Issues and Options
The SAT was developed to predict academic success and the various forms of the IQ test supposedly measure intelligence. However, do these tests achieve the goal of assessing aptitude or do they merely test one's knowledge of mainstream white culture? Do those who score poorly on the tests have less innate intelligence or are the tests "culturally loaded?"
Those who believe that the tests are culturally loaded feel that the language and content of the questions provide an unfair advantage to children of the dominant culture. They believe that the tests do not measure aptitude, but rather they test the knowledge of mainstream white culture. Proponents of this view recognize that minority and non-minority children grow up in different environments, learning different skills and values that are necessary to success in their respective culture. Thus, they assert that the creation of one "universal", unbiased test is impossible. Those opposing the culturally loaded argument believe that the tests are valid and that they accurately measure intelligence and predict future achievement.
In order to examine whether test bias exists in the SAT and IQ test, we examined three tests aimed at African-Americans. Our mission was to probe whether blacks score higher on these tests than whites. The Dove Counterbalance General Intelligence Test (1968), known as the Chitling Test, was developed to show that blacks and whites are fundamentally opposed in their manner of speech. An example of a question from the Chitling test is : "Bo Diddley is an? A) game for children, B) down-home cheap wine, C) a down-home singer, D) a new dance E) a moejoe call (Dove, 1968)." Another test developed to measure the intelligence of blacks was the Black Intelligence Test of Cultural Homogeneity (BITCH) (Robert L. Williams, 1972). This test is composed entirely of words, terms, and expressions which are particular to black culture. Consider the following question from the BITCH: "If a judge finds you holding wood (in California), what's the most he can give you? A) indeterminate (life), B) a nickel, C) a dime, D) a year in county, E) $100 (Jensen, 1980)." A third test aimed at measuring black intelligence is the S.O.B. test. This test continued in the tradition of the BITCH test and thus was coined the Son of a Bitch Test. A sample question from this test is: "Running a game? A) writing a bad check, B) looking at something, C) directing a contest, D) getting what one wants from another person or thing (Psychology Today, May 1974)." Not surprisingly, blacks scored significantly better on all three of these tests than did whites.
These questions might appear ludicrous if you are not familiar with black culture. However, test questions on the SAT may appear equally irrelevant to African-Americans. An example of a culturally biased question from the SAT is: "Runner:Marathon A) envoy:embassy B) martyr:massacre C) oarsman:regatta D) referee:tournament E) horse:stable. (Herrnstein and Murray, 1994)" This question seems more likely to be answered correctly by upper class children (who are predominantly white) because they are more likely to know what a regatta is.
Of course, not all questions on the SAT or IQ tests are biased. Indeed, the tests are composed of mostly neutral questions. Yet, very few questions are geared toward black culture and more questions are biased toward mainstream culture. Therefore, it seems blacks are at a clear disadvantage when taking the same SAT and IQ tests that whites do.
In The Bell Curve, Herrnstein and Murray acknowledge that culturally loaded questions exist. However, they query whether the presence of culturally loaded questions actually affect black test scores. The data presented in The Bell Curve demonstrates that blacks do better on questions that appear to be biased than on those that seem to be neutral. Furthermore, Herrnstein and Murray assert that blacks and whites agree upon which questions are most difficult. Most importantly, they conclude that the tests are valid.
They include evidence that suggests cognitive ability tests are good predictors of educational achievement.
The many opponents of Herrnstein and Murray argue that the tests might accurately predict success in school and in the work place, but point out that these institutions are also dominated by white culture. Therefore, they suggest that one's test score merely reflects an ability to succeed in a white-dominated society. Although these tests might be valid in that they accurately predict future success, it can be argued that this success is not based on intelligence, but rather on one's ability to comprehend and practice the values of the dominant culture.
The disparities in test scores between the various racial groups is well-documented. Some claim that racism is the dominant factor in these differences, while others assert that they are the result of genetic discrepancies between the races. However, an often overlooked educational factor is the socio-economic background of a student. For example, James Fallows charges that bias does exist in standardized tests, particularly in the SAT, yet he contends that "the bias is not racial so much as economic, and the overall point is so bald that it can hardly be ignored." To illustrate his point, he included the following data from the 1974 College Board:
Fallows argues that the above patterns "strongly suggest that what the tests measure is exposure to upper middle class culture- perhaps even the culture of the professional class of the east coast". If he is correct (and the data seems to support him), then perhaps Affirmative Action advocates and minority rights watchdogs have been barking up the wrong tree. If socio-economic background is a better predictor of test scores, why are we concentrating so much on racial issues?
The ETS (maker of the SAT) argues that the SAT provides lower-income students a chance to go head-to-head with students from wealthier backgrounds.
Thus, the SAT supposedly provides students from a small, rural, public high school a chance to compete with students from an elite prep school such as Phillips Exeter Academy for admission into competitive colleges and universities.
But does it? Similar to Fallows, Crouse and Trusheim more recently noted the continuing differences in SAT scores along class lines:
The clear correlation between SAT scores and social class did not escape Ralph Nader and Allan Nairn, who said, "If ETS scores really measure an important aspect of a person's 'merit' or, as ETS specifically calls it, 'scholastic aptitude', then merit in the United States is distributed according to parental income" (Nairn et al., 1980, p. 204).
One example of the many advantages of higher parental income involves SAT preparation courses. These courses, which tend to raise their participants' scores, provide an advantage to wealthier students because only they can afford to take them. Poorer students are left to prepare on their own, if they prepare at all. This does not seem "fair", but are we to eliminate the right of parents to do all they can to support their children's education?
Race and socio-economic status are not separate entities. Rather, they are intertwined because blacks and Hispanics both earn less income and score lower on standardized tests on average than whites and Asians. Harris and Amprey note that "Black Americans are still subject to racism, and proportionately more blacks than whites are poor. Therefore, being black very often leads to lower academic achievement but not merely because of skin color" (Harris and Amprey, 215). Yet, they concluded that the primary factor in test scores is socio-economic background: "...the biggest differences, with regard to academic achievement, have been and will probably continue to be caused by class differences" (Harris and Amprey, 215).
We have initiated policies that favor racial minorities. However, if the primary determinant in test scores is socio-economic background, where does this leave poor white students? Jay MacLeod's Ain't No Makin' It tracked the educational and vocational/professional attitudes and achievement of white and black students from the same low-income housing project. He found that blacks generally had a more positive attitude toward school because they felt their social status was gradually rising. Because of programs such as Affirmative Action, they foresaw a potentially bright future for themselves if they behaved well and performed well in school. On the other hand, the whites by and large felt trapped by generations of perpetual poverty and they tended to harbor negative attitudes toward school. They perceived no benefit in academic achievement and seemed to accept as fate that they would be in jail, end up on welfare, or both. Ultimately, the blacks in this study tended to be slightly more successful in the "real world".
This example suggests that we need to reevaluate our policies that seek to help minorities. Perhaps we should concentrate less on racial background and more on socio-economic background if we are going to make exceptions for categories of students. This approach would not diminish educational opportunities for minorities; rather, it would look to compensate for the factors that have most influenced their education. Regardless of any policy suggestions, the data demonstrates that socio-economic background has a greater affect upon test scores than racial background.
Academic and intelligence assessment of some type is necessary to place students in classes, for college admissions, and to provide information to employers. It is not universally accepted, however, that standardized tests are the best form of assessment. If we conclude that tests are culturally loaded, we are left with three options with which to attempt to provide a fair means of assessment: 1) eliminate standardized tests as a means of measuring aptitude and intelligence, 2) revise the existing tests in an attempt to create greater cultural neutrality, or 3) create new unbiased tests. Government processes such as affirmative action and quotas are also policy measures used to compensate for racial inequalities. These options are examined in the external policy section of this web site.
Many argue that tests are inevitably biased and that test questions always contain content relevant to a specific cultural group. For example, the SAT is often thought to favor white, upper-middle class students. However, efforts to change the tests are stalled by the fact that the tests do measure what they intend to. Therefore, it is argued that alternate forms of academic assessment are necessary. One example of an alternative means of assessment is to use school grades. It would be impossible to fairly evaluate each student on the basis of grades because each school has different requirements and standards. Yet, proponents of an increased emphasis upon grades also support the development of a national curriculum. E.D. Hirsch provides a persuasive argument in favor of universal educational standards: "we cannot conclude from the present state of affairs that deprived children would be predestined to low achievement under a different school curriculum" (Hirsch, 1987). Nevertheless, this system does not account for discrepancies in teacher's standards.
Another option that was proposed in order to reduce the degree of cultural bias on tests is to modify the content of the existing tests. One measure taken in 1973 to reduce bias was to gear 1/6 of SAT test questions to minority students (Washington Times, Feb. 23 1992). Currently, the SAT consists of 85 verbal and 65 math questions. On the math section only 5 questions typically have cultural reference and one of these five is aimed at African-Americans. Although this is an important step in providing neutral forms of assessment, it is not sufficient. To provide a completely fair means of assessment, the test should contain an equal number of questions geared towards minority culture and white mainstream culture (Washington Times, Feb. 23 1992). The number of minority-based questions should not be proportional to the number of minorities taking the test. Instead, it should be equal to the number of mainstream-oriented questions in order to achieve cultural neutrality.
Some argue that all of the current test formats are inherently biased. These individuals believe that multiple choice tests are not the ideal form of academic and aptitude assessment (Los Angeles Times, Nov. 24, 1991). For example, Governor Pete Wilson of California pushed for student-centered, performance-based exams in 1991. These tests were to consist of essays, oral exams, and complex math problems, rather than multiple choice questions. It is argued that essays and oral exams allow the student to critically examine material and demonstrate his/her analytical reasoning ability. Multiple choice tests often merely test a knowledge of white, upper-middle class culture. Yet, essays and oral exams allow students to use their cultural knowledge and experiences to answer the questions. Unfortunately, multiple choice questions look for specific word answers that are often foreign to minority students.
Government 375: Educational Reform and Ideology