Smart phones and child injuries

Looking at your smart phone may in fact be more interesting than watching your small child, at the margin at least.  It seems that some child injuries may be going up for this reason.  Here is a new paper (pdf) by Craig Palsson at Yale:

From 2005 to 2012, injuries to children under fi ve increased by 10%. Using the expansion of ATT’s 3G network, I find that smartphone adoption has a causal impact on child injuries. This eff ect is strongest amongst children ages 0-5, but not children ages 6-10, and in activities where parental supervision matters. I put this forward as indirect evidence that this increase is due to parents being distracted while supervising.

Here is Palsson’s other work, I hope he sends me a copy of his Haiti research once it is available.


Lot of other things changed during the same period, so attributing it to one device seems a little strange.

Oops, while I was typing this my grandson just ran out into the street.

Got to leave. If I just had a smartphone with auto dictation this would not have happened because I wouldn't have been sitting at my computer.

"I just had a smartphone with auto dictation this would not have happened because I wouldn’t have been sitting at my computer"

Because who can be expected to pay attention to children if there's a screen around?

Woo hoo! Pretense for [further] regulation of smartphones!!!

Wow >> !

there's lots of very weak social science research out there but this one is breathtakingly bad.

however, this guy has chops and goes full throttle with outrageous claims of 'causation' backed by nonsense statistical conjuring.
He has a bright future in academia. (...Publish Something -> Anything !)

Mad sociology is not a real mad science.

Gentleman - looking forward to your elaboration on what makes this "breathtakingly bad".

"outrageous claims of ‘causation’"

He is measuring the correlation between 3G network rollout and childhood injuries. It is a design created with the intent of capturing causation. Perhaps it is open to some specific criticisms, but you haven't provided any.

There is lots of bad research out there, but you can't just regurgitate the same criticisms every time without thinking about whether they apply to this example. Actually, you can, but that doesn't make it true.

Why people get so worked up for another 4-5 inch screen when TV has been "always" there?

Smartphones allow you to not pay attention in more dangerous locations.

Holy God these are some bad comments. Do you honestly think the paper just correlated the time series of accidents and smart phone usage? Do you honestly think people should just not discover things about the world because you are afraid a third party will use that information to pass laws you don't like? Do you honestly think difference-in-differences estimation, techniques well known to even undergraduates at third rate universities, is "nonsense statistical conjuring"? Awful all around.

Fine Mr Scientist, could you please explain to the unenlightened masses Figure 1 on page 24? According to my high school level math, injuries per year for US children under 5 > 30K times the world population.

What could I know? it's written with Latex, therefore true.

"Do you honestly think the paper just correlated the time series of accidents and smart phone usage?"

The paper didn't even come close to doing that. It makes no attempt to measure smart phone usage. It only correlates accidents and smart phone *availability* in that city.

Which is a much stronger design, because it is much less likely to be confounded by personal characteristics.

While most of the criticisms seem shallow, I think your intuition is wrong. The difference between usage and availability may be important. Think about the mom who has her cellphone in her pocket and the mom who's trying to get her work emails done while she's at the park.

Also, I get the backlash. Anyone who's had kids in this time period has, if anything, noticed an uptick in parental supervision rather than some move to free-range kids. So it's sort of counter-intuitive.

"techniques well known to even undergraduates at third rate universities"

Ah, those darn third rate universities! Turning out hopelessly third rate graduates that can't even understand difference-in-differences estimation. What the heck's the point? We'd be better off just letting the Ivy Leaguers make all the observations and decisions for the rest of us, pretty much an extension of the way it actually is right now.

The first question is whether his conjecture is consistent with behaviour you see in the streets. It certainly is. So I wouldn't reject it out of hand; far from it.

What should be done about it? Perhaps let Darwinism run its course?

Ooh no, I know. Let's have a ginormous tax on mobile phones.

That would be great. Nowadays every broke idiot has a phone that is just as good as mine. Put a $5K/year tax on owning one and either having one would regain some status, or at least the broke idiots would start paying their own way and my taxes could go down.

Your taxes won't go down.

Not to start a fight, because I agree. But at the same time, I don't really understand why anecdotal experiences are the things we use to tell us whether we should believe studies that don't suffer from all the biases that any of us have with our limited experiences and subjective interpretations of events. Shouldn't it be the other way around -- we are more likely to update our opinions because of studies, rather than reject or accept a study because of our experiences?

It is funny- just last week I saw a woman walk right into a street sign post while fingering and looking at her smartphone, and if she hadn't hit the post, I am not even sure she wouldn't have walked out into traffic. I had never personally witnessed such a happening before about 5 years ago, but have seen similar accidents on three occasions now.

Proof that parenting shouldn't be left to amateurs. One can only hope that laws are passed that nobody can use a smart phone while parenting. Put the phone away and get out there and mollycoddle! /sarc

We can't put the genie back in the bottle, whatever the advocates of Pigovian taxes on smart phones may think.

Rather, we should try moving in the opposite direction: we should try to make smart phones so beguiling that they're not ony more interesting than toddler wrangling, but that they distract people from the process that produces the toddlers in the first place.

That accomplished, make sure that every potential single mother of three, and every potential sperm donor thereto, has the latest and most feature-laden phone by the time she reaches puberty...

It would be interesting to know the longer term trend for both severity and frequency of injuries. Perhaps this is just a blip in the overall downtrend.

This post seems to be directed more towards food for thought than factual inform. Regardless, this is clearly an under-developed argument that needs a lot more research before claiming causation. Of course, the issue hasn’t been around long enough to complete a firm study, especially with the amount of dynamic factors that influence the argument. However, it is good food for thought and something to keep in mind as the technological advances continue to grow exponentially.

On the critical side of the study, as I stated above, smartphones have not been out long enough to confirm such a study. In addition, there is of course a very wide range of important, dynamic factors that play a part in every situation involved in the study. This study appears to be based off of a single statistic that does not seem to take into account any of the outlying, crucial factors that play a large role in the outcome of the statistic being used. It is very hard to accurately conclude that this increase in injuries among children is caused by the expansion of ATT’s 3G network, or even linked together at all. However, it likely does play a role in this statistic and will be very interesting to see play out over the next few centuries.

On the defensive side, this study does seem to be quite broad with the intentions of focusing specifically in the future. As I stated earlier, this post seems to be more directed to spark thought in those who have yet to stumble across this probable correlation. I’d be very interested to see how this study will accurately confirm the hypothesis in the future when more crucial, controlled statistics can be included.

Comments for this post are closed