[zsgoldberg]

They can't put the percentage or denominator in the headline? Seems borderline click-baity to me. Even a percentage probably wouldn't cut it -- I can't tell from either whether it's lower or higher than I should expect. Obviously I hope for 0 sexual assaults, but it'd be nice if the headline communicated meaningful information

[vo2maxer]

For more detailed information, the article links to the full report which includes methodology: https://www.uber-assets.com/image/upload/v1575580686/Documen...

[farisjarrah]

From the report:

    The vast majority (99.9%) of Uber trips end without any safety-related issue at all. 
    For example, for the trips in 2017 and 2018:
        •1.4% of trips had a support request of any kind, most frequently for issues such as lost items, refunds, or route feedback.
        •0.1% of trips had a support request for a safety-related concern, and the majority of those concerns were about less-severe safety issues such as complaints of harsh braking or a verbal argument.
        •0.0003% of trips had a report of a critical safety incident,8 which are the incidents referenced in this report

[ocdtrekkie]

Trying to downplay sexual assaults as a small percentage is a way of trying to dismiss that it happens, and that it happens a lot.

[tuesdayrain]

But if the percentage is actually small then it doesn't happen a lot. Looking at the literal number of occurrences instead of the percentage just gives you a biased impression, because any number of assaults > 0 can seem like a lot.

[tzs]

For comparison, it looks like the probability of a random passenger on a 737 MAX dying in an MCAS related accident was about 10 times the probability that a random passenger would report a sexual assault on their Uber ride.

[samcheng]

They grounded the 737 MAXes because that probability of death was too high.

If the probability of assault in an Uber is within an order of magnitude, that suggests to me that it is reasonable to expect some additional scrutiny / regulation.

[nwah1]

Maybe there should be government cameras in all the vehicles, and other sensors that can detect what is going on and provide it all directly to the NSA.

[samcheng]

That seems a bit obscene! I'm sure there are much smaller / fairer measures that shouldn't be dismissed out of hand.

[tyingq]

Nobody is trying to diminish it. Good statistics for crime help provide context, comparisons, measure progress, etc. All of which helps you do better in addressing it.

[erikpukinskis]

Raw numbers provide comparisons and measure progress too. Not sure what you mean by context.

I’m of the same mind as your OP: the goal is zero violence, in raw numbers. Shooting for a low seeming percentage seems to be most often just a way to hide the fact that specific small groups are consistently receiving the lions share of violence, which is unacceptable to me.

[tyingq]

Context would mean many things. A percentage, for example, would let you see if there are areas that exceed the average. Or driver demographic slices (age group, for example). Or time of day/week. Or comparisons to other related industries. If it's higher than regular taxi service, that might help you see if they are doing something you aren't.

It's hard to improve things you don't measure with actionable data.

[erikpukinskis]

I just don’t understand how a percentage counts as actionable data but a count does not?

[tyingq]

Because percentages can be compared across, locations, for example. Raw numbers cannot. 10 assualts in city A vs 20 assaults in city B tells you squat, unless you know how many rides in each.

Similar for other comparisons. Like assaults by driver demographic, like age group. Or comparisons to traditional taxi service...if they have lower incident rates, maybe they have a practice you should adopt.

[satyrnein]

Actionable is having a percentage AND the percentage of the most likely alternative, such as a taxi. Then, you can decide which is safer to take home tonight.

[bshipp]

3400 assaults is a ridiculously high number regardless of the denominator, IMHO.

If anything, the headline isn't clickbaity enough.

[kshacker]

What we should be asking for is analysis of these 3000-3400 cases to see what were the patterns? Maybe there are 3 patterns we can identify as being inherently risky and then educate customers to avoid those.

I agree the number is surprisingly high. It appears 10 people are getting assaulted every day. That's a scary thought. However, at the same time, I am happy that this data exists and is published so that we can now ask for year over year comparisons moving forward.

[827djdjs6]

Trying to equate adding context with 'downplaying' a topic is what's slowly turning people off to most activism. You're not doing anyone favors by advocating for ignorance.

[macspoofing]

>Trying to downplay sexual assaults as a small percentage is a way of trying to dismiss that it happens

Why?

[rolltiide]

since half of those sexual assaults are accusing passengers it really does not give you any insight into the safety of the service as a passenger

the passengers being accused of sexually assaulting a driver or another passenger

for any uber user looking for validation of danger, don't take shared rides, don't ride with friends, don't be a driver and your chances of avoiding sexual assault have just doubled.

[fellowniusmonk]

Article is pay walled, is that number assaults on passengers, drivers or both? Does article it break it down? Seems lower than baseline population numbers either way.