Reviews selected studies that answer a variety of research questions on the topic of social media, with a focus on meta-analyses and reviews of literature.
Question #1: Is self-esteem connected to social media use?
Do people use social media with the intention of comparing themselves to others? Does self-esteem increase or decrease as a result of social media exposure? “Humans are thought to possess a fundamental drive to compare themselves with others” (p. 206) and the researchers thought that social media exposure was the perfect mechanism to see this social comparison in action. Two types of social comparison include upward and downward. Upward comparison occurs when you compare yourself to someone perceived as better than you; downward comparison occurs when comparing yourself to someone who seems lesser in some respect. Intuitively, we feel better about ourselves when engaging in downward comparison but may suffer a hit on our self-esteem with upward comparison. Due to the reality that social networking sites “provide the perfect platform for meticulous self-presentation” (p. 207) via carefully curated content and selfies, users of social media are likely to experience upward comparison.
The researchers reported on two studies in this article: one correlational and the other experimental. Variables under study in the correlational study included self-esteem, Facebook usage, and upward or downward social comparison. Not surprisingly, increasing Facebook use was correlated with decreasing self-esteem and greater social comparisons, especially upward comparisons. In the experimental study, the researchers sought to identify a causal link between social comparison and self-esteem. Social comparison’s effect on self-esteem was measured by asking participants to read fictitious profiles on a social media platform that conveyed personal attributes about a user as well as the user’s popularity in the social network. In this study, self-esteem was negatively impacted by upward comparison with the fictional user’s profile as well as the user’s popularity, however, there was a stronger effect for popularity in the network. Thus, the likes, comments, and upvotes by the fictional user’s network proved more salient as a social comparison measure than the self-posted user content.
Takeaways: results from the first study were appropriately couched as correlational and therefore incapable of providing cause-effect proof. Results of the experimental study, though not surprising in light of what we already know about the human tendency toward social comparison, were significant. However, what is not clear is whether stronger results would have been obtained if participants viewed profiles of real people within their own network rather than those of fictional users. State self-esteem (which is fluid) was measured rather than trait self-esteem (stable): it would be interesting to see the implications for trait self-esteem after upward social comparisons. Further, a longitudinal look at changes in self-esteem over time, rather than in a static condition, could be instructive.
Citation: Vogel, E. A., Rose, J. P., Roberts, L. R., and Eckles, K. (2014). Social comparison, social media, and self-esteem. Psychology of Popular Media Culture, 3(4), 206-222.
Question #2: Can social media use decrease loneliness?
One of the identified benefits of social media use is the opportunity to expand social networks well beyond the bounds of face-to-face interactions. Thus, social media should decrease loneliness through these extensive, and sometimes worldwide, social networks. The researchers wanted to discover whether the SM platform, image-based vs. text-based, affects loneliness, happiness, and SWL (satisfaction with life). Facebook, Snapchat, Yik-yak, Instagram and Twitter were assessed in the study.
The study’s findings confirmed the hypotheses that image-based SM platforms, such as Snapchat and Instagram, are associated with lowered feelings of loneliness and increased feelings of happiness and satisfaction with life. Text-based platforms like Twitter did not achieve the same results. The researchers proposed that given the fact that pictures and videos of friends can convey the “presence” of that person (as if he or she is there) explain the benefit of image-based SM platforms for feeling connected. Text-based platforms, in spite of the immediacy of communication, lack intimacy “that is needed to more accurately replicate face-to-face conversations” (p. 164).
Takeaway: since the sample was drawn from an undergraduate student population (mean age = 22.55), and young adults are more likely to use SM than older adults, their attitudes regarding SM and mental health may skew more positive. It would be helpful to run the same study using middle-aged and older adults.
Citation: Pittman, M., & Reich, B. (2016). Social media and loneliness. Why an Instagram picture may be worth more than a thousand Twitter words. Computers in Human Behavior, 62, 155-167.
Question #3: Can people become clinically addicted to social media?
Paper A: Factors and outcomes associated with social media addiction
Technically, no. That’s because the American Psychiatric Association has not yet codified “social media addiction” within their Diagnostic and Statistical Manual of Mental Disorders (DSM). However, in the 2013 edition of the DSM, the APA included the first behavioral addiction (gambling disorder) and internet gaming disorder is currently under study and could be included in a future DSM edition. So, addiction to social media could potentially be added to the DSM at some point. Even though it is not yet a clinical diagnosis, certain components of addictive behavior can be observed in some people. These components include salience, tolerance, mood modification, relapse, withdrawal, conflict.
The authors wrote this paper to address sometimes contradictory findings from social media research and to accomplish two goals: to identify factors and outcomes associated with social media addiction. Their review of literature gathered studies published in English from 2008-2019 on the topic of social media addiction, with a final total of 132 studies reviewed.
Major findings from cited studies:
User characteristics that predispose to SM addiction (distal factors: indirect effect on behavior):
- Young
- Female
- More time spent, more frequent use, and more intense use of SM
- Subscribe to more than one SM account
- Use SM to create or maintain relationships and for entertainment
- Pre-existing anxiety, depression or ADHD and other psychological disorders
- Pre-existing addiction to other technologies, such as the internet and cell phones
- Accessing SM at night or on the weekends
- Big Five personality traits have inconsistently been associated with SM addiction
- Shyness and narcissistic tendencies
User characteristics that predispose to SM addiction (proximal factors: direct effect on behavior):
- Low self-esteem
- Use SM to gain recognition or gratification
- Escapism
- Lack of social support
- Low life satisfaction
Outcomes from SM addiction:
- Increased stress symptoms
- Lower sleep quality
- Lower performance in physical activities
- Increased depression and anxiety
- More likely to develop other technology addictions
- Lower satisfaction with and quality of social relationships
Features of some social media platforms, such as likes, updates, friend counts, and notifications, are associated with addiction.
Ahmed, E., & Vaghefi, I. (2021). Social media addiction: A systematic review through cognitive-behavior model of pathological use. Proceedings of the 54th Hawaii International Conference on System Sciences, 6681-6690.
Paper B: Prevalence of SM addiction and the role of culture
Since 2004, a growing body of research has focused on the problem of social media addiction. Various scales have been developed to measure this type of addiction, with the Bergen Social Media Addiction Scale (BSMAS) being the most prevalent. Studies using the scale have found SM addiction prevalence ranging from single digits up to 40% of respondents. The wide-ranging prevalence rates are due in part to differences in classification schemes. Of secondary interest to the researchers was to identify if SM addiction is more prevalent in collectivist vs. individualist cultures.
A total of 49 studies met criteria for inclusion in the meta-analysis, for a total of 63 independent samples and 34,798 respondents across 32 countries. Prevalence rates for individual studies ranged from 0% to 82%. The pooled prevalence rate for SM addiction across all studies was 24%. The prevalence was greater in Asia, Africa, and the Middle East, and lower in North America and Western/Northern Europe. SM addiction was more prevalent in collectivist cultures (31%) than individualist cultures (14%). Not surprisingly, cut-off scores (threshold for addiction) influenced classification. Using stricter thresholds lead to fewer respondents being identified as addicted. Further, people of younger ages were more likely to be classified as addicted to SM across these studies.
Takeaway: meta-analyses serve an important contributing role to understanding findings of diverse studies with varying numbers of subjects. However, meta-analyses are not without their own methodological problems, which must always be acknowledged. The findings of the current paper reveal some important facts that broaden our understanding of SM addiction and can inform future directions of research on this topic.
Citation: Cheng, C., Lau, Y-C, Chan, L, & Luk, J. W., (2021). Prevalence of social media addiction across 32 nations: Meta-analysis with subgroup analysis of classification schemes and cultural values. Addictive Behaviors, 117, 106845.
Question #4: Is social media harmful to mental health?
Paper A: Longitudinal Study
The researchers sought to overcome limitations of prior research that used cross-sectional between-subjects data to pinpoint harm done by adolescent social media use. Prior longitudinal study of this topic has yielded mixed results. The present study employed a longitudinal design to follow adolescents over an 8-year period, looking for within-subjects changes over time.
Time spent on SM sites increased from age 13 onward, with an average of 2 hours per day on SM by young adulthood. Girls spent significantly more time on SM sites than boys, and also reported more depression and anxiety symptoms.
Social networking was not predictive of future depression one year later for boys or girls. However, at age 13, girls’ SM use was correlated with depressive symptoms. Adolescents who rapidly increased their SM use also experienced a rapid increase in anxiety symptoms. In sum, the researchers found only a moderate association between social networking and depression and anxiety. There did not appear to be an association between increasing or decreasing SM use and a corresponding change in depressive or anxiety symptoms within individual participants.
Takeaways: time use on SM alone is not a good predictor of depression and anxiety. Increasing or decreasing time use of SM did not lead to a requisite change in emotional health. Depression and anxiety are caused by multiple factors and SM is only one of those factors. Comparing between-subjects and within-subjects results over an 8-year longitudinal study is quite noteworthy. However, data were collected using self-report measures which can lead to distortions.
Citation: Coyne, S. M., Rogers, A. A., Zurcher, J. D., Stockdale, L., & Booth, M. (2020). Does time spent using social media impact mental health? An eight-year longitudinal study. Computers in Human Behavior, 104, 106160.
Paper B: Meta-Analyses
With the accumulating body of research (the number of publications on SM doubled between 2013 and 2018) assessing SM’s impact on key psychological indicators, the authors sought to review meta-analyses of SM effects to date. Specifically, they reviewed meta-analyses of correlational studies on SM and its associations with three topical areas: well-being, academic achievement, and narcissism. What the authors discovered contradicts the hype that SM is destructive to adolescents.
“Seemingly conflicting findings published in the literature can be frequently attributed to sampling error, measurement error, or other biasing influences that compromise empirical studies. Such conflicting findings are prevalent in studies of SM” (p. 61). Through meta-analyses, artifacts of individual studies can be corrected and problems of insufficient power can be addressed.
Four meta-analyses of SM and well-being were reviewed. “Well-being” was measured using a collection of variables such as self-esteem, loneliness, thin body ideal, and life satisfaction. Findings revealed a negative correlation between these well-being indicators and SM, however, effect sizes were small. SM use accounted for only 1% of the variance in well-being. Thus, these meta-analyses found that 99% of the variance in well-being and SM could be accounted for by something other than SM use. A stronger correlation was obtained between SM use and a thin body ideal. A positive correlation was found for SM use and self-esteem, based on the size of one’s social network.
Three meta-analytic studies were reviewed for SM and academic achievement. All meta-analyses found negative correlations between SM and achievement, and SM accounted for less than 1% of the variance in achievement. Further, when objective measures of achievement (such as grades reported by the school) were used, there was no significant relationship between SM and achievement.
On the topic of narcissism and SM use, three meta-analyses were reviewed. Effect sizes achieved were greatest for grandiose narcissism and SM use, which is unsurprising given the components of SM that enable a narcissist to curate content shared about the self. However, there were still only small to moderate correlations.
In sum, there was a weak link between SM and well-being indicators (four meta-analyses) and academic achievement (three meta-analyses). A stronger link exists between SM use and narcissism, but still with only small to moderate associations across three meta-analyses.
Takeaway: although meta-analyses have their own limitations and potential biases, they represent an important step toward understanding the true relationships between SM and key variables. SM is one of the “usual suspects” credited with downward trends in adolescent well-being. However, cherry-picking data and over-emphasizing significant yet weak correlational evidence for SM harm will divert us from identifying the true culprits in adolescents’ lives.
Citation: Appel, M., Marker, C., & Gnambs, T. (2020). Are social media ruining our lives? A review of meta-analytic evidence. Review of General Psychology, 24(1), 60-74.
Question #5: Do narcissistic people use social media more and differently than non-narcissists?
The researchers undertook this meta-analysis due to conflicting findings regarding the use of social media by people classified as narcissistic. Could these conflicting findings be due to the presence of moderator variables? They chose to include the following moderator variables: birth cohort (generation), culture, and SM platform. The studies chosen for this analysis classified narcissism as either grandiose or vulnerable (click here to read my white paper about these types).
The first variable of interest was social media use, measured by time spent on SM, frequency of status updates, number of friends, and number of selfies. Birth cohort was identified by age of participant, culture was identified by country or region.
The meta-analysis included 62 samples from 29 studies (both published and unpublished) of over 12,000 participants. All studies were cross-sectional. The average ages for samples ranged from 14-35. The majority of studies had samples from Western countries (67%), with the remaining coming from European, Asian and other countries. Social media platform was primarily Facebook (65%), and a few samples each used Twitter and Instagram. Over three-quarters of studies used self-report data, and slightly more than half used data from undergraduate students with the remaining studies using M-Turk and high-school students.
As expected, grandiose narcissists spent more time on SM, made more status updates, had more friends/followers, and posted more selfies than vulnerable narcissists or non-narcissists, although effect sizes were small. For grandiose narcissists, time spent on SM was moderated by generation (Millennial), platform (Facebook), and culture (U.S.). The relationship between narcissism and status updates and number of friends/followers was significant among Russian samples yet Asian and European samples were not significantly different than U.S. samples.
Takeaway: this meta-analysis confirmed what has been proven in many studies: grandiose narcissists behave differently on SM platforms than do non-narcissists or vulnerable narcissists. Moderator variables added more complexity to the analyses and raised new questions, especially regarding cultural differences, that need further investigation. We learned little about vulnerable narcissists except that their SM behavior is more similar to that of non-narcissists than of grandiose narcissists. Given the number of studies to date focusing on Facebook and narcissism (22), more research is needed about narcissism and other SM platforms.
Citation: McCain, J. L., & Campbell, W. K., (2018). Narcissism and social media use: A meta-analysis. Psychology of Popular Media Culture, 7(3), 308-327.
Question #6: Does SM encourage risky behavior in adolescents?
According to the authors, SM is ubiquitous among adolescents, “with recent reports indicating that 93-97% of 13-17-year-olds use at least one social media platform” (p. 259). With adolescents having access to social media sites on computers, tablets and smartphones and, therefore, outside of their parents’ purview, SM can be a vehicle to foment risk-taking behavior among this population. The authors cited two previous meta-analyses of risky behavior linked to SM use. One study found “a small, positive association between online exposure to sexually explicit materials and engagement in risky sexual behavior” (p. 260). The other study found a medium effect size for social media engagement and increased alcohol use. The focus of the present meta-analysis was to explore findings of published and unpublished studies regarding the relationship between SM use and risky behavior in general among adolescents. They expected to find a small to medium effect size or association between the variables.
The analyses included 27 studies assessing risky behavior, 14 studies examining substance use, and 14 studies focusing on risky sexual behavior. These studies were published between 2011 and 2018, and included participants between 12.6 and 18 years of age. As predicted, small-to-medium effect sizes were found linking SM use to each of the three types of risky behavior. After accounting for publication bias, the final r value was slightly higher for each type. The average age of participants in the sample proved an important moderator between SM use and risky sexual behavior, with a higher correlation between the two variables in samples with a lower average age.
The modeling effect for learning risky behavior from peers applies to using SM. “Exposure to social media content related to positive portrayal of risky behaviors contributes to the favorable attitudes toward risky behaviors as being normative and socially desirable” (p. 269). Adolescents who look to improve their digital social status (getting more likes and followers) may view risky behavior as a means to achieve this goal.
Takeaway: The effect sizes obtained after including unpublished studies revealed a healthy though moderate link between risky behaviors and SM use for adolescents. We can confidently say that for some adolescents, SM provides a platform that encourages risky behavior that may ultimately result in harm to themselves or others. Adolescent risk-taking is nothing new, but SM is a new vehicle to promote this behavior.
Citation: Vannucci, A., Simpson, E. G., Gagnon, S., & Ohannessian, C. M. (2020). Social media use and risky behavior in adolescents: A meta-analysis. Journal of Adolescence, 79, 258-274.
Question #7: Do men and women become addicted to different types of online platforms?
Prior studies of internet addiction (IA) have yielded conflicting findings regarding the prevalence of IA among men and women. IA is a broad label and many studies have attempted to pinpoint addictive behaviors to specific aspects of internet use (such as cybersex, gambling, or internet gaming).
Based on findings of previous studies about gender differences in internet addictions, the researchers focused their meta-analysis on internet gaming disorder (IGD) and social media addiction (SMA). They hypothesized that males will have higher rates of IGD and females will have higher rates of SMA.
A total of 49 studies with 53 samples (82,440 participants from 21 countries) were analyzed for IGD and 38 studies with 41 samples (58,336 participants from 22 countries) were included in the SMA meta-analysis.
The moderator variable of world region was included in analyses of IGD and SMA. Most countries followed the pattern of higher IGD among males and higher SMA among females, with a few exceptions. In India, an inverse pattern of addiction was found. In Germany and Taiwan, males had higher rates of both IGD and SMA.
Results indicated that males are more likely to experience IGD than females, with a moderate effect size (.462). As hypothesized, females experience higher rates of SMA but with a smaller effect size (-.202). The authors explained these findings by stating that “data suggest that gender-related differences in IGD and SMA likely involve biological, psychological, and social factors” (p. 15). Gender-based addictions may develop using different vehicles of internet use as “females may use social media to fill voids when depressed and social needs are not met” whereas gaming cues may “elicit higher cravings in males” and “males may experience feelings of success and achievement from engaging in online games” (p. 14).
Takeaway: although internet addictions are not presently recognized as diagnosable mental disorders in the DSM-5, internet gaming disorder is currently under consideration for a future edition. Gambling disorder was the first behavioral addiction ever included in the DSM (2013 edition). Although only one behavioral addiction is codified in the DSM at this time, practitioners continue to treat a number of behavioral addictions that lead to real and potentially negative consequences for patients. This meta-analysis has helped to identify the unique types of internet addictions experienced by men and women, which may (hopefully) positively impact future educational efforts and interventions to curb this growing trend.
Citation: Su, W., Han, X., Yu, H., Wu, Y., & Potenza, M. N. (2020). Do men become addicted to internet gaming and women to social media? A meta-analysis examining gender-related differences in specific internet addiction. Computers in Human Behavior, 113, 108680. doi:10.1016/j.chb.2020.106480.
Reviews authored by Jennie Dilworth, Ph.D
All rights reserved.