Mediated roles of generalized trust and perceived social support in the effects of problematic social media use on mental health: A cross‐sectional study

Abstract Background Current literature lacks evidence concerning how problematic social media use associates with mental health. To address the gap, the present study used mediation models to examine whether generalized trust and perceived social support (PSS) are potential mediators in the relationship between problematic social media use and mental health. Methods The sample comprised Iranian adults (n = 1073; 614 females). The participants completed a number of scales to assess problematic social media use (Bergen Social Media Addiction Scale), generalized trust (Generalized Trust Scale), PSS (Multidimensional Scale of Perceived Social Support, happiness (Oxford Happiness Questionnaire Short Form), depression and anxiety (Hospital Anxiety and Depression Scale), and mental quality of life (Short Form‐12). Results Problematic social media use had negative effects on happiness and mental quality of life via the mediators of generalized trust (bootstrapping SE = 0.017; effect = −0.041; 95% CI = −0.079, −0.012) and PSS (bootstrapping SE = 0.023; effect = −0.163; 95% CI = −0.211, −0.119). Problematic social media use had positive effects on anxiety and depression via the mediators of generalized trust (bootstrapping SE = 0.022; effect = 0.064; 95% CI = 0.026, 0.113) and PSS (bootstrapping SE = 0.024; effect = 0.052; 95% CI = 0.009, 0.102). Conclusions Problematic social media use, generalized trust and PSS are important factors for an individual's mental health. Health‐care providers may want to assist individuals regardless of having mental health problems in reducing their problematic social media use and improving their generalized trust and social support.


| INTRODUC TI ON
The World Health Organization 1 has defined mental health as the 'subjective well-being, perceived self-efficacy, autonomy, competence, intergenerational dependence, and self-actualization of one's intellectual and emotional potential, among others… It is, however, generally agreed that mental health is broader than a lack of mental disorders' (p. 5). Therefore, mental health encompasses a broad spectrum with both positive and negative outcomes. From a positive perspective, mental health includes happiness and a good mental quality of life. From a negative perspective, good mental health includes the absence of disorders such as anxiety and depression. 2 Literature further shows and suggests that mental health can be a basis and a reciprocal consequence of an individual's overall health, including physical wellness and social burden. 2,3 Indeed, a large-scale study conducted in Taiwan showed that mental health problems (ie psychological symptoms) might lead to lowered quality of life in various domains (social, physical and environmental) among 1080 healthy workers. 4 Therefore, improving mental health has been proposed to be an important topic for public health. 2,3 Before health-care providers can design effective programmes to improve mental health among general populations, they need to identify relevant factors that are related to good and/or poor mental health. Previous empirical studies have demonstrated that poor mental health is related to lower generalized trust, 5,6 lower perceived social support (PSS) 7 and problematic social media use. 8 Consequently, higher levels of both generalized trust and PSS are related to better mental health (including increased happiness, elevated mental quality of life, reduced anxiety and decreased depression). On the other hand, problematic social media use is related to poorer mental health.
Generalized trust (aka general trust or trust) is defined as 'a willingness to be vulnerable to the actions of others' (p. 1) 9 and is the basis of societal factors that help maintain daily living of an individual. Also, generalized trust supplements PSS in contributing to good social relationships. Indeed, evidence shows that trust makes an individual anticipate positive rather than negative outcomes from others' actions. [10][11][12] Consequently, a positive relationship is then established based on the positive interaction between two individuals.
A positive relationship may further assist an individual in obtaining health benefits, including good psychological health, low levels of distress and happy mood. [13][14][15][16][17] Perceived social support is defined as 'perceptions of the extent to which individuals from one's social network are available to provide social support' (p. 85) 7 and is the basis of a good social relationship. A greater level of PSS often leads to satisfactory social interaction and consequently results in improved mental health and the avoidance of psychological distress. 18,19 Moreover, PSS has been reported as having larger effects concerning mental health than actual received social support. 20 Therefore, assessing how an individual perceives social support may provide some insights for health-care providers to an individual's mental health.
Problematic social media use, a term that is sometimes interchangeably used with social media addiction or social media overuse, currently does not have a consensus in its definition among researchers. 21 Although there is no clear definition of problematic social media use, current literature has demonstrated consistent findings in the relationship between problematic social media use (or other related terms) and mental health. More specifically, problematic social media use has been related to psychological distress, including stress, anxiety and depression. 8,22,23 Individuals with problematic social media use may suffer from fear of missing out which may result in mental health problems for some individuals. 24 Apart from the relationships with mental health problems, the literature reports that problematic social media is also related to poor trust and low levels of PSS. 25,26 Problematic social media use has become a public issue worldwide over the past decade for a number of reasons. First, the growth of global social media use has been rapid given the advances in technology on the internet and smartphone. 27,28 Data from 2020 reported that 3.6 billion people worldwide who had accessed the internet were social media users. 29 More specifically, many social media platforms (eg Twitter, Instagram, Facebook) have been developed with the advances of internet/smartphone technology. These very popular and attractive sites are designed to keep users engaged and for many individuals their use can be habitual and/or excessive taking time away from activities that may be more important such as educational and/or occupational duties. Consequently, excessive social media use, like other types of excessive use such as excessive internet or smartphone use, may cause a variety of health problems, including sleep problems, musculoskeletal discomfort, mood issues and impaired daily functions, as well as addiction-like consequences in a small minority of individuals. [27][28][29][30][31][32][33][34][35][36] Empirical evidence also shows that poor mental health is related to low levels of generalized trust and PSS, as well as problematic social media use. 7,13,22 However, previous research has rarely examined simultaneously whether generalized trust, PSS, and problematic social media use are all directly related to mental health. To address the literature gap regarding direct or indirect relationships, the present study used a phased cross-sectional design to examine whether generalized trust and PSS were mediators in the relationship between problematic social media use and different types of mental health (including happiness, depression, anxiety and mental quality of life). The present authors hypothesized that generalized trust and PSS would be mediators because prior research has shown that problematic social media may lead to reduced trust and lowered PSS. 25,26 2 | ME THODS

| Participants and procedure
The study sample was recruited from health centres and health houses in Qazvin province, Iran. A two-level cluster sampling design was utilized. More specifically, (a) three health centres and two health houses were randomly selected in each city (Qazvin, Takestan, Abyek and Buin Zahra) in Qazvin province and (b) using a convenience sampling approach, participants were selected from those individuals who had been referred to health centres and health houses to receive primary health care. To be eligible, participants had to be adult (ie 18 years of age or older) and agree to participate.
Qazvin University of Medical Sciences' ethics committee approved the study, and all study participants gave their written informed consent (IR.QUMS.REC.1397.122).
After participants had provided informed consent, they completed a baseline questionnaire of problematic social media use. The same participants completed the first follow-up measures on PSS and social trust one month after the baseline assessment. Finally, the participants completed a measure relating to depression, anxiety, quality of life and happiness three months after the baseline assessment. All the assessments were completed between May and November 2019.

| Generalized Trust Scale (GTS)
The GTS comprises six items rated on a five-point Likert scale to assess generalized social trust. 12,39 A sample item of the GTS is 'Most people are basically honest'. A higher score on the GTS indicates a greater level of generalized trust. Because there was no previous Persian validation of GTS, the GTS was translated into Persian following international guidelines. 40,41 The Cronbach's alpha in the present study was 0.89.

| Multidimensional Scale of Perceived Social Support (MSPSS)
The MSPSS, 42 including the Persian version, 43 comprises 12 items rated on a seven-point Likert scale to assess PSS comprising three dimensions, including four items related to family, four items related to friends, and four items related to significant others. A sample item of the MSPSS family dimension is 'My family really tries to help me'; a sample item of the MSPSS friends dimension is 'My friends really try to help me'; and a sample item of the MSPSS friends dimension is 'There is a special person who is around when I am in need.' Higher scores on the MSPSS indicate a stronger level of PSS. The Cronbach's alpha in the present study was 0.82.

| Oxford Happiness Questionnaire Short Form (OHQ-SF)
The OHQ-SF, 44 including the Persian version, 45 comprises eight items rated on a six-point Likert scale to assess happiness. A sample item of the OHQ-SF is 'I feel that life is very rewarding'. Higher scores on the OHQ-SF indicate a higher level of happiness. The Cronbach's alpha in the present study was 0.87.

| Hospital Anxiety and Depression Scale (HADS)
The HADS, including the Persian version, 46

| Statistical analysis
After using descriptive statistics to examine the sociodemographic features of the participants, Pearson correlations were applied to understand how the studied variables (ie problematic social media use, generalized trust, PSS, happiness, anxiety, depression and mental quality of life) were related. Following this, four mediation models were constructed and all shared the same controlled variables (ie age, gender, marital status, having children or not, place of resident, occupational status and years of education), mediators (ie PSS and generalized trust), and independent variable (ie problematic social media use). The differences among the four mediation models were the use of different dependent variables, that is, each mediation model had a different type of mental health (ie happiness, anxiety, depression and mental quality of life; see Figure 1). With the use of 5000 bootstrapping resamples, all the mediation models were assessed using Model 4 in PROCESS macro for SPSS. 50 The PROCESS macro based on the linear regression models was used to construct mediation models with the use of the bootstrapping method. More specifically, the associations between variables (ie the associations between independent variables and mediators; between mediators and dependent variables; and between independent variables and dependent variables) were estimated using the ordinary least square method. The significance of the direct effects was examined using independent t tests; the significance of the indirect effects was examined using the bootstrapping method (ie if the 95% samples of the bootstrapping resamples do not cover 0, the mediating effect is significant). Moreover, all the independent variables, mediators and dependent variables were treated as numeric variables in the mediation models. This treatment is appropriate given that all the independent variables, mediators, and dependent variables were summed total scores of validated psychometric scales.

| RE SULTS
The participants' mean age was 36.57 years (SD = 10.21), and their mean number of years spent in education was 9.87 years (SD = 4.68).
Nearly half of the participants were married (n = 571, 49.3%) and employed (n = 524, 48.8%; Table 1). it is likely that the overall mental health at a population level can be improved.
An association between PSS and mental health was clearly observed in the present study. Moreover, such finding is consistent to the literature. More specifically, when individuals perceive a greater level of social support, they often have satisfactory social interaction, good mental health, and little psychological distress. 18,19 PSS may minimize the feeling of loneliness and improve leisure activity performance (eg having partners to engage in leisure activities that cannot be done alone) for an individual. 19 Given that loneliness and lack of leisure are risk factors contributing to poor mental health, a high level of PSS is likely to prevent an individual from developing poor mental health.
Regarding the clinical significance of the present study's findings, the guidance proposed by Cohen 56 was applied: r = .1 indicates small effect; .3 indicates medium effect; and .5 indicates large effect. Therefore, the Pearson correlations shown in Table 2    Note: Age, gender, marital status, having children, place of resident, occupational status, and years of education were adjusted for all the models. Boot = Bootstrapping; Boot SE = bootstrapping standard error; CI 95% = 95% confidence interval using 5000 bootstrap samples; LLCI = lower limit confidence interval; ULCI = upper limit confidence interval; SE = standard error.

TA B L E 3
Problematic social media use effect on happiness, anxiety, depression, and mental quality of life with mediators of perceived social support and generalized trust Therefore, health-care providers may want to assist individuals (irrespective of whether they have mental health problems) in reducing their problematic social media use and improving their generalized trust and social support.

ACK N OWLED G EM ENTS
The authors would like to thank the participants and Dr Mehran Alijanzadeh.

CO N FLI C T O F I NTE R E S T
The authors have declared that no competing interests exist.

DATA AVA I L A B I L I T Y S TAT E M E N T
There are no data available for sharing because of conditions imposed by the Ethics approval under which this research was conducted.