Children under 2 can't learn much from watching TV, but older kids can

• In a large study that used a nationally representative dataset, researchers reported differences in groups of children who watch TV at a younger age compared to those who don't.
• They found differences in these groups in the categories of math, reading, and vocabulary by age six.
• Some data shows that watching a lot of TV before age three seems to be an issue. 
• Watching TV at older ages doesn't seem to matter.
• Overall, this isn't to say that a lot of screen time is fine. "Cribsheet" author Emily Oster says we just don't really know yet.
• Visit INSIDER's homepage for more stories. 

Parental confession: I have never thought of television as a learning opportunity. My kids watch a bit of TV and it is heavily concentrated in time periods in which I need to get something done. At the end of the day on the weekend, when you've spent an entire day with the kids and need to cook dinner, it is awfully nice to send them off to watch TV for half an hour. The pull of the Baby Einstein videos for me was not that it would teach Finn anything, but that they might hold his attention for longer at a younger age.

If some quiet distraction is your goal, then your question is probably not whether TV is a learning opportunity, but whether it is detrimental. Does TV rot your child's brain?

Many studies say yes. For example, a 2014 study shows that preschoolers who watch more TV have lower "executive function"—meaningless self-​control, focus, etc.9  An earlier study, from 2001, shows obesity is higher among girls who watch more TV.

These are just exemplars—many, many research papers correlate more television with bad outcomes. Among the most influential is a 2005 paper by Frederick Zimmerman and Dimitri Christakis. Using a large, nationally representative dataset, their goal was to relate television-​watching at early ages to test scores among children ages six to seven. The researchers categorized the children into four groups based on how much TV they watched in two age ranges: under three years old, and three to five years old. "High" TV-watching was more than three hours a day; "Low" was less than that.

Twenty percent of children fell into what they called the "High-​High" group: more than three hours of TV a day both before age three and between ages three and five. Twenty-​six percent fell into the "Low-​High" group: less TV before age three, more from ages three to five. Fifty percent were in the "Low-​Low" group, and just 5 percent fell into the "High-​Low" group.

The authors reported the differences among the groups in math, reading, and vocabulary test scores at age six. Their results suggest that watching more TV under the age of three lowers test scores; not a huge amount, but by the equivalent of a couple of IQ points. If you are looking in this data for evidence that TV is bad, which is what the authors argue, high watching before age three seems to be an issue.

However, watching TV at older ages doesn't seem to matter. When the authors compared, say, the kids who watched only a little TV before age three and then a lot between ages three and five to the children who watched little TV before age three and little later, they found their test scores to be no different. If anything, the kids who watched more TV later had higher test scores than those who watched less.

Second, although the authors try to control for this, it is very difficult to adjust for all the other differences between kids who watch a lot of TV and those who do not. The majority of the kids in the sample—75 percent— watched less TV between birth and age three; the ones who watched more must have been unusual in some ways. How do we know it was the TV and not these other things that matter? We can't, which is why this is a hard result to interpret.

Some researchers have tried to do a better job adjusting for this second issue, in particular. In my view, the best causal evidence on this comes from a 2008 paper by two economists, one of whom is my husband (but really! I think it is a good paper for other reasons!). In fact, I like this paper so much that I also talked about it in Expecting Better. It's a good example of how to think about generating causal conclusions for a complicated question. It's also helpful for actual decisions about TV.

In the study, Jesse and his coauthor, Matt, took advantage of the fact that television was introduced to different areas of the United States at different times. This variation meant that, when television was first introduced in the 1940s and '50s, some kids had access to TV when they were children and some did not. Since the timing of when people got TV in their area was not related to other parenting inputs, a lot of the concerns raised about other papers could be avoided.

The idea was to see how having TV access as a young child related to test scores when kids were in school at slightly older ages. Jesse and Matt found no evidence that more exposure to television at an early age negatively affected later test scores. This suggests the correlations in other data may be just that—correlations, not causal effects. Of course, TV in the 1940s and '50s differed from TV today, but kids in this time period did watch a lot of television, so the amount of TV isn't very different.

These studies all focus on TV. But in the current parenting climate, screen time has expanded. Your kid can now watch TV on your phone or iPad, but also play games and apps and do all manner of other things. Is this type of screen time like TV? Should it be limited?

We basically have no idea. There are a few studies, but they have pretty big flaws. An example is one paper—not even a paper, more of an abstract—that got a lot of press for showing that language delays were more common in children who had more exposure to a phone between the ages of six months and two years. But this has the same problem, probably even more extreme, as the paper on TV discussed before. What other features of the family correlate with a lot of phone time for a six-​month-​old? Is it not possible those features are what are associated with language delay?

This isn't to say that a lot of screen time is fine. We just do not really know.

From CRIBSHEET: A Data-Driven Guide to Better, More Relaxed Parenting, from Birth to Preschool by Emily Oster.  Reprinted by arrangement of Penguin Press, part of the Penguin Random House company. Copyright (c) 2019 by Emily Oster. 

Join the conversation about this story »

NOW WATCH: How athletes will train in the future