The Measurement Error Elephant in the Room: Challenges and Solutions to Measurement Error in Epidemiology

Abstract

Measurement error, although ubiquitous, is uncommonly acknowledged and rarely assessed or corrected in epidemiologic studies. This review offers a straightforward guide to common problems caused by measurement error in research studies and a review of several accessible bias-correction methods for epidemiologists and data analysts. Although most correction methods require criterion validation including a gold standard, there are also ways to evaluate the impact of measurement error and potentially correct for it without such data. Technical difficulty ranges from simple algebra to more complex algorithms that require expertise, fine tuning, and computational power. However, at all skill levels, software packages and methods are available and can be used to understand the threat to inferences that arises from imperfect measurements.

Seasonal Impacts of Particulate Matter Levels on Bike Sharing in Seoul, South Korea

“…First, the investigation was restricted to additive errors. Nonetheless, multiplicative errors can be expressed as additive errors through transformation 14 . Secondly, while 𝑅 !…”

Exploring causality from observational data: An example assessing whether religiosity promotes cooperation

“…In this latter scenario, the extent of this residual confounding necessary to alter the study's conclusions can be estimated, using the methods described above. Without a gold standard measure to validate against, methods to correct for measurement error can be difficult to apply (Innes et al, 2021 ); we therefore acknowledge this potential measurement error as a limitation of our study, and return to it when interpreting our results, but predominantly focus on methods for exploring confounding and selection bias. Despite this potential measurement error due to self-reported data collection, prior work in the Avon Longitudinal Study of Parents and Children has shown that self-reported measures of medical history and mental health – which may also be subject to social desirability bias – are comparable with ‘gold standard’ measures, such as medical records and clinical interviews (Golding et al, 2001 ).…”

Shoulder Range of Motion Measurements and Baseball Elbow Injuries: Ambiguity in Scientific Models, Approach, and Execution is Hurting Overhead Athlete Health

“…If prospective observational studies are carefully designed, 47 with adequate (large) sample sizes, causal questions can be answered with robust scientific rigor. 48 These studies must require careful controlling of confounders 30 and other biases (e.g., data missingness, 49 misclassification, 50 immortal time, 51 and measurement error 52 ) to truly investigate the potential for a causal relation. To account for the multitude of confounders and biases, researchers need to develop “scientific models before statistical models.” 53 In other words, one must design a plausible causal model and identify methods to control for appropriate confounders and biases before proceeding with data collection or statistical analyses.…”