Academic Profile & Strategy
In addition to being a mathematics discipline, statistics is a widely used tool applied at many different levels for clarification and decision-making of a noticeably varying quality. The popularity and considerable legitimacy of statistical methods lies in the fact that they can be used to prove and identify regularities (for instance, the connection between observable variables) or to reveal a lack of such regularities, alone on the basis of observations that are encumbered by “noise” or random variation.
Unfortunately, however, one can also use – often successfully – terminology from statistics to hide connections or to give pseudo reasons for connections that are more or less speculation or vaguely stated. This type of abuse can be difficult to uncover, because rejecting a statistical argument normally requires complete knowledge of the model or method being refered to. The problem of "quackery", which has no doubt existed as long as it has been possible to describe statistical methods disconnected from their mathematical foundations, has not diminished with the dissemination of effective statistical programme packages – on the contrary.
It is perhaps most important to emphasise one critical aspect of the Center’s profile, both when it pertains to research and teaching: Statistical methods must be used critically and objectively. A negative attitude towards other people’s use of statistical methods can, in many cases, be just as constructive as a positive attitude towards, for example, a new class of models. Choosing a model or method with an eye to achieve certain results is just plain bad statistics. With professional use of statistical methods, choosing a model, in most cases, is dictated by the problem at hand, apart from unessential details. One important requirement for a statistical model is that it must not be more complicated than the problem and the data structure call for.
In regard to teaching, it is our humble ambition to do it as well as possible within the limitations that exist regarding course time constraints and student mathematical qualifications. In our opinion, even though teaching in most courses (Business Administration and Management Science courses aside) must necessarily be presented either partially or completely without the related mathematical system of concepts, it is still important that it is done – or are at least managed – by instructors who are up-to-date with the basis of and most recent developments within the discipline. This is especially necessary due to the rapid development within the discipline regarding automatisation and computing (neural networks, data mining etc.) and their interplay with the possibilities mentioned above for abusing the methods.
Last updated by Birgit Rasmussen 30/03/2009
