Vol. 3, Issue 1 (2018)
Improving SAT reading scores by mitigating the confirmation bias
Author(s): John Leddo, Bhadra Kadangal, Shyla Bisht, Neha Rajeev
Abstract: Standardized testing is pervasive in our education system and is often associated with high stakes as schools use test results in their admissions and graduation/certification decisions. One of the most common standardized test formats is multiple choice, where test takers are presented with a question and multiple answer choices and then asked to select “the best answer.” This type of format and instructions could lead test takers to succumb to a form of the confirmation bias (Wason, 1960) where test takers look for evidence to support a preferred answer while ignoring evidence that may suggest the answer is incorrect. Because test makers often present multiple plausible answer choices, i.e., are at least part true, test takers can be induced to focusing on evidence that supports incorrect answers, thereby lowering their test scores. The present study investigated whether combatting the confirmation bias by instructing test takers to look for evidence that rejected answer choices were wrong as well as evidence supporting selected answers would lead to improved test scores. The testbed chosen was SAT reading because it is a universally known test used in college admissions. 30 students in Northern Virginia participated in a study where they were given two SAT reading passages containing 26 questions or one-half of the number of questions on a reading section of the SAT. Ten students were assigned to each of three conditions: a control group that received the standard instructions given to SAT students, a group that was given the standard instructions and asked to write down evidence to support their selected answer choice to each question, and a group that was given the standard instructions and asked to write down evidence to support their selected answer choice and evidence for why their rejected answer choices were false. Results showed that students receiving standard test instructions scored the same as those asked to provide reasons for their selected answer while those who were also asked to provided evidence why they rejected non-selected answer choices scored significantly higher than the other two groups. Results suggest that even without prior training or preparation, test takers can overcome the confirmation bias and dramatically increase SAT reading scores. Future research could investigate whether this applies to other types of reading.