Exploring the idea of ‘peak car‘ a little further, here is a chart showing the percentage of people with full driving licences by age and how this has changed over time. It seems that driving is morphing from something that younger people do to one that older people do and from a male thing to a much more gender-neutral thing. Notice how the percentage of 17-20 and 20-30 year olds has fallen since the mid 1990s and how the number of 50-60, 60-70 and 70+ year olds has been rising steadily; the percentage of people 60-70 year olds overtook the percentage of 20-30 year olds in 2001 and the 70+ year olds are about to do the same.
The two smaller charts break this data down by gender. The percentage of 70+ men overtook 20-30 men in about 2003. The percentage of women with driving licenses grew fast from a much lower base in 1975 getting much closer to parity with men over time. The recent decline in licences held by younger women has been less pronounced than for younger men. Source data from the Department for Transport.
Pedestrian Liberation 
There’s something up with that figure. At the beginning, all the classes are increasing, also there are multiple, non-overlapping categories that add up to more than 100%. So these are not straight-forward percentages.
Looking at the original XLS, the figures just don’t make sense at all. They’re not percentages. They can’t be millions (or there’d have 300M+ in the UK). They’re not percentiles.
If there is some sense to that DoT data, then perhaps the figures here need better labels to help confused people like me. But those figures seem extremely weird and odd.
Columns B-I are percentages of the people in each group but be aware that the ’17+’ coloumn (B) covers all ages aggregated and should be ignored for this chart. Cell C10 indicates that 27.54% of 17-20 year olds had a licence in 1975/6 which is what I show on the chart. Cell C26 shows that 35% of males from 17-20 had a license in 1976 etc. This is what they show on their chart of ” Full car driving licence holders by age and gender: Great Britain, 1975/76 to 2010 “. Column J shows the total number of people with driving licenses as a number (which I don’t use but matches with the percentage in a plausible way). I think that Column K shows the sample size for the group they used to estimate of license percentage (I don’t use that either). Does that make more sense now.
That’s how I understand their labelling too. However, add together columns C10 to J10 – or even just E10+F10… Maybe I’m having an embarrassingly mega brain-fart here, but numbers that add up to much more than 100, for non-overlapping data, don’t seem like percentages to me.
Their chart should show the problem even more clearly. It’s explicitly labelled as “Percentage”, yet appeas to show that there are far more than 100% of drivers who are aged 40+ – however, it also can’t be a cumulative distribution, as it’s decreasing past 40.
There’s either a major problem with the labelling in the original data, or else with the actual data.
I’ve emailed the DFT.
Ah, I get the sense of it now. It’s not percentage of drivers who are that age bin, but percentage of population in that age bin who hold a licence.
Yup. that’s the one. Possibly it would be good for them to have a ‘this is what the data is about worksheet’ as well.