Kallie Clark’s research examines education systems, specifically how policies and practices within education systems serve to facilitate postsecondary inequities for disadvantaged youth.
Reframing discussions of college quality and performance
Determining whether a college is a “good” college depends on who is asking the question. From the student and family’s perspective, a “good” college is one where their child is likely to graduate, and one that provides them with a foot forward as they transition to adulthood. For families and students the question of determining whether a college is a good college boils down to how does this college compare to my other options.
From a college’s perspective, a “good” college is one that does well by the students they serve, given the constraints the school faces. Colleges that serve a large number of students with social, economic, and academic challenges can’t be expected to graduate students at the same rate as a school serving more privileged students. To complicate matters, schools that serve disadvantaged students often have limited funding to dedicate to the many needs of their students. So, from a college’s perspective, the question of a good college boils down to how does a college do compared to their peer colleges serving similar students.
The student and family perspective of comparing colleges
Students with similar four-year college completion prospects can have vastly different outcomes when attending different colleges.
The colored dots represent students of different prospects of completing college based on their high school grades, test scores, and student characteristics. Purple dots represent students with above average likelihood of completing college, blue dots represent students with average college completion prospects, and orange dots represent students with below average college completion prospects. Students with above average and average college completion prospects are graduating at vastly different rates depending on the college. Students with below average college completion prospects seem to struggle at every college in the figure.
The college perspective of comparing colleges
Even amongst elite, four-year colleges students with similar ACTs and GPAs can have vastly different college graduation outcomes.
Variation in student outcomes is an important theme even amongst elite colleges and universities. This figure illustrates the wildly varying levels of success of Chicago Public Schools graduates enrolled at several elite universities in Illinois. Attending a very-selective college did not guarantee students similar college graduation prospects. College A underperformed for both students with modest and good prospects of graduation. Alternatively, College D clearly over-performed for similar students.
Moving Away From Standardized Test Scores
For decades preparing to apply to four-year colleges meant preparing for the SAT or ACT. In recent years more and more four-year colleges have moved away from requiring standardized test scores as part of the application process. In this study we compare whether grades or test scores serve as better predictors of college completion. Each fo the gray lines in the figures below represent the a high school in Chicago. The dark line in each depicts the overall relationship between GPA and college completion, and ACT and college completion. In both cases, as GPA and ACT increase, so does the probability of earning a four-year degree. However, when you look at the curve of each gray line individually, you see the curve for ACT becomes much flatter, meaning that within high schools GPA is a much better predictor of college completion prospects than ACT.
Once the characteristics of the colleges students attend are accounted for, ACT loses much of its predictive validity. The predictive validity of GPA, on the other hand, remains strong, even when accounting for the colleges to which students attend. Our analysis also revealed students’ college completion prospects differed significant by high school of attendance, even after accounting for individual student characteristics, GPA, ACT score and college characteristics.
