On Bias

Please note: this page is a work in progress

Central to our project - and to the SSH field, we argue - is the need to scrutinise the term ‘bias’: what do we mean with the concept ‘bias’?

Instead of a straight-forward definition with a straight-forward solution to the ‘problem’ of bias, we found that bias is often used as a heuristic, a short-hand to refer to a collection of different yet interrelated concepts, including harmful language, source selection, representation among others. In order to get to an answer to our question that is actionable and reusable, we are currently developing a ‘bias taxonomy’. This taxonomy aims to provide an overview of these concepts brought together under the term ‘bias’, visualise their connections and their place within the lifecycle of a historical dataset.

The taxonomy is currently in development. To get an idea of what types of concepts are affiliated with ‘bias’, take a look at our Resources List, in the column ‘concepts used’. We invite anyone to contribute to both the list in terms of resources and/or concepts!