Data Science Is No Panacea for High-School Math Education
Science, technology, engineering, and mathematics are the fastest-growing fields in terms of both student interest and job opportunities. For example, in California, the number of bachelor’s STEM degrees increased at a rate more than triple that of other degrees between 2010–11 and 2016–17. This is for good reason: studies show that STEM majors enjoy higher salaries and lower unemployment. The growth of STEM fields makes K–12 mathematics education more relevant than ever. Students without strong mathematical foundations will be shut out of these higher-paying and faster-growing fields. Hence, improving K–12 education, in particular for lower-income students and students of color, is of the utmost importance.
Given that context from the job market, the billion-dollar question is: why does the United States rank 36th out of the 79 countries included in the Programme for International Student Assessment math rankings? Those results followed two massive education-reform initiatives, No Child Left Behind and the Common Core state standards. Neither one lifted the United States into the top tier of performers globally.
There is no simple explanation for U.S. performance in these rankings, but to improve that performance, it is crucial to understand a key fact of U.S. math-education-reform initiatives: there is a hyper-focus on math curriculum and not enough attention paid to teacher recruitment, training, and retention. We know that a student’s success in math rests heavily on having a highly qualified teacher. A robust math curriculum is useless if teachers are not equipped with the material and training to deliver it well. Top-performing countries on the PISA exams, such as Japan, South Korea, Estonia, the Netherlands, and Poland, have varying curricula (with Estonia’s and Poland’s still influenced by the Soviet system), demonstrating that success in math education is less about changing curricula and more about who is teaching it and the training and support they get.
Increasing the number of highly trained math teachers addresses another education crisis that the math curriculum cannot address alone: capacity and access. According to the U.S. Department of Education’s Office for Civil Rights, advanced mathematics is offered at only 65 percent of high schools, and calculus is offered at only 50 percent of high schools. Moreover, the 5,000 high schools with more than 75 percent Black and Latino student enrollment offer advanced math and calculus at a significantly lower rate than that of high schools overall.
The increased importance of STEM fields for future career options, economic growth, and national security places particular emphasis on topics such as algebra and calculus. In particular, calculus is part of the curriculum in all STEM majors; students who complete a calculus course in high school have a significant advantage for pursuing STEM coursework and job opportunities during college. Calculus and advanced algebra are also at the heart of the “machine learning revolution” that led to recent breakthroughs in artificial intelligence, and an understanding of these topics is a key skill for work in data science. Far from being relics from the “Sputnik era,” calculus and algebra are more important than ever in K–12 education.
Unfortunately, recent efforts at “education reform,” including the (in progress) proposals for the California Mathematics Framework, devalue such fundamental mathematical courses. In particular, some have advocated replacing them with “data science,” asserting that this subject is more relevant than the “antiquated curricula” of algebra and calculus courses in our modern world. These advocates also claim that data science is somehow “a more equitable alternative to calculus” and can be a tool for addressing educational gaps. Both claims are false.
Claims about the relevance of data science confuse the importance of the field itself with what can be taught in a K–12 course. Much as a high-school first-aid course does not prepare one for a career in medicine, a high-school data-science course can only give students a superficial taste of the area. Indeed, such a course is more properly called a “data-literacy” course than data science—it can be very beneficial to students but should not be considered an alternative to basic mathematical courses. The field of data science builds on mathematics, statistics, and computer science, and a thorough data-science education requires foundations in all three fields. For this reason, taking advanced math courses (algebra II, precalculus, and calculus) is a much better preparation than high-school data science, even for students who are interested in data-science careers. Nearly 1,700 STEM researchers, educators, and practitioners signed an open letter decrying the proposals to devalue foundational mathematics. Signatories include winners of their field’s highest honors (including the Nobel Prize, the Fields Medal, and the Turing Award), as well as leaders in the field of data science itself.
These experts know that the mathematical maturity gained from working through problems is crucial for STEM preparation. It is true that, these days, we all have a powerful calculator in our pocket. But this does not mean that one can be a data scientist without knowing how to multiply. Mathematics is different from literature, in that different topics rely upon each other. While it is possible to read Angelou without first reading Shelley, one cannot understand least-squares regression without first understanding the Pythagorean theorem. As an associate provost and the dean of engineering at the University of California, Berkeley, recently wrote jointly, “the pervasiveness of computers means that we should focus more on mathematical reasoning, not less.”
Some advocates claim that data science is more equitable than other fields of math. To put it mildly, this claim is not justified by research. Remember, closing education gaps requires improved teacher recruitment, training, and retention. While material can always be improved, education gaps were not created by the curriculum and cannot be addressed via curricular changes. Moreover, creating “data-science pathways” as alternatives to the standard pathway can and will have a particularly harmful impact on disadvantaged students. Such pathways emphasize proficiency with computational tools such as spreadsheets over the mathematical concepts (functions, equations, symbolic manipulation, and logical reasoning) that are crucial prerequisites for more advanced math and that also build the type of thinking needed for coding. Hence, in practice, data-science pathways will become lower tracks by another name. Such “implicit tracking” can be more pernicious than explicit tracking: less-resourced students or students of color might end up choosing this track under the false impression that it leads to career opportunities, while students with more means and access to college counseling will realize that the traditional pathways keep more options open. Indeed, this seems to already have been the case, with wealthier districts in California such as Beverly Hills and Cupertino signaling their rejection of the California Mathematics Framework revisions.
Too often with math-education initiatives, education reformers do not think about the unintentional consequences for creating a de facto lower track in mathematics. For example, low-income students of color in this track will be shut out of programs such as Questbridge and Thrive Scholars. Both nonprofit organizations provide low-income students with financial support and other resources that ensure they graduate from the best colleges in the country. Such programs, as well as STEM-specific programs including Berkeley’s SEED, are interested in accepting students who take the advanced mathematics courses that lead to calculus because they know the best colleges in the country look for calculus on students’ transcripts, and that such courses prepare students for STEM success. These courses also help students prepare for the SAT and ACT. While one can argue that programs and colleges should not use calculus or standardized exams for admissions, it is important for K–12 education to prepare students, especially low-income students and students of color, to be successful in the world as it exists today, rather than in an ideal world that may or may not exist in the future. Not all students are interested in STEM, and not all students need to learn calculus in high school, but all students deserve honesty about the consequences of different educational pathways. Students and parents are best equipped to make this tradeoff, but they should get accurate information.
The United States has had more than its share of curricular experiments, often done on low-income students or students of color, with mixed results at best. Promoting data science at the expense of algebra and calculus is yet another experiment backed by dubious evidence. The vast majority of subject-matter experts reject it, since it won’t provide students with the foundations for STEM success. While well-resourced students will find ways to bypass it, such a “reform” will mostly harm the students it purports to help. Some advocates claim that K–12 data-science courses are easier than algebra and calculus and provide better preparation for the data-intensive high-paying jobs of the 21st century. However, one maxim remains as true in this century as it was in the past: “If something sounds too good to be true, it usually is.”
The post Data Science Is No Panacea for High-School Math Education appeared first on Education Next.