How valid are motorcycle safety data?

The recent increases in motorcycle crashes across many jurisdictions have led to a resurgence in motorcycle safety research.  Much of the research has applied the general data analysis designs and methods of road safety, analysing road authority crash databases with or without comparisons with exposure data to estimate crash or injury rates.  Yet motorcycle crash data and motorcycle exposure data have particular limitations that can affect the validity of the results and the conclusions drawn.  The limitations in motorcycle crash data can affect estimates of the total size of the problem, severity profiles, and the relative importance of different types of crashes and of different age groups and causative factors.  This is most evident in analyses related to unriders (unlicensed riders or unregistered motorcycles).  The usefulness of third party insurance data and hospital data is also examined in this paper.  In calculating crash or injury risk, the denominators commonly used are licences held, registered vehicles or distance travelled.  The paper discusses the advantages and disadvantages of each type of denominator in developing measures of motorcycle safety.