Which collisions to look at?

One of the aims of the 20mph scheme is to reduce the frequency and severity of collisions. To capture both of these dimensions I’ll focus on the collisions in which there is a casualty who is killed or severely injured (KSI collisions). Of course, reducing collisions overall would be a good target to aim for as well but reducing KSI collisions should be where the real benefits of the 20mph limits lie.

What time frame to look at?

For my baseline I’ll look at the collisions in 2023-2025. While some of this data will include the effects of changes in the Highland and Borders council areas, much of Scotland was yet to fully adopt 20mph limits at that time. I’ll treat 2026 as a preliminary post-implementation year but wait until the 2027 data is available before drawing any solid conclusions.

What should my denominator be?

Again, this was discussed in my previous post. I want to capture the exposure of drivers to each speed limit. Ideally, I’d like to have the distance driver per km of road in each speed limit. As far as I can see, this isn’t currently published and the traffic volume data seems to lag behind collisions when it comes to being published. I’ll do my best to get this data, perhaps with an FOI, but in its absence I can use the volume data split by major/minor road or simply use the total length of roads at each speed limit. Since I can’t commit exactly at the moment, I’ll update here when I have a clearer picture as well as in a separate post before I start the analysis.

Which speed limits to consider?

This is an important step - I want to separate the signal from the noise without throwing out too much data. I don’t expect the 20mph policy to have much effect on roads over 40mph, and indeed I want to insulate my results from any changes that happen on these faster roads. Given that the vast majority of new 20mph roads will have been previously 30mph, I’ll look at the number of KSI collisions on those speed limits combined. This is also the methodology used by the Welsh Government in their collision reporting and the independent Office for Statistics Regulation were happy with that. More than good enough for me.

I’ll also look at the trend in collisions on 40mph roads as a separate group to act as a kind of control. I don’t think it’s likely to be a clean comparison, but worth looking for any significant increases in 40mph collisions could tell us if collisions are being displaced to elsewhere on the network.

What effect size will I look for?

With roughly 900 KSI collisions a year on roads with 20 and 30mph limits, random chance alone would move the annual count by around 7% either way, so I’ll treat any reduction under 10% as indistinguishable from normal variation As a reference point for a large effect I’m using Wales, where collisions on 20 and 30mph roads fell about a fifth after its national switch. Because Scotland’s roll-out is staggered and partial, I expect a smaller change than that, so anything in the 10–20% range would still be consistent with the policy working. The caveats here are that the Welsh figures include all collisions² whereas I’m looking for KSIs, and any effect should stand out against the trend on 40mph roads, though with only around 100 KSI collisions a year there, that comparison will be noisy and is really a check for obvious displacement.

When will I report?

I’ll use the UK Dept for Transport’s STATS19 final data release for each year. This seems to arrive in late July the following year. I’d expect to be able to put something on thie site within a month of each release. I’ve set a reminder in my calendar and I’m sure you will too.

Do I have a prediction?

Based on the experience in Wales I think a significant reduction in KSIs is likely. I’m less confident about the scale of that reduction given the differences in the nature of the road networks in Scotland and Wales but if I had to put a number on it I’d go for a 10-20% reduction. Any more than that and I’ll be delighted.

What data are we looking at?

I’ve decided to concentrate on the main entry points and routes through the city. There’s a continuous counter placed anywhere you see this icon on the map below¹:

This gives counts at the following locations:

Both Millburn Road and Academy Street have two counters each, one for eastbound and one for westbound traffic. Since this analysis isn’t interested in the directional distinction, they have been combined into one count for each street.

Cycling volume history

The first number we want to look at is just how busy each counter is with cycle traffic. Figure 1 shows all of the cycle traffic volume data since collection started in March 2021. Millburn Road is highlighted orange, and you can mouse over the other lines to see which counter they correspond to.

A couple of things stand out on the chart before we address Millburn Road. The first is that the data in the 2021-2022 period looks quite messy. I think it’s reasonable to say that the travel patterns in the early post-pandemic years is sufficiently different to where we are now to give that data a separate look another time.

The second thing is the steep drop in counts at all sites in the winter 2024/25 period. I don’t know enough yet to explain it, but I do want to say I’m working on figuring it out and will publish what I find. It’s worth noting that cycle traffic appears to be overall lower after this blip. Until I know more about whether that’s real or an artefact of data collection, I’ll refrain from commenting on any trends over time.

Figure 2 shows the data starting in January 2023:

It’s clear that the Millburn Road counter has consistently been among the highest counts in the city. Other sites with similarly high cycling volume are Huntly Street (coloured ochre) and Ness Walk (coloured blue). The significant dip in volume in 2023/24 on Ness Walk was caused by its closure to upgrade the cycling infrastructure there (yay!).

Since 2025 is the most recent full year for which we have data², the median count at each site in that year also gives a useful picture.

This is a clearer way to show the main takeaway from this data: Millburn Road is, realtive to the rest of the city centre, busy with bicycles.

The Mode Shift Share

A really important measure when looking at any mode of transport is its share of the overall traffic in the data. Mode share (sometimes called modal share) is the kind of measure you could name a blog after.

Figure 3 and Table 2 show the mode share companions to our volume data:

You may have had the same initial gut instinct as me - yikes. Millburn Road plummets down the table to nestle between Castle Road and Bank Street, neither of which is a big hitter in the cycle volume league above. Millburn’s previous contenders, Ness Walk and Huntly street stay at the top of the charts, with their mode share towering above all others.

However, if we take the raw volume into account as well as the mode share it actually highlights Millburn Road as being uniquely good at getting bikes through. Ness Walk and Huntly Street have high cycle volumes and high mode shares in part because relative to the other sites they have low numbers of motor vehicles. Millburn Road carries a similar volume of cyclists but with much more motor traffic. This is really clear to see if we look at the motor vehicle counts as well:

Cyclists on Millburn road are typically coexisting with a volume of motor traffic that is an order of magnitude bigger than comparable sites in the city.

Conclusion

It’s fairly clear that historically, Millburn Road is unique within Inverness city centre for its ability to transport large numbers of cyclists alongside heavy motor traffic. This in itself is not entirely surprising as it contained the only section of segregated cycle lane next to a busy road in the city but it does highlight the importance of that lane.

To me, this makes the Highland Council’s decision to remove a section of it right at the most complex junction seem to go against their stated vision to make active travel an attractive and realistic choice for more people, more often, for more of their everyday journeys. I hope they will take a good look at this data as they plan for how they will move forward at that site. I think it makes a good case for improving and indeed extending the cycle lane along that corridor to enable even more people to arrive in the city by bicycle.

🚩 No link to the data source.

The article states:

New figures released by the Scottish Government show 570 collisions on 20mph roads across Scotland ended in injury last year.

Given the lack of a link in the article to the source of data, one would hope that typing “Scottish Government road collision statistics” into your favourite search engine would get you on the way. Unfortunately, what you’ll find is Transport Scotland’s Road Casualty Statistics which are great, but as far as I can tell don’t break down collisions by speed limit.

Instead, the closest match I can find to data quoted in the article is published by the UK Government’s Department for Transport under the stats19 reporting system. It’s a great example of government open data, being comprehensive, well documented and very clean. Hats off to the nerds who keep this data in such good shape, and an extra special nod of thanks for writing and maintaining an R package to go along with it.

Wherever this data comes from it isn’t sufficiently linked or sourced - does this invalidate the article? No, but it starts the article on the wrong foot. To me it says ‘we’re not confident enough in these numbers to let you see them for yourself’ and ‘we don’t care enough about this to get the details right’.

🚩 There’s no denominator.

Ok, fine, technically there is a denominator as we’re given the number of collisions per year but in this context that’s not meaningful enough. Here’s an example of the text of the article:

Last year, 254 crashes on 40mph routes resulted in a driver of [sic] passenger being injured. This was compared to 165 injury collisions on 50mph routes and 268 on 70mph dual carriageways. The figures highlight a five year high for collisions on 20mph routes. A total of 515 were recorded back in 2021, followed by a reduction to 488 two years later.¹

While these are real numbers from a real dataset, they don’t contain enough information to draw any conclusions about the effect of 20mph zones. The most glaring issue I have with this is that if there are more roads with 20mph than last year then more collisions will happen on those roads. Saying that the number of collisions is at a five year high doesn’t tell you much if you don’t take into account the underlying conditions behind that number.

A more informative figure here could be “collisions / kilometer of road type” or even better “collisions / vehicle kilometer on that road type”. The former of these would account for changes in the make-up of the road network while the latter would also account for changes in vehicle volumes. It turns out that the data to make these comparisons is trickier to get hold of (more on that later) but that doesn’t excuse choosing meaningless numbers to make a point.

🚩 “Only 20mph can beat 30mph here”

The next statistical sin feels a lot like the dodgy graphs on the back of election leaflets. Here’s Tim Eagle MSP quoted in the article:

20mph zones account for the third highest number of vehicle crashes that result in an injury, with more collisions than in 40mph, 50mph and 70mph areas

If we take the figures quoted earlier and graph them you can certainly see Mr Eagle’s point:

You may have noticed that the values for 30mph roads and 60mph roads aren’t referenced in the articles. Lets pull them out and add them to the graph:

Oh.

So here’s the thing - it’s OK to exclude some data from your comparisons if you have justification for doing so. In fact I’m going to do just that later in the blog. No justification is given in the article for not reporting the numbers for 30mph and 60mph, and I can’t even charitably think of one that anyone might give in good faith. You could argue, in fact that given most new 20mph limits are replacing existent 30mph limits that the comparison between those two is the most important one you could make.

If we accept the logic of the article that the collisions number is a useful proxy for safety (we shouldn’t), then turning 30mph zones into 20mph zones would be an obvious step to improve safety. The whole picture is completely different with the inclusion of that data. Omitting to mention the counts for 30mph and 60mph without justification is at best misleading and at worst wilfully dishonest.

🚩The Secret to Stats is the Same as the Secret to Comedy

…Timing.

Lets look at one of the claims in the headline:

20mph speed limit branded ‘ploy against drivers’ as collisions reach five year high

Of course, it is factually correct that collisions in 20mph zones has reached a 5 year high. Here’s a graph to prove it:

It does indeed show numbers at a 5 year high, but does that support the claim that 20mph zones aren’t improving safety? If you want to show that on a time series graph you can annotate the graph to show where in time the intervention you’re talking about takes place, like in this classic from xkcd. Let’s add in the deadline for the roll-out of 20mph zones in Scotland, March 2026:

The article is using data to infer the effects of a policy that is taken before that policy has been enacted. It’s a bit like calling your GP to complain that your antibiotics aren’t working before you’ve started taking them. The Scottish Greens, who championed the policy in the Scottish Parliament surveyed all local authorities and found that only 5 had completed the roll-out in June 2026. To be clear you cannot blame anything found in this data on the 20mph roll-out because, as far as the data is concerned it hasn’t happened yet.

A Better Use of the Data

I don’t want to just sit here and criticise the article, I want to see if there’s something we can genuinely learn from the data if we treat it with more respect. I’ll add links to my primary data throughout.