Applied Statistics 2016 / van Duijn

Improving scientific integrity through statistics education

Marijtje van Duijn
University of Groningen, Department of Sociology, Groningen, The Netherlands

Some years ago, the Dutch academic world was confronted with several cases of fraud, revealing that the prevailing research culture allowed questionable research practice. As a consequence, finding means to prevent this kind of misconduct has received a lot of attention. Stricter rules and procedures have been established for researchers to improve transparency and verifiability, with increased interest in replication studies. This seems to have sparked the discussion about the use and usefulness of p-values.

Moreover, awareness about the need to teach scientific integrity at all levels of the academic curriculum has increased. Discussing principles of scientific integrity like reliability, verifiability, and impartiality, will provide fruitful outcomes in terms of awareness of ethical norms, but will not necessarily make students aware of the consequences of misconduct and their implications for advancing scientific insights. They should also learn that questionable research practice will threat the validity of the results.

In a small team of statisticians working in a faculty of social and behavioral sciences as statistics lecturers and statistical consultants/collaborators, we have been teaching an intensive course called Applied Statistics for first-year Research Master students over the past ten years, increasing statistical awareness and applied skills in a first research project.

As a response to the fraud case, we have developed lectures and workshops for teaching scientific integrity through statistics, for Bachelor, (Research) Master students, and PhD students. Well-known methodological and statistical concepts are connected to principles of good scientific research, focusing on statistical conclusion validity and external validity. For PhD students attention is given to study design and reporting of results, and the checking of results (e.g. by peers/reviewers).

An outline of the lectures/workshop and our experiences with the various audiences will be presented.