Evaluating the Effectiveness of Norway’s “Speak Out!” Road Safety Campaign: The Logic of Causal Inference in Road Safety Evaluation Studies

Abstract

The effects on road safety of the “Speak out!” road safety campaign are evaluated. The campaign, which began in Sogn og Fjordane County in Norway in 1993, is targeted toward teenagers and calls on car passengers to act as back-seat drivers and “Speak out!” to drivers about unsafe driving. The campaign’s effects were evaluated by means of two before-and-after studies and and a multivariate Poisson regression analysis. The results of these evaluation studies were very similar. The number of teenagers 16 to 19 years old who were killed or injured was reduced by about 10 percent; the number of occupants in this age group who were killed or injured was reduced by about 15 percent; and the number of car passengers who were killed or injured was reduced by about 30 percent. The number of killed or injured car drivers 16 to 19 years old did not change. Only the reduction among car passengers was statistically significant at the 10 percent level. It is nevertheless concluded that the “Speak out!” campaign has probably been effective in reducing the number of teenagers killed or injured in Sogn og Fjordane. This conclusion is based on a careful discussion of the logic of causal inference in nonexperimental evaluation research. Seven criteria are proposed for attributing causality to the relationship between a measure and changes in the dependent variable that the measure is intended to influence. The majority of these criteria were met in evaluations of the “Speak out!” campaign.