Related Experiment Video
Updated: May 18, 2026

Inverse Probability of Treatment Weighting (Propensity Score) using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
Research on injury compensation and health outcomes: ignoring the problem of reverse causality led to a biased
Natalie M Spearing1, Luke B Connelly, Hong S Nghiem
1Australian Centre for Economic Research on Health, The University of Queensland, Herston, Queensland 4029, Australia. nataliespearing@hotmail.com
Objective:
This study highlights the serious consequences of ignoring reverse causality bias in studies on compensation-related factors and health outcomes and demonstrates a technique for resolving this problem of observational data.
Study Design And Setting:
Data from an English longitudinal study on factors, including claims for compensation, associated with recovery from neck pain (whiplash) after rear-end collisions are used to demonstrate the potential for reverse causality bias. Although it is commonly believed that claiming compensation leads to worse recovery, it is also possible that poor recovery may lead to compensation claims--a point that is seldom considered and never addressed empirically. This pedagogical study compares the association between compensation claiming and recovery when reverse causality bias is ignored and when it is addressed, controlling for the same observable factors.
Results:
When reverse causality is ignored, claimants appear to have a worse recovery than nonclaimants; however, when reverse causality bias is addressed, claiming compensation appears to have a beneficial effect on recovery, ceteris paribus.
Conclusion:
To avert biased policy and judicial decisions that might inadvertently disadvantage people with compensable injuries, there is an urgent need for researchers to address reverse causality bias in studies on compensation-related factors and health.
Related Concept Videos
Causality in Epidemiology
Bias
In statistics, a sampling bias is created when a sample is collected from a population, and some members of the population are not as likely to be chosen as others (remember, each member...
Bias in Epidemiological Studies
Strategies for Assessing and Addressing Confounding
Confounding can be addressed at both the design phase of a study and through analytical methods after data...
Cause and Effect
Confounding in Epidemiological Studies
