Hacker News Reader

Yup... while the trend of increasing pedestrian death rates is worrying and should definitely be addressed, the blog is not doing itself a favor by trying to make the numbers look worse than they already are. An 18% increase from 2019 to 2022 should already be enough to convince anyone that something has to be done about it.

A 18% increase combined with a high number of cases.

Otherwise you get silly things like their claim that New Hampshire has seen a whopping 150% increase in pedestrian deaths from 2021 to 2022.

They went from 2 to 5.

The error in the title is they're not talking about death rate but number of fatalities.

since "2019, the last pre-pandemic year, pedestrian fatalities have surged 18 percent in just three years – nine times faster than U.S. population growth," the report said.

Comparing the number of fatalities to population growth makes sense because one would increase naturally with the other.

The ratio is meaningless, though. If the US population increased by 0.001% and the pedestrian fatality rate went up 9%, the fatalities would have increased 9000x times faster than US population growth.

Thats correct.

A high ratio between the two means that population growth doesn't explain the increased pedestrian facilities and therefore this is something to worry about.

The ratio itself does not mean that the rise in pedestrian fatalities is significant. Consider 0.00000000000000001% population growth and a 0.000001% rise in fatalities, or any number of zeroes that you like.

The per capita or per 100,000 rise would be meaningful. You don't get the per capita rise by dividing the rise in fatalities by the rise in population growth.

No, it doesn't mean that the rise in pedestrian fatalities is significant, it means that there is something that is causing a rise in pedestrian fatalities (and that it isn't simply population growth).

I agree the numbers used in the article are not the most explicit, though.

> Comparing the number of fatalities to population growth makes sense

It only makes sense if we compare the number of fatalities to number of pedestrians growth.

More people doesn't necessarily means more pedestrians.

For example: a car ban like there are in many European city centres, leads to a growth of the number of pedestrians in the area, not linked to a growth in the general population.

EDIT: counting population growth as a proxy for pedestrian growth is a spurious correlation.

Population can grow and pedestrian could go down in number and vice-versa, it is been observed in the past 20 years as a result of the fluctuation of urban areas population VS suburb/non-urban areas population.

Depends on how you look at the problem. if you look at it as a health problem it makes sense.

Sorry, I don't understand what you mean.

I'll try anyway to express my though again.

As an example: The analysis by the Governors Highway Safety Association, a nonprofit that represents state highway safety offices, found that pedestrian deaths during the COVID-19 pandemic had spiked significantly, as speeding and impaired and distracted driving proliferated

Now it would be incorrect to say that COVID-19 is dangerous for pedestrians, even though it happened due to COVID-19, the root cause has been speeding and impaired and distracted driving.

In the same way population growth is not directly correlated linearly to an increase of the number of pedestrians fatalities.

If, for example, a family of 2 moves into a new city but they never leave the house because they are both old and disabled, the population has grown, but the risks for pedestrians haven't.

In the case I used in the previous comment, more people walking doesn't correlate with more fatalities, it's in fact the opposite. There are more people walking, true, but in a much safer space for them.

So population growth makes sense as an health problem only if it produces more pedestrians, in particular more pedestrians at risk.

For example an higher driver/pedestrian ratio, which in the end is the same thing of more pedestrians. To use a bad analogy, for which I apologize in advance, it doesn't matter if there are more "targets" for cars to hit or more cars are hitting more of the same number of "targets", the result is roughly the same.

If we look at it as an health problem the only conclusion that makes sense is that the less pedestrians and other road vehicles cross their paths, the more pedestrians fatalities will go down, regardless of population growth.

There are also definition issues. Is a bicycle rider a pedestrian? Is someone waiting for a bus or uber a pedestrian? Is someone walking in a parking lot to thier car a pedestrian? Definitions can range from anyone on foot to only those deliberately going somewhere without a vehicle. Do we include people walking where they are legally not allowed so to do? Without a consistent and stated definition, any stats are just fluff for headlines.

> Is a bicycle rider a pedestrian?

> Is someone waiting for a bus or uber a pedestrian?

> Is someone walking in a parking lot to thier car a pedestrian?

I'm quite sure they are are literally not, the article uses the term "pedestrian" in the title but then it uses "walkers" in the text, it looks quite clear that they are referring to people walking on public spaces as a means of travel.

  pedestrian
  /pɪˈdɛstrɪən/
  noun: pedestrian; plural noun: pedestrians

  a person walking rather than travelling in a vehicle.

Sure, but it provides some grounds to get money for a study on it. Pedestrian deaths are basically free from the CDC, and population is free from the Census. More detailed data would require studies, unless someone is publishing a nation-wide count of pedestrians that I'm not aware of.

There's always better data, and making a study that doesn't have any spurious correlations is extremely expensive. Pedestrian growth is also not necessarily linearly correlated to pedestrian fatalities. If cities become more walkable, it could be that there are more pedestrians walking shorter distances. I would expect that to impact fatalities per pedestrian less than having more pedestrians walking the same amount.

This data just says that the growth in population isn't likely the direct cause of the effect. It could be a difference in the demographics of the population (new people walk more/less). It could be a difference in the existing pedestrians (i.e. % of pedestrians is constant, but people are walking at night or with headphones, etc). It could be a difference in the existing drivers (more dangerous cars, more dangerous driving, more DUIs, etc).

Very good point. They actually say "a staggering 18 percent increase over the number of walkers who died in early 2019". Doesn't quite sound as good as "9 times!".

> the 1.18 times more deaths divided by 1.02 more people

This should have been the title of the post, or, if not, then a very close approximation of it.

What's worse is that going by the title alone I had suspected some clickbait-y foul play, because many of today's articles that rely on numbers or stats have start resorting to those tactics.

Thank you. This sort of thing is a real problem nowadays... It's actually gotten to the point that for 95% of posted content I will check the comments first to make sure it's not total bogus.

Totally bogus internet content has increased infinite % compared with the hedgehog population when compared to 1929

Well, in this case it's not total bogus, the increase in pedestrian deaths is real, but that's exactly what will happen unfortunately if you try to make the numbers look worse than they are: people will dismiss the whole article wholesale...

It's sufficiently untrue that the author or publisher should know that it is untrue and misleading. I conclude they did it on purpose for clicks. However, I am not a scientific sample.

Thank you very much for that nice explanation! I'm almost torn between "what a stupid idiocy!" and "what a clever deception!"

Now hang on a minute

Really well earned use of 'hang on' there, sounds like the article is bollocks

Presumably the numbers are correct, but the title is statistical bullshit.

Technically correct but designed to mislead. Statistical assholery.

This sort of "lying but with some plausible deniability on top" is SOP in the world of "journalism bordering on activism", a category which the site clearly falls into whether one agrees with the policy they're peddling or not.

Welcome to streetsblog.

They have a clear anti-car agenda, they seem to be pro public transportation but even that's questionable

https://nyc.streetsblog.org/2021/05/07/opinion-roll-back-cuo...