It’s that dreaded time of year again for teachers: standardized testing time. The Illinois Standards Achievement Test (ISAT) begins February 28th, and I get the joy of having five third-graders who will be taking this exam. For those of you who are not familiar with education, these standardized tests are required by No Child Left Behind in order to make schools accountable for students’ learning. These exams are otherwise known as “high-stakes tests” because a whole lot depends on their outcome. The results of these tests can determine funding, and low scores, or “failing,” can cause disciplinary action against the school or district.
You have probably heard debate about the benefits and disadvantages of such accountability and of NCLB in general. There are a lot of problems with the act, however, in my view, the biggest problem with using standardized, high-stakes tests is the biases these tests carry.
I teach English Language Learners (ELLs) with cognitive impairments (IQs less than 70). My five third-graders will have to take an exam that is biased against them because of language, culture, and ability. Additionally, the girls will be subjected to a gender bias.
The results of these tests are made public: schools either “pass” or “fail,” and broad generalizations are then drawn. While I believe in freedom of information, the media needs to be more aware of how they deliver the results and what they actually mean because the general public is not aware of the intricacies of testing bias and validity. A test is not a valid measure unless it is actually testing what it says it is testing (more on that later).
While there are ways for educators to help combat these biases and to prepare students for a biased test, it is very important for the general public to be aware that they exist. Too often I hear people make uneducated remarks regarding students who do not fit the dominant culture and their abilities. Therefore, I have written a brief overview of some of the problems that students will face when they take these tests.
One of the major problems with English high-stakes tests is validity for ELLs. For example, a test may be designed to assess students’ reading comprehension; however, if a student struggles with English, said test is really assessing their English language skills (there is a separate assessment for that). This is also true for content (science, social studies, and math) tests. Math tests usually include less language bias because of the lack of reading. Obviously, ELLs score the lowest on word problems where it is necessary to be able to read English in order to complete the math.
A more valid test might be one in the students’ native language. This is particularly true for students who are enrolled in bilingual programs where they are being taught these skills in their native language while learning English. ELLs can receive accommodations in order to help them, but these are by no means sufficient for achieving true test validity.
Language bias is also seen with students who speak a different dialect of English (such as African American English); they receive no accommodations.
Cultural bias affects anyone who is not from the dominant culture. Most obviously affected by this are immigrant students. However, this bias is also prevalent against low-income students, minority students, inner-city students, and rural students. I once saw an essay question asking children to write a persuasive essay about why fishing is fun. That seems at first glance to be a pretty innocuous question, as I was raised with a father who loves to fish. However, how many children from the inner city go fishing? How would they know if it is fun or not? A student may be able to produce a well-thought-out essay in general, but if they are not familiar with the prompt, they will not succeed according to the exam.
This bias is usually seen in predictive standardized testing taken during the high school and college years. Patricia Rosser did a study in the early ’90s regarding the SATs and women. The SATs are supposed to predict students’ success in college; however, Rosser found that women in general score lower on the SAT than men, yet receive higher grades in college (Rosser, 1992). How is the SAT then a valid measure of students’ collegiate success?
This bias is pretty self-explanatory. My students have IQs in the 60s and are learning from a functional academic curriculum, yet they are being compared to their regularly developing peers. As with ELLs, students with disabilities are allowed to have certain accommodations on the test, however, these do very little in the way of leveling the playing field.
In conclusion, all standardized tests are normed, and all the population taking the test should be included in the norm. However, this is not always the case. Students are given tests that are not valid measures of their academic achievement because they were not included in said norming group. These test results are then misused to punish schools who do not perform well and to propagate stereotypes about the intelligence of students who are minority, low income, inner-city, etc.
By no means am I advocating that certain students be exempt from standardized testing; assessment has its place in education. However, in order for these tests to reliably and validly measure and compare students’ academic successes, they need to be un-biased.
Butler, F. A., & Stevens, R. (2001). Standardized assessment of the content knowledge of English language learners K-12: Current trends and old dilemmas. Language Testing, 18(4), 409-427.
Froese-Germaine, B. (2001). Standardized testing + high-stakes decisions = educational inequality. Interchange, 32(2), 111-130.
Rosser, P. (1992). Sex bias in college admission tests: Why women lose out (4th ed.). Cambridge, MA: National Center for Fair and Open Testing (FairTest).
“Exams are looming” by James Stanier on Flickr