How Much Evidence Do We Need?
It seems that not a day goes by without our reading some harsh indictment of Facebook, its subsidiary Instagram, Twitter, You Tube, and other social media platforms. They are accused of warping young minds, distorting election results, spreading misinformation about diseases, and generally imperiling modern society.
An article by columnist Ishaan Tharoor in the Washington Post last October bore the headline “The indisputable harm caused by Facebook.” Tharoor wrote that “Facebook and the other apps it owns…. are now increasingly seen through the prism of the harm they appear to cause. They have become major platforms for misinformation, polarization and hate speech. At the same time, [Mark} Zuckerberg [the Facebook founder and owner]and his colleagues rake in billions of dollars each quarter in profits.”
But what is the evidence that we are actually harmed by Facebook, Instagram, WhatsApp, and other social media platforms? Last September the Wall Street Journal accused Instagram of having done research demonstrating it is “toxic” to teenage girls. This was part of a larger series of articles, based in part on testimony by a whistleblower, alleging that Instagram and its owner Facebook have hidden internal documents showing how much harm they do.
The charge that Instagram harms young women is at first glance plausible. Numerous sources have shown that rates of depression and other mental health problems are rising among young people in the U.S. and elsewhere. For example, a 2019 Pew Research report showed that 17% of U.S. teenage girls had experienced at least one episode of major depression in 2017, compared to 7% in 2007. At around the same time, Instagram users rose at a rapid pace. So, is logging on to Instagram and scrolling for hours through its feed of photos and videos causing teenagers to become depressed?
The Data Are Lacking
In October, Laurence Steinberg, a professor of psychology at Temple University, wrote about this in the New York Times and his findings are startling: “Amid the pillorying of Facebook that has dominated the latest news cycle there is an inconvenient fact that critics have overlooked: No research – by Facebook or anyone else – has demonstrated that exposure to Instagram, a Facebook app, harms teenage girls’ psychological well-being.”
What is the research that the Wall Street Journal says showed that Instagram harms teenagers, especially girls? “Facebook conducted surveys and focus groups to which people were asked to report how they thought they had been affected by using the Instagram app,” Steinberg explained. “Three in ten adolescent girls reported that Instagram made them feel worse about themselves.” Now surveys and focus groups are important aspects of research attempting to understand how things affect us, but as Steinberg points out, they rarely can establish cause and effect relationships. That is, the research Facebook apparently did on its own that was leaked to the Wall Street Journal is merely suggestive and does not establish that Instagram causes depression or any other mental health disorder in anyone, including teenage girls. First of all, note that the data cited suggest that seven in ten adolescent girls did not indicate that looking at Instagram altered their self-esteem. Second, it could be that depression increases the chances that someone will spend time looking at Instagram and not the other way around. Perhaps lonely, depressed people seek answers to their problems on social media. Third, as Steinberg points out, a myriad of other factors could mediate any relationship between Instagram and mental health.
Steinberg noted that there is some research looking at relationships between social media use and mental health. “Of the better studies that have found a negative correlation between social media use and adolescent mental health,” he wrote, “most have found extremely small effects—so small as to be trivial and dwarfed by other contributors to adolescent mental health.”
We are not attempting here to defend Facebook, Instagram, or any other social media platform. [Full disclaimer: Critica president Jack Gorman once owned about $2000 in Facebook stock, which he sold several months ago because of conflict of interest concerns]. We at Critica have had our own negative experiences with Facebook. One of our projects is to engage people who spread misinformation about COVID-19 vaccines on online platforms, including Facebook and Twitter. We have seen our posts, which attempt to provide correct information about health and science, deleted by Facebook’s algorithms while misinformation about vaccines, sometimes containing what we consider to be akin to hate speech, are left standing by those algorithms. Facebook officials have acknowledged that what we post should not be deleted and tried to help us with this problem, but it is a daunting one because with millions of posts circulating on its platforms it is not possible for artificial intelligence to correctly pick out with 100% accuracy which ones are worthy to let stand, and which are misinformation and should be deleted.
We recognize that because some of our work depends on our posting on Facebook, it could be argued that we are incentivized to be reluctant to be critical of it. We are still fully cognizant of the problems with misinformation, hate speech, and political polarization that circulate throughout social media and agree that these pose a threat to society in multiple ways. Nevertheless, the example of the charge that Instagram harms teenage girls’ mental health should make us pause and ask the most basic question of whether we know that social media is in fact causing harm or instead is reflective of adverse circumstances rather than their cause.
This is not merely an academic question because if social media is indeed causing the kind of harm to which it is accused, then it is perfectly appropriate for us to request government regulation as a remedy. Let us say, for instance, that it could be demonstrated that Instagram causes depression in 30% of teenage girls who use it. We would be remiss if we didn’t demand some intervention to prevent such a large number of young women developing such a serious disorder. We might ask if we would have such severe political polarization and hate speech leveled against society’s marginalized people if it weren’t for Facebook? If we didn’t have social media platforms to spread misinformation about vaccines, would more people be vaccinated today against COVID-19 today, and would fewer people have died? Would we have fairer elections throughout the world if there were no social media platforms spreading disinformation? Clearly, if the answer to any of these questions is “yes,” then social media urgently needs very serious regulation.
Meaningful Research is Very Difficult to Accomplish
It will not be easy to acquire such data, however. Sticking with the claim that Instagram causes depression in young women, how would we go about documenting a true causal relationship? Steinberg calls in his New York Times op-ed piece for randomized controlled trials, but it is not easy to imagine what those would look like. We would need to take a group of young women who were free of depression as the research participants and randomize them to have some exposure to Instagram versus no exposure or exposure to something else. Then we would see whether the group exposed to Instagram developed worsened self-image or even depression. You can see immediately what the problem with this design is, however, because implementing the exposure and deciding how much exposure is not at all straightforward. Some participants will already have considerable experience with Instagram and others less or none, so how to control for that in the experiment is tricky. Trickier still is figuring out how much exposure in the experiment is both practical and meaningful. Simply having the exposure group look at one Instagram post or stay on Instagram for a specified time period may be insufficient to show an effect; people who use Instagram can do so on a regular basis and it may be that prolonged and/or repeated exposure is needed to induce a mental health problem. Finally, even if we were able to come up with a viable randomized study design, there are obviously important ethical issues in attempting to actually see if something can make people feel worse about themselves or develop depression.
So designing a meaningful experiment to show that a social media platform can cause a mental illness is not going to be easy. How much more difficult will it be, then, to show that social media actually causes adverse behaviors like not getting vaccinated or harms minority groups or causes elections to be unfair?
Social scientists are working on designing experiments that will help us answer these questions in ways that are practical but do not pose ethical issues. When we contemplate regulating social media, we must ask a number of difficult questions. What exactly would we be regulating? Banning teenagers from using Instagram? Banning certain types of Instagram posts? Banning Instagram entirely? There is even the possibility that under some circumstances, looking at Instagram could be helpful for mental health. Perhaps depressed teenagers go to Instagram seeking answers for their problems and maybe some of them actually get help in making connections there to other people or to groups that specialize in mental health issues.
One place to start would be to implement regulations mandating that social media companies make their platforms available to scientists working on these issues. Last summer Facebook banned a group of New York University investigators from using its platform to conduct research on political ads and disinformation. We see that as emblematic of a general hostility to research and transparency on the part of social media companies like Facebook, something we believe can only be addressed by a legislative intervention. Ryan Calo, professor of law at the University of Washington, has called for just such action. “Congress holds the power to stay Meta’s [the new name for the Facebook company] hand when it comes to threatening legal action or blocking accountability research…It could mandate transparency.”
Accusations about harms caused by Facebook and other social media platforms are being made at present but in some cases we still lack sufficient data to know with a reasonable degree of certainty that social media is indeed causing those harms. But the accusations are serious and therefore we cannot be reticent about demanding the evidence to decide if they are warranted. If social media companies like Facebook refuse to allow researchers to gather the information society needs to judge their safety, then we believe it is appropriate for governments to take action to make them. It may be that the end result of good internet research is that social media is not the purveyor of the harm it is accused of; we need to find that out.
criticascience 