A recent study from DTU Compute has highlighted a safety risk associated with election posters in Denmark. Researchers have found that the presence of these posters correlates with an increased number of traffic accidents, suggesting that they may distract drivers.
“Be careful when there’s an election, and you are driving a car and see an election poster. People may get distracted by their favorite candidates (or perhaps the opposite), enough to significantly show up in accident statistics during parliamentary elections. It is common knowledge with mobile phones, but there are also other things that can distract driving to a level where it causes accidents. Drive when you drive," says Associate Professor Anders Stockmarr.
Model accounted for many variations
In the study, Anders Stockmarr, industrial PhD student at DTU Compute and COWI A/S Albert Skovgaard Bisgaard, and Magnus Hamann Poulsen, who has a master’s degree from DTU Compute, used a so-called seasonal autoregressive-moving-average model with exogenous variables (seasonal ARMAX) to analyze Danish traffic accident data (registers from Statistics Denmark) from January 1, 1997, to December 15, 2009. This type of model is a statistical tool for understanding and predicting time series data, which in this case is the daily number of traffic accidents.
The model accounted for variations in weather conditions (data from Danish Meteorological Institute), and general traffic volume, indicated through public holidays and so-called autoregression. The analysis excluded accidents where other known risk factors for accidents were present, such as drunk driving, narcotics, unfit drivers, and low visibility conditions (fog).
The findings revealed that during general election periods, the presence of election posters was associated with a statistically significant increase in traffic accidents (p = 0.003). The p-value is a measure of whether the association between (here) traffic accidents and election posters is real, or due to chance. A p-value below 0.05 is generally considered statistically significant. On average, the study estimated an additional 1.18 accidents in Danish urban areas per day during these periods. In total, that amounts to 127 extra accidents for the investigated period of time.
”Our analysis has considered all known risk factors, and indeed unknown dittos through autoregression, making us quite confident in that there is a real connection between election posters and an increased risk of traffic accidents,” says Anders Stockmarr.
An opportunity to look more broadly at traffic disrupting elements
The study is the first one to document a statistically significant association between traffic accidents and election posters. These results suggest a need for a discussion of the appropriateness of the current traffic legislation, and potential further regulation of election posters to enhance road safety, says Anders Stockmarr.
For future research, a natural extension will be to explore other similar types of traffic disturbances and compare our findings across different countries, as election posters themselves are actually not something that appear in general in democratic elections throughout the world.
“We know that e.g., commercials cause inattention in traffic, but whether they also cause additional traffic accidents is largely not documented today. While commercials or visual campaigns thus might/might not, similarly impact traffic safety, their non-repetitive nature poses a scientific challenge for consistent analysis, that would be interesting to incorporate into the framework that we have developed,” Anders Stockmarr says.
The study is published in the journal Traffic Safety Research.
