We all wonder whether high-stakes standardized tests are “fair” to our students. Our students are unique individuals, yet on testing day with pencil in hand they are suddenly all the same. Whether they differ ethnically, racially, socioeconomically, no one seems to care.
As an instructor in an urban school district, I see some of my students struggle greatly with terms and concepts on standardized tests that are not always directly related to the curriculum. Ronnie Reese, a well-known writer and social justice advocate notes in his article Minority Testing Bias Persists (Huff Post Black Voices, 2013) that this concern has been around for a long time.
Anyone remember the 70s show “Good Times?” In one episode, the character Michael Evans, a young genius, fails an eighth grade IQ test on purpose to point out the racial bias in the test. In response to the question “Which word belongs with cup: wall, saucer, table, or window?” one of his classmates incorrectly chose table. However, this choice makes perfect sense to a child who places a cup on a table and has no experience with saucers. Michael further points out references to “residences” and other middle-class terms.
That was then, this is now, but the formats, terminologies, and designs of standardized tests still confer advantage to some and disadvantage to others. Today the biggest gap between the two will be exposed with computer-based tests, which penalize students who have not had sufficient practice navigating the technology. Many middle class test writers have swallowed the middle class notion that our children are more technology savvy than we are. They assume computers are so ubiquitous that every child can use them. This assumption is false. I can tell you from experience that even my best students need to be carefully guided when using a computer for research.
But testing bias doesn’t just affect the socioeconomically disadvantaged. When I was teaching at a private school for gifted students, I remember a teacher coming to me with a story about one of her students. The class was taking a standardized test for second grade and one of the questions asked, “Which man is working?” The answers were in picture form and showed men eating, sleeping, chopping wood, and reading a book. She explained to me that her student chose the man reading a book. When she asked the child about it, the student said his dad was a professor and read books for work. His father also chopped wood, to relax after work. If a professor’s child can succumb to a hidden test bias, then all children are susceptible.
Title IX has not completely erased the gender gap in standardized testing either. In the 1970’s a widely used vocational test was produced in pink and blue. It asked boys if they would like to be president and asked girls if they would like to be married to the president. This disparity continues in the Armed Services Vocational Aptitude Battery (ASVAB) and other vocational aptitude tests that are frequently designed to determine abilities and aptitudes in male-dominated fields.
In addition, young women have routinely had lower scores on college admissions tests, despite having higher overall GPA’s in both high school and college. A 2007 survey by FairTest, The National Center for Fair and Open Testing, showed that the ACT and SAT routinely under predict the abilities of young women to succeed in college. Indeed, in 1976 it was discovered that the young women were out-scoring their male counterparts in the verbal sections. In response, there was a concentrated effort to “balance out” the results by manipulating test items so they would be more familiar to males. Ever since, females have scored lower than males. The survey further notes that the multiple choice format, limited time response, and guessing penalty are at odds with the way that women typically approach problem solving.
We must ask ourselves, as educators, whether standardized tests measure educational quality at all. They certainly don’t measure educational equity, not as long as testing bias persists, and no matter how well crafted a test is it can’t remove all remnants of bias from its design.
By Linda Kennedy, Columbus Education Association