Sam Behseta, Professor of Statistics, California State University, Fullerton; Cherie Ichinose, Associate Professor of Mathematics Education, California State University, Fullerton; David Drew, Professor of Education, Claremont Graduate University

By most metrics, statistics is still a young discipline, especially in contrast to mathematics and most scientific fields of research and discovery. More importantly, though, statistics is a dynamic field, continuously responding and adapting to the new realms and variations of data.

In the epilogue to their triumphant book Computer Age Statistical Inference, Brad Efron and Trevor Hastie showcase a triangular schematic in which the evolution of the field is contextualized, first through applications, motivated by the need for devising effective processes for data collection and data summarization. Subsequently, and inspired by probability theory, there is an overwhelming effort devoted to building theory and appropriate mathematical language, syntax, and framework for statistical inference, as the field was also gradually taking a turn toward computation and responding to the ever-increasing demands for modeling and data analysis. Inexorably, with the advances in modern computing, especially in the first half of the 1950s, there was a natural gravitation toward the center of the triangle, where each of the three components of applications, theory, and computing would end up playing a critical role in what is now perceived as the discipline of statistics.

We sometimes use this graph in our classes when the discussion, almost organically, pivots toward the role of statistics in data science. When viewed through the lens of the triangular presentation, one is hard-pressed to identify the relationship between the two. As such, we are left with the following assertion: Statistics is at the core of data science.

When we add to this story the overwhelming interest in big data, analytics, and predictive modeling, we cannot escape the thought that, quite possibly, while the knowledge and skills in mathematics and computing remain essential, it is statistical thinking that will play the pivotal role in unraveling the information embedded in complex data structures.

This whole idea—coupled with the fact that statistical concepts are bound to have a more prevalent presence in the K–12 curricula across the nation—led us to surmise that the discourse about teaching and learning statistics will have to be prioritized by, at the very least, those involved in the overlapping worlds of statistics and education.

The digital revolution and big data have transformed our society and economy. Statisticians, even those fresh from their undergraduate studies, are in demand and command high salaries. There is a growing need for professionals who can teach those undergraduates, but there is an even greater need for K–12 teachers who can introduce elementary and secondary students to statistics. Recently, we have called for a proliferation of graduate programs to train those statistics teachers and conduct research about teaching statistics.

In the mid-20th century, industry powered the United States, and the high-school curriculum was attuned to corporate needs. Students in the sciences studied the mathematics sequence through precalculus in high school. Their homework calculations were performed on a slide rule. (For those too young to remember a slide rule, it was a computing device that looked like a ruler and was based on logarithms.)

But there still is a major gap in the high-school and elementary-school curricula. Big data and statistics increasingly drive executive decisions in business and nonprofits, yet most students do not encounter statistics until college. Most of the instructional content in elementary and secondary schools is mired in the curriculum of the 1950s. Even those who recognize this curriculum gap are at a loss to find educators who can teach statistics. In fact, some college students major in education in part to avoid math courses.

American school mathematics prior to 1800 consisted of the knowledge one needed for everyday life—essential arithmetic. In 1957, a major shift in mathematics education occurred with the Soviet Union launching Sputnik. The United States was forced to examine how and what mathematics were being taught in the public schools. The 1960s launched a reform for a more rigorous school mathematics curriculum. Mathematics, ‘new math,’ was more about the study of mathematical structures through abstract concepts.

The Commission on Mathematics of the College Entrance Examination Board made the case for additional curriculum reform by including probability and statistics. Unfortunately, the results of a 1965 CEEB survey showed the absence of probability and statistics in the curriculum of most schools. Further, rigor of the new curriculum was not completely understood by the teachers or parents. As a result, the reform efforts of the 1960s were unsuccessful and dependency on basic drill and memorization practices, once again, became the norm in school mathematics.

Roughly 30 years later, school mathematics came under fire yet again. American students, regardless of grade level, generally underperformed in mathematics compared to their counterparts from other developed countries. A commission was formed by the National Council of Teachers of Mathematics and an agenda for action was created to standardize school mathematics curriculum to focus on problemsolving and teaching and learning for understanding. Further, it recognized with the advancement of technology, computational tasks should now be done by machines rather than by hand. NTCM emphasized the importance of familiarity with basic statistics for every citizen. This movement resulted in significant reforms across the country.

For example, in California, with current Common Core State Standards, appropriate components of probability and statistics are being introduced in earlier stages. Elementary students are required to graph and make inferences from data. Secondary students are required to perform more advanced analysis and testing. Nonetheless, there is a lack of implicit evidence showing the implementation of such standards in school classrooms. This may be in part due to the lack of teacher preparation in statistics because most of the top institutions in the country lack programs related to statistics education. In California, which supports the largest system of public universities, not one university has a program dedicated to statistics education with any focus on statistics literacy.

We call on the higher education community to respond to this need and create programs for training PhD educators who have an interest in developing ideas about statistics pedagogy and its manifestations in the K–12 curricula. The ASA’s Guidelines for Assessment and Instruction in Statistics Education is a great starting point, as it recommends teaching statistical thinking, focusing on conceptual understanding, integrating real data with a context and purpose, and using technology to explore concepts and analyze data.

Personally, we have worked with single-subject teachers in middle and high school. But there also is a need for elementary teachers to have instruction about statistics and how to teach it. We believe communicating the concepts of variation, estimation, statistical modeling, and statistical prediction in the K–12 curriculum is pivotal to achieving quantitative literacy in our nation’s school systems. Such a reformed curriculum would hopefully help with closing the gap in our students’ appreciation of statistical thinking and decision-making under uncertainty. This can only be achieved if we train the next generation of academics with expertise in statistics education.

Editor’s Note: A version of this article also appeared in the Journal of Humanistic Mathematics.

Share Your Views!
Do you have an opinion you would like to share with colleagues? Send us a letter. Letters should be 600 or fewer words and must include your name. Send your letter to the editor or any other news you would like to share to Megan Murphy.