Why We Need Data on Our Profession
Keith Crank, ASA Research and Graduate Education Manager
Due to the economic downturn, many (most?) colleges and universities are looking to cut costs. At the University of Central Florida, this means eliminating the Department of Statistics (among others). At other universities, it could mean merging a statistics department with another discipline. Even worse, it could mean breaking up a statistics department and sending individual faculty to different departments.
I’m sure most of you agree that for our profession to remain strong and grow, we need statistics and biostatistics departments at our major colleges and universities, and these departments cannot be subsumed under another discipline. Unfortunately, administrators are not always aware of the importance of the statistics discipline to the welfare of the university. So, it is important for us to educate them and our students.
How do we convince university administrators of the importance of statistics? Obviously, we provide them with data. (And we interpret it for them.) There are three types of important data: student, faculty, and community. Let me explain.
By student data, I mean data on the number of courses and students taught each year. This has probably increased over the past 5–10 years, and that should be pointed out. Student data should also include information about how many other disciplines require at least one statistics course for their majors. (When a statistics course is one of multiple options, interpretation of the data would include showing that the number of students who choose a statistics course is increasing and explaining why the statistics course is more relevant to the chosen major than other options.)
According to the Conference Board of the Mathematical Sciences (CBMS), undergraduate course enrollment in statistics courses taught by statistics department faculty increased almost 80% between 1990 and 2005. More than two-thirds of that increase was in elementary-level courses. Most of these courses would still need to be taught, even if the statistics department was eliminated. So, any financial benefit obtained by eliminating a statistics faculty position would be offset by the need to hire other faculty.
When I talk about faculty data, I mean the non-teaching contribution of statistics faculty to the rest of the university. This would primarily be consulting-type activities through a department consulting lab, service on student dissertation committees, or less formal consulting with students or faculty from other disciplines. Data of this type from a consulting lab should be collected routinely. Some effort may be required to compile data on faculty service on dissertation committees, but this information is presumably provided at least yearly for faculty performance reviews. The less-formal consulting may be difficult or impossible to track. For faculty data, interpretation is important. If there are no statistics faculty members, the quality of the theses students write and the quality of the research being produced by other faculty members will decline. This, in turn, will affect the reputation of the university.
Community data refers to the impact of the department, its faculty, and its students beyond the university. Do faculty members consult with local business and industry? Is the consulting lab open to (and used by) the larger community? Have your graduate students organized something like Purdue’s STATCOM to provide statistical consulting for local government and nonprofit organizations? Do your graduates go to work for local companies?
When a department is threatened with elimination, local data is important. But, we also need data at a national level. For departments that want to expand (or avoid getting downsized), national data can provide the justification. Is the number of faculty sufficient for the size of the university? Do you offer enough course options for your students? National data is also important for statistics groups that need to justify becoming a department and for departments that currently exist to get resources from their deans.
There are multiple places that collect portions of the data mentioned above. The ASA collects salary data and, for the last two years, has collected other department information (such as number of faculty, students, and degrees awarded) from PhD-granting statistics and biostatistics departments. The American Mathematical Society (AMS) has been surveying departments for many years, and for at least the last 10 years, it has been separating information from statistics and biostatistics departments. And the Conference Board of the Mathematical Sciences conducts a survey on undergraduate programs every five years (with the next one scheduled for 2010).
The response rates to these surveys from statistics and biostatistics departments needs improvement. Although responding to the many requests for data is time consuming (and, admittedly, some of the requests are for the same information), the importance of the data in maintaining and expanding our disciplinary departments is immense. If you receive a request for data from the AMS, ASA, or CBMS, remember how important it is and submit your completed questionnaire in a timely manner.
To contact me, send an email to firstname.lastname@example.org. Questions or comments about this article, as well as suggestions for future articles, are always welcome.