Well-being: Problematic Use

Raw text

Transcribed text as recorded, including headers and formatting fragments.

Exhibit 40 Well-being: Problematic Use I am continuing to share updates on our 4 priority well being topics. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210347 We aligned to focus on these core well-being topics Rationale for selection Highest-rated mission-driven critique among stakeholders with lower cross-functional investment at the company. -- Bullying & harassment Highest-rated mission-driven praise among stakeholders with lower cross-functional investment at the company. -- Connections This is a big bet topic. While emotional topics did not rank highly in our data (besides amongst academics) there is increasing interest from leadership and impact related questions have come up at recent senate hearings. -- Emotional impact This is a big bet topic. While this topic did not rank highly in our data (besides amongst the media), this is an area where many stakeholders' opinions are loosely held, and thus has a greater chance of being influenced by media cycles or events. This topic has come up recently in popular media (e.g., The Social Dilemma) as well as senate hearings. -- Problematic use Limited work on suicide and self-injury also taking place [gum] Last time I gave an update on Bullying and Harassment and also published a one pager. And I'll cover Problematic use today. We had picked this as a big bet topic. Over the past year or so we've seen this topic gain traction. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-IG-AG-00210348 The prevailing critiques focus on how social media companies deliberately manipulate people by designing "addictive" products to trick them into "mindlessly scrolling" Political Hearings - Focus on manipulation: "So, Mr. Zuckerberg and Mr. Dorsey, you profit from your company's hooking users to your platforms by capitalizing on their time. So yes or no, do you agree that you make money off of creating an addiction to your platforms ...?" (Rep Johnson, March 25, 2021 of the House Committee on Energy & Commerce) Popular media - focus on designing addictive products: "A lot of people think that Google is just a search box and Facebook is just a place to see what my friends are doing. What [they] don't realize is that there are entire teams of engineers whose job it is to use your psychology against you." (The Social Dilemma, Netflix 2020) We often hear/read how Facebook INTENTIONALLY creates products to be addictive -- often to increase profit for the company -- or how Facebook manipulates people by leveraging human psychology against users (aka Social Dilemma) CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210349 "Addiction" vs. "Problematic Use" "Social media addiction" is a colloquial term, not a formal clinical diagnosis. We and other experts use the term "Problematic Use" as a more accurate term to describe habitual or undesired use of Facebook. Definitions vary widely, but internally we define problematic use as a lack of control over social media use that leads to negative life outcomes. [Talking points] The first thing I want everyone to remember is that "Social media addiction" is a colloquial term. To date no formal clinical diagnosis relating to "social media addiction" [exists] in either the Diagnostic and Statistical Manual of Mental Disorders or the International Classification of Disease, so there is no psychometrically validated tool to detect or measure it. Rather, we -- and a number of external researchers -- use the term Problematic Use which is defined as "a lack of control over social media use that leads to negative life outcomes, examples include disruptions in productivity or sleep, or suffering social relationships." Of all the well-being or societal impact related topics, this is probably one of the hardest to study. And I will walk you through those -- the first is the wide range of definitions of problematic use that can make the research on this topic difficult to parse. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210350 The lack of an agreed-upon definition of problematic use results in varied estimates. Internal research suggests it affects ~3% of US Facebook MAP. This increases to over 10% if the definition is broadened to include feelings of guilt about Facebook use (see figure). External estimates are even more wide-ranging, from ~2.8% to as high as 48%, depending on definition (not shown). Irrespective of the definition, there is substantial cross-country heterogeneity in estimates. Internal research using the definition I just mentioned, finds that approximately 3% of US Facebook MAP report problematic use. However, this estimate increases to 10% when the definition is expanded to capture feelings of guilt around time spent on Facebook. External estimates of problematic social media use are even more wide-ranging, from ~2.8% to as high as 48%, reflecting the variety of ways that problematic use is defined and measured across the academic and practitioner literatures. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG—AG-00210351 A wide range of factors drive the impact of social media on someone's life. Research rarely captures this. As with a lot of topics, research on problematic use has been yet unable to make causal inferences. It remains unclear if problematic use is caused by social media or if it's a symptom of other psycho-social problems. For instance, we know that certain personality characteristics, such as narcissism and neuroticism, are associated with problematic use (Gerson et al, 2016; Atroszko, 2018). Even robust studies that use experimental techniques -- including a forthcoming paper from Stanford, Harvard, and NYU researchers -- rarely overcome expectation effects and sampling challenges. Another limitation -- which isn't unique to this area -- is causality. But what is unique here is because problematic use rests on negative life outcomes, it's difficult to know if social media is causing those outcomes, or if those outcomes drive people to use social media. For example, does social media disturb people's sleep? Or do people use social media when they can't sleep. A number of external experts -- including Amy Orben, a fellow researcher in Cambridge who is quite well known in this field -- calls out that the relationship is probably bi-directional and complex. Even some recent studies by very well known researchers (Gentzkow and Alcott, who also extensively study polarization) -- that models self-control and habit formation on social media via an experiment -- but even the researchers themselves note limitations related to sampling and contamination effects. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-IG-AG-00210352 Time spent is a common -- but poor -- proxy for identifying problematic use. Self-reported vs. actual time spent on Facebook. External data relies on stated time spent. This is flawed for a number of reasons: There is no clear threshold past which time on social media becomes "problematic". Many external studies rely on self-reported time spent, which is notoriously unreliable (see figure). Evidence from internal research suggests that there are more reliable behavioral proxies for identifying problematic use: "Passive" consumption, Frequent low-engagement sessions, Disproportionate night-time usage, Repetitive app checking, Receiving and responding to more push notifications. The other limitation of external research is that they rely on stated time spent. This is flawed for a number of reasons: There is no clear threshold past which use becomes "problematic". For example, people who spend "excessive time" on social media may find that time consistently beneficial, such as those who use it for job searches, small business promotion, or connecting to friends and family. In other words, the focus should be on quality or types of interactions over quantity. Many external studies rely on self-reported time spent, which is notoriously unreliable. This has been shown in both our internal AND by external researchers. This is particularly problematic because there is research showing that measures of well-being are often associated with mis-reporting of time spent. Rather, our internal research suggests that certain browsing patterns are more reliable behavioral proxies for identifying problematic use, such as the ones listed here: "Passive" consumption, Frequent low-engagement sessions, Disproportionate night-time usage, Repetitive app checking, Receiving and responding to more push notifications. All this is to say that it is IMPERATIVE that we not take every published finding at face value because there are several limitations. And that a lot of what is written in the media are loosely held opinions. That said, we DO know that some people are at higher risk of problematic use. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG—00210353 Problematic use is more common among teens/young adults, but isn't necessarily a "teenage girl" issue. Problematic use IS more common among teens and people in their 20s. This is consistent with younger people having problems with self-regulation. People commonly believe that women and girls report more problematic use of social media. Findings are mixed across external studies, however internal research leveraging surveys and log data finds that men are 1.3x as likely as women, on average, to report feelings of problematic use of Facebook. Both internal and external research DOES find that problematic use to be more common among teens and younger people. But this alone isn't surprising -- it's highly consistent with well-documented research on how the adolescent brain self-regulates. (when I publish my one-pager, I will link to this. It's very interesting) There are some loosely held opinions that women and girls are more likely to report problematic use. Findings in external studies are inconsistent on this point. Some report higher levels of problematic use among women, others find higher levels among men. These studies often leverage self-reported data. However, our internal research -- which leverages log data -- we find that men are more likely than women to report feelings of problematic use. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210354 Appendix CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210355 A wide range of factors drive the impact of social media on someone's life. While the causal direction remains unclear, we know some people are at higher risk. Overall, it remains unclear if problematic use is caused by social media or if it's a symptom of other psycho-social problems. Personality traits such as introversion, narcissism, and high levels of neuroticism are highly correlated with higher risk of problematic use (Gerson et al, 2016; Atroszko, 2018). Social insecurities and low self-esteem, loneliness, fear of missing out, and fear of negative social evaluation (Jung et al, 2017; Baker et al, 2019) are also associated with higher risk. Major life events in the past two months were also found to be associated with feelings of problematic use (Baker and Chen, 2018). [Talking points] Impact of social media is confounded with many other factors (e.g. impacts of smartphones, internet use, gaming, shopping) and varies from one person to the other based on personality traits, intention of use, major life events, and preconceived perceptions about social media. The direction between social media use and negative life impacts is inconclusive (e.g. are lonely people using social media more or is social media using leading to loneliness?) Most research points to associations between time spent using social media and negative life impacts rather than causation. We know there is a percent of our users who are more vulnerable but we do not know whether products designed to help them are effective in supporting them gain more control over their use or if they are even aware of these products. CONFIDENTIAL - PROTECTED BY CONFIDENTIALITY AGREEMENT MT-lG-AG-00210356