Bias in assessment can affect both students and instructors. As instructors, it's important to reduce unintentional bias in our assessments and to understand how a perceived threat of bias can affect student performance.
Confirmation Bias
One of the most common types of bias that may affect our assessments is confirmation bias. Confirmation bias occurs because people tend to focus on evidence that “confirms” their existing beliefs or theories and dismiss evidence that does not support these beliefs or theories. The problem with this type of bias is that it often occurs outside of our conscious grading process. For example, you may have two students: student A and student B. Student A works very hard, participates in class, and turns in all work on time. Student B, on the other hand, is frequently on her phone during class and submits work late. If these two students submitted the same work, confirmation bias might mean you grade student A's work higher than student B's because you unconsciously look for evidence to support your belief that student A is a harder worker (and deserves a better grade) than student B. While this is the most common sort of confirmation bias we try to guard against, there is a more sinister type that can creep in if an instructor believes that certain groups of students are smarter, harder working, etc. than other students based on group membership alone.
Strategies to Minimize Confirmation Bias
Because we're not fully aware of all of our potential biases, it's important to understand the mechanisms behind these judgment processes, which can have a considerable influence on students’ grades and class experiences.
- One of the best ways to guard against confirmation bias is to grade “blind,” or to block the names of the students you are grading until after you’ve assessed their work. Knowing how to identify biases helps ensures that, consciously or unconsciously, confirmation bias does not taint faculty’s evaluations of students’ work.
- Creating clear and well-defined rubrics is another way to ensure that your grading is more 'objective' and less likely to be affected by confirmation bias.
- While not typically feasible, asking multiple people to assess students' work can also help to reduce the effects of confirmation bias in grading (Steinke & Fitch, 2017).
By knowing how we internalize and process information, we can take the needed steps to be proactive in preventing unintentional bias.
Teaching Confirmation Bias: Examining Urban Legends
A fun and interesting example that can be used as a class activity is to have students quickly research and discuss an urban legend or celebrity rumor of the day that is accepted as factual. Afterward, you can describe the story of the Beatles and the rumored car crash death of Paul McCartney. What started as prank article in a college newspaper about Paul dying and the Beatles replacing him with a double to hide his death from the public became a phenomenon. The rumor convinced people to infer the distorted and barely audible voice at the very end of “Strawberry Fields Forever” as “I buried Paul.”
They believed the cover of the Abbey Road album, the iconic image of the four band members crossing the road, supported the rumor. In the front is John, whose white suit represented God. Ringo follows, wearing a preacher’s old-fashioned frock coat. After him comes Paul or the look-alike, wearing the conservative suit, which many viewed as corpse dressed for burial. The fact that he was out of step with the others, shoeless and holding a cigarette often called a “coffin nail” was viewed as supporting evidence. George is last, appropriately, because his work shirt and jeans indicate he’s the gravedigger. Nearby, a Volkswagen has a license plate that reads LMW 28IF, which means “Linda McCartney Weeps” (Linda was Paul’s wife) and that Paul would be 28 if he were still alive."
For more information on how this story can be a useful teaching tool, view, "Teaching Confirmation Bias Using the Beatles" by John A. Minahan, PhD.
Stereotype Threat
Stereotype threat is a psychological threat in which members of a particular group are (or
believe they are) at risk of conforming to or validating stereotypes about their group. Stereotype threat can create create doubt and anxiety for students if they fear negative stereotypes about one or more of their group memberships will be reinforced by their performance (APA, 2006). Reminders of negative stereotypes can negatively affect a student’s performance on assessment measures. For example, several studies have confirmed that merely mentioning that a stereotype exists lowers students’ grades compared to a control group who did not have the stereotype brought up prior to a test (eg., Steele & Aronson, 1995; Spencer, Steele & Quinn, 1999).
To address the effects of the stereotype, it's essential to have inclusive classrooms. By creating an "identity-safe" classroom faculty can intentionally acknowledge and value students' identities (Steele & Cohn-Vargas, 2014). For example, international students may not be aware of or comfortable with American idioms. However, when faculty acknowledge the sensitivity of the topic, students at risk of stereotype threats can dissociate themselves from the negative stereotypes and better realize their academic potential (McGlone, 2007). Two example strategies for mitigating stereotype threat include using positive imagery (asking students to think of a time that they've done well in the topic being tested, for example) and addressing stereotype threats directly (warning students the stereotypes exist, dispelling myths, and discussing these myths directly with students).
For more information on how stereotype threat impacts academic performance, view the American Psychological Association article, "Stereotype Threat Widens Achievement Gap."
Further Reading
Aronson, J., & Williams, J. (2004). Stereotype threat: Forewarned is forearmed. Unpublished manuscript, New York University, New York.
Bowen, Natasha K. Wegmann, Kate M. Webber, Kristina C. (2013).
Enhancing a brief writing intervention to combat stereotype threat among middle-school students. Journal of Educational Psychology, 105(2), 427-435.
Nickerson, R.S. (1998). Confirmation Bias: A Ubiquitous Phenomenon in Many Guises.
Review of General Psychology 2(2), 175-220.
References
American Psychological Association (2006). Stereotype threat widens achievement gap. https://www.apa.org/research/action/stereotype. Retrieved January 2019.
McGlone, Matthew. (2007). Communicative Strategies for Mitigating Stereotype Threat Among Female Students in Mathematics Testing.
Conference Papers -- International Communication Association. Annual Meeting, 1-27.
Spencer, S.J., Steele, C.M. & Quinn, D. M. (1999). Stereotype threat and women’s math performance. Journal of Experimental Psychology, 35(1), 4-28.
Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69(5), 797-811.
Steele, D., and B. Cohn-Vargas.
"Creating an Identity-Safe Classroom." Edutopia. 21 October 2015
Steinke, P. & Fitch, P.
Minimizing Bias When Assessing Student Work. (2017). Research and Practice in Assessment 12, 87-95.