Supporting theoretically-grounded model building in the social sciences through interactive visualisation

Abstract

This is the accepted version of the paper.This version of the publication may differ from the final published version. Permanent repository link AbstractThe primary purpose for which statistical models are employed in the social sciences is to understand and explain phenomena occurring in the world around us. In order to be scientifically valid and actionable, the construction of such models need to be strongly informed by theory. To accomplish this, there is a need for methodologies that can enable scientists to utilise their domain knowledge effectively even in the absence of strong a priori hypotheses or whilst dealing with complex datasets containing hundreds of variables and leading to large numbers of potential models. In this paper, we describe enhanced model building processes in which we use interactive visualisations as the underlying mechanism to facilitate the construction and documentation of theory-driven models. We report our observations from a collaborative project involving social and computer scientists, and identify key roles for visualisation to support model building within the context of social science. We describe a suite of techniques to facilitate the exploration of statistical summaries of input variables, to compare the quality of alternative models, and to keep track of the model-building process. We demonstrate how these techniques operate in coordination to allow social scientists to efficiently generate models that are tightly underpinned by domain specific theory.

The Novel Coronavirus and Enforcement of the New Separate System in Prisons

“…The grounded construction process operates with three goals (Turkay et al, 2017). First, we seek to construct an outcome variable that provides an accurate measure of prison response to COVID-19 by reducing the impact of missing data.…”

“…The exclusion of questions with a "Yes" response rate is also designed to hedge against missing data or potential measurement error. We further decrease the potential impact of measurement error by excluding questions with a "NA" response greater than 30% (Turkay et al, 2017).…”

Jails in the Time of Coronavirus

“…The research team utilizes a grounded process for the construction of the outcome variable in this analysis. The grounded construction process operates with three goals (Turkay et al, 2017). First, we aim to develop an outcome variable that provides an accurate measure of COVID-19 response by reducing the impact of missing data.…”

Hunting High and Low: Visualising Shifting Correlations in Financial Markets

“…This idiom makes MDS suitable for visualising changes in correlations [GF00] and other network topologies [LS08, LSSdN08] over time. In a static setting, MDS has been used as a layout to communicate similarities between data dimensions as represented through their correlations [YPH*, TSL*17]. However, due to the stochastic nature of the computation, and its susceptibility to errors in data and over‐sensitivity to inherent structures [BGM12], MDS needs to be incorporated with care.…”