This report investigates the psychometric properties of the Hazard Perception Test (HPT) and the relationship between novice drivers' performance on the HPT and subsequent involvement in police-reported casualty accidents. The main fmdings of the study are: 1) The test in its current form has a low reliability, which in part is due to the relatively short length of the test, and partly due to some items eliciting responses that are inconsistent with the responses to other items. 2) The HPT measures were evident in identifying fatal and serious injury accident involvement for novice drivers generally and identifying involvement in 'other' injury accidents for 18 year old female drivers specifically. 3) With the HPT instrument in its present state of short test length and containing some potentially problematic items and/or administration methods, the appearance of the HPT measure as a statistically significant contributor to the prediction of fatal, serious and other injury accidents is a positive outcome for the test. If the latent trait that this test is assessing can be measured with more precision, by replacing poor performing items and/or adding more items, the test's predictive powers are likely to increase. (Author/publisher)
Abstract