An analysis of motorcyclist injury severity under various traffic control measures at three-legged junctions in the UK

Chih Wei Pai, Wafaa Saleh

Research output: Contribution to journalArticle

53 Citations (Scopus)


The effects various junction control measures have on accident frequencies among vehicles have been investigated by previous studies through the use of statistical modelling techniques but such effects on motorcyclist injury severity, given that a junction-type accident has occurred, have received little attention in literature. This paper attempts to estimate econometric models of motorcyclist injury severity under different control measures at three-legged junctions in the UK, as a function of demographic, vehicle and environmental factors. Separate ordered probit models were estimated for unsignalised and signalised junctions, using the data extracted from the STATS19 accident injury database (1999-2004). Also examined in the models are collision partners' aggressiveness toward motorcyclists and the impacts various crash configurations have on injury severity. The modelling results uncover several combined factors that were deadly to motorcyclists: for example, injuries tended to be much more severe while motorcyclists involving in approach-turn collisions at signalised junctions than at unsignalised junctions. This study ultimately offers insights into potential countermeasures that could be undertaken to help lessen motorcyclist injury severity at three-legged junctions in the UK.

Original languageEnglish
Pages (from-to)832-847
Number of pages16
JournalSafety Science
Issue number8
Publication statusPublished - Oct 2007
Externally publishedYes



  • Motorcyclist injury severity at three-legged junction
  • The ordered probit model
  • Traffic control measure

ASJC Scopus subject areas

  • Chemical Health and Safety
  • Safety, Risk, Reliability and Quality
  • Public Health, Environmental and Occupational Health
  • Health Professions(all)
  • Human Factors and Ergonomics
  • Safety Research

Cite this