2013 ASA President Marie Davidian is William Neal Reynolds Professor of Statistics at North Carolina State University. She is a Fellow of the ASA, Institute of Mathematical Statistics, and American Association for the Advancement of Science and an elected member of the International Statistical Institute. Davidian has served as editor of Biometrics and earned several awards, including the Award for Outstanding Statistical Application.
Montserrat “Montse” Fuentes is head of the department of statistics at North Carolina State University. She earned her bachelor’s degree in mathematics and music (piano) from the University of Valladolid (Spain) and her PhD in statistics from The University of Chicago (1999). An ASA Fellow, Fuentes is the editor of the Journal of Agricultural, Biological, and Environmental Statistics and has authored more than 55 publications.
Pam Arroway earned her PhD in statistics from Iowa State University in 1999. She joined the department of statistics at North Carolina State University shortly after graduation and has been there ever since. She is currently assistant department head and co-director of graduate programs for one of the largest and oldest statistics programs in the country. Her current interests are in pipeline, mentoring, and work force issues in the field of statistics, as well as statistics education.
Recent media reports of alarming accusations of professional misconduct and questionable research practices by members of academia have inspired renewed discussion about professional ethics among both scientists and the public. In this article, we offer our thoughts about this important issue as it pertains to our discipline, with an eye toward students and statisticians just embarking on their careers and on how we—as a profession—train future statisticians.
The statistics discipline is concerned with the collection and interpretation of information (data), a task on which we are uniquely prepared to collaborate with investigators in other disciplines. In general, investigators are focused on pursuit of the truth. Excited investigators have hypotheses to verify, ideas to defend, and the hope that data may be collected that support these ideas. Excited statisticians want to assist their collaborators by providing elegant data analyses that answer meaningful questions. For knowledge in all disciplines to advance in an uncompromised manner, these efforts must be carried out in an honest and principled fashion, and we play a key role in ensuring that this is the case.
A statistician strives to present only what is reasonable to infer, to make limitations transparent, and to not be pressured by “special interests.” Unfortunately, some collaborators might not think highly of statistics or statisticians and view us as “staff” who run software packages and, hence, should have little input into the final interpretation of analyses. Some researchers might view statistical methods as a “bag of tricks,” where one can apply different methods until a favorable conclusion is obtained. It is our role to educate these investigators and to ensure that what is reported is reasonable and defensible, with the limitations clearly stated.
In other situations, collaborators might want to conduct a study to test a hypothesis, but have limited resources. It may be a waste, or downright unethical, to conduct a study that is too small (underpowered) or poorly designed. It is the statistician’s obligation to be steadfast and to teach collaborators what can reasonably be achieved within the resource constraints and to encourage them to consider the scientific relevance of a possible result given the limitations.
Given our unique role, a guiding tenet of our profession is that it be practiced independently and with principle, honesty, integrity, and fairness, and we must train the next generation of statisticians to do the same. Many graduate programs in statistics or biostatistics have ethics training integrated into the curriculum in an informal way. But, we are sometimes disappointed by our students’ behavior and later say, “I didn’t know I had to tell him/her not to do that.”
Ethics training for statisticians must be deliberate. Both the National Institutes of Health and National Science Foundation require students supported by grants to receive some sort of responsible conduct of research (RCR) training. RCR training should include how to work with nonstatistician researchers, the publishing process, and ethics and responsibility in teaching. The ASA has developed ethical guidelines, but how many statistics or biostatistics programs (graduate or undergraduate) are educating students deliberately about these guidelines? Training to work with nonstatistician researchers should be integrated into every course involving data. The integration must be purposeful, not just assumed and left to when it comes up in an example. Programs should be able to identify which courses will contain ethical training for students. The ASA’s ethical guidelines are a great starting point for discussion about where students will get this training in the curriculum.
In academia, the pressure to publish is considerable in any discipline. There is an eternal struggle of quantity versus quality, which can naturally lead to a temptation to send out work that is incomplete or to promote methods that are not sufficiently well studied. The objective of research and its publication is to advance knowledge. Thus, prior to submitting our papers, we should ask ourselves if our work represents a publication-worthy advance; if we have portrayed the advantages and disadvantages fairly and accurately; and if we have done our best to communicate our work clearly and effectively, rather than expecting editors/referees to do it for us. When acting as referees, we should resist the temptation to be competitive. The goal is to evaluate honestly and fairly whether the contribution is substantial enough to be communicated, regardless of its impact on our own work.
In teaching statistics or communicating statistical ideas, we carry a responsibility for completeness, honesty about what can and cannot be done, explaining the assumptions used, and staying up to date.
A frequent challenge for statisticians is authorship. For collaborative work, if one has made a contribution to a paper, he/she should be listed as an author. In some disciplines, there is a protocol for authorship. In others, there is no tradition of granting authorship to statisticians, and we may have to fight for recognition. Programs training statistical researchers should teach students about the written or unwritten rules of our profession for determining authorship and order of authors.
In teaching statistics or communicating statistical ideas, we carry a responsibility for completeness, honesty about what can and cannot be done, explaining the assumptions used, and staying up to date. Developing and maintaining a good course is hard work and time consuming, but promoting statistics to nonmajors (and majors) through clear and thoughtful teaching benefits the entire discipline. An easy, entertaining course and light workload may please students in the short term, but fails them in the long term. Our goal is to communicate statistical thinking and principles in the best way we know how.
Students, it is important to get into this habit of working with honesty and integrity early in your career. Stephen Vardeman and Max Morris wrote an excellent article on this topic for The American Statistician, titled “Statistics and Ethics: Some Advice for Young Statisticians.”
These authors not only discuss key principles of ethical conduct, such as working independently on assignments when asked to do so and treating all students fairly when acting as a teaching assistant, but they also provide guidance on personal responsibility to one’s own professional development. In particular, coursework should be treated as an opportunity to learn and hone technical and applied skills and to cultivate an ability to work independently as preparation for your later work habits, not just a prerequisite for an “A.”
Think about taking demanding (not “easy A”) courses that will expand your knowledge and challenge your abilities, or courses beyond the minimum requirement for your degree. Your dissertation is an exercise in learning to work and think independently. Do not expect your advisor to “assign” the next task. Don’t just get results, interpret them. Do the results make sense? Identify the next step yourself, investigate new approaches on your own, try new simulation scenarios, etc. By doing so, you will be developing skills that are critical to being a good statistician and building the confidence to defend ethical statistical practice when called upon.
The ASA’s curriculum guidelines for undergraduate statistics curricula do not mention ethics training. We do not expect any of our colleagues would dispute that ethics training for students is important. At NCSU, we require all PhD students to take a course about statistical research that has some specific ethics content, although we certainly do not address all items in the ASA Ethics Guidelines. At the moment, we do not include any explicit ethics training in our undergraduate program. Integrating ethics training in our curricula as a discipline will enhance integrity of our profession and better prepare our students for the challenges they will face. And, most importantly, we must all act as role models.
This is a really informative article. I love the in-depth description of why ethics is so important to people learning about statistics.