Motorcycles and Fatalities
Task: Please consider the following questions and write down whether you think the answer to each is TRUE or FALSE.
1. There is a strong negative correlation between the number of vehicles on the road, as the number of vehicles goes up the number of fatalities goes down. This shows that more vehicles make the roads safer. TRUE or FALSE?
The correct answer is false.It's not correct to say that these two variables are correlated - in statistical terms we are looking for a plot of one variable against the other. Even if there were correlation, it is not safe to suggest there is a causal link. Vehicle safety standards, people's use of the roads, the use of safety equipment, congestion have all changed over time and could explain any of the trends we see.
2. If I toggle the variables so that I am looking at motorcycles only, the trends follow each other closely since the 1940s. As the number of bikes goes up, the number of fatalities goes up, as the number of bikes goes down the number of fatalities goes down. TRUE or FALSE?
The correct answer is false. I think we can see what the data are telling us about trends, but it's difficult to know what to conclude. If we conclude that motorcycle deaths are in proportion to the number of vehicles does that tell us not to worry about things (deaths are a feature of the number of bikes, not the lack of effort in injury prevention). If this is the case, why have the number of all vehicle deaths declined as the number of vehicles has increased? If we risk a tea-leaf analysis, a big fall was seen in the 80s when CBT was introduced which might tempt us to conclude that legislating bikers off the road is the safest thing to do. The point is that there are limits to the conclusions we can draw from a data set. But it still important that we start our thinking from the best available understanding of the data.