Randall Pruim is chair of the department of mathematics and statistics at Calvin College in Grand Rapids, Michigan. He is the author of the book Foundations and Applications of Statistics: An Introduction Using R (2011) and a member of the American Statistical Association and Mathematical Association of America’s Joint Committee on Undergraduate Statistics.

When I was asked to reflect on how the way we teach statistics has changed over the past 10 years, my first thought was to get some data to inform me. But for an association of statisticians, we have surprisingly little data available to answer questions about how we (collectively) teach and how well it is working. For the most part, I’m left to fall back on a small, biased convenience sample of what I know best, combined with some personal reflection and speculation.

So, how have things changed? There have been several trends that have affected statistics teaching over the past decade and seem likely to continue to have an effect for the next 10 years.

Student Preparation

More students are learning more statistics before they go to college. That much is clear. The number of students taking the AP Statistics exam has nearly tripled over the last decade—66,000 in 2004 to 184,000 in 2014. In addition, the Common Core State Standards Initiative has increased the pressure to include more statistics throughout the U.S. school curriculum. It is reasonable to expect this trend to continue for the foreseeable future.

It’s less clear what impact this is having. The most obvious changes I have seen in my students are an increased familiarity with a wider range of graphical displays of data and an increasing number of students who do not take any statistics in college because they have “gotten out of statistics” by taking AP Statistics in high school.

Some institutions are reporting an increasing number of students beginning their college statistics with a course that has intro stats as a prerequisite.

But I don’t know of any reliable data from which to learn how AP Statistics is shaping student impressions of the statistics field as a career option, how many AP students take courses at the university level, how many AP students repeat intro stats (by which I mean a course with no statistics or calculus pre-requisites) rather than begin with a more advanced course, how many university statistics courses AP students take, or whether AP students do better or worse than their non-AP counterparts.

In addition to changing the population of students who enter college statistics courses, the increased emphasis on statistics in elementary and secondary school increases the importance of training in statistics and the teaching of statistics for prospective teachers. The recently published “Statistical Education of Teachers” (SET) was commissioned by the ASA to clarify the recommendations for the statistical preparation of teachers proposed in the Conference Board of the Mathematical Sciences’ “Mathematical Education of Teachers II” report.

SET outlines the opportunities and challenges to preparing the next generation of teachers, describes best practices, and suggests model curricula and assessment items.

Advances in Technology

Perhaps the most obvious change over the past 10 years has been the advancement of technology to support the teaching of statistics. With the exception of the ubiquitous Texas Instruments calculators, all forms of technology that might be used in support of teaching statistics have improved markedly. Installation of statistical software has become much simpler, interfaces have improved, and online alternatives such as RStudio Server, SAS OnDemand, and StatCrunch make it possible to teach entire courses in which all statistical analyses are performed “in the cloud.”

The number and quality of other technological resources have also increased, and there are now freely available online tools that make it easier than ever to perform analyses, prepare reproducible reports, visualize statistical methods, access interesting data sets, and conduct student-designed experiments. The challenge is for instructors to select from and take advantage of what has become available.

Curricular Focus

Looking back, we will see this past 10 years as the decade when simulation-based inference became possible for the masses. Calls for a shift to this approach by George Cobb and others have been answered by improved technology and the availability of textbooks such as Statistics: Unlocking the Power of Data and Introduction to Statistical Investigations, Preliminary Edition to support it. The next decade will tell us whether this approach, with its promises of improved conceptual understanding, will become the dominant approach to introducing statistics to students, or whether it will be an approach zealously espoused by its advocates but ignored by the majority of instructors.

Meanwhile, rumbling in the distance are calls for other sorts of change. In her 2013 James R. Leitzel Lecture, Ann Watkins listed three approaches she sees as competing with the currently dominant curriculum for the future of intro stats: the simulation-based approach already mentioned, a modeling approach, and a Bayesian approach. Technology will likely play a key role in determining whether a Bayesian approach is viable, but a modeling approach—one that emphasizes multivariable and complex data early—depends more on a mind set shift among statistics educators.

Collectively, we invest an enormous amount of time and energy in intro stats. But there are signs we are beginning to invest more in other areas of the curriculum, as well. Enrollments in statistics majors and minors are up at many institutions. Many smaller institutions that a decade ago did not offer these have introduced them. Other institutions have revised their major and minor programs or introduced additional flavors—generally moving in the direction of programs with a more applied focus and often making the development of computational skills more central to the program.

Indeed, for the next decade, the most important curricular question for undergraduate statistics programs may well be how to provide students with the computational skills and conceptual frameworks necessary to learn the tools that will allow them to “think with data.” For a more thorough discussion of the opportunities and challenges for the undergraduate statistics curriculum, see the guidelines.

The Importance of Making Connections

Statisticians have a long history of collaborating with people from a variety of other disciplines in academia, industry, and government. Our track record in making these sorts of connections in statistics education is not as strong. It is interesting to note the similarity in the themes of the 2005 and 2015 U.S. Conference on Teaching Statistics. In 2005, the theme was “Building Connections for Undergraduate Statistics Teaching,” and in 2015, simply “Making Connections.” A review of the presentations at each of these conferences suggests progress here has been measurable, but slow.

Statistics educators face an interesting and important challenge when trying to make these connections. There is a general sense in many disciplines that statistical training is important for students. But curricular pressures; a level of dissatisfaction with the current state of statistics courses; a large number of students and faculty in other disciplines who do not feel comfortable working with data, mathematics, or computers; and the siloing of education within academic departments all contribute to a state that is less than ideal.

Interestingly, one commonly proposed solution is to have students take fewer statistics courses and move statistical topics into courses within other disciplines. See Nicholas Schlotter’s “A Statistics Curriculum for the Undergraduate Chemistry Major” in the Journal of Chemical Education for an explicit call for this in the context of chemistry. To counter this trend, it is crucial that statistics educators collaborate with educators in client disciplines to ensure the courses we offer serve students well in the context of their major programs and career goals and the principles and methods being taught in statistics courses are being appropriately used and reinforced in other courses.

The Arrival of Data Science

We may not know exactly what it is or how long it will go by that name, but data science has arrived. No sooner had Hal Varian proclaimed statistics “the sexy job in the next 10 years” than the wrangling began over what data science is and whose it is. In truth, Varian’s comments already hinted at this.

Now we really do have essentially free and ubiquitous data. So the scarce complementary factor is the ability to understand that data and extract value from it. I think statisticians are part of it, but it’s just a part. You also want to be able to visualize the data, communicate the data, and use the data effectively.

The impact of Big Data; the emphasis on communication and visualization; the necessity of new computational tools; and the connection to real, important, large-scale problems in business, science, and society are only beginning to be felt in statistics education. But if we respond now to the voices calling us to move in these directions, this has the potential to be the source of the biggest changes to statistics teaching in the next decade.

Looking Forward

The next decade promises to be an important one for statistics education. We appear to be at a golden moment of opportunity and the stakes are high. This is because movement in (statistics) education is slow, while movement in the world of data and technology is much faster. We can’t slow down the latter, so we will need to pick up the pace of the former. Data are plentiful and statistical skills are in demand, but if we don’t deliver the desired skills to the people who need them, someone else will.

FURTHER READING

American Statistical Association Undergraduate Guidelines Workgroup. 2014. Technology Innovations in Statistics Education 1(1).

Conference Board of the Mathematical Sciences. 2012. The mathematical education of teachers II. Providence RI and Washington DC: American Mathematical Society and Mathematical Association of America.

Franklin, C., A. Bargagliotti, C. Case, G. Kader, R. Schaeffer, and D. Spangler. 2015. The statistical education of teachers. Alexandria, VA: American Statistical Association.

Lock, Robin H., Patti Frazer Lock, Eric F. Lock, Dennis F. Lock, and Kari Lock Morgan. 2013. Statistics: Unlocking the power of data. Hoboken, NJ: Wiley.

National Governors Association Center for Best Practices & Council of Chief State School Officers. 2010. Common core state standards for mathematics. Washington, DC.

Schlotter, Nicholas E. 2013. A statistics curriculum for the undergraduate chemistry major. Journal of Chemical Education 90 (1):51–55.

Tintle, Nathan, Beth L. Chance, George W. Cobb, Allan J. Rossman, Soma Roy, Todd Swanson, Jill VanderStoep. 2014. Introduction to statistical investigations, preliminary edition. Hoboken, NJ: Wiley.

Varian, Hal. 2009. Hal Varian on how the Web challenges managers. The McKinsey Report, January 2009.

Watkins, Ann. 2013. Statistics isn’t mathematics: So how’s that working out? James R. Leitzel Lecture at the Mathematical Association of America MathFest 2013.