The Effect of Student Race on Teachers’ Instructional Decisions

• In Event: Examining the Role of Racial-Ethnic Biases in Teacher Perceptions and Decision Making

Abstract

A body of research establishes that racial biases affect teachers’ expectations for their students (e.g., Author et al., 2019; Gershenson et al., 2016; Author et al., 2016), but less research investigates how these biases affect classroom practice and instructional decisions. To add to the evidence with this internet-based experiment we used data from 90 general education grade 4 – 8 mathematics teachers in this internet-based experiment study The majority of participants were White teachers teaching in urban public schools (51.1%) in the U.S. A little more than half were female (53.3%), and on average, teachers had taught for ten and a half years, which did not differ by condition. The survey platform randomly assigned participants to review a roster in which 80% students on the roster were identified as either Black or white. Teachers were asked to review the information in the roster, as well as a sample of work from “Student 3,” who was identified as a Black male in the predominantly Black condition or a white male in the predominantly white condition.

Embedded in the example of student work was a mathematical error common to many students who are first learning to add unlike fraction (1/4 + 3/8 = 4/12). Additionally, the student work included an error related to following directions and several errors in language mechanics. Teachers scored Student 3’s work, were asked to identify the cause of the errors present in his work, and then share the written feedback they would give him.

Next teachers were told to carefully review four options of curriculum materials to follow up the fifth-grade district fraction assessment, and then choose the curricular approach that was “most appropriate for this class.” The curriculum materials were all written as teacher guides, featured a fraction addition task, and included directions and some scripted language for teachers. They all concluded with the same three “practice problems.”

Finally, teachers performed the Teachers’ Implicit Association of Academic Achievement Task (TIAAAT) adapted from Author et al.’s (2016). This test is structured in the same way as the more commonly used Black-White Implicit Association Task. However, unlike the IAT, which measures general preferences for Black and White faces, the TIAAAT specifically measures associations between faces of Black and White students and images/words signaling academic achievement and failure.

We find that teachers were more likely to choose low-quality curriculum materials and propose “easy” instruction for Black students compared to White students. Additionally, teachers more often attributed errors made by Black students to students and errors made by White students to instruction. However, this effect was attenuated when teachers had lower implicit racial biases. Additionally, we find that teachers were more likely to endorse student-centered instruction for Black students. We do not find evidence that student race affected the grades or feedback provided to students. Mixed findings suggest that racial bias may affect select areas of instructional decision making rather than instruction writ large.

Author

• Cynthia Pollard, Harvard University