Racial and ethnic disparities in advanced mathematics and science achievement occur as early as kindergarten, much earlier than previously thought, according to a new study led by a Penn State College of Education researcher. The findings suggest that economic and educational policies designed to increase diversity in science, technology, engineering and math (STEM) fields may need to begin by elementary school.
“Our conjecture is if you’re not displaying a high level of math or science achievement during elementary school, you’re probably less likely to take advanced STEM-related course work in middle or high school and less likely to participate in the STEM workforce,” said Paul Morgan, Harry and Marion Eberly Faculty Fellow, professor of education and demography, and Population Research Institute associate.
Not only do students who display high levels of science and math achievement contribute to society by majoring in STEM fields, Morgan said, there is “longstanding concern about underrepresentation among different demographic groups.”
The researchers followed about 11,000 U.S. students from kindergarten through fifth grade. Trained field staff continuously assessed each student’s science, mathematics and reading achievements. The researchers wanted to see if a small set of student, family and school factors like family socioeconomic status (SES) and student science, mathematics and reading achievement by kindergarten could explain racial and ethnic disparities in STEM achievement.
The new study extends previous work on science achievement disparities by focusing on students who are in the top 10% of the achievement distribution in science and math. This is significant, the researchers noted, because advanced STEM achievement (e.g., performing above the 90th percentile) by elementary school predicts scientific innovation in adulthood as indicated by being listed as an inventor on a technology patent application. Racial and ethnic disparities in later patent rates are also evident by elementary school.
In their recent study, the researchers sought to answer two questions: 1) are students of color less likely than white students to display advanced science or mathematics achievement during elementary school? If so, how large are the observed gaps? 2) Do antecedent, opportunity and propensity factors explain the lower likelihoods that students of color display advanced science or mathematics achievement during elementary school?
They found large racial and ethnic gaps in advanced math and science achievement were already occurring by elementary school. They reported their findings in Journal of Learning Disabilities.
The researchers found that about 13% to 16% of white students versus 3% to 4% of Black or Hispanic students displayed advanced science or math achievement during kindergarten. At the end of fifth grade, about 13% of white students and 22% of Asian students displayed advanced mathematics achievement. The contrasting percentages were 2% and 3% for Black and Hispanic students, respectively.
“I don’t think anyone had known how large these disparities are or how quickly they emerge during schooling,” Morgan said.
For the second question, the researchers examined how antecedent, opportunity and propensity factors explained racial and ethnic gaps in advanced math and science achievement. Antecedent factors include a family’s socioeconomic status (SES), the language spoken in the home and the student’s disability status. Opportunity factors are aspects of learning contexts in homes and schools that facilitate skills acquisition including parenting quality and the school’s economic and racial or ethnic composition. Propensity factors are the student’s own characteristics that facilitate skill acquisition and maximize the ability to benefit from learning opportunities, including prior achievement, behavior and executive functioning.
The researchers observed that children’s kindergarten mathematics achievement was the strongest predictor of whether they displayed advanced mathematics achievement in later grades. Kindergarten reading and science achievement also were positive and significant predictors of advanced mathematics achievement across subsequent elementary grades.
The researchers also found that antecedent factors (e.g., family SES) and other propensity factors, such as whether the children display high levels of math and science achievement, substantially explained the differences between white and Hispanic students. However, those factors accounted for some but not all of the disparities between white and Black children’s scores. The average percentage difference in relative odds of advanced science achievement from first to fifth grade between Black and White students, taking explanatory factors into account, was reduced from 80% to 49%. In contrast, the average percentage difference in relative odds of advanced science achievement between Hispanic and White students was reduced from 71% to 18%.
“We could almost fully explain the gap between Hispanic and white children. However, and though substantially reduced, the gap between Black and white children was still significant,” said Morgan. “The field doesn’t know fully why these gaps are occurring. It requires further study to understand how we can better support talented students who are Black in displaying advanced math and science achievement.”
According to Morgan, the results of the study help identify areas where policies and programs can address STEM underrepresentation. First, he said, greater attention should be given to the early onset of STEM disparities during elementary school.
“It suggests to us that waiting until middle school, high school or college may be too late,” said Morgan.
One reason the disparities may be occurring, Morgan said, is that students from more privileged families are more likely to access supports and services outside of the formal school system such as high-quality childcare, enrichment activities, tutoring and summer programs. In contrast, research has shown that Black and Hispanic families often lack access to private tutoring or other supports and instead are more likely to solely rely on schools to develop their children’s talents. According to Morgan, universal screenings — educational assessments that are given to all students — might be helpful in identifying and supporting talented students of color.
“The study really argues for implementing economic and educational policies to address social inequalities by early childhood,” said Morgan.
Morgan’s co-authors are Eric Hengyu Hu, a doctoral student in the Penn State College of Education; George Farkas, professor of education at the University of California, Irvine; Marianne M. Hillemeier, professor emerita of health policy and administration and demography in the Penn State College of Health and Human Development; Yoonkyung Oh, assistant professor at the University of Texas Health Science Center at Houston; and Cecelia Gloski, a doctoral student in the Penn State College of Education.