Comment by brownbat | Hacker News Reader

Comment by brownbat | Hacker News Reader

Parent Full thread

I'd like to know some estimate of the rate of disabling or deadly accidents per mile ridden to appreciate the scale of the problem.

The lack of gear and training probably makes accidents worse than they would be. But how frequently and how severe?

> "Some of these injuries have been life-threatening; others have left people permanently disabled."

How many does "some" represent here? Definitely some fraction of the total injuries, which "could be as high as 1,000 per month."

Why "could be?" How uncertain is that figure and what other reasonable estimates of injury rates are there??

On the one hand, the "vast majority end up getting discharged with cuts and scrapes," but on the other hand, "some people may need lifelong care."

How many? And why are you hedging with "may?"

There are some specific numbers in the article-- but they are for individual emergency rooms, without the additional context necessary to extrapolate from that ER over any large regions. And then the information from different ERs ranges so widely as to appear to contradict each other.

I left the article with two takeaways:

1) I assume this is a very real problem.

2) I hate how we reward nonspecificity and anecdotal evidence in journalism. The problem's serious enough to at least take a shot at metrics that would help us to compare it to other social problems, like motorcycle or car or pedestrian injuries.

> On the one hand, the "vast majority end up getting discharged with cuts and scrapes," but on the other hand, "some people may need lifelong care."

> How many? And why are you hedging with "may?"

That's a quote from a doctor about specific cases that went through a specific hospital he works at. That's why.

I've posted the full quote below, for context. Ironically, the journalist went through the effort of giving some very concrete numbers:

> Doctors in Dell Seton's emergency room say they're seeing about 10 injuries a day.

> "The vast majority end up getting discharged with cuts and scrapes, maybe a broken bone," said Ziebell. "But some injuries are significant."

> The hospital has seen 37 severe traumas since April: eight head injuries, 23 orthopedic injuries, four facial injuries and two "other" injuries. In October alone, doctors say 18 people were admitted to either the intensive care unit or surgeries with overnight hospital stays.

> "The folks that had severe head injuries, they're in for a long course of rehab," Ziebell said. "Some people may need lifelong care, like a nursing home."

> Ironically, the journalist went through the effort of giving some very concrete numbers:

C'mon. I addressed that in my very next sentence after the one you quoted.

A "concrete number" that is an arbitrary slice of the problem is more misleading than helpful.

The concrete numbers provided here, about individual ERs, with no indication as to what slice of the problem we're talking about, are essentially random noise when considering the problem as a whole.

> The concrete numbers provided here, about individual ERs, with no indication as to what slice of the problem we're talking about, are essentially random noise when considering the problem as a whole.

Again, the article lays this out concretely:

> Because rentable electric scooters are so new, federal and local officials haven't started tracking accidents, and the companies have declined to release any stats. But some numbers are beginning to appear.

Then, they list the concrete numbers they do know.

Huh. I don't feel like that was a "concrete" rebuttal to the point I actually made, especially when you really look at the "numbers" they cite next in the article, but I want to take a step back.

I think it's getting lost in the noise, so to reiterate, my main point is that I'd just like journalism to include a little more analysis, to put numbers in context, and to avoid relying extensively on emotional appeals as a substitute for chalkboards or backs-of-napkins when dealing with complex social problems.

Based on your username, maybe I'm wrong, but I feel like you know what it's like to wrangle data and communicate findings. Sometimes you don't have all the data you want, but you have to make sense of what you have access to, get signal from noise, and make reasonable extrapolations even under a lot of initial uncertainty.

Now imagine you have access to everything the journalist had. You could ask additional questions at the hospitals or conduct additional searches in the public record.

If we asked, "what else could we learn about this important trend?" is your response really, "well, if the companies don't just tell us and there's no fully funded federal study... we just can't possibly know anything aside from the fact there were some number greater than fifty headlines in six months."

When we stop there I get terrified that Paulos was just right, we struggle to use numbers in public settings so nobody tries. https://en.wikipedia.org/wiki/Innumeracy_(book)

I feel like we could have, say, talked about the populations served by these ERs, or the fraction of overall ER cases that involve scooters and how that's changing over time. Anything to give us a sense of why these ERs vary in their numbers by a factor of ten, to determine how representative they are, or anything to give us a trendline. Even headline counting COULD be rigorous if you add more methodology than this article's "quick search." The gun violence project is one example of that approach.

The journos I really respect take these sorts of next steps. They do more with less. Sometimes they have to hedge and caveat, but they take a shot. It's not impossible, so I just wish we held more journalists to that higher standard, that's all.

Reply on news.ycombinator.com