Well-Being Across Social Media Platforms

We examine the well-being, depression, anxiety, and stress of users of different social media platforms.

New technologies have the power to revolutionize society, and any technology widely adopted by most people is likely to affect them in numerous ways that may be good and/or bad. Therefore, it’s not surprising that people are concerned about whether social media affects people’s well-being, and if it is especially risky for certain vulnerable subgroups of people, like our youth, as described in a recent U.S. Surgeon General Advisory.

In today’s post, I examine how mental health and well-being vary across users of different social media and communication services. Given that the Neely Center Social Media Index survey is descriptive and not experimental, we cannot make any claims about whether social media use causes improved or worsened well-being. Different kinds of users choose to use each of these platforms and resulting differences in well-being may reflect differences in those user bases.  Nevertheless, we can examine whether the use of different social media and communication services are related, on average, to measures of well-being.

Measuring Well-being

Well-being is a multi-faceted concept that cannot adequately be assessed with one single measurement or question. Generally, psychologists define well-being as a construct consisting of relatively more positive than negative affect, and a positive subjective assessment of how well a respondent’s life is going at some point in time. There are many academic debates about the actual components that comprise well-being, but that is beyond the scope of this report. Instead, we included questions from three psychometrically-validated and widely-used surveys that assess Satisfaction with Life, Comprehensive Thriving, and Depression, Anxiety, and Stress. Each of these relate to happiness, health, and other variables that should be related to how emotionally and mentally well or unwell people are.

Satisfaction with Life

The Satisfaction with Life scale is among the most widely validated measures in the well-being space, and correlates with better mental, physical, and social health outcomes. Another benefit of using the Satisfaction with Life scale is that it has been included in many international and nationally-representative surveys, including Gallup’s tracking survey, which allows for comparing well-being in the current sample to those assessed at different time points and in different countries. The scale asks respondents to indicate their level of agreement or disagreement with 5 statements about their global satisfaction with life, and 7 statements about their satisfaction with specific domains of their lives.

Before seeing how these life satisfaction differs across social media and communication services, let’s examine what life satisfaction looks like for our nationally representative sample of US adults. American adults are substantially more satisfied than dissatisfied with their lives, both in our current sample and in a related survey conducted by Gallup in January 2023.  For the purposes of this graph, neutral answers (i.e., those where respondents neither agree nor disagree with each statement) were removed, so scores will not add up to 100%.

Before we examine how satisfied US adults who use each social media and communication service are, I want to note that responses to each of these life satisfaction questions are highly correlated (Spearman’s ⍴ = .61). What this means is that when respondents say that they are satisfied with their mental health, they are also likely to say that they are satisfied with their physical health, their family lives, and each of the other domains listed above. I am including the plots for all of the questions because there may be slight differences across services, but the overall patterns should be quite similar for all of the services.

The following plots display how US adults who use each service compare to all US adults. Negative scores are represented in yellow and indicate that a service’s users are more dissatisfied than the overall US adult population. Positive scores are represented in green and indicate that a service’s users are more satisfied than the overall US adult population. Most services’ user bases do not significantly differ from the general US adult population, per common definitions of statistical significance, but I call attention to those that do differ by these criteria. In some cases, there may be numerical differences, but due to reasons related to sample size, platform use, and measurement error, these may not be statistically significant. Therefore, in the subsequent descriptions, I highlight the statistically significant and nearly statistically significant differences, but omit discussion of numerical differences that are unlikely to be meaningful.

The only services with user bases that significantly differed from the national average in agreement with the statement “I am satisfied with my life” were LinkedIn, Facetime, and Email. The user average satisfaction with life on LinkedIn, Facetime, and Email were more satisfied than the general US population. Numerically, TikTok, Discord, and Snapchat have slightly greater dissatisfaction than the US population, but these differences are within the statistical margin of error.

When looking at the statement “I am satisfied with my mental health,” Discord, Snapchat, and TikTok users were significantly more dissatisfied than the general US adult population. In contrast, LinkedIn users were more satisfied than the general US adult population. No other services’ user bases differed from the national US adult average.

When looking at the statement “I am satisfied with my financial situation,” Discord, Snapchat, TikTok, and Instagram users were significantly more dissatisfied than the general US adult population. No other services’ user bases differed significantly from the national US adult average.

When looking at the statement “I am satisfied with my job or other daily activities,” TikTok users were significantly more dissatisfied than the general US adult population. No other services’ user bases differed significantly from the national US adult average.

When looking at the statement “I am satisfied with my physical health,” on average, Discord and TikTok users were more dissatisfied than the general US adult population. No other services’ user bases differed from the national US adult average.

When looking at the statement “I am satisfied with my social life,” on average, Discord and TikTok users were significantly more dissatisfied than the general US adult population; LinkedIn and Facetime users were significantly more satisfied than the general US adult population. No other services’ user bases differed from the national US adult average.

When looking at the statement “I am satisfied with the amount of leisure time I have,” Snapchat and TikTok users were significantly more dissatisfied than the general US adult population. No other services’ user bases differed significantly from the national US adult average.

When looking at the statement “I am satisfied with my family life,” TikTok users were significantly more dissatisfied than the general US adult population. No other services’ user bases differed significantly from the national US adult average.

When looking at the statement “In most ways, my life is close to my ideal,” TikTok users were more significantly dissatisfied than the general US adult population. No other services’ user bases differed significantly from the national US adult average.

Comprehensive Thriving

Respondents also indicated their agreement or disagreement with some questions from the Comprehensive and Brief Inventories of Thriving. These scales were designed to take the Satisfaction With Life scale’s subjective assessment of how a respondent’s life is going, and then also assess the relative positive and negative affect they experience. These measures predict important health outcomes (including doctor’s visits, heart disease, hospitalization, and longevity).

When looking at the statement “My life is going well,” only LinkedIn’s user base differed from the general US adult population. Specifically, LinkedIn users were more likely to say their lives were going well.

When looking at positive affect, no platforms’ users differed significantly from the general US adult population. LinkedIn users were marginally more likely to say they felt happier most of the time than people in the general population were, but this difference was not statistically significant.

When looking at negative affect, Online Gaming, Snapchat, and TikTok users felt significantly more negative than did the general US adult population. On the other side, though, LinkedIn users felt significantly less negative than did the general US adult population.

Another way to evaluate how these different social media and communication services compare to each other is by ranking them relative to each other on each of the dimensions of satisfaction with life and thriving described above. In the heat map below, you’ll see that the user bases of LinkedIn, along with the various direct messaging apps (i.e., Facetime, Email, WhatsApp, and Text Messaging), generally have higher levels of well-being than the user bases of some of the broadcast- and short-form video-oriented platforms like Snapchat and TikTok.

While these are some of the more popular tools to assess well-being, they are not the only ways. Another method involves asking people about specific experiences and symptoms they may have. We look at this in the next section.

Depression, Anxiety, and Stress Symptoms

Well-being, as a construct, tends to be focused on the positive elements of people’s subjective experiences. That conceptualization correlates negatively with the experience of mental illness symptoms. One approach clinicians use in assessing patients’ mental health is by asking them about how often they experience symptoms related to different mental illnesses. The Depression Anxiety Stress Scale is one such general use instrument used in mental health assessment. These symptoms may be thought of as acute and more specific negative experiences, beyond the more holistic “negative affect” captured in the satisfaction with life and comprehensive inventory of thriving scales reviewed in the previous section. 

We included a shortened form of this scale in our survey, where we asked respondents to indicate how often they experienced 6 different symptoms in the previous week. For simplicity’s sake, I computed the percentage of people who did not report experiencing the symptom, and the percentage of people who reported experiencing the symptoms sometimes, often, or almost always in the past week.

Before diving into comparing symptoms user bases of different social media and communication services report, it’s useful to assess the symptoms reported by the general US population, which, as reported above, mostly feels satisfied with their lives. Although they are generally satisfied, we find that the majority of US adults also report feeling agitated and down-hearted, and that it was difficult for them to wind down and work up the initiative to do things at least some of the time in the past week. 

Next, let’s compare how symptom frequency varies across the user bases of the different social media and communication services. No service’s user base reported experiencing agitation significantly more or less frequently in the past week than the general US adult population. Numerically, Pinterest, Twitter/X, and Reddit users reported marginally more frequent agitation than the general population in the previous week.

When examining feeling “down-hearted and blue”, we see that 9 apps and services did have user bases reporting feeling significantly more down-hearted and blue in the past week than the national average. Online gaming, Reddit, and Discord users reported the most frequent rates of feeling down-hearted and blue, but TikTok, Twitter/X, Pinterest, Instagram, and Snapchat users all reported rates of feeling down-hearted and blue greater than the national average. No app or service’s users reported significantly less frequent feelings of being down-hearted and blue.

When examining feeling worried about situations in which the respondent might panic and make a fool of him/her/themself, we see that 8 apps and services did have user bases reporting significantly more worry in the past week than the national average. Snapchat, Reddit, Online gaming, TikTok, Twitter/X, Instagram, and Facetime users reported the most frequent rates of fearing panicking and making a fool of themselves relative to the national average. Additionally, these services’ users reported this symptom significantly more than text message, email, and WhatsApp users. The comparison to email and text messaging as relatively ubiquitous services is helpful in disentangling differences that may be more specific to engaging in online platforms.  Again, no app or service’s users reported significantly less frequent feelings of fear of panicking and making a fool of themselves.

When examining feeling scared without a good reason, we see that 6 apps and services did have user bases reporting significantly more fear in the past week than the national average. TikTok, Online gaming, Snapchat, Reddit, Pinterest, and Instagram users reported the most frequent rates of feeling scared relative to the national average. Discord users were marginally more fearful than the national average, but with the relatively smaller user base, this difference within the margin of error. Again, no app or service’s users reported significantly less frequent feelings of being scared.

When examining feeling it was difficult to wind down, we see that 5 apps and services did have user bases reporting significantly more difficulty in winding down in the past week than the national average. Discord, Online gaming, LinkedIn, TikTok, and Snapchat users reported the most frequent rates feeling it was difficult to wind down relative to the national average. WhatsApp and Reddit users also reported elevated rates of difficulty in winding down, but the difference between their rates and the national average was just slightly less than the margin of error. Again, no app or service’s users reported significantly less frequent feelings of it being hard to wind down.

When examining feeling it was difficult to work up the initiative to do things, we see that 4 apps and services did have user bases reporting significantly more difficulty in the past week than the national average. Discord, Reddit, Online gaming, and LinkedIn users reported the most frequent rates feeling it was difficult to work up the initiative to do things relative to the national average. Again, no app or service’s users reported significantly less frequent feelings of it being difficult to work up the initiative to do things.

Another way to evaluate how these different social media and communication services compare to each other is by ranking them relative to each other on each of the depression, anxiety, and stress symptoms described above. For consistency’s sake, I’ve reverse-scored symptom frequency so that less frequent symptoms means higher ranks and more frequent symptoms means lower ranks. In the heat map below, you’ll see that the user bases of email, text messaging, Facetime, Facebook, and YouTube tend to report fewer symptoms than Discord, Reddit, Online gaming, TikTok, and Snapchat users.

Correlates of Well-being

Across the various measures of well-being examined above, there is a general tendency for social media and communication apps that are used relatively more by younger people and by women to have lower average user well-being. To see whether certain individual demographic characteristics that other research has shown to be related to well-being, we can examine those correlations in our sample, and then see whether these demographic characteristics do indeed correlate with platform usage.

In the table below, each row contains each of the questions described in the analyses earlier in this report, each column represents demographic variables, and then the table’s cells contain the correlation between the column’s demographic variable and responses to the question displayed in each row. If the correlation value in each cell is larger than +/- 0.063, then it is “statistically significant.”

Replicating past research, age is positively correlated with well-being and negatively correlated with depression, anxiety, and stress symptoms. As found in past research, women report slightly lower satisfaction with life and slightly more depression, anxiety, and stress symptoms. Higher education and political conservatism each also show slight positive correlations with well-being and slight negative correlations with depression, anxiety, and stress symptoms.

Now that we’ve established that these demographic variables correlate with our well-being measures, let’s examine how platform users differ in these characteristics. In the table below, we see that the platforms with the youngest average users and with relatively more women, like TikTok, Snapchat, and Pinterest, also tend to be the ones with the worst well-being outcomes described earlier in this report. On the other hand, platforms with the most educated users, like LinkedIn, tend to be the ones with the best well-being outcomes.

These correlations may suggest that some social media and communication apps could drive changes in well-being for different users, but they may also suggest that some apps appeal more to some people who already have higher or lower well-being. It is also possible that well-being, demographics, and platform use are not directly correlated with each other, but instead with some other extraneous or confounding variable. Future analyses will dig into this, particularly as more longitudinal data is collected.

Summary 

While there are many debates about whether and/or how social media and technology are affecting people’s well-being, our data can only speak to how the average well-being of users on the different apps and services differ from each other, and from the national estimates. What we see here is that social media and communication apps that emphasize broadcast short-form video and have larger proportions of younger users, like TikTok, Snapchat, and Instagram, also tend to have the worst outcomes when it comes to various measures of well-being. On the other hand, social media and communication apps that emphasize direct personal connections, like Facetime, Email, or Text Messaging, or that have more educated user bases, like LinkedIn, tend to have the best well-being outcomes. Again, these analyses should not be interpreted as anything causing anything else; these are descriptive and correlational data that imply these variables are related in some way, but do not permit inferring any causality.

Additionally, this research should be viewed as the starting point, and not the totality of what we will be able to conclude as we collect more data and do more longitudinal analyses. These estimates provide us an overall baseline of well-being for users of different social media and communication services, and as we track these users over time, we will learn how usage of different platforms may correspond to increases or decreases in well-being. As more longitudinal data are collected, we will also be able to examine whether certain subgroups of users’ are more likely to have their well-being change in relation to changes in usage of these different services or their self-reported experiences with them. The current analyses represent an initial view of well-being across social media and communication services, and we look forward to sharing more deeper analyses over time.