High School Never Ends: Standardized Tests in Hiring Practices

By: Eli Scottsat_test

“It is a little confounding how a test somebody took when they were 17 predicts success in a competitive workplace when they’re 22,” said Kevin Monahan, a career services dean at Carnegie Mellon University. Monahan was referring to the controversial practices of many white-collar employers to ask potential candidates for standardized test scores such as the SAT when applying for employment.

Consulting firms and investment banks, including Bain & Co., and McKinsey & Co., and investment banks like Goldman Sachs, Group Inc. still encourage applicants to submit their standardized test scores from high school. For Boston Consulting Group Inc., this is not a new practice; the firm especially focuses on the math portion of the SAT. Many companies don’t claim that SAT results can predict job performance, but they do use standardized tests to assess an applicant’s “critical thinking, problem-solving skills, and quantitative abilities.” In fact, Cvent, Inc., a publicly held software-as-a-service company, emphasizes SAT scores without any concern for how they correlate to job success. For Eric Eden, Cvent’s vice president of marketing, “knowing [the SAT is] a standardized test is really enough for [him].” However, the presupposition of the “standardization” of standardized tests, particularly the SAT, may prove deleterious for the employment prospects of certain demographics based on the inherent problems with the SAT.

Kyle Ewing, head of global staffing at Google, claims that there is “very little correlation between SAT scores and job performance.” While there may be academic support for the argument that cognitive ability can predict job performance, the evidence linking high SAT scores to workplace success is lacking. But the SAT’s lack of predictive success is not its only limitation; there are inherent economic, gender, and racial biases in the structure and composition of the test itself that result in applicants being discriminated against in the hiring process.

Despite The College Board’s claims that there is no significant socio-economic bias in the test, independent researchers have found that students with higher family incomes tend to perform better on the SAT. The National Center for Education Statistics conducted a study on the achievement levels of students from high, medium, and low socio-economic statuses. Thirty-two percent of students from a high socio-economic background earned a score of 1100 or greater on the SAT, while only 9 percent of students from a low socio-economic status achieved the same score, according to education author Rebecca Zwick. This divergence can be attributed somewhat to a more stable lifestyle and posh upbringing, but most of the disparity in SAT scores stems from unequal access to test preparation materials and coaching programs. In fact, achievement test scores were shown to improve by 0.25 standard deviations with the help of coaching, such that a student at the 50th percentile was raised to the 60th percentile. Of the three sections, Critical Reading and Writing both show a 130-point disparity between the top and bottom income brackets of <$20,000 and >$200,000, respectively, and this data elucidates a definite socioeconomic gap in achievement on the SAT.

The proof of this economic bias being exclusive to the SAT lies most strongly in the fact that first-year college GPA, high school class rank, and extracurricular involvement show no correlation to family income, whereas achievement on the SAT is a direct linear correlation to family income. Moreover, family income and status is so relevant to achievement on the SAT that parents’ roles in the economic system can predict success: top professionals’ children outperform white-collar workers’ children and these children outperform blue-collar workers’ children.

Gender bias in the SAT is less prominent than economic and racial disparities, but it persists nonetheless. Males tend to perform better than females on the mathematics portion despite similar overall abilities. This gap is presumably caused by the structure of the SAT. Because women are more likely to doubt their answers and re-check them, the stress caused by the test’s time constraint can result in males outperforming females.

Lastly, and most importantly, extreme racial disparities in SAT achievement exist, and the performance gaps between races are steadily increasing. While the average SAT score for all test-takers was 1498 out of 2400, the averages for Hispanic Americans and African-Americans were 1354 and 1278, respectively. Test design only exacerbates these disparities. by utilizing two practices that are racially biased against non-Whites.

The first of these inhibiting features is the usage of an experimental section every year to sample possible questions for an upcoming test year. Despite its seeming neutrality, the Educational Testing Service’s mandated policy is racially biased because it perpetuates the status quo. Jay Rosner of the Princeton Review Foundation analyzed experimental sections and found that the ETS required an appropriate question to parallel the results of the entire test such that “if high-scoring test-takers—who are more likely to be white (and male, and wealthy)—tend to answer the question correctly in pretesting, it’s a worthy SAT question.” This process of point bi-serial correlation has been used to show that “black questions,” questions in which more African-Americans answer correctly than whites, never make it out of the experimental section. Thus, the SAT contains solely “white questions.”

The SAT also employs Differential Item Functioning, which means that the test makers use questions they know will have different success rates for different racial groups. The quintessential example of Differential Item Functioning is the inclusion of the oarsman-regatta analogy, a question that test designers knew would favor whites because of its allusion to cultural norms absent in urban and diverse environments. Thus, minorities do not necessarily perform worse on the SAT on the account of different upbringings, but the SAT presents structural biases in its questions.

As evidenced by the cultural and gender biases of the SAT, the usage of the SAT in consideration of hiring practices for white-collar jobs has a predictable outcome: nearly 97 percent of corporate senior executives in the United States are white. Accordingly, only 5 percent of professionals are African American, despite the fact that African Americans represent 12.7 percent of the total workforce. Although there may be other confounding factors affecting the hiring of minorities in corporate jobs, many companies and colleges, in fact, are taking note of the discrimination in the SAT. Google abolished its formerly score-fixated hiring policies two years ago, moving toward a more egalitarian hiring process based on interviews. In this manner, because of the cultural and gender biases inherent in the SAT, standardized tests should not be included in the criteria for job selection for white-collar employment because this practice disproportionately precludes minorities, economically disadvantaged people, and women from white-collar corporate employment.