Predicting the effectiveness of road safety campaigns through alternative research designs

• Published in: Journal of safety research 2016-12, Vol.59, p.83-95
• Main Authors: Adamos, Giannis, Nathanail, Eftihia
• Format: Article
• Language: English
• Subjects: Automobile Driving; Behavioral modeling; Campaigns; Communication; Communication campaign; Design; Driving Under the Influence - prevention & control; Experimental design; Greece - epidemiology; Information Dissemination; Intention; Mental Fatigue - epidemiology; Mental Fatigue - etiology; Research design; Road safety; Safety - statistics & numerical data; Seat Belts - utilization; Traffic accidents & safety
• ISSN: 0022-4375; 1879-1247
• DOI: 10.1016/j.jsr.2016.10.003

A large number of road safety communication campaigns have been designed and implemented in the recent years; however their explicit impact on driving behavior and road accident rates has been estimated in a rather low proportion. Based on the findings of the evaluation of three road safety communication campaigns addressing the issues of drinking and driving, seat belt usage, and driving fatigue, this paper applies different types of research designs (i.e., experimental, quasi-experimental, and non-experimental designs), when estimating the effectiveness of road safety campaigns, implements a cross-design assessment, and conducts a cross-campaign evaluation. An integrated evaluation plan was developed, taking into account the structure of evaluation questions, the definition of measurable variables, the separation of the target audience into intervention (exposed to the campaign) and control (not exposed to the campaign) groups, the selection of alternative research designs, and the appropriate data collection methods and techniques. Evaluating the implementation of different research designs in estimating the effectiveness of road safety campaigns, results showed that the separate pre-post samples design demonstrated better predictability than other designs, especially in data obtained from the intervention group after the realization of the campaign. The more constructs that were added to the independent variables, the higher the values of the predictability were. The construct that most affects behavior is intention, whereas the rest of the constructs have a lower impact on behavior. This is particularly significant in the Health Belief Model (HBM). On the other hand, behavioral beliefs, normative beliefs, and descriptive norms, are significant parameters for predicting intention according to the Theory of Planned Behavior (TPB). The theoretical and applied implications of alternative research designs and their applicability in the evaluation of road safety campaigns are provided by this study. •Experimental, quasi-experimental and non-experimental designs are tested for predicting intention and behavior.•The above designs are being tested in three campaigns addressing: drink driving, seat belt usage and driver fatigue.•The construct that mostly affects behavior is intention, whereas the rest constructs have a lower impact on behavior.•Behavioral beliefs, normative beliefs, and descriptive norms are significant parameters for the prediction of intentio