A Different Way to Look at Screen Time

Questions about screens remain among the most asked in emails I get from parents. They’re such a big part of our lives — a source of benefit and a source of stress. It feels like there is an endless parade of concerns about some particular study, or whether some particular screen-time behavior is okay (FaceTiming Grandma, for example).

I want to take a bigger step back and try to give some more perspective on this whole debate. It’s easy to get in the weeds — what does this one study say? Should I freak out? — and also easy to flip-flop: a week in which screens give you a feeling of sanity by distracting your toddler while you cook dinner, followed by a weekend in which your child is a total jerk about wanting more screen time and you’re ready to throw the iPad in the trash. My argument here is we’ll have an easier time navigating this if we can take a broader perspective.

Our big-picture goals:

1. Why the data on screen time is unlikely to provide us any answers (meaning that most of the time “New study says blah blah blah” should be ignored)

2. How to think about opportunity cost, and planning

3. Whether the content or type of screen matters

Let’s go!

Take me to the bottom line

Are screens bad? And why is this question hard?

Why can’t we just have an answer to the question of whether screens are bad or not, and in what way? Is that too much to ask?

Yes. Yes it is.

To see why this is such a hard problem, let’s think about the ideal experiment, which we could do if we were the dictators of the world and everyone was forced to bend to our will. In this case, we would randomly allocate a large number of babies at birth to different screen exposures. Some would get none, some would get a limited amount, maybe varying by age; perhaps some would get assigned to many hours a day. In this ideal experiment, our sample size would be very large, we could have many different treatments, and we’d follow everyone for a long time and measure all kinds of outcomes.

The first thing to say is that even this ideal experiment is not actually perfect, because technology changes. If we started this experiment in 2004, with the intention of following people through the age of 20, we wouldn’t have known about phones and iPads. The answers we would get about early-life screen exposure might be totally irrelevant to the current system.

Second, this experiment is impossible. Not only would it be prohibitively expensive and difficult to monitor, but almost no one would actually want to enroll, and, if they did, we’d be quite worried they wouldn’t follow through on their experimental proposal. We would likely struggle to really believe the results we were getting. The simple fact is that this experiment will never be run.

With the data we have, then, there are two options for how to learn about this question. The “next best” option to an experiment is some kind of natural experiment — some naturally occurring variation in exposure that results in randomness even if it is not done explicitly by researchers. The best example of this is a paper co-written by my husband, Jesse Shapiro, that relies on differences in timing of television introduction to various locations around the U.S. in the 1940s and 1950s. I’ve talked about that paper in most of my books, and I think it’s a great example of this kind of analysis. The paper concludes that television didn’t lower test scores for kids.

However good that paper is, though, it’s a stretch to take those conclusions directly to the current moment, just given how different screen exposure is now. And there aren’t a lot of other good examples of this kind of natural variation in the more recent context.

Instead of this, most papers you read on this topic use observational data. That is: researchers have survey data on kids in which they know about their screen exposure, and they measure some, or many, outcomes. These papers come in many forms. Sometimes they ask about screen exposure in the past (a retrospective study). Sometimes they have data about screen exposure and they follow kids over time (a prospective cohort study). Sometimes they have a sample of children with some particular issue and they collect a control group without the issue and compare the screen exposure (a case-control study).

All of these study types share the same basic problem: they can establish correlations between screen exposure and outcomes but cannot establish causality. There are two significant issues that prevent us from drawing causal conclusions. The first is that the characteristics of families where kids watch a lot of screens differ in other ways — parental education, income, other resources, school quality, etc. — from those who watch less. It is difficult or impossible to separate these other factors (in economics, this issue is called “omitted variable bias”). The second issue is that, in some cases, the outcomes researchers are looking at are driving screen exposure, not the other way around. Parents of kids who have autism or sensory disorders may use screens to regulate behavior, producing a correlation but not suggesting causality (this problem is called “reverse causality”).

Both of these problems are very significant, because screen time is an important choice for many families and wrapped up in all kinds of other features of the family.

In the end, I would argue that we learn almost nothing from papers that use data like this. From a practical standpoint: every time you read a panic headline about screens, look at the data source. If screen exposure isn’t randomized (which it basically never is because — see above — it’s impossible), you shouldn’t read farther.

The most important idea: opportunity cost

A natural follow-on question to the above is: if none of the panic headlines about screens are real, should I conclude that any amount of screens is fine?

The answer to this is no.

The reason is opportunity cost. In economics, when we think about the cost of anything, we think about both the obvious up-front cost (which we call the “accounting cost”) and also the foregone other uses of the money (the “opportunity cost”). The true economic cost of a banana is the 79 cents I paid for it, plus the happiness I’m giving up by not getting my actual favorite fruit.

This concept is very useful in thinking about the value of time. The “cost” of spending an hour on one activity is, effectively, about what else that time would be used for. Here’s me, back in 2013, talking about this idea more broadly. My best personal example of this is from my mother: in my childhood she once told me that she had our groceries delivered rather than shopping because she had a “high opportunity cost of time.”

In the case of screens, the opportunity cost of screen time is the other stuff your child would be doing. We know kids benefit from being read to, from reading, from talking to adults, from playing outside, from sleep. To the extent that screens get in the way of that, they do have a cost. That doesn’t mean that they shouldn’t be used! It just says that we should consider what else would be happening.

For me, this influences a huge share of how I think about screen time.

For example, when my kids are on an airplane, screen time is pretty much unlimited. Because what else are they going to be doing? We have more screen time on weekends and vacations (and sick days) because there are fewer other activities to crowd out.

Being explicit about the opportunity-cost idea has also helped us think about changes over time. Generally, my kids have had some television most days before dinner (to give you a sense of the neurotic routine-based life we’ve got over here: raw veggies are served at 5 p.m., and television starts at 5:12. Yes, 5:12. No, I don’t remember why). But my older kid has more to do these days — more homework, more violin practice, more sports — and these dominate. If everything is done, there is screen time. If not, there isn’t. This isn’t seeing screen time as a reward or lack of it as a punishment; it’s simply a reflection of where screens fall in the hierarchy of activities.

A thought experiment that may be helpful here is to remove the emotional valence we associate with screens and imagine that the activity we’re talking about is your kids sitting in their room staring at the wall. It’s completely neutral. Not especially beneficial, not especially harmful. The question is: when does wall-staring make sense, and when does it not?

The most important approach: planning

If you do not make a plan about screens, you will fight about them all the time. Your kid loves screens. I’m sorry, but they do. They’re fun, entertaining. YouTube unboxing videos have a pull that I cannot personally explain, but it is real.

A solution to this is a plan on screen time. Decide when it will be allowed — when, how long — and then hold to that. This will involve some upsetness. That may happen up front, if you want to make a plan that is limited relative to what is happening now. It will certainly involve some unhappiness down the line when you take away the screens. Your child may cry. They may tell you they hate you and you’re the worst parent.

I am here to tell you: this is okay. You’re not the worst parent. This is just a thing they say! You cannot take it too seriously. If you hold the line, they will get used to things pretty quickly. Just say no!

Also related to planning, there is very good evidence that screens within an hour before bed disrupt sleep. If at all possible, this is one time that you should mostly avoid using them. The occasional family movie night? Sure. But ideally, screens before bed would usually be off-limits. Oh, and that’s true for you too, adults.

Are all content and all mediums the same?

Given a plan about screens, how much should you care what your kids are watching or doing on screens?

Generally, research on content doesn’t suggest it’s that important. Watching more violent shows or playing more violent video games doesn’t seem to lead (causally) to more violence in the long term, and while some information may be learned from educational television, it’s not nearly as good as actual education. (Put differently: if you want your kid to learn to read, you’ll get there a lot faster with 20 minutes of reading lessons a day and 40 minutes of mindless television than with an hour of “educational” television.)

There are some content points, though.

Some curation for age-appropriateness is necessary. Common Sense Media is a great source for information on this, since it breaks down details rather than just providing a single age rating. So if you care more about language than scariness, for example, you can evaluate that.

Related to this: YouTube is extremely popular, and the YouTube algorithm may feed your child things you do not want them to see. YouTube Kids is a good option for very young children, but older kids want the real thing. This is tough, and it’s one of these issues every family is going to deal with differently. You could say no YouTube. You could allow YouTube and watch with them (or make sure they are near you). You could work with them to make sure that if they do see something uncomfortable, they talk to you about it. There are no easy answers here.

In terms of medium — iPad versus TV, say — we do not have particular evidence pointing to differences.

Phones we will leave for another day, except to say that they are a different ball game, in part because they are always there. Limiting screen time becomes increasingly difficult if the screen is with your child at all times. Add on social media, and you’ve got a much more complex problem.

The bottom line

There are no clear and easy answers about screens. Our interactions with them — and our understanding — is going to continue to evolve. Even what we mean by screens will evolve. The iPad has replaced the flat-screen TV, which replaced the set-top box of the 1950s. In another five years, we’ll probably all be sitting on the couch wearing our Apple Vision Pros.
All this change and uncertainty means we can’t guarantee that we’ll make the right choices, for us or our kids. All we can guarantee is that we make thoughtful choices — hopefully something ParentData helps you do.
So please go forward with an understanding of opportunity cost, a plan for planning, and an avoidance of panic headlines!
And finally: if you want to subscribe to an entire newsletter all about screens and kids, I highly recommend Jacqueline Nesi’s Techno Sapiens.