No. cyclists in Melbourne before and after introduction of mandatory helmet law between 90 & 91 (Finch et al)

The effect of the mandatory helmet law on cycling is one of those debates that goes on endlessly because of the absence of hard evidence one way or the other. We simply don’t know how many kilometres of cycling, if any, are foregone because of the helmet laws.

If only a State Government would grant a trial helmet exemption so a proper ‘before and after’ study could be undertaken. But it’s hard to see that happening – there’s no “politics” in it.

However there was an opportunity to undertake a ‘natural experiment’ at the time the mandatory helmet laws were introduced in Australia in the early 1990s. Fortunately, according to an influential paper by D. Robinson, Head injuries and bicycle helmet laws, at least two such studies were undertaken.

These have been cited frequently by opponents of the law to show conclusively that it had a significant negative effect on cycling rates. One was done in NSW by the State’s Roads and Traffic Authority and one by Monash University Accident Research Centre in Victoria.

These studies are frequently cited and re-cited in debates and, as so often happens, succeeding generations of discussants quote others who’ve quoted others who’ve….There’s always the risk of the Chinese Whispers effect – the message gets distorted with each retelling.

My starting view is the deterrent effect of the law (less cycling) probably out-weighs the protective effects it confers (fewer head injuries). Given all the debate about helmets engendered by the disappointing performance of the Brisbane and Melbourne bikeshare schemes, I thought it would be useful to test my view by going direct to the source material: not someone else’s summary of them, but the actual documents themselves.

The first study I sought out is titled An observational survey of law compliance and helmet wearing by bicyclists in NSW – 1993, by N. Smith and F. Milthorpe, published by the NSW Roads and Traffic Authority in 1993.

Astonishingly, I couldn’t find an electronic version of the study anywhere. It’s referenced hundreds of times, but the study itself eluded me. Then I reluctantly looked for a hard copy, searching the catalogues of two major universities, as well as the NSW transport agencies. I know it exists because it’s been cited so many times, but where? So I’ve had to let that pass for the moment – if someone has an electronic copy, could they please make it available?

No such problems with locating the Victorian study, fortunately. It’s got a direct, no-nonsense title: Bicycle use and helmet wearing rates in Melbourne, 1987 to 1992: The influence of the helmet wearing law. It’s by C. Finch, L. Heiman and D. Neiger, and was published in February 1993 by Monash University Accident Research Centre.

Helmets became mandatory in Victoria on 1 July 1990 (the first state in the world). Finch et al counted both the change in the number of cyclists pre and post the law as well as the change in the time spent cycling. They counted cyclists at the same 65 observation sites in Melbourne on four occasions – 1987 and 1990 (the 1990 count was taken five weeks before the new law came into effect), and 1991 and 1992 (after the law).

On each occasion, counting was undertaken over a two-week non-holiday period, seven days a week, between 8am and 6pm. The 1987/88 survey was done in summer, but the others were all done in May/June, a considerably cooler and wetter time of the year in Melbourne.

The exhibit shows that the number of adult cyclists (18 years and over) counted at the observation sites fell in the first year after the law was introduced but returned to almost the pre-law level by the second year. The number of young children (5-11 years) who cycled was declining prior to the new law, but flattened slightly in the first year after and by the second year had returned to the old level.

The big loss was in the high school-age youth segment (12-17 years). Their numbers were flat in the years leading up to the law, but declined sharply in the first year and remained at that lower level in the second year.

So the helmet law didn’t have much effect on the numbers of either adults or primary school-age children who cycled. That’s not a message that’s come through particularly clearly from the frequent references I’ve seen to the Finch et al paper!

Although there were not quite as many of them in absolute terms as adults, the youth segment was a significant group. The age range spans just six years (12-17) but accounted for circa 40% of all cyclists in Melbourne immediately prior to the new law.

However the decline in youth riders has to be evaluated against the main purpose of the law. The improvement in helmet use was spectacular. The adult helmet wearing rate rose from 36% in 1990 to 84% in 1992. The rate for children increased from 65% to 77% over the same period.

And for youth, who clearly didn’t like helmets, the rate rose almost threefold, from a desultory 21% just prior to the new law to 59% of those who were still riding by the second year after the intervention. Of course this won’t cut any ice with those who reckon helmets are pointless (or in some cases positively bad for you!), but it will mean a lot for the much larger group who think helmets have a protective benefit.

I think there are in any event good reasons to question the long-run effect of the new law on youth cycling rates. This cohort was the only one who went from voluntary helmets to compulsory helmets (literally overnight). Subsequent cohorts may not have had as strong an aversion to wearing a helmet since it was always ‘just the way things are’ during their primary school life. Anyone born after 1985 wasn’t even in primary school when helmets were made mandatory (and anyone born after 1978 wasn’t in high school).

It’s also relevant that high school age youth are notoriously fashion and status conscious. This is a fickle business – what was unfashionable in the early 90s might not be unfashionable today, or even five or ten years later. Back in 1990 the Rosebank Stackhat was ugly by any measure. Since then, helmets have become designer items and an Australian won le Tour last year (wearing a helmet). It no longer follows automatically that helmets are as uncool among a majority of today’s 12-17 year olds as they were twenty years ago.

Further, while I haven’t seen any numbers (so I’m not going to die in a ditch over it), I also wonder if cycling has the same allure for youth today that it did in 1983 when Nicole Kidman starred in BMX Bandits. Competing leisure pursuits like game consoles, as well as the increasing propensity of parents to drive their children to school and other destinations, may have reduced their per capita use of bicycles significantly independent of anything to do with helmets. In other words, any aversion to helmets might be very much a second order effect.

But it’s the spectacular growth in adult cycling over the last twenty years that really makes me wonder if the helmet law is a serious constraint on cycling. Bicycle sales have gone gang busters. Rissell and Wen report 18% of Sydneysiders aged 16 and older who were randomly sampled in October 2010 say they had ridden at least once in the last month (and 11% in the last week).

Even if helmets were a serious disincentive twenty years ago (and Finch et al’s work suggests that was only true on average for 12-17 year olds in Melbourne), the alternative interpretation is Australians have adapted to them; we’ve regularised them; we’ve accepted them; we’re simply not bothered any more. Virtually everyone’s forgotten we once had a choice.

Still, that’s only the paper by Finch, Heiman and Neiger. Perhaps the elusive Smith and Milthorpe paper tells a different story, so I’ll reserve judgement until I see a copy of their paper. If someone has an (electronic) copy, could they please provide a link so all can access it?

NOTE: there’s a possibility someone will be tempted to argue that Finch et al mention a cycle rally in 1992 that supposedly “inflated” the numbers at one observation point in 1992, thereby giving a higher adult cycling count than the “true” one.

I don’t think inconvenient results can be ‘cherrypicked’ away. The authors address this question:

From a statistical point of view….an occurrence such as this is a true observation, well within the bounds of “normal” behaviour for that time period, and cannot be excluded from the analysis.

The fact is we don’t know all the myriad other factors that might also have influenced the overall counts – up or down – in all four years or at any of the 65 observation points. It can be argued Finch et al’s sample size was too small to accommodate occurrences like the rally, but in that case the entire study would have to be rejected. I don’t see how that would help the case of those who argue against mandatory helmets.