The Importance of Evidence
The NTIA’s focus on “reliable sources of concrete information” in their request is laudable. Though all policies should strive to be grounded concretely in the available evidence and research, for a few reasons it is particularly important in the issue of minors’ social media use and mental health. First, an effective solution requires an accurate understanding of the problem. Without a solid basis of research and evidence, policymakers are likely to misdiagnose the problem and focus their efforts on the wrong solution set. Time spent on the wrong solution is time lost. When it comes to an issue such as minors’ mental health, that lost time imposes a massive opportunity cost.
Beyond the opportunity costs incurred, misplaced policies can create new harms. After China officially recognized “internet addiction” as a mental disorder, abusive treatment camps followed.1Ben Westcott, “Children in China Locked Up for As Long As 10 Days at Internet Addiction Camp,” CNN, July 8, 2020, https://www.cnn.com/2020/07/08/asia/china-court-abuse-internet-addiction-intl-hnk/index.html. This is an extreme example, but something that policymakers should keep in mind as we consider what approach we would like to take as a nation. One major consequence, as noted by a number of human rights and LGBTQ+ groups that opposed the Kids Online Safety Act (KOSA), is how children’s online safety bills could harm or be weaponized against LGBTQ+ children.2“More Than 90 Human Rights and LGBTQ Groups Sign Letter Opposing KOSA,” Center for Democracy and Technology, November 28, 2022, https://cdt.org/press/more-than-90-human-rights-and-lgbtq-groups-sign-letter-opposing-kosa/.
A coarse understanding of the issue may also lead us to an inappropriate one-size-fits-all solution that erases all benefits of social media along with potential harms. Teenagers themselves recognize the benefits of social media use. As a Pew report notes, majorities of teens credit social media with providing a space for connection and support.3Monica Anderson, Emily A. Vogels, Andrew Perrin, and Lee Rainie, “Connection, Creativity, and Drama: Teen Life on Social Media,” Pew Research Center, November 16, 2022, https://www.pewresearch.org/internet/2022/11/16/connection-creativity-and-drama-teen-life-on-social-media-in-2022/. Even in the famous “Facebook Files,” the number of teens who said Instagram made them feel better about themselves was almost double those who said it made them feel worse.4“One in Five Teens Say that Instagram Makes Them Feel Worse about Themselves, with UK Girls the Most Negative,” in Georgia Wells, Jeff Horwitz, and Deepa Seetharaman, “Facebook Knows Instagram Is Toxic for Teen Girls, Company Documents Show,” Wall Street Journal, September 14, 2021, https://www.wsj.com/articles/facebook-knows-instagram-is-toxic-for-teen-girls-company-documents-show-11631620739?mod=article_inline. Note: Even though the title of the slide says 1 in 5 teens feel worse about themselves, notice that the chart shows 41 percent of teens in the US feel better about themselves because of Instagram—that’s 2 in 5 that benefit from it. This does not discount the teens who say it made them feel worse, but it does show that a one-size-fits-all approach is not well-suited to the issue.
Second, government actions on this issue impact speech and therefore require a solid basis in evidence to pass judicial scrutiny. It is not enough to correctly note, as does the US Surgeon General’s Advisory on Social Media and Youth Mental Health, that the United States “has a strong history of taking action” to ensure the safety of products that have a “widespread adoption and impact on children” such as toys or transportation.5Social Media and Youth Mental Health: The US Surgeon General’s Advisory, Office of the Surgeon General, Department of Health and Human Services, (May 23, 2023), https://www.hhs.gov/sites/default/files/sg-youth-mental-health-social-media-advisory.pdf. Any policy seeking to regulate or otherwise enforce safety standards on social media platforms will invariably impact speech.6Netchoice LLC, et al, v. Attorney General, State of Florida, et al., USCA 11 Case: 21-12355 (11th Cir. 2022). https://www.govinfo.gov/content/pkg/USCOURTS-ca11-21-12355/pdf/USCOURTS-ca11-21-12355-0.pdf. As such, these policies will receive much greater judicial scrutiny than the Standard Consumer Safety Specification for Toy Safety or Federal Motor Vehicle Safety Standards. In order to withstand this scrutiny, the government must be able to prove that its action:
- serves a compelling interest—one of the highest order—when it targets a specific type of content or subject matter;
- is so carefully drafted (so narrowly tailored) that it restricts no more speech than is necessary to serve that interest (no alternative ways exist that would restrict less speech);
- clearly defines key terms to avoid being unconstitutionally vague; and
- is not overbroad (it does not bar a substantial amount of protected expression along with speech that legitimately can be restricted).7Clay Calvert, “Protecting Minors Online: Some Tips for Lawmakers from Yesteryear’s War on Violent Video Games,” AEIdeas, September 26, 2023, https://www.aei.org/technology-and-innovation/protecting-minors-online-some-tips-for-lawmakers-from-yesteryears-war-on-violent-video-games/.
Without a solid basis of evidence that social media usage causes harm to minors and a nuanced understanding of how those harms occur, policymakers will find themselves ill-equipped to write a law or guidance that can withstand judicial scrutiny. This has already proven to be the case. Of the states that have passed bills attempting to impose safety standards or regulate minors’ usage of social media platforms, three have been challenged and blocked.8Memorandum Opinion and Order Net Choice, LLC v. Tim Griffin, in his Official Capacity as Attorney General of Arkansas, No. 5:23-CV-05105 – Doc. 44 (W.D. Ark. 2023); Order Granting Mot. for Preliminary Injunction, Net Choice, LLC v. ROB BONTA, Attorney General of the State of California, No. 22-cv-08861-BLF – Doc. 74 (N.D. Cal. 2023); Order Granting Plaintiffs’ Mot. for Preliminary Injunction, Free Speech Coalition, Inc., et al. v. Colmenero, No. 1:2023cv00917 – Doc. 36 (W.D. Tex. 2023). As such, a full consideration of the evidence currently available to us is necessary before taking any further action.
The Current State of Research
Mental health issues are inherently complex. They are not monocausal.9Jacqueline Nesi, “Does Social Media Cause Teen Mental Health Issues?” Techno Sapiens, January 30, 2023, https://technosapiens.substack.com/p/does-social-media-cause-teen-mental. People react differently to the same stressors due to myriads of factors including genetics, social situations, environmental factors, etc. Thus, like any social phenomenon, the relationship between social media and teen mental health is almost certainly complex. Unfortunately, much of the research we currently rely upon in this area tends to “treat individuals as faceless aggregations, which obscures the detail necessary to link complex phenomena like social media use and depression,” for example.10Aaron Brown, “The Statistically Flawed Evidence That Social Media Is Causing the Teen Mental Health Crisis,” Reason, March 29, 2023, https://reason.com/2023/03/29/the-statistically-flawed-evidence-that-social-media-is-causing-the-teen-mental-health-crisis/.
The bulk of current research on the impact of social media usage on teen mental health is correlational—typically cross-sectional, though sometimes longitudinal.11Renata Maria Silva Santos, Camila Guimarães Mendes, Guilherme Yanq Sen Bressani, et al., “The Associations Between Screen Time and Mental Health in Adolescents: A Systematic Review,” BMC Psychology 11, no. 127 (April 20, 2023), https://doi.org/10.1186/s40359-023-01166-7. Though the research is mixed, when considered all together, there is a general consensus that “adolescent children and young adults showed a small to very small association between screen time and depressive symptoms, varying between different devices and uses.”12Santos, Mendes, Bressani, et al., “The Associations Between Screen Time.” While these studies are helpful for an initial gauging of an issue, they do not provide a comprehensive, nor detailed, understanding of the relationship.
One primary concern with the current research is the lack of consistent definitions and measurements of social media usage, making it challenging to draw clear causal relationships between social media usage and mental health outcomes. The categories tend to be coarse. Some of the most refined simply measure time spent “social networking” without a distinction between sites and type of use.13Sarah M. Coyne, Adam A. Rogers, Jessica D. Zurcher, Laura Stockdale, McCall Booth, “Does Time Spent Using Social Media Impact Mental Health?: An Eight Year Longitudinal Study,” Computers in Human Behavior 104 (March 2020), https://doi.org/10.1016/j.chb.2019.106160.
Further, nearly all of the most highly cited research relies upon self-reported screen time measurements. However, these reporting methods have been shown to be systematically biased and unreliable. For example, communications professor Michael Scharkow compared self-reports of internet use with computer log files, which show everything that a computer has done and when, and found that “survey data are only moderately correlated with log file data.”14Michael Scharkow, “The Accuracy of Self-Reported Internet Use—A Validation Study Using Client Log Data,” Communication Methods and Measures 10, no. 1 (March 24, 2016): 13–27, https://doi.org/10.1080/19312458.2015.1118446; Douglas A. Parry, Brittany I. Davidson, Craig J. R. Sewall, Jacob T. Fisher, Hannah Mieczkowski, and Daniel S. Quintana, “A Systematic Review and Meta-Analysis of Discrepancies between Logged and Self-Reported Digital Media Use,” Nature Human Behaviour 5, no. 11 (November 2021): 1535–47, https://doi.org/10.1038/s41562-021-01117-5. His paper has since been replicated and broadly confirmed. A systemic review of 106 experiments by six researchers found that “self-reports were rarely an accurate reflection of logged media use and that measures of problematic media use show an even weaker association with usage logs. These findings raise concerns about the validity of findings relying solely on self-reported measures of media use.”
The current body of research also lacks a consistent definition and measurement of mental health across studies.15Steinsbekk, Nesi, and Wichstrøm, “Social Media Behaviors.” Some studies seek to capture anxiety, depression, or even just well-being generally. In a study attempting to better examine the issue by using gold standard mental health evaluations, researchers noted that the varying conceptualizations of mental health may explain why studies often report inconsistent and even conflicting results.16Steinsbekk, Nesi, and Wichstrøm, “Social Media Behaviors.” Self-reported measures of mental health issues share many of the same problems as self-reported social media usage. The questions are often subjective and may reflect the respondents’ current mood more than their overall state of mental well-being. As Aaron Brown puts it, “checking that you ‘sometimes’ rather than ‘rarely’ feel helpless is only distantly related to how depressed you are.”17Brown, “The Statistically Flawed Evidence.”
Finally, it is important to keep in mind that the research currently available lacks clarity on the causal direction of these associations or whether there exists a third confounding variable. In other words, poor mental health could be driving social media usage. Or, a third confounding variable could be driving them both. For example, the potential for screen use to replace other activities important for mental health, including sleep and physical activity, is also being investigated by researchers.18Denver M.Y. Brown and Matthew Y.W. Kwan, “Movement Behaviors and Mental Well-Being: A Cross-Sectional Isotemporal Substitution Analysis of Canadian Adolescents,” Frontiers in Behavioral Neuroscience 15 (October 5, 2021), https://doi.org/10.3389/fnbeh.2021.736587.
Thankfully, we are not limited to the current research and can improve upon what we have in order to improve our understanding of the problem. Researchers should conduct longitudinal studies to test specific theories of harm by measuring types of usage and online behaviors. They should complement this with the use of psychiatric interviews to note mental health issues. One study, conducted among Norwegian children, has already met many of these criteria.19Steinsbekk, Nesi, and Wichstrøm, “Social Media Behaviors.” The study in question followed a birth cohort of Norwegian children ages 10–16. The researchers captured symptoms of depression, social anxiety, and general anxiety through psychiatric interviews. The study does rely upon self-reports to measure social media usage. However, researchers improved upon previous studies by asking detailed questions about usage. As such, they were able to capture nuances, like whether participants used social media in a “self-oriented” or “other-oriented” way. Ultimately, the researchers find no connection between “self-oriented” and “other-oriented” use of social media and future symptoms of depression and anxiety. They do offer a jumping-off point for future research, however, by noting that this study is not comprehensive. There could very well be other specific behaviors not captured by this study that do relate to future mental health issues. Researchers should dedicate time and effort to nuanced and complex studies such as this to better develop our understanding of the issue. The US federal government, via the National Institutes of Health, is already on this path. As this RFC notes,
In accordance with the CAMRA Act, [the NIH] supports biomedical and behavioral science research to study the health impacts of digital media exposure on youth, which may include the positive and negative effects of exposure to and use of media, (such as social media, applications, websites), to better understand the relationships between media and technology use and individual differences and characteristics of children and to assess the impact of media on youth over time.2088 Fed. Reg. 67,733.
As researchers develop a better understanding, we should assess what we currently know and take action accordingly. What we can tell from the research currently available to us is that certain types of social media usage may have a negative impact on the mental health of certain adolescents sometimes.21Amy Orben, Andrew K. Przybylski, Sarah-Jayne Blakemore, and Rogier A. Kievit, “Windows of Developmental Sensitivity to Social Media,” Nature Communications 13, no. 1649 (March 2022), https://doi.org/10.1038/s41467-022-29296-3; Ine Beyens, J. Loes Pouwels, Irene I. van Driel, Loes Keijsers, and Patti M. Valkenburg, “The Effect of Social Media on Well-Being Differs from Adolescent to Adolescent,” Scientific Reports 10, no. 10763 (July 1, 2020) https://doi.org/10.1038/s41598-020-67727-7. We know that t/here has been a rise in adolescent mental health issues.22Centers for Disease Control and Prevention, Youth Risk Behavior Survey Data Summary and Trends Report: 2011–2021, April 27, 2023, https://www.cdc.gov/healthyyouth/data/yrbs/pdf/YRBS_Data-Summary-Trends_Report2023_508.pdf. We know that parents of adolescents are worried about the potential negative impact that social media could have on their children.23Risa Gelles-Watnick, “Explicit Content, Time-Wasting are Key Social Media Worries for Parents of U.S. Teens,” Pew Research Center, December 15, 2022, https://www.pewresearch.org/short-reads/2022/12/15/explicit-content-time-wasting-are-key-social-media-worries-for-parents-of-u-s-teens/. These deserve to be addressed, but our response must be in line with the evidence.
So, perhaps the most important question we can ask ourselves is: what type of intervention is supported by the current research?
Where to Go from Here?
As noted above, direct regulation of social media platforms and the content they host will face strict scrutiny in courts and may create additional problems for the very people the policies are trying to help. Given the current evidence and constitutional constraints, there are no silver-bullet policy solutions. The best path forward is incremental and composed of a multitude of small and medium-sized solutions. This approach has the significant advantage of adaptability, which is important given the constantly changing nature of social media technology, the changing preferences of minors, the constant innovation in devices on which most minors and adults access social media tools, and all the other unforeseen technological and cultural shifts in the future. By contrast, any technology-specific law in this space risks cementing into place practices that could be outdated within just a few years. A multifaceted approach can be adapted as research continues to be conducted on the link between social media use and the mental health of minors.
We outline a set of incremental solutions that could be implemented now or in the short term to aid in making social media a safer experience for minors. These are not implemented in silos, but instead represent a multi-sided and multi-factored approach.
There are a multitude of tools available to teens and parents that can help mitigate risks that minors face online.24Taylor Barkley, Aubrey Kirchhoff, and William Rinehart, “Key Resources for Parents,” Benchmark, The Center for Growth and Opportunity, March 7, 2023, https://www.thecgo.org/benchmark/key-resources-for-parents/. These are provided by a range of sources including social media companies, Wi-Fi router manufacturers, third party software, and even internet service providers.25“Children Online Safety Tools,” Competitive Enterprise Institute, accessed November 3, 2023, https://cei.org/children-online-safety-tools/. For example, the eero Wi-Fi router offers an eero Plus service for $9.99 a month that at the internet access point offers a wide range of services and insights for caregivers. According to the eero Plus website, users can turn on “preset filter levels” based on age ranges, block or set time restrictions for certain categories of content or applications, add custom websites to a blocked list, and monitor usage for individual users and their devices.26“eero Plus,” eero Help Center, accessed November 8, 2023, https://support.eero.com/hc/en-us/sections/115000878703-eero-Plus. However, most tools are included with the platform and offered free of charge. Apple, Google, and other software and device manufacturers include services that also can set time limits on certain apps, filter or block content, and allow for cooperation between caregivers and minors.27“Google Families | Empowering Kids to Safely Connect, Play, and Learn Online,” Google Families, https://families.google/intl/en_us/; “Families,” Apple, https://www.apple.com/families/. These tools don’t incur an extra charge because they are market responses to customer demand and subsidized by other sectors of the business.
Compelled use of these tools would raise First Amendment issues. However, government education on the availability and features of such tools would be welcome and helpful. Currently, the number of options available may overwhelm caregivers. With proper awareness and education, the abundance of tools could become an asset that allows parents and caregivers to choose which ones are the best fit for them and their families.
There are also a number of non-profit organizations dedicated to educating minors, caregivers, educators, and policymakers on tools and practices that can mitigate the risks of social media.28Barkley, Kirchhoff, and Rinehart, “Key Resources for Parents.” These organizations share many of the same values, recommendations, and goals and should be commended by policymakers whenever they are speaking about methods that minors and caregivers can use to stay safe online.
Nonetheless, there is evidence that caregivers, educators, and minors themselves do not know about these organizations and tools. And if they are aware of them, they often feel overwhelmed or do not know how to effectively use the tools. Companies, non-profits, and policymakers should call more attention to these organizations that are standing ready to help.
States are also passing laws that direct departments of education and other agencies to incorporate social media digital literacy education into school curricula. Florida was the first mover in 2023, and more states are poised to follow.29James Sheets and Taylor Barkley, “Clicks and Codes: Social Media Awareness Legislation in the States,” The Center for Growth and Opportunity, November 7, 2023, https://www.thecgo.org/benchmark/clicks-and-codes/.
Companies are making design and feature decisions that are intended to help minors stay safe online. To the extent that they fall short, the companies should be held to a high standard on this front by all interested parties. If it is apparent that a company, large or small, is not doing enough to prioritize the safety of its youngest users, they should be called out and challenged to make fixes. Indeed, such pressure from civil society groups, caregivers, and policymakers seems to have driven recent responses. For example, in December 2021 the day before the Senate testimony of Adam Mosseri, Head of Instagram, the platform “released a set of tools meant to promote user health.”30Thomas Barrabi, “Instagram Chief Adam Mosseri Defends App, Calls for Regulation at Senate Hearing,” Fox Business, December 8, 2021, https://www.foxbusiness.com/politics/instagram-adam-mosseri-senate-hearing.
Companies should also ensure that users are aware of the tools and services they provide and that caregivers and minors can confidently use them to good effect. Ultimately, the marketplace is one of the surest forms of accountability for private companies. As noted by our colleagues in a previous comment to NTIA, “markets play a crucial role in disciplining company behavior and holding them accountable to their customers,” through the mechanisms of “competition, reputation, customer feedback, pricing, and transparency,”31Neil Chilson and William Rinehart, “Public Interest Comment on the National Telecommunications and Information Administration (NTIA) AI Accountability Policy Request for Comment,” The Center for Growth and Opportunity, June 13, 2023, https://www.thecgo.org/research/public-interest-comment-on-the-national-telecommunications-and-information-administration-ntia-ai-accountability-policy-request-for-comment/.
The Role for Policymakers
Policymakers at all levels, including the NTIA, the Task Force, and its members, should use their platforms and public roles to highlight the tools and resources currently available to minors and their caregivers. Policymakers should not merely point out the problems and risks that minors face on social media, but also the benefits of social media, exemplifying best practices. A good example of an educational approach by a policymaker is from Governor Spencer Cox of Utah. Governor Cox recently launched a public awareness campaign about the risks and benefits of social media that includes links to resources.32“Unmask the Dangers of Social Media,” Utah Department of Health and Human Services, accessed November 3, 2023, https://socialharms.utah.gov/. The NTIA and Task Force could recommend federal agencies and policymakers to pursue a similar awareness campaign. Such an approach would maintain the benefits of social media and engage minors and caregivers as active participants in finding a solution while holding social media companies to account in the court of public opinion.
Social media offers many benefits and risks. A one-size-fits-all government solution is not possible given constitutional constraints, the current inconclusive state of the research connecting social media use to rising negative mental health trends, and the shifts in technology and user habits. Companies and policymakers must consider the many different family settings, mental health statuses, maturity levels, locations, and personal proclivities of minors whenever making changes to products or policies. The surgeon general is correct that action in this area is urgently needed. However, the solutions will require cultural adaptation, energy and investment from civil society and companies, and policymakers that target proven harm while preserving civil liberties and innovation. Such an approach is not a silver bullet, but it will ultimately make the internet safer for minors.