There is long a recorded gender divide among American high school students in terms of SAT test scores. According to a study by ETS, girls have consistently under-performed boys in both SAT math and reading during 1967-2004. The gender gap was 30 score points for both subjects before the mid 90s. An encouraging trend came at that time when changes started to take place in American K-12 education. As a result, the average SAT scores increased across the subjects and genders by 20 points except for boys’ reading that only changed minimally. As a result, the gender gap in reading shrank but that in math persisted.
The cross-board increases came from recentering in 1995 SAT test scores, a measure that arbitrarily moved the average back to around 500. In the 30 years prior to 1995, these averages had degraded to 425 in verbal and 475 in math. As an effect of the change, a 500 in verbal prior to 1995 corresponds to 580 in 1995, and 500 in math prior to 1995 means 520 in 1995.
A handful researches have studied the issue but reached no conclusion. AEI, a leading think tank in Washington, DC, attributes the gap to the student quality in general. “One possible explanation…would be that high school boys are better students on average than high school girls and are better prepared in mathematics than their female classmates.” Another popular view takes the issue to stereotype bias, a psychological phenomenon whose scientific cause is yet unknown.
Both explanations failed to account the trends in American high school classrooms. The afore-mentioned ETS research found that girls had spent increasingly more hours and studied in more courses in math subjects in school since 1990s. On the contrary, boys’ effort in math has been stagnant if not declining. As a result, arguably, girls outperform boys in terms of math grades. In spite of girls’ growing outperformance in class, the divide has persisted over the period.
What, then, is the real cause? The answer, as I recently found, lies in a gender bias in the SAT test design. In my summer SAT prep camp in Naperville, I came into this fantastic idea to motivate students to practice CRM model. Standard SAT requires students to choose one and eliminate three out of four choices for each question. Instead, I allowed the students to choose two and eliminate two, and record their results as R(2). Then, they are allowed to make their final choice and the result is recorded as R(1). I collected both results and graded them separately in that a result scores 1 if it contains the final correct answer. Typically, to distinct between the last two choices is the harder than among all four.
While the adaptation helped many students in learning the CRM model, I noticed a surprising and consistent pattern among the students. Girls score consistently higher than the boys in R(2), but fall behind in R(1). I concluded that even though girls are better in choosing two, their ability to eliminate the last one is a fraction of that of the boys. The girls revealed to me that they tend to stuck in the similarity of the last two. The boys, on the other hand, do not have much difficulty in telling difference of the two in R(1). In fact, they see differences of all four and eliminate two in R(2).
The cause of it, as I further analyzed, is the gender difference in terms of critical thinking, including both comparing and contrasting among things. Girls are better comparing things, while boys are better contrasting. In R(2) stage, students started with all four choices. Girls compared them with the reading materials find two most obviously similar ones as their choices. Boys, on the other hand, saw one or two choices are obviously irrelevant to the reading topic as they understood. As girls generally read more careful then boys, they excelled in R(2). In the next R(1) stage, girls seemingly proficient with similarity could not tell as much the difference between the last two as did the boys who are obsessed with contrasting.
All SAT tests are designed in the same way that you have to pick one exact rightness out of three distinct wrongsome. When many of the wrongs are made very confusing, the ability to contrast helps much more than that to compare does. Girls who are made by our nature mother to see the commonalities more easily than the difference are thus discriminated by such a test design.
Mind2learn has since turned out its training mechanisms to cope with the situation, but I still hope that SAT may eventually find their way to correct such a deficiency.