This report reviewed both theoretical and empirical evidence which suggested that as economic activity rises, the level of road loss will also rise. Evidence of this relationship is provided through the most frequently used indicators of economic activity, employment and Gross Product. It is argued that as these measures have provided consistent results, they should be utilised in future models. It is argued that measures of income and price indexes do not reflect the level of activity within the economy, and it is therefore recommended that such measures should not be used. Additionally, this report reviewed statistical recommendations that have been made in regard to modelling techniques used in road safety research. Recommendations for future models include: 1) In general, the use of rate variables should be avoided; 2) The level of data aggregation should be considered and made explicit; 3) Time series analysis is preferable to cross-sectional analysis; 4) careful consideration should be given to the form of modelling to be used; and 5) If regression techniques are used, the assumptions made should be examined for violations. (A)
Abstract