The Plural of Anecdote...

My blog pal Vik posted some interesting thoughts on safety over at his blog and I strongly suggest you read through his post. Now you may or may not agree with what Vik has to say but he's a thoughtful guy and he's very open to the possibility that not everyone will agree with him 100% of the time on 100% of what he says.

I actually agree with Vik a very large percentage of the time and I certainly agree with the bulk of his latest post but I'm going to pick at a nit that I don't think is a nit. I think this nit is actually key to explaining why certain subjects like helmet use and helmet laws or segregated cycle paths and vehicular cycling seem to generate vast amounts of conversation and passion, yet very few people ever actually seem to change their minds on these matter based on these conversations.

I think the problem is in the understanding of the anecdote. Vik describes people who "quote a statistically invalid anecdote to support their un-logic" and then, like the good engineer that he is, he trots out his better data. And he may toss in some anecdotes of his own, like the story of his cousin's wife who died on her first attempt at sky-diving. But he's using that as an example of a personal, invalid data-point he's tossed out, because he's logical and logic is better.

And that, I think is the crux of the problem. Because the data that's out there, the data that matters, is not the data on the efficacy of helmets or probabilities of a cyclist being struck from behind. The data that matters is what has been learned about how human beings assess risk, react when startled, and most importantly, make decisions. We do not, in general, use data to make rational decisions. We use data to rationalize what we have already decided. This may not be logical, but it is very human.

The phrase often used to dismiss anecdotes is the infamous line "the plural of anecdote is not data." The interesting back story to that quote reveals that the original line actually was the opposite "the plural of anecdote is data" but I'd argue that the plural of anecdote is narrative. Narrative is how we structure our lives. Our brains do not retain data, our brains retain stories.

Every human society has a story-telling tradition going back thousands of years and while I certainly will not dismiss the accomplishments of logic and engineering, I think that those who dismiss the anecdotal, dismiss too much. Engineers may dismiss anecdotes, but people remember anecdotes every day, retell them and propagate them. Data sits in tables and appendices to reports. Anecdotes roam and rule the world.

I think Vik is certainly right about fear and the culture of fear. Too many scary stories are being retold every day. For data to effectively counter the fear, data is not enough. We need to tell a story, a compelling story that will be remembered, saved and propagated. Data should support this compelling story, but data alone is not enough.

I have no cease-fire for the helmet wars and I'm not going to tell you what color jacket to buy. And I've got a little library of data and reports that I dig into now and then. When asked about safety I still tend to point people the very handy website How Not To Get Hit By Cars. And when someone tells me of their fear or tells me some scary story, I try, I really try, to listen. And then, most often, I tell them some story. It usually involves someone being happy on a bike.

Keep 'em rolling,

Kent "Mountain Turtle" Peterson

Issaquah WA USA