What is Measurement Error in the Social Sciences: Forms, Impacts and Adjustments?


Bio: I am an Associate Professor in Quantitative Criminology with a background in Social Statistics. Most of my research has focused on the analysis of unwarranted disparities in criminal justice decisions, for what I have collaborated with the Crown Prosecution Service, the Sentencing Council for England and Wales, and the Parole Board. More recently I have been working on the problem of measurement error in police statistics, where I have been exploring its prevalence, impact, and strategies to adjust for it.

Measurement error is a pervasive - yet often unacknowledged - problem in the Social Sciences. It is present in many of the survey and administrative datasets commonly used across different fields. Here we will focus on a series of examples (exam results; perceptions of institutional legitimacy; police statistics; and self-reported labour status) to illustrate the different forms that measurement error can take; and how they can impact our findings. We will also review a series of adjustment methods commonly employed in the literature (latent variable estimation and regression calibration); which require multiple measures of the same concept. We will end by introducing some less well-known adjustment methods (simulation-extrapolation and multiple overimputation); which could be used as sensitivity tools. That is; to assess the potential impact of measurement error when all we have is an educated guess about the form and prevalence of the measurement error mechanisms affecting our data.