Researchers find the most plausible cause of wellbeing decline in youth is increased screen time

GettyImages-622069096.jpg
A new paper analyses wellbeing and lifestyle data from over a million US youth

By Alex Fradera

Have young people never had it so good, or do they face more challenges than any generation? Our current era in the West is one of high wealth and relatively free of deprivation, meaning minors enjoy material benefits and legal protections that would be the envy of those living in the past. But there is an increasing suspicion that all is not well for our youth, and one of the most popular explanations, among some experts and the popular media, is that excessive “screen time” is to blame (all the attention young people devote to their phones, tablets and laptops). However, this is a contentious theory and such claims have been treated sceptically by some scholars based on their reading of the relevant data. 

Now a study in the journal Emotion has provided another contribution to the debate, uncovering strong evidence that adolescent wellbeing in the USA really is experiencing a decline, and arguing that the most likely culprit is the electronic riches we have given them. 

The background to this is that from the 1960s into the early 2000s, measures of average wellbeing went up in the US, and this was especially true for younger people. This reflected the fact that these decades saw a climb in general standards of living and avoidance of mass societal traumas like full-scale war or economic deprivation. However, the “screen time” hypothesis, advanced by researchers such as San Diego State University’s Jean Twenge, is that electronic devices and excessive time spent online may have reversed these trends in recent years, causing problems for young people’s psychological health. 

To investigate, Twenge and her colleagues Gabrielle Martin and Keith Campbell dived into the “Monitoring The Future” dataset based on annual surveys of American school students from grades 8 (age 13-14), 10, and 12 that started in 1991. In total, 1.1 million young people answered various questions related to their wellbeing. 

Twenge’s team’s analysis of the answers confirmed the earlier, well-established wellbeing climb, with scores rising across the nineties, and into the later 2000s. This was found across measures like self-esteem, life satisfaction, happiness and satisfaction with individual domains like job, neighbourhood, or friends. But around 2012 these measures started to decline, year on year, up to 2016, the most recent year available. The average effects size of -.14 is typically considered small, but it’s quite something for this to happen over just four years; this is twice as large as the rate of change normally observed between birth cohorts.

Twenge and her colleagues wanted to understand why this change in average wellbeing has occurred, but it’s very hard to demonstrate causes in non-experimental data such as this. In fact, when Twenge previously used the Monitoring the Future data to suggest a screen time effect, sceptical commentators were quick to raise this problem: they argued that her causal-sounding claims rested on correlational data, and that she had not adequately accounted for other potential causal factors (Twenge’s critics also describe her repeated analysis of the same data as a “salami slicing” strategy). This time around, Twenge and her team make a point of saying that that they are not trying to establish causes as such, but that they are assessing the plausibility of potential causes. 

First, they explain that if a given variable is playing a causal role in affecting wellbeing, then we should expect any change in that variable to correlate with the observed changes in wellbeing – if not, it isn’t plausible that the variable is a causal factor. So the researchers looked at time spent in a number of activities that could plausibly be driving the wellbeing decline. Less sport, and fewer meetings with peers correlated with lower wellbeing, as did less time reading print media (newspapers) and, surprisingly, less time doing homework (this last finding would appear to scotch another popular hypothesis that it is our harrying of students with assignments that is causing all the problems). Additionally, more TV watching and more electronic communication both correlated with lower wellbeing. All these effects held true for measures of happiness, life satisfaction and self-esteem, with the effects stronger in the 8th and 10th-graders. 

Next, Twenge’s team dug a little deeper into the data on screen-time. They found that adolescents who spent a very small amount of time on digital devices – a couple of hours – had the highest wellbeing, even more than those who never used them. However, higher doses of screen-time were clearly associated with lower happiness – those spending 10-19 hours per week on their devices were 41 per cent more likely to be unhappy than lower-frequency users, and those who used them 40 hours a week or more (one in ten of teenagers) were twice as likely to be unhappy. 

The data was slightly complicated by the fact that there was a tendency for kids who were social in the real world to also use more online communication, but by bracketing out different cases it became clear that the real-world sociality component correlated with greater wellbeing, whereas greater time on screens or online only correlated with poorer wellbeing.

So far, so plausible. But the next question is, are the drops in average wellbeing happening in tandem with trends toward increased electronic device usage? Superficially, it looks like it – after all, 2012 was the tipping point when more than half of Americans began owning smartphones. Twenge and her colleagues dug deeper and also found that across the key years of 2013-16, wellbeing was indeed lowest in years where adolescents spent more time online, on social media, and reading news online, and when more youth in the US had smartphones. And in a second analysis, they found that where technology went, dips in wellbeing followed: for instance, years with a larger increase in online usage were followed by years with lower wellbeing, rather than vice versa. This doesn’t prove causality, but is consistent with it. Meanwhile, TV use didn’t show this tracking – TV might make you less happy, but this isn’t what seems to be driving the recent declines in young people’s average happiness.

A similar but reversed pattern was found for the activities positively associated with wellbeing – for example, years where people spent more time with friends were better years for wellbeing (and followed by better years). Sadly, the data also showed face-to-face socialising and sports activity had declined over the period covered by the survey.

There is another explanation waiting in the wings that Twenge and her colleagues wanted to address: the impact of the great recession of 2007-2009, which hit a great number of American families and could potentially be affecting adolescents through instability for the family. The dataset didn’t include economic data, so instead the researchers looked at whether the 2013-16 wellbeing slump was tracking economic indicators. They found some evidence that some crude measures, like income inequality, correlated with changes in wellbeing, but economic measures with a more direct impact, like median family income and unemployment rates (which put families into difficulties), had no relationship with wellbeing. The researchers also note that the recession hit some years before we see the beginning of the wellbeing drop, and before the steepest wellbeing decline, which occurred in 2013.

The researchers conclude that “electronic communication was the only adolescent activity negatively correlated with psychological wellbeing that increased at the same time psychological well-being declined.” This isn’t bullet-proof evidence to be sure, and I suspect that sceptical figures in the field will be keen to address alternative explanations – such as unassessed variables playing a role. But the new work does go further than before and suggests that screen time should still be considered a potential impediment to young people’s flourishing.

Decreases in psychological well-being among American adolescents after 2012 and links to screen time during the rise of smartphone technology

Alex Fradera (@alexfradera) is Staff Writer at BPS Research Digest

31 thoughts on “Researchers find the most plausible cause of wellbeing decline in youth is increased screen time”

  1. I can’t believe how blind these researchers are. The reason there’s a decline is because parenting is in Decline. period. As a teacher the only person to blame when I have a difficult student is the parents.

    Like

    1. I would argue that I have never seen a teenager come off a gaming device happier than when s/he went on it. They tend to be spaced out, irritated at having been asked to stop playing or annoyed that they have ‘died’ in the game. The euphoria and joy you see when kids come off a football/hockey/netball pitch, win or lose, demonstrates a totally different mind state. Its simple to me – time on devices makes no one really happy.

      Liked by 1 person

  2. Video games, TV, radio, cinema, books, newspapers, pamphlets, cave paintings… you have to go back an awful long way to find an era when young people weren’t ‘corrupted’ by something….lock up that damned abacus!!! they once said in frustration…

    Liked by 1 person

    1. ‘Corrupted’ in the sense you imply and decline in well-being are not the same thing.

      BTW, the abacus is unlikely to cause either corruption or decline in well being.

      Like

  3. Everyone has their own view on this. For me it’s economical because I feel that impacts myself most. I can’t get a decent job and can only afford the same level of education that student loans permit, devaluing the education that I have. I am obviously not a teenager, but it’d be worth spreading the study wider to see.

    Like

  4. I’m not sure about the way the economical variables were discounted where the screen time ones were not.
    The reasoning presumes an immediate feedback rather than a delayed effect. I can imagine that good wellbeing would continue for a while through financial hardship, but after too much hardship people would start to feel the effects more keenly over time?

    Like

  5. The BPS should get an epidemiologist to review their blog posts about observational data. The problem of confounding means that no crude association can be interpreted as a causal one without extra causal argumentation/modelling. The finding of an absence of a crude association between two variables also does not mean that there is no causal association – again a confounder can mask the occurrence of a true association, as illustrated in some versions of the Simpson’s paradox.

    Liked by 1 person

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s