Logo Utrecht University

Circular Data Modelling

Introduction

Circular data are measurements with a periodical nature, that is, measured on a circle where the distance between the scores 0º and 350º is equal to the distance between 0º and 10º. The main instruments providing circular data are the clock (0-24 hours; e.g. crime occurrence times) and the compass (0-360º; e.g. the sway of body parts in a moving room experiment). Furthermore, in the behavioural sciences also circumplex scales, such as Russel’s mood scale or Leary’s interpersonal circumplex, provide scores expressed as angular positions on the circle.

kompas

To analyse circular data special statistical techniques are required. The toolbox for circular statistics is, however, limited and in fact several research questions of empirical scientists can currently not be analysed. Therefore, researchers with circular data often turn to one of two unsatisfactory solutions:

  • The periodical nature of the data is ignored without knowing the impact this may have on the validity of conclusions,
  • The research questions are simplified to be able to evaluate them with the limited set of tools that are known and available to the researcher.

With this project, new tools for circular data will be developed and presented to applied researchers such that they will be able to properly analyze the research questions that they have.