Bayesian adjustment for measurement error in an offset variable in a Poisson regression model

Fatal car crashes are the leading cause of death among teenagers in USA. The Graduated Driver Licensing (GDL) programme is one effective policy for reducing number teen fatal crashes. Our study focuses on Michigan during 1990–2004 excluding 1997, when GDL started. We use Poisson regression with spatially dependent random effects to model county level crash counts. develop a measurement error account fact that total teenage population used as proxy driver population. To best our knowledge, there no existing literature considers adjustment an offset variable. Furthermore, limited work has addressed errors context spatial data. In modelling, Berkson applied adjust error-prone variable Bayesian paradigm. Markov chain Monte Carlo (MCMC) sampling implemented rstan. assess consequence adjusting error, we compared two models and without error. found effect time indicator becomes less significant measurement-error adjustment. It leads conclusion reduced drivers can help explain, some extent, effectiveness GDL.