The findings of this study indicate an association between high frequency of SNS use and a large network size on the one side and poor mental health on the other side, among Swedish young female adults. Furthermore, it shows that this association is modified by the level of PES. Women who used SNS on an almost hourly basis had increased odds of experiencing poor mental health than when using the platforms less frequently, as well as when having more than 600 contacts on SNS compared to fewer. Women with high PES seemed largely protected from the detrimental effects of having more than 600 SNS contacts.
Despite the widespread knowledge that women, especially young women (38), are more frequently burdened by depression than men (39), it seems somewhat surprising that very few studies have specifically looked into the difference of impact between genders in SNS use and mental health (2, 3).
4.1 Early Age Gender Difference
In a recent, longitudinal UK study, girls and boys aged 10–15 showed decreasing happiness and increasing active SNS usage (40). Additionally, socio-emotional difficulties (emotional, peer-relationship and conduct problems) rose for girls but declined for boys. The increased use of SNS at age 10 was associated with decreased happiness and increased socio-emotional difficulties in later years thereafter (up to age 15), in girls only. This study showed how at a very young age, use of SNS can affect girls’ well-being in later adolescent years (40). It is thought that since the adolescent years shape one’s future physical and mental health to a large extent (41), many adolescents experiencing mental health issues during this crucial developmental period will go on to experience such issues in later years, especially in their 20’s (42). Hence, a decline in well-being due to SNS use occurring from an early age onwards, could escalate and result in mental health issues in later young adult years. However, it may be, that this trend reverses beyond the young adult years, as women aged 27 – 34 in this study were slightly less likely to indicate poor mental health (Table 2).
Another study of an adolescent population with a mean age of 15 years found that passive Facebook use increased depressed mood in girls only, whereas active, public (posting content on their profiles) Facebook use was associated with depressed mood in boys. However, active public and private use in girls yielded positive outcomes when they perceived online social support (43). Together, these studies suggest that SNS use affects children already at a young age, that these effects vary by gender, and by the way in which the sites are used.
4.2 Problematic Facebook Use & a Gender Perspective
A meta-analysis on Problematic Facebook Use (PFU), or Facebook addiction, suggested that females were more prone to demonstrate PFU behaviour than males (44). One definition of PFU is such Facebook usage that disrupts everyday life at school, work or with relationships, by causing distress in cognitive functioning and/or well-being (44). In a meta-analysis of 23 studies with adolescent and young adult populations, PFU was positively associated with depression, anxiety and psychological distress in general (45). Marino et al. (2018) found that time spent on Facebook correlated with PFU (44). Females in the present study who reported using SNS almost hourly could in fact be described as exercising PFU behaviour. Of course, frequency of SNS use does not completely equate to time spent on Facebook, but the two variables go hand in hand. PFU individuals tended to have larger friendship networks compared to non-PFU individuals, and females with PFU also tended to send more friend requests and private messages (46). It has been repeatedly shown that women use SNS for communicating, maintaining friendships and accessing social information, whereas men tend to use the platforms more for gaining information and playing games (47-49). A deeper understanding as to why women use SNS more and are potentially also detrimentally affected by them could be provided by a gender perspective (50). Essentially, gender roles are transposing onto the use of SNS, so that women are more attracted to use SNS for social connectivity than men (43, 47). For example, the behaviour to compare oneself on SNS is more common among women than men (51). Indeed, increased Facebook use has been linked to increased social comparison, fuelling envy, and resulting in depression, which is partially grounded in causal evidence (22). A difference between genders has also been detected, in that adolescent females who tended to compare themselves on SNS had worse depression outcomes (52). Interestingly, when the number of SNS contacts was analysed in this study as a continuous variable, there was a statistically significant association between each additional contact and being a case for poor mental health in the female population, with an incremental chance of 0.1%, OR 1.001 (CI = 1.00-1.001). Though a small value, it is nonetheless meaningful, considering the number of contacts are as many as several hundred for many SNS users. Also, females have been shown to emotionally respond worse to negative images or messages, which may provide further explanation for the poorer health outcomes (50).
Our finding regarding the gender difference in the association between number of SNS contacts and mental health appears to be the first such finding reported in the scientific literature.
4.3 No Gender Differences
Some studies do not show differences in gender, such as Kross et al. (2013) (10), where declines in subjective well-being, associated with frequent Facebook use, were not moderated by gender. Gender was also accounted for by Verduyn et al. (2015) (14), however it did not moderate the relationships (14). Although SNS use was associated with body image and eating disorders, this meta-analytic association did not differ between men and women in studies that examined gender (53).
It may be that a true association between SNS use and mental health has been clouded by combining gender in analyses, contributing to the mixed findings in the literature. Although most studies controlled for gender, a deeper analysis as in the above studies (sections 4.1 and 4.3) has been the exception.
4.4 PES & SNS Use
Women are, in the absence of emotional support, significantly more prone to depression than men (54). This was corroborated by the findings in this study. Similarly, in another Swedish study of late adolescents, women benefitted more than men in their psychological well-being from high-quality, trusting friendships (55). The findings in our study showed that high PES protected very frequent SNS users from a negative impact on their mental health. However, the current findings may be an artefact in that persons with low PES inherently use SNS more than individuals with high PES, as a meta-analytic study showed low PES led to loneliness, which subsequently led to increased Facebook use (16). A further study, which aimed to investigate PFU, performed a 3-way interaction analysis between PFU, neuroticism and well-being, which differed according to gender. Women high in PFU and neuroticism were at 17 times higher odds of having low mood compared to men low in neuroticism and PFU (56). It has been shown that neuroticism and PES are related, in that females high in neuroticism also perceive lower emotional support (57), suggesting a similar finding to ours regarding the interaction between frequency of SNS use and low PES.
Most PES inquiries are concerned with online-derived support rather than offline ditto (2). For example, one such inquiry found that larger Facebook networks were associated with increased perceived online social support as well as life satisfaction (21). Further analysis in the above-mentioned study found a negative curvilinear relationship between the number of SNS contacts and perceived social support (58). This meant that from a certain number and up to a certain number of Facebook friends, these provided perceived social support, but beyond these points, no or too little support was derived; the authors suggest this may be due to the time and effort devoted to friends becoming too much or too little, thereby contributing detrimentally (58). Further comparable studies are sparse, so that the finding of effect modification by PES on the association between SNS use and mental health appears to be a novel finding.
4.5 Methodological Considerations & Limitations
The main limitation of this study is the cross-sectional design, which does not exclude causal effects in both directions between the main variables, and it is not unlikely that the causal mechanism may be bidirectional. Since non-response was not negligible, selection bias could be an issue and theoretically leading to either overestimation or underestimation of the found associations. However, it seems unlikely to be the major explanation of the findings. A strength of this study is the comparatively large sample size, which reduces the risk of random error. The main exposure variables were developed by one of the authors due to lack of well-validated alternatives, and are rather straight-forward questions on frequency of SNS use and number of contacts which are unlikely to be misunderstood by the respondents or to be an issue of other types of response bias. Since self-reported SNS use may be under-reported when compared to actual use, there is a risk of differential misclassification (i.e. if some high users with poor mental health were erroneously classified as non-high-users), and thus, our findings may represent an under-estimation of the true effect (59). The outcome variable was measured by a very well-validated instrument, which also reduces the likelihood of misclassification. The associations were controlled for the most obvious potential confounders; age, occupation, relationship status, as well as PES, which does not exclude bias from confounding, but ought to render it a less likely explanation of the found associations. Finally, the SNS measures used in this study were not specific to certain SNS. As these sites vary in their purpose of usage, interface and content, it would be beneficial to include separate SNS survey questions. Facebook is the most commonly studied SNS (2), hence also the literature presented here is biased towards Facebook. Having said that, most Swedes use Facebook (4), therefore it is probable that these results are most applicable to Facebook.