A challenge for adequacy-for-purpose views of data modeling

Abstract

Alisa Bokulich and Wendy Parker (2021) provide an account of data modeling they call the pragmatic-representational view (PR view). According to this view, data models are akin to theoretical models in that they should be evaluated based on their adequacy for particular purposes. In this paper, I present a challenge for the PR view. I argue that a separation between data generators and users can prevent adequacy-for-purpose from being a good evaluative tool. I analyze an example from microbiome bioinformatics to illustrate my critique of Bokulich and Parker’s view and then propose a tripartite disambiguation of the term ‘data model’.