Home
About the Department
Faculty & Staff
Course Offerings
The Major
The Minors
Student Opportunities
Activities
Graduate & Other Programs
Prizes & Awards
Resources
Math Calendar
Departmental News
Center for Women in Mathematics
Mathematics & Statistics

Strengthening bridges between statistics and the natural sciences

Johanna Hardin, Nicholas Horton (Smith College) and Danny Kaplan (Macalester College) organized a Mellon Foundation funded workshop entitled Strengthening bridges between statistics and the Natural Sciences.

Liberal arts colleges have been the source and test bed for reform of introductory statistics for more than a decade. These changes have been internal within the statistics community, motivated by concerns about the effectiveness of pedagogy. The results have been much greater emphasis on statistical literacy, use of real and more complex data, less emphasis on abstract probability, and additional use of computer technology. Modern textbooks, which are now used nationwide, reflect the changes.

Despite the successes in improving statistics education to reflect the internal values of the field, the changes do not always address directly the evolving needs of other disciplines. There has been an increasing sophistication of the use of statistics in fields as diverse as political science, economics, neuroscience, psychology, sociology, biology, and environmental science. In molecular biology, for example, technology has made possible the collection of huge data sets which have brought new statistical analysis challenges. Researchers are now tackling far more complex problems and using sophisticated methodologies that have only recently started to be incorporated into the undergraduate statistics curricula. The methodological changes, absent support from the statistics curriculum, have constrained the involvement of undergraduate students in real-world research projects.

In addition to the increasing complexity of methods, the last two decades have seen a dramatic increase in recognition of the importance of quantitative arguments. A number of national reports have called for change in the quantitative education of science students (e.g., the Bio2010 and MAA CRAFTY reports). Last year's recommendations from HHMI-AAMC on the scientific preparation of pre-medical students "release[s] the student from specific course requirements, but ... require[s] each institution to identify the instructional means by which the necessary competencies can be gained." In regard to quantitative matters, the necessary competencies are very different from those gained from the traditional year of calculus. They include the ability to:

It's clear that statistics is central to the quantitative education of scientists. In order to be able to fulfill their proper role, statistics curricula need to engage the specific requirements of today's science disciplines. It's doubtful that the conventional curriculum, however sophisticated the pedagogy, addresses the above list in any but the most abstract way. The problem is not that statisticians are unaware of the needs of the sciences --- indeed many statisticians work in intensive collaborations across diverse areas of science. Rather statistics, like many academic disciplines, has evolved according to internal standards and, at most liberal arts colleges, statistics is more closely aligned with mathematics than science.

In January, 2011, we brought together close to twenty faculty, including half dozen scientists from allied fields (biology, geology, psychology, and neuroscience). The primary focus was on the applied statistics curricula, with the main audience statisticians from the Mellon 23 institutions, augmented by a number of scientist colleagues from their institutions. Hosting the workshop at the Claremont Colleges facilitated participation from all of the local colleges, in particular, by inviting natural scientists from Pomona, Harvey Mudd and Scripps to participate in break-out sessions designed specifically to address needs of a specific discipline.

The workshop sought to reinforce the bridges between statistics and the natural sciences, enhancing the scientifically relevant practice of applied statistics into the curriculum. These efforts will help realign the introductory and intermediate statistics curricula with science curricula to provide a solid foundation for quantitative education and to ensure that the next generation of researchers have appropriate quantitative capacities.

The specific goals of the workshop included review of the introductory and intermediate statistics curricula to foster greater connection with efforts underway in the natural sciences. Participants brainstormed and refined other curricular (partial credit, J-term or summer programs) and co-curricular (journal clubs, statistical consulting structures, mentored research programs) programs that are appropriate within a liberal arts setting. In addition, we proposed ways for statisticians and scientists to create improved structures to facilitate better communication.

We began with a group discussion on how statistics is used in other fields and how it could be better incorporated into the classroom and research projects of the allied fields. Each of the allied scientists spoke briefly about their own work and how statistics was involved (or not involved). Deb Nolan (UC Berkeley), gave a keynote address on “The Future of Statistics.” In her talk she described some of hands-on data analysis projects that allow undergraduates from all fields to learn quickly and creatively in order to solve important problems.

The following day, the statisticians presented some of the novel ways they were incorporating real scientific problems into their teaching or using their statistical background to collaborate with colleagues outside of statistics. Rob Gould (UCLA), and Roxy Peck (Cal Poly San Luis Obispo), then gave an invited address on the connections between statistical education and the outside community. Rob focused on using technology and teaching with more focus on problem solving. Roxy talked about using data approaches to think about probability.

Among the many ideas generated at the workshop was a plan to share course materials for courses that might be unique or focused on a particular aspect of collaborating with our allied science colleagues. As an example, Danny Kaplan, one of the organizers, created a website to share the materials he has used in his epidemiology course at Macalaster (free account needed to access the site). Each of the workshop participants was given an account to access the site. http://www.causeweb.org/wiki/mosaic/index.php/Epidemiology_at_Macalester

Other resources shared at the workshop included flowingdata.com (a website for data visualization, infographics, and statistics) as well as kaggle.com (a website to use in courses for data analysis competitions).

Additional ideas that came out of the sessions included building consulting relationships between statisticians and allied science colleagues. And most importantly, participants were universally positive that the workshop helped build connections between the scientists and the scientists, the statisticians and the statisticians, and both groups. We each left with new ideas and a group to work through some of them. We believe that the workshop successfully generated ideas and created contacts that will naturally form additional research, teaching, and consulting collaborations.

Resources