadian university hiring patterns is consistent with British and US observations. For several disciplines the study compared the fraction of female hires with the corresponding fraction of recent doctoral graduates, finding the same pattern for all disciplines: the data suggests discrimination against women prior to about 1970, and subsequently in favour of women.

Hence NSERC’s initial restriction of its University Faculty Awards to women (subsequently aboriginal men have been included) illustrates the consequences of uncritical acceptance of the assumption of discrimination without regard to issues of merit and fairness. Prior to the exclusion of men, women won 57 fellowships and men won 363. It seems hard to avoid the conclusion that disqualifying the category of overwhelmingly successful applicants must have an effect on the operation of the merit principle in the awarding of these fellowships.

In considering the ethnic composition of university faculty, a number of factors should be considered when discrimination is suspected. As an example, in many Western countries, and in particular in Canadian cities such as Toronto, immigration is rapidly changing the demography. On this ground alone, one cannot expect any but the most recent recruits to a workplace to approximately reflect the applicant pool. This is especially true of university faculty, for which the turnover can be very low. Leaving aside the effects of ‘equity’ hiring, the current Canadian university faculty composition appears generally to reflect the availability of suitable candidates at the time of hiring.

Even if one accepts the notion of proportionality, there remains the problem of statistical fluctuation. Calculations show that, for as few as 50 employees, random sampling from a population with just two categories fails the EEOC’s 80 percent criterion one trial in five. For multiple categories or smaller groups the probability that deviations from proportionality are solely due to random fluctuations mounts rapidly.

Considering all these factors, it makes no sense to compare the composition of the "student body" with that of a relatively unchanging faculty as a method of identifying systemic discrimination - the most recent extension of the proportionality fallacy.

Turning now to Dr. Chun’s grievance, we note the conclusions of U of T’s investigator that he was exploited in his role as a research associate, but that there was no racial discrimination in the disputed appointment competitions. Thus our concern lies with imputations of systemic discrimination made in the report of the Academic Freedom and Tenure Committee of the Canadian Association of University Teachers (CAUT) and of both systemic and intentional racism made by the OHRC investigator: both manifest the conceptual and procedural difficulties we discuss above.

One would expect a CAUT committee charged with guarding academic freedom to be aware of the difficulties inherent in investigating systemic discrimination, and of human rights tribunals’ penchant for trampling on academic freedom. But its report is oblivious to these difficulties. It embraces the proportionality fallacy and the concomitant idea that anything resulting in deviation from proportionality suggests racism. It approves the "guilty until proven innocent" mentality of human rights tribunals. While confessing to knowledge of only one side of the case, it is replete with unsupported speculation.

The OHRC report is a disgrace: it divines racism in a troubled relationship between an individual and the University, when alternative explanations are at least as plausible. Typically, it reads racism into a referee’s frank but supportive comment on Chun’s facility with English. It makes much of a computer virus incident while ignoring the implication of one of the contested hiring competitions, in which the search committee chair was the son of a Japanese Canadian interned during World War II, and the successful individual has an Albanian Muslim heritage. Alluding to the ethnicity of the older members of the Physics Department, it speaks of "cronyism," when the ethnic composition reflects the availability of qualified candidates at the time of hiring. It ignores the implications of recent hiring data, which show that, in the last 10 years the so-called "under-represented minorities" were awarded 25 percent of the positions, even though they accounted for only 10 percent of the applicants. Worst of all, it fails miserably to address the central question: Was Dr. Chun demonstrably better qualified than the successful candidates?

The report’s claim that U of T’s internal investigation was "flawed" is unsupportable. The U of T investigator, Professor Yip, is of Asian origin. He is experienced in the resolution of grievances, both as an officer of the University’s Faculty Association, and as a member of the University’s Grievance Review Panel. Furthermore he was assisted by legal counsel familiar with human rights law who is now an Ontario Appeals Court judge. Professor Yip observes that the OHRC report presents no facts not already covered in his own report and that, as both a researcher and one who has handled faculty grievances, he is familiar with "looking for a pattern [of discrimination]." "My investigation ... did not see a pattern of discrimination that can be supported by fact," he wrote.

In conclusion, a charge of systemic discrimination cannot be substantiated without a reliable and disinterested analysis. First, alternative explanations of an unexpected pattern of participation or employment must be examined. Second, a specific mechanism of bias should be demonstrated in the particular case. Arguments based merely on an assumed necessity of proportionality between ‘populations’ and the participation of specific groups (in education or employment) are usually fallacious.

It’s time that Canadians questioned both the ways in which the concept of systemic discrimination has been elevated to the status of a "substantive theory" by our judicial and other organizations and the fairness of the methods used by human rights tribunals to investigate claims of both intentional and systemic discrimination.