## I'm not a statistician, so......

fat and old
Posts: 3760
Joined: Sat Aug 30, 2014 12:06 pm
Location: Mill Park

### I'm not a statistician, so......

When measuring things such as injuries or deaths with regards to cycles is "per xxx million kilometers travelled" the unit of measurement? e.g. from Scintillas paper (certainly not unique)

Averaged over the years 2002 to 2005, the number of bicyclist fatalities per 100 million km cycled

Why not per capita?

Sparx
Posts: 255
Joined: Wed Dec 12, 2012 3:57 pm

### Re: I'm not a statistician, so......

Guess it helps you work out your probability of being killed based on how much you ride

2 people - one rides to the shop to buy milk vs. one who does 15000km a year.

human909
Posts: 9123
Joined: Tue Dec 01, 2009 11:48 am

### Re: I'm not a statistician, so......

fat and old wrote:When measuring things such as injuries or deaths with regards to cycles is "per xxx million kilometers travelled" the unit of measurement? e.g. from Scintillas paper (certainly not unique)

Averaged over the years 2002 to 2005, the number of bicyclist fatalities per 100 million km cycled

Why not per capita?

It depends what you are trying to measure. Per capita isn't a very good measure because a country that has 0 cyclist would have 0 cyclist fatalities. Likewise per capita the Netherlands has a high fatality rate of cyclists. But that is because there are so many people cycling.

You need to account for the RATE at which the activity is done. Distance or time are both justifiable choices for transport. IMO distance gives pretty skewed results because each distance affects transport choice, but it is a justifiable measure.

(It is skewed because people who cycle are more likely to travel shorter distances than people who drive. So the actual likelihood of a fatality is overstated for the slower modes of travel.)

find_bruce
Moderator
Posts: 8011
Joined: Mon May 09, 2011 8:42 pm
Location: Sydney

### Re: I'm not a statistician, so......

It is used to distinguish between actions which change the activity and actions which change the relative incidence of an occurrence associated with the activity.

To use a real world example, if MHL reduced fatal head injuries on a per capita basis, you are unable to say whether this is (1) due to the protection from a foam lid or (2) a reduction in cycling. By using distance traveled you get a measure of the extent of protection & whether cycling reduced or not.

The complication - distance or time traveled by riders is not recorded in a useful way. Sites such as strava record both measures, but they are a subset of riders which may not be a representative sample