|Source: Channel NewsAsia|
Given that road accidents increased over 20% from about 7,100 in 2001 to 8,600 in 2010, is this incident symptomatic of greater carelessness on the part of motorists?
While the number of accidents may be affected by how careful motorists are, other factors play a part as well. In this case, there were more accidents simply because there were more vehicles on the road in 2010 as compared to 2001. In fact, the number of accidents per vehicle, as shown in the graph below, appears to be on the decline (dotted line). This means that each motorist gets into fewer accidents on average, and motorists might actually be more careful than before.
Nevertheless, the rates for each year fluctuate quite abit. These are shown by the green dots. In some years they are much higher than previous years', and others much lower, and the downward trend is not clear. In fact, a separate statistical test I conducted found that the downward trend (the dotted green line) arises only because of chance. So there is no evidence that the accident rate is on a steady decline.
Then again, not all accidents are the same. Some, like the tragic one above, lead to fatalities. Others result in injuries - not a good thing of course, but nonetheless an improvement over the former. Lastly, other accidents don't harm anyone, and only see a dented bumper or damaged street light. If motorists are becoming more careful, we should see the number of fatalities and injuries per accident fall as the accidents shift from being fatal to injurious to non-harmful.
To see if this was the case, I first looked at the number of fatalities, or deaths, per accident. As shown in the chart below, this figure hit a high of 0.033 in 2003 before falling 32%, quite an amount, to 0.022 in 2010. Despite what happened with the Ferrari, it does appear that motorists are becoming more careful. Perhaps motor insurance premiums should fall in response.
A different picture arises when we look at the injury rate. The average number of people injured in each accident fell between 2001 and 2005, but rose steadily after that. As a result, we obtain a U-shaped trend in the chart below (dotted red line). So much for cheaper insurance.
Putting these together, we find some evidence that motorists have become more careful from 2001-2005 as both fatality and injury rates declined. The good news did not last as injury rates rose steadily in recent years. What can we make of this? Perhaps motorists are indeed becoming more careful (fewer fatalities), but not quite careful enough (more injuries).
An aside: notice that I've drawn "lines of best fit" for each chart to spot trends. To do this, we need data for every year, and a few clicks on Excel (more thorough statistical methods are also available).
Nevertheless, trends are sometimes "spotted" just by comparing rates in two years. This is not the correct way to go about it. For example, you can "prove" that injury rates have increased by comparing rates in 2005 and 2010 (left chart below). However, you can "prove" that injury rates have declined by just by changing the year of comparison to 2006 (right chart below).
At the end of the day, as the saying goes, "there are three kinds of lies: lies, damned lies, and statistics". Don't believe an argument just because there are numbers, keep questioning, and get more experience and practice by reading this blog more often :)