The Search for New College Applicant Metrics is Valid, but Assessment Centers Are Not the Answer

Cameron Hecht; Twitter Blog Post #2; Due Date: 10/6/2014; Reaction to Twitter post by @StoibsKd on 10/5/2014

In his article, “Throw Out the College Application System”, Adam Grant, professor of management and psychology at the Warton School at the University of Pennsylvania, argues for the use of “assessment centers” to measure high school students’ aptitude for college. Assessment centers, he describes, have been used to assess individual aptitude in a number of domains, beginning with aptitude to be a spy in 1944 (Grant, 2014). At assessment centers, individuals complete interviews and tasks, both individually and in groups. Participants are assessed not only on their task performance, but also on important personal attributes such as “leadership skills, motivation and optimism” (Grant, 2014). Although assessment centers have been used to assess job applicants (including potential school principals and administrators), they have never been used to assess college applicants. Grant argues that such centers may provide a measure of student aptitude that is not revealed in grades or standardized test scores, where some students with high aptitude fail for various reasons. He claims that with such a measure, “colleges might be less likely to reject the next Walt Disney” (Grant, 2014).

While Grant’s proposal is innovative and well founded in its goal to stop high-aptitude students from falling through the cracks, it contains aspects that, if not addressed, could be particularly detrimental for low-income students. Perhaps the biggest problem with implementing assessment centers as key predictors of college success is the potential for this approach to lead to greater stratification of colleges based on family income. Stratification by income is already prevalent across colleges. One recent study, for instance, suggested that the majority of high-achieving students in the bottom income quartile in the U.S. do not even apply to a selective college or university (Hoxby & Avery, 2013). Such stratification results not only from discrepancies in application behavior, but also other factors such as information seeking. A recent article released by the American Council on Education describes how students from high-income families tend to make more use of college rankings information than students from low-income families. This causes proportionally more high-income students to attend selective colleges, reap the financial benefits of having attended such a college, and pass on to their children an array of income-related advantages when they apply to college (ACE, 2014). Grant’s proposal threatens to exacerbate this stratification by adding yet another income-related advantage. Similarly to standardized tests that are used to determine college aptitude, assessment centers’ measures would likely become studied and used to inform the design of expensive preparatory courses that afford advantages to students whose families can afford such courses. By conferring an additional advantage to students from high-income families, the already high correlation between college selectivity and family income threatens to be increased.

Another aspect of Grant’s proposal that may detriment low-income students relates to students’ differential experiences of diagnostic situations based on group affiliation and related stigmas. A large body of social psychological research has investigated the phenomenon of “stereotype threat”, the perceived risk of personally confirming a negative stereotype about a group to which one belongs (Steele & Aronson, 1995). Research on stereotype threat has found that individuals tend to perform worse on assessments when they perceive the possibility of confirming a negative stereotype about their group. It has been specifically found that when an assessment is described as a measure of intellectual ability (compared to when it is not described as such), low-socioeconomic status (SES) participants’ performances on tests suffer, presumably because they are distracted by the fear of confirming the stereotype that people from poor backgrounds have low intellect (Croizet & Claire, 1998). If assessment center testing were to become a known diagnostic of college aptitude, low-income students’ performance at these centers may suffer due to stereotype threat. This possibility illustrates another way in which Grant’s proposal may hold unintended negative consequences for low-income applicants.

Finally, the funding required for students to attend assessment centers, which Grant (2014) claims “universities and colleges could provide”, might be better spent in ways that could improve academic outcomes for high school students from low-income families, improving their academic performance and ultimately decreasing college stratification by making low-income students more competitive applicants. Research has revealed a number of successful interventions for improving academic outcomes for low-income or otherwise disadvantaged students in college (e.g., see interventions targeting single-parent college students in Goldrick-Rab & Sorensen, 2010). If modifying these types of interventions to meet the needs of low-income high school students could improve these students’ aptitude and competitiveness when applying to college, then spending money instead on adding another metric for college applications may incur too great an opportunity cost to be justifiable.

Grant’s goal of finding ways to assess students on dimensions other than academic performance or test scores is appropriate. Certainly there are high-aptitude students whose ability is not reflected by traditional metrics, and further ideas and research on ways to assess currently unmeasured applicant attributes could undoubtedly be fruitful. Unfortunately, the potential unintended effects of Grant’s proposal on low-income students, if not addressed, seem to constitute too high a cost to outweigh the additional information that assessment centers would allow applicants to provide to colleges. Furthermore, even if these unintended effects can be addressed in a satisfactory manner, it is not clear that Grant’s expensive proposal would be the best way to spend federal or college-provided dollars. In other words, adding an assessment center metric would at worst directly detriment low-income students and at best incur an opportunity cost to improving outcomes for low-income students through intervention. Grant is right to consider new ways to assess applicant aptitude, but for the reasons described above, assessment centers are not a satisfactory way to achieve this goal.