The next part of the course is intended to make us think about the amount of effort we should put into preventing SMIDSY events (a driver pulling out of a side road in front of an oncoming vehicle).

In a few words/bullet points give your throughts on which

• road users
• road user behaviours
• event types

we should deal with first.

For the purposes of this exercise, you only have to choose between SMIDSY (car pulling out in front of bike) and SVLOC (motorcycle losing control, usually on bend).

Explain what information you need in order to help you decide. Please feel free to say exactly how you think this decision should be made (use some stats, use our experience of which is most common, use our knowledge of which problem we can cure etc. etc. etc.)

Next, please indicate which event type (SMIDSY or SVLOC) you think is responsible for the most fatalities in Devon, giving reasons for your answer.

If you answered SMIDSY you would be incorrect. 

Surprisingly enough, it isn't (as far as we can tell) the most common cause of motorcycle fatalities in Devon. Quite a few of the top five event types appear to be entirely biker related (taking a corner too wide and hitting a car travelling in the opposite direction, misjudging overtaking).

However, if you ask any two wheeled road user, they will almost certainly tell you that SMIDSY is the biggest problem on the roads. You don't need to experience many near-SMIDSY events to make you very sensitive about the matter. As professionals, we need to balance this perception against the evidence that however unpleasant SMIDSYs are (and however many non-injury and minor-injury crashes there are), they don't kill as many people as SVLOC.

If you answered SVLOC you would be correct.

As far as we can tell this is the single most common cause of death among motorcyclists in Devon.

However, if you ask any two wheeled road user, they will almost certainly tell you that SMIDSY is the biggest problem on the roads. You don't need to experience many near-SMIDSY events to make you very sensitive about the matter. As professionals, we need to balance this perception against the evidence that however unpleasant SMIDSYs are (and however many non-injury and minor-injury crashes there are), they don't kill as many people as SVLOC.

An attempt to provide the inconclusive with a conclusion:

(a) As far as we can tell, we know something about the event types that cause the most fatalities in Devon amongst motorcycles.

(b) And as we understand it, motorcyclists account for almost 1 death in 4 on Devon's roads, yet only 2% of the traffic.
 

This information may tell us what our problems are; and is essential to have when deciding on interventions. 

However we also need to balance the need for action against the likelihood of success.  If SMIDSY is a major problem, and we think we have an effective intervention, then maybe that is our priority area. 

If SVLOC is a bigger problem, but we don't think we have an effective intervention, then we shouldn't spend money just because "something has to be done", as that money will not then be there for other problems that we can fix.

Need for "hard facts": 

This should help persuade you that in making these decisions we need some hard facts (statistics) about where our efforts should go. The statistical information helps to tell us how big a problem we have and evidence based practice tells us which problems we can fix. Experience helps us balance the competing demands on our resources.

Uncertainty:

Finally, being a Statistician means never having to say you're certain.  So, apologies if this is throwing mud in an already murky pool, BUT..

You could also take the view that fatalities are relatively rare events, and you'd rather not base all your effort on fatality data. Or you could decide that you are worried about all injuries not just fatalities. If you look at these data, you might get a different impression of your priorities. If you are responsible for an urban area you might have a whole different set of problems.

Either way, the aim of this course is to try to persuade you that data should be helping in the decision making process. We also have to be honest and state that no data (bar graph or table) can make the decision for you. But whilst you make the decision, it should use the best available hard evidence.