The Wald test is a way of testing the significance of particular explanatory variables in a statistical model. (H. KyngaÈs and M. Rissanen).
On Friday 3rd May 2013 a bus from the National Transport Corporation lost its breaks which resulted in 10 deaths and many injured. A similar incident occurred in 2009 and as such a common belief has grown in the media and the population that servicing of the buses was the cause. Others thought it was the portion of the road that had been badly designed.
Initially I thought that by looking at the statistics some sort of conclusions could be drawn about the hypothesis of maintenance of the buses but there are too many aspects that I don’t know to make such a conclusion with certainty.
Anyway (if you are still reading…) this post looks at the road accident statistics for buses and private cars for the past few years and some interesting facts can be noted.
- As of 2011 there were more than 400 000 motor vehicles on the road. Private cars accounted for more than 33% but surprisingly buses accounted for only 1% (around 3000 in total)
- The number of casualties from accidents has risen from 3000 in 2002 to close to 4000 in 2011. That’s a 26% increase over a decade. By comparison the number of motor vehicles on the road has increased by 45% in that same period.
- The government has over the years produced two analysis of the road statistics data that I could find published online. In one done in 2006 by the Ministry of Public Infrastructure it came to the conclusion, based on the number of vehicle type on the road as compared to the number of casualties in which a particular type of motor vehicle was present, that buses were the highest likely to be involved in an accident which causes casualties.
- The number of buses on the road has increased by 18% from 2002 to 2011. However the most telling fact is that the kilometers traveled by a bus per day on average has increased from 157 km in 2003 to 176 km in 2011 (see graph below).
- Where I disagree with these findings is that if a simple calculation is done a bus on average traveled in 2006 (167*365) more than 60 000 km a year. If one assumes that on average a private car travels about 50 km per day, that’s about three times less than a bus. So conclusion such as those formulated in that report cannot be based on the number of vehicles on the road alone. It has to take into account other factors such as distance traveled and the age group driving these vehicles because no nineteen year old drives a bus.
- Data about road accidents are not so easy to collate and thus any analysis has its limitations. This report outlines the limitations to data collection.
1. Not all accidents are reported to the police
2. Classifying the severity of accidents is not always easy as records on injuries vary between hospital and police records.