Variation in Correlations between Test Scores and Grades

• In Event: Accountability, Inequality in the Administration of Education Policy

Abstract

Standardized tests and teacher-assigned grades both represent assessments of student performance, but they measure different things. Standardized tests such as the End of Course (EOC) tests in public high schools in North Carolina measure mastery of content knowledge and skills, which might be considered the “hard skills and knowledge” appearing in course curricula. Those factors also play a role in the grades teachers assign to students, but others do as well, including timely completion of assignments, willingness to follow directions, and cooperation with adults and peers, among many others, that might encompass “soft skills” and reflect differences in cultural capital, social capital, emotional intelligence, and opportunities to learn (Westphal, Becker, Vock, Maaz, Neumann, & McElvany, 2016). The link between teacher-assigned grades and classroom socioeconomic composition: The role of classroom behavior, motivation, and teacher characteristics. Contemporary Educational Psychology, 46, 218-227.
Because school contexts vary in median family resources (including socioeconomic status, racial background, neighborhood disadvantages/advantages) and educational opportunities (teacher quality, school climate, peer culture, other resources), it is possible that the correlation between students’ test scores and teacher assigned grades might be different from one school to the next (see Fahle & Reardon, 2018). How much do test scores vary among school districts? New estimates using population data, 2009–2015. Educational Researcher, 47(4), 221-234.. We suspect that in schools serving student populations with the advantages of high socioeconomic status (SES), content knowledge is both common and valued, and students possess many of the soft skills as well—attention in class, completion of homework, enthusiasm for school, and so on. This results in grades that are strongly correlated with test scores. In lower-SES and concentrated minority schools, on the other hand, soft skills may play a stronger role in teacher assigned grades. Teachers may reward academic improvement, class participation, effort, and collaboration with better grades because the educators perceive the student has improved, but that progress may not be captured by the items appearing on a standardized test. This may be particularly true when the teacher and the student share cultural capital, or when a teacher of any background teaches a culturally relevant curriculum employing practices culturally relevant pedagogy. In this environment, students’ standardized test scores may not closely align with grades, resulting in a weaker correlation between test scores and grades.
Using student-level data from all public high schools in North Carolina in the 2011-2012 school year, we calculated the correlation between students’ scores on EOC exams and their grades for the state-required Biology course for each school. Our results indicate that as the percent of students in poverty in a school increases, the correlation between Biology test scores and grades decreases, supporting our hypothesis that test scores and grades are more similar in high-SES schools. Moreover, as poverty increases, schools' test-grade correlations exhibit more variation, further supporting the argument that other factors are influencing the relationship between grades and tests among high-poverty schools.
Further analysis using measures of teacher quality and school context are currently underway using the Biology, English, and Algebra courses required by the state. Preliminary analyses indicate that some of the variation in these correlations can be attributed to teacher and school characteristics.
High school test scores and grades are used for decisions regarding promotion, college admission, scholarships, employment, and more. If the median family background in a school is influencing the factors that determine grades, researchers need to explore that relationship so that the meaning of grades and tests can be better understood and used fairly and appropriately.

Fahle, E. M., & Reardon, S. F. (2018). How much do test scores vary among school districts? New estimates using population data, 2009–2015. Educational Researcher, 47(4), 221-234.

Westphal, A., Becker, M., Vock, M., Maaz, K., Neumann, M., & McElvany, N. (2016). The link between teacher-assigned grades and classroom socioeconomic composition: The role of classroom behavior, motivation, and teacher characteristics. Contemporary Educational Psychology, 46, 218-227.

Author

• Jason Giersch, University of North Carolina, Charlotte