A meta-analysis of the problematic social media use and mental health

Abstract

Background:

Although previous meta-analyses were conducted to quantitatively synthesize the relation between problematic social media (SM) use and mental health, they focused on Facebook addiction.

Aims:

The purpose of this meta-analysis is to examine this relation by extending the research scope via the inclusion of studies examining problematic use of all platforms.

Method:

One hundred and thirty-three independent samples (N =244,676) were identified.

Results:

As expected, the mean correlations between problematic SM use and well-being are negative, while those between problematic SM use and distress are positive. Life satisfaction and self-esteem are commonly used to represent well-being, while depression and loneliness are usually used to indicate distress. The mean correlations of problematic SM use with life satisfaction and self-esteem are small, whereas those of problematic SM use with depression and loneliness are moderate. The moderating effects of publication status, instruments, platforms and mean age are not significant.

Conclusions:

The magnitude of the correlations between problematic SM use and mental health indicators can generalize across most moderator conditions.

Keywords

Mental health well-being distress meta-analysis problematic social media use

Social media (SM) is defined as ‘Internet-based channels that allow users to opportunistically interact and selectively self-present, either in real-time or asynchronously, with both broad and narrow audiences who derive value from user-generated content and the perception of interaction with others’ (Carr & Hayes, 2015, p. 50). Popular SM platforms are such as WhatsApp and social networking sites (SNSs; e.g. Facebook). Widespread use of SM means that problematic SM use has become a serious problem. For example, the Bergen Facebook Addiction Scale (Andreassen et al., 2012) conservatively estimated that 3.3% of Italian users were addicted to Facebook (Biolcati et al., 2018). Vangeel et al. (2016) found that 7.1% of secondary school students in Belgium were addicted to Facebook, with compulsive users spending 2 hours 38 minutes on a school day and 4 hours and 35 minutes on a holiday. A French study found a 10% rate of Facebook addiction (Chabrol et al., 2017). A study found that 18% of Turkish college students were classified as disordered SM users (Kircaburun, Demetrovics, & Tosuntaş, et al., 2019). A rate of 26.2% was found in college students in US, 29.4% in Singapore, and 44.5% in China (Tang et al., 2017).

Empirical evidence revealed that compulsive social networking site (SNS) use was related to physical (Moqbel & Kock, 2018) and mental health (Frost & Rickwood, 2017; Pontes, 2017; Ryan et al., 2014). Researchers have paid increasing attention to excessive SM use, and have performed meta-analyses (Marino et al., 2018a, 2018b) about the relationships of compulsive SM use with mental health, psychological distress and well-being. While these meta-analyses have improved the understanding of the relation between SM addiction and mental health, they do not address some salient issues. For example, these meta-analyses focused narrowly on Facebook addiction. The present meta-analysis extends the concern to problematic use of all SM platforms.

Problematic SM use

Several terms are used to describe problematic SM use. Addiction (Ryan et al., 2014) is one common term. Chamberlain et al. (2016) defined the core factors of behavioral addiction as inability to control of use, functional impairment and continuing involvement in the behavior regardless of its negative impacts. Some researchers (Caldiroli et al., 2018; Miele et al., 1990) used the term ‘dependency’ to describe problematic SM use. Dependence was defined as an indispensable behavior to achieve goals, while addiction refers to failure to control leading to impairment of personal or work lives (Ferris & Hollenbaugh, 2018). Hence, addiction has an absolutely negative effect, while dependence does not necessarily. Other terms, such as compulsive use (Aladwani & Almarzouq, 2016; De Cock et al., 2014), excessive use (Wang et al., 2016), and disordered use (van den Eijnden et al., 2018), were also used. Problematic use was chosen in this study because it is broad enough to incorporate different levels of excessive use (Lee et al., 2017).

Empirical studies

Previous empirical studies were conducted in various research contexts, and have different findings about the strength of the relation between problematic SM use and mental health. For example, Kircaburun (2016) sampled 1,130 Turkish secondary school students and found that the relation between problematic SNS use and self-esteem was r = −.09. Turel and Qahri-Saremi (2016) also found a small relation between problematic Facebook use and self-esteem at r = .01 for a pilot sample of 60 undergraduate students, and r = −.05 for 341 Facebook users from a large university in North America. A moderate correlation (r = −.24) was found in Aladwani and Almarzouq (2016) who used a sample of 407 undergraduate students in Kuwait. Biolcati et al. (2018) also found support for a moderate effect. On the other hand, a large correlation between problematic Facebook use (r = −.43) and self-esteem was found in Baturay and Toker (2017) who sampled 120 college students in Turkey.

Findings about other mental health indicators were also inconsistent. For example, the relation between problematic SM use and depression was from small (r = .13; Kircaburun, 2016) to large (r = .45; Błachnio et al., 2015). As both empirical studies varied in research situations, and research findings were not consistent, moderator effects are worth investigating.

Publication status

Publication bias refers to the unrepresentativeness of included studies that can be caused by availability and accessibility (McShane et al., 2016). For example, conference papers have more limited availability than journal papers. The inaccessibility of relevant studies (e.g. unpublished manuscripts) can lead to unrepresentative data in a meta-analysis. To explore this possibility, the mean correlations among publication outlets were examined.

Study country

Caldiroli et al. (2018) suggested that the problem of technology addiction was especially serious in China, South Korea and Taiwan. As the prevalence of technology addiction varies with country, the relation between problematic SM use and mental health may vary with country or culture. To examine the possible country effect, Marino et al. (2018a) examined the country effect on the relation between problematic Facebook use and psychological distress, and found that the correlation was likely to be higher in studies from Western countries compared to that from Asian countries. As Marino et al. (2018a) had a small number of effect sizes, and thus low generalizability of findings, the country effect is worth re-investigation.

Measures of problematic SM use

Researchers have used several instruments to measure problematic SM use. The most popular measure is the Bergen Facebook Addiction Scale (Andreassen et al., 2012), which assesses six key components, namely salience, mood modification, tolerance, withdrawal, conflict and relapse. Each component is initially assessed by three items. After the deletion of items with relatively low item-total correlations, one item for each component is retained. The one-factor solution is supported by the confirmatory factor analysis. The Bergen Social Media Addiction Scale (Andreassen et al., 2017) is a modified version to measure problematic SM use in general by replacing ‘Facebook’ with ‘social media’ in each item.

The Facebook Intrusion Questionnaire (Elphinston & Noller, 2011) is composed of eight items measuring cognitive salience, behavioral salience, interpersonal conflict, conflict with other activities, euphoria, loss of control, withdrawal and relapse. Each item is assessed on a seven-point Likert scale. A unidimensional model was supported by exploratory factor analysis (Elphinston & Noller, 2011).

The Social Media Disorder Scale (SMDS; van den Eijnden et al., 2016) was developed in the Netherlands, and is based on the DSM-5 criteria, namely preoccupation, tolerance, withdrawal, persistence, displacement, problem, deception, escape, and conflict. Initially, three items are developed for each of the nine components, and the item with the highest factor loading within each of the nine criteria is selected.

Other researchers (Baturay & Toker, 2017; Hong et al., 2014) adapted the Internet Addiction Test (IAT, Young, 1998) to measure problematic Facebook use or general problematic SM use. As these problematic SM use measures assess different components, the relation between problematic SM use and mental health may vary as a function of measures of problematic SM use.

Measures of mental health

Empirical studies have examined several positive indicators of mental health, such as self-esteem (e.g. Choi & Lim, 2016), life satisfaction (e.g. Hawi & Samaha, 2018), well-being (e.g. Verma & Kumari, 2016), happiness (e.g. Satici & Uysal, 2015), and positive affect (e.g. Satici, 2018). Self-esteem is the most common indicator of well-being. The most popular instrument to measure global self-esteem is the Rosenberg Self-Esteem Scale (Rosenberg, 1965) which consists of 5 positively-worded and 5 negatively-worded items. Due to brevity and easy administration, this scale has been adapted into more than 50 different languages (Schmitt et al., 2005). A shorter instrument to measure global self-worth is the Single Item Self-Esteem Scale (SISES; Robins et al., 2001) comprised of the item, ‘I have high self-esteem’, on a 7-point Likert scale.

Life satisfaction is another common indicator of well-being, and the most popular measure is the Satisfaction with Life Scale (Diener et al., 1985), which consists of 5 items with 7 response categories ranging from 1 to 7. Thus, the total score ranges from 5 to 35. Total scores of 5 to 9 are viewed as ‘extremely dissatisfied’, 15 to 19 as ‘slightly dissatisfied’, 21 to 25 ‘slightly satisfied’, and 26 to 30 ‘satisfied’ (Pavot & Diener, 1993).

Negative indicators of mental health can be represented by anxiety (e.g. Durak, 2018), depression (e.g. Worsley et al., 2018), loneliness (e.g. Yu et al., 2016), suicidal ideation (e.g. Jasso-Medrano & López-Rosales, 2018), distress (e.g. Laconi et al., 2018) and negative affect (e.g. Satici, 2018). Depression is the most examined indicator, with common measures such as the Patient Health Questionnaire-9 (PHQ-9; Kroenke et al., 2001), the Center for Epidemiologic Studies Depression Scale (CES-D; Radloff, 1977), the depression subscale of the Depression Anxiety Stress Scales-21 (DASS-21; Lovibond & Lovibond, 1995), the depression subscale of Short Depression-Happiness Scale (SDHS; Joseph et al., 2004), Hamilton Depression Rating Scale (HAM-D; Hamilton, 1960, 1967), Montgomery–Åsberg Depression Rating Scale (MADRS; Montgomery & Åsberg, 1979), and Beck Depression Inventory (BDI; Beck, & Steer, 1987). The PHQ-9 assesses 9 depressive symptoms in primary care settings over the last two weeks with 4 response categories, 0 for ‘not at all’, 1 for ‘several days’, 2 for ‘more than half the days’, and 3 ‘Nearly every day’. Thus, the total score of the PHQ-9 ranges from 0 to 27. The cutoff scores for mild, moderate, moderately severe and severe depression are 5, 10, 15, and 20, respectively (Kroenke et al., 2001). On the other hand, the CES-D assessing depressive symptom for general population over the past week consists of 20 items on a 4-response scale, ranging from 0 to 3. The total score is from 0 to 60. An arbitrary cutoff score of 16 or higher is considered to indicate possible depression requiring clinical assessment (Radloff, 1977).

The depression subscale of DASS-21 measures depressive emotional state over the past week, and consists of 7 items, each on a 4-point scale, ranging from 0 to 3 (Szabó, 2010) Scores of 10, 14, 21, and 28 indicate mild, moderate, severe and extremely severe depression (Lovibond & Lovibond, 1995). The depression subscale of the SDHS consists of 3 items with 4 response categories. The items are, ‘I felt dissatisfied with my life’, ‘I felt cheerless’, and ‘I felt that life was meaningless’. Cutoff scores for the depression subscale are not provided in Joseph et al. (2004).

The HAM-D has 6-, 17-, 21-, and 24-item versions, and is often used to measure treatment effect instead of state of depression (Santen et al., 2008). The most popular version, consisting of 17 items, is on a 3- or 5-point scale. The multidimensional factor is supported, but the number of factors varies across studies (Bagby et al., 2004). Another popular measure for assessing change of depressive symptoms in clinical trial research is the MADRS, which comprises 10 items chosen by reliability and validity from an original set of 17 items from the Comprehensive Psychopathological Rating Scale (Montgomery & Åsberg, 1979). The factor structure of the MADRS varied across patient groups (Ketharanathan et al., 2016). As both HAM-D and MADRS were usually used in clinical research, their scores are strongly correlated (Heo et al., 2007).

The BDI consisting of 21 items on a 4-point scale is a self-reporting measure to assess depression for adolescents and adults (Beck & Steer, 1987). The latest version of the BDI, the Beck Depression Inventory–II (BDI–II), is derived from DSM-IV (American Psychiatric Association, 1994), with the items of Body Image Change, Work Difficulty, Weight Loss and Somatic Preoccupation in the BDI replaced with Agitation, Worthlessness, Loss of Energy, and Concentration Difficulty in the BDI–II. The items of Changes in Sleeping Pattern and Changes in Appetite have seven response categories, and the remaining 19 items have four (Beck et al., 1996).

The most prevalent measure to assess loneliness is the UCLA Loneliness Scale (Russell et al., 1980), consisting of 10 positive-worded and 10 negative-worded items, each item on a 4-point Likert scale. The UCLA Loneliness Scale has sound psychometric properties (Hartshore, 1993), and the 3-factor (isolation, relational connectedness, and collective connectedness) structure was supported (Dussault et al., 2009).

Platform

Some researchers focused on problematic use of SM in general. For example, Worsley et al. (2018) examined the relation between problematic use in general and depression for 1029 university students in UK, and found that the relation was r = .27. Some researchers focused specifically on problematic Facebook use. Steggink (2015) used a sample of 315 users recruited online with mean age of 28.74 years, and found that the correlation between problematic Facebook use and depression was r = .10. Whether the magnitude of correlation varied with platform was unknown, and this meta-analysis addressed this possibility.

Participant age

Few longitudinal and cross-sectional studies have been conducted to examine the age effect on the relation between problematic SM use and mental health. van den Eijnden et al. (2018) adopted a three-wave design with a one-year interval between adjacent assessments for a sample of 543 teenagers with mean 12.9 years at study inception. The correlations of problematic use of SNSs assessed at time 1 with life satisfaction assessed at times 2 and 3 for boys were r = −.21 and r = −.11, respectively. The corresponding correlations of time 2 problematic use of SNSs with time 2 and 3 life satisfactions for boys were r = −.33 and r = −.09, respectively. For girls, the correlation of time 1 problematic use of SNSs with time 2 and 3 life satisfaction were r = −.43 and r = −.33, respectively and those of time 2 problematic use of SNSs with time 2 and 3 life satisfaction were r = −.48 and r = −.55, respectively. The age effect seemed to be supported in the cross-sectional study. Kanat-Maymon et al. (2018) found that the relation between problematic Facebook use and self-esteem for an online adult sample with mean age of 33.36 years was r = −.52. The corresponding correlation for a sample of 80 undergraduate students was only small, at r = −.05.

The age effect on the relations of problematic Facebook use with psychological distress and well-being were not supported in Marino et al. (2018a). The number of effect sizes in that meta-analysis was small and non-significant findings can be caused by low statistical power.

Participant gender

Griffiths (2000) revealed that technology addicts are usually male. Some researchers examined the relation between problematic SM use and mental health especially for males. For example, Kim and Park (2015) found that the correlation between problematic SNS use and self-esteem among 213 male middle-school students was r = −.38. Some researchers examined the moderating effect of gender on the relation between problematic SNS use and mental health. For example, van den Eijnden et al. (2018) examined the moderating effect of gender for teenagers, and found that the correlation between problematic SNS use and life satisfaction was r = −.48 for girls and r = −.33 for boys. Walburg et al. (2016) selected 115 boys and 171 girls, and found that the correlations of problematic Facebook use with depression and suicidal ideation were r = .37 and r = .23 for boys, and r = .10 and r = .20 for girls.

Marino et al. (2018a) examined the gender effect represented by the proportion of female users on the relations of problematic Facebook use with psychological distress and well-being and the gender effects were not significant. Re-examination of the gender effect is prominent for three reasons. First, the small number of effect size in Marino et al. (2018a) could lead to a low statistical power. Second, empirical studies rarely address the issue of gender effect on the relation between SM addiction and mental health. Third, empirical studies mentioned above seemed to demonstrate a potential gender effect.

Previous reviews and meta-analyses

Frost and Rickwood (2017) reviewed the relation between Facebook addiction and mental health based on 5 cross-sectional and 1 longitudinal studies. They concluded that Facebook addiction was associated with poor mental health. Ryan et al. (2014) identified three studies (Hong et al., 2014; Koc & Gulyagci, 2013; Uysal et al., 2013), and reported that Facebook addiction was related to depression and anxiety. Another review by Keles et al. (2019) that identified three articles also supported that addiction to SNSs was related to depression.

Marino et al. (2018a) identified 23 samples that examined the relations of problematic Facebook use with psychological distress and well-being. The mean correlation between problematic Facebook use and psychological distress was r = .29, and the correlation corrected for attenuation was ρ = .34. The correlations for specific factors were r = .30 and ρ = .35 for depression; r = .29 and ρ = .33 for anxiety; r = −.19 and ρ = −.22 for general well-being; r = −.16 and ρ = −.19 for life satisfaction. Marino et al. (2018b) meta-analyzed 8 correlations between problematic Facebook use and self-esteem, and found that the mean correlation was r = −.23. No moderator analyses were conducted for this correlation.

As two previous meta-analyses (Marino et al., 2018a, 2018b) specifically focused on problematic Facebook use, their conclusions may not be valid for studies addressing problematic use of other SM or general problematic use of SM. The current meta-analysis aimed to conduct a comprehensive analysis of accumulating empirical evidence obtained by studies examining the association between problematic SM use and mental health.

Method

Literature search

To identify relevant studies, the ERIC, PsycINFO and ProQuest Dissertations and Theses Global databases were searched using SM terms (namely, Facebook, Twitter, Instagram, MySpace, ‘social media’, ‘online social network*’, and ‘social network* site*’) and problematic use terms (addict*, abuse, misuse, overuse, intrusion, ‘problematic use’, ‘excessive use’, ‘compulsive use’, ‘pathological use’, ‘disordered use’) through July 19, 2020. The ERIC, PsycINFO and ProQuest databases yielded 157, 1,553 and 287 articles, respectively. The reference lists for eligible articles and previous meta-analysis (Marino et al., 2018a, 2018b) were then examined. The author screened each article by reviewing the title and abstract. The full texts of studies passing the initial screening were then identified to determine eligibility based on three inclusion criteria. First, studies should provide sufficient statistics to compute the correlation between problematic SM use and mental health. Second, studies should report the sample size. Lastly, the study should be published in English. Four unpublished datasets in Marino et al. (2018a) were not available because they did not report titles, sources or manuscripts.

Analysis

The Pearson Product-Moment correlation between problematic SM use and mental health was coded. The distribution of r depends on the population correlation, ρ, and sample size. Unless the sample size is sufficiently large, the distribution of sample correlation is skewed (Card, 2012). To normalize the sample correlation, the correlation r was converted to Z_r using the Fisher’s transformation equation. The inverse variance (N-3) was used as a weight to compute the mean correlation. Random-effects were used.

Results

This study included 123 articles presented in the Appendix. Andreassen et al. (2016) and Andreassen et al. (2017) analyzed the same data, yielding 122 studies. Of these 122 studies, 111 were published in journals, 5 in doctoral dissertations, 3 in Master theses, 2 in conferences, and 1 in a book chapter. Eleven studies each consisted of 2 samples, and thus 133 independent samples involving 244,676 participants were analyzed in the subsequent analyses. Table 1 presents the descriptive statistics of the 133 samples. When multiple indicators of mental health, multiple platforms, or multiple measures of problematic use SM were assessed, all relevant effect sizes were coded. The summary of all effect sizes were presented in Table 2.

Table 1.

Descriptive statistics of the 132 independent samples included in the meta-analysis.

variable	k	Min	Max	Mean	SD
N	133	55	154981	1839.67	13536.57
age	125	13.02	50.13	21.89	6.22
female	129	0	1	0.59	0.17
ES	133	–0.57	0.59	0.10	0.27

age = mean age of the sample; female = proportion of females in the sample.

Table 2.

Summary of samples examining the problematic social media use and mental health.

Study	Country	N	Age	FM	MH	MHMeas.	SM	SMMeas.	ES
Abbasi and Drouin (2019)	miscellaneous	742	27.44	0.64	negative affect	PANAS	Facebook	FIQ	0.13
AbuDamous (2020)	US	264	36.85	0.54	self-esteem	SISES	SM	Bergen	−0.05
Akin and Akin (2015)	Turkey	370	20.20	0.53	life satisfaction	SWLS	Facebook	Bergen	−0.39
Al Mamun and Griffiths (2019)	Bangladesh	300	19.50	0.39	depression	PHQ-9	Facebook	Bergen	0.26
Aladwani and Almarzouq (2016)	Kuwait	407	20.04	0.54	self-esteem	Rosenberg	SM	Meerkerk (2007)	−0.24
Andreassen et al. (2016)	Norway	23533	35.80	0.65	anxiety, depression	HADS	SM	Bergen	0.34, 0.19
Andreassen et al. (2017)	Norway	23532	35.80	0.65	self-esteem	Rosenberg	SM	Bergen	−0.25
Apaolaza et al. (2019)	Spain	346	18.73	0.52	self-esteem, social anxiety	Rosenberg, Social Anxiousness Scale	WhatsApp	Faber and O’Guinn (1992)	−0.42 0.44
Atroszko et al. (2018)	Poland	1157	20.33	.52	self-esteem, loneliness, social anxiety	Atroszko et al. (2018), Short Loneliness Scale, LSAS	Facebook	Bergen	−0.10, 0.13, 0.19
Aydin et al. (2020)	Turkey	111	30.14	0.55	depression	Montgomery-Asberg Depression Rating Scale	SM	Bergen	0.20
Balcerowska et al. (2020)	Poland	1099	21.44	0.72	well-being	Ultra-Short Protocol for Measuring Subjective Well-being	SNSs, Facebook	Bergen	−0.15 −.13
Balci and Gölcü (2013)	Turkey	892	21.10	0.59	loneliness	NA	Facebook	Balci and Gölcü (2013)	0.35
Balci and Tiryaki (2014)	Turkey	486	17.60	0.51	loneliness	NA	Facebook	Bergen	0.04
Baturay and Toker (2017)	Turkey	120	21.46	0.53	self-esteem, psychiatric well-being	Rosenberg, GHQ-12	Facebook	IAT	−0.43, −0.29
Bérail et al. (2019)	miscellaneous	932	21.25	0.73	social anxiety, loneliness	LSAS, UCLA	Youtube	IAT	0.32, 0.28
Biolcati et al. (2018)	Italy	755	25.17	0.80	social loneliness, life satisfaction	SELSA-S, SWLS	Facebook	Bergen	0.27, −0.24
Błachnio and Przepiórka (2018)	Poland	360	22.22	.64	life satisfaction	SWLS	Facebook	FIQ	0.01
Błachnio and Przepiórka (2019)	Poland	597	21.22	0.68	loneliness, self-esteem, life satisfaction	De Jong Gierveld Loneliness Scale, Rosenberg, SWLS	Facebook	FIQ	0.15, −0.12, −0.02
Błachnio et al. (2019), #1	Italy	317	24.66	0.67	self-esteem, life satisfaction	Rosenberg, SWLS	Facebook	FIQ	−0.24, −0.01
Błachnio et al. (2019), #2	US	238	NA	0.73	self-esteem, life satisfaction	Rosenberg, SWLS	Facebook	FIQ	−0.23, −0.17
Błachnio et al. (2015)	Poland	672	27.53	.65	depression	CES-D	Facebook	FIQ	0.45
Błachnio et al. (2016)	Poland	381	20.73	0.63	self-esteem, life satisfaction	Rosenberg, SWLS	Facebook	Bergen	−0.16, 0.03
Boer et al. (2020)	miscellaneous	154981	13.54	.51	life satisfaction	Cantril Ladder	SM	SMDS	−0.20
Brailovskaia and Margraf (2017)	Germany	179	22.52	0.77	depression, anxiety	DASS-21	Facebook	Bergen	0.22, 0.32
Brailovskaia, Rohmann, Bierhoff, Margraf, and Köllner, et al. (2019)	Germany	349	50.13	0.70	depression, mental health	BDI-II, Lukat et al. (2016)	Facebook	Bergen	0.23, −0.30
Brailovskaia, Schillack, Margraf, et al. (2018)	Germany	520	22.42	0.75	depression, anxiety, happiness	DASS-21, SHS	Facebook	Bergen	0.40^a, 0.42^a, −0.25^a
Brailovskaia et al. (2018)	Germany	122	22.70	.83	mental health	Lukat et al. (2016)	Facebook	Bergen	−0.27
Brailovskaia, Velten, Margaf, et al. (2019), #1	Germany	531	21.63	.75	depression	DASS-21	Facebook	Bergen	0.43
Brailovskaia, Velten, Margaf, et al. (2019), #2	US	909	37.24	0.48	depression	DASS-21	Facebook	Bergen	0.55
Brown (2015)	US	55	19.50	0.82	self-esteem	Rosenberg	Facebook	Facebook Compulsion Inventory	−0.17
Burnell and Kuther (2016)	US	256	25.41	0.62	self-esteem	Rosenberg	SM	Al-Menayes (2015)	−0.26
Cargill (2019)	US	224	33.01	0.82	anxiety	STAI	SM	IAT	0.29
Casale and Fioravanti (2015), #1	Italy	193	22.45	0	social anxiety	SIAS	SNSs	GPIUS2	0.44
Casale and Fioravanti (2015), #2	Italy	207	22.45	1	social anxiety	SIAS	SNSs	GPIUS2	0.22
Chabrol et al. (2017)	France	456	20.50	0.76	depression, social anxiety	CES-D, SASA	Facebook	IAT	0.31, 0.30
Chae et al. (2018)	South Korea	253	21.50	0.64	depression	CES-D	SNSs	SNS Addiction Proneness Scale for College Students	0.42
Chavez and Chavez (2017)	Philippines	119	15.00	NA	loneliness	UCLA	Facebook	Balci and Gölcü (2013)	0.21
Chen et al. (2020)	Hong Kong	308	23.75	0.68	anxiety & depression	HADS	SM	Bergen	0.21
Choi and Lim (2016)	South Korea	419	25.98	0.49	self-esteem	Rosenberg	Facebook	Koc and Gulyagci (2013)	−0.22
Crowell (2014)	US	380/381	33.56	0.80	self-esteem, life satisfaction	Rosenberg, Quality of Life Enjoyment and Satisfaction Questionnaire Short Form	Facebook	Bergen	0.24, −0.20
Cudo et al. (2020)	Poland	235	21.79	.63	self-esteem, depression	Rosenberg, PHQ-9	Facebook	FIQ	−0.05, 0.27
da Veiga et al. (2019)	Portugal	404	21.65	0.73	depression, anxiety	BSI	Facebook	Bergen	0.29, 0.23
De Cock et al. (2014)	Belgium	1000	43.00	NA	self-esteem, loneliness, depression	Rosenberg, Rasch-Type Loneliness Scale, DML	SNSs	Bergen	−0.29, 0.22, 0.38
Dempsey et al. (2019)	US	291	20.03	0.58	life satisfaction, depression, social anxiety	SWLS, PHQ-9, SIAS	Facebook	Bergen	−0.07, 0.20, 0.30
Dhir et al. (2018), #1	India	1554	14.63	0.46	depression, social anxiety	Salokangas et al. (1995), SASA	Facebook	Bergen	0.38, 0.33
Dhir et al. (2018), #2	India	1144	14.88	0.44	depression, social anxiety	Salokangas et al. (1995), SASA	Facebook	Bergen	0.31, 0.22
Durak (2018)	Turkey	451	15.00	0.48	social anxiety	SASA	SM	SMDS	0.58
Durak and Seferoğlu (2019)	Turkey	580	22.90	0.60	social anxiety, loneliness	LSAS, UCLA	SM	SMDS	0.37, 0.31
Gerhart (2017)	US	413	19.50	0.48	life satisfaction	SWLS	SNSs	Gerhart (2017)	0.11
Giota and Kleftaras (2013)	Greece	143	23.80	0.58	depression	Questionnaire of Self Evaluated Depressive Sympomatology	SNSs	GPIUS	0.26
Guven (2019)	US	188	NA	0.78	self-esteem, life satisfaction, positive affect, negative affect	Rosenberg, UCLA, PANAS	SM	FIQ	0.23, −0.08, −0.16, −0.33
Hawi and Samaha (2017)	Lebanon	364	21.10	0.48	self-esteem, life satisfaction	Rosenberg, SWLS	SM	FIQ	−0.23, −0.03
Hawi and Samaha (2019)	Lebanon	512	21.23	0.44	self-esteem, life satisfaction	Rosenberg, SWLS	SM	FIQ	−0.20, −0.04
Holmgren and Coyne (2017)	US	442	18.86	0.52	depression	CES-DC	SM	PUMP	0.29
Hong et al. (2014)	Taiwan	241	20.00	0.41	self-esteem	Rosenberg	Facebook	IAT	−0.12
Hou et al. (2018)	China	714	19.80	0.62	loneliness	UCLA	Weibo,Wechat	Microblog Excessive Use Scale, WeChat Excessive Use Scale	0.14, 0.03
Hussain and Griffiths (2019)	miscellaneous	638	32.03	0.47	depression, anxiety	DASS-21	SM	Bergen	0.32, 0.38
Hussain et al. (2019)	UK	69	23.09	.68	self-esteem, depression, anxiety	Rosenberg, DASS-21	Facebook	Bergen	−0.12, 0.24, 0.16
Jasso-Medrano and López-Rosales (2018)	Mexico	374	20.01	0.59	depression, suicidal ideation	CES-D, Positive and Negative Suicidal Ideation Inventory	SM	Social Network Addiction Questionnaire	0.25, 0.16
Jeri-Yabar et al. (2019)	Peru	212	20	0.45	depression	BDI	SNSs	NA	0.46
Kanat-Maymon et al. (2018), #1	Israel	337	33.36	0.55	self-esteem	Rosenberg	Facebook	Bergen	−0.52
Kanat-Maymon et al. (2018), #2	Israel	80	22.91	0.91	self-esteem	SISES	Facebook	Bergen	−0.05
Karakose et al. (2016)	Turkey	712	16.00	0.58	loneliness	UCLA	Facebook	Bergen	0.13
Khattak et al. (2017)	Pakistan	200	NA	0.50	depression	BDI-II	Facebook	Bergen	0.24
Kim and Park (2015)	South Korea	213	14.00	0	self-esteem	Kang (2007)	SNSs	Cho and Suh (2013)	−0.38
Kircaburun (2016)	Turkey	1130	14.67	0.59	depression, self-esteem	CDI, Rosenberg	SM	Arslan and Kirik (2013)	0.13, −0.09
Kircaburun, Demetrovics et al., (2020)	Turkey	344	20.76	0.82	depression, self-esteem	SDHS, SISES	SNSs	Xanidis and Brignell (2016)	0.22, −0.21
Kircaburun et al. (2019)	Turkey	827	20.36	0.60	self-esteem	SISES	SM	SMDS	−0.06
Kircaburun, Griffiths, Billieux et al. (2019)	Turkey	470	16.29	0.60	depression	SDHS	SM	Bergen	0.32
Kircaburun et al. (2020)	Turkey	460	19.74	0.61	depression, loneliness	SDHS, UCLA	SNSs	Xanidis and Brignell (2016)	0.34, 0.26
Kırcaburun et al. (2018), #1	Turkey	804	16.20	0.48	depression, self-esteem	SDHS, SISES	SNSs	Xanidis and Brignell (2016)	0.37, −0.15
Kırcaburun et al. (2018), #2	Turkey	760	21.48	0.60	depression, self-esteem	SDHS, SISES	SNSs	Xanidis and Brignell (2016)	0.22, −0.11
Koc and Gulyagci (2013)	Turkey	447	21.64	0.22	depression	GHQ-28	Facebook	Koc and Gulyagci (2013)	0.28
Laconi et al. (2018)	France	822	21.60	0.55	distress	BSI	Facebook,Twitter	Bergen	0.33, 0.22
LaRose et al. (2011)	US	364	17.76	0.70	self-esteem, loneliness	SISES, UCLA	SNSs	Caplan (2010) and Meerkerk et al. (2009)	−0.23, 0.06
Lee-Won et al. (2015)	US	243	19.50	0.72	social anxiety	Fenigstein et al. (1975)	Facebook	Koc and Gulyagci (2013)	0.18
Li et al. (2018)	China	463	19.12	0.53	life satisfaction	SWLS	Wechat	WeChat Addiction Scale	0.10
Lim et al. (2020)	US	1231	37.28	0.54	depression	PHQ-9	SM	SMDS	0.30
Lin et al. (2017)	Iran	2676	15.54	0.44	depression, anxiety	DASS-21	SM	Bergen	0.21, 0.17
Lin et al. (2020)	Iran	1791	27.20	0.70	anxiety, depression	HADS	SM	Bergen	0.31, 0.23
Liu and Ma (2018)	China	301	26.92	0.27	self-esteem	Rosenberg	SM	NA	−0.16
Majid et al. (2019)	Pakistan	378	NA	0.70	social anxiety	SIAS	SNSs	Moqbel and Kock (2018)	0.30
Malik and Khan (2015)	Pakistan	200	NA	0.50	self-esteem	Rosenberg	Facebook	Bergen	−0.18
Martinez-Pecino and Garcia-Gavilán (2019)	Spain	233	15.12	0.47	self-esteem	Rosenberg	Instagram	Marino et al. (2017)	−0.14^a
Menning et al. (2020)	Germany	700	25.60	0.76	distress	BSI	SNSs	SNSDQ	0.06
Milošević-Đorđević and Žeželj (2014)	Serbia	2014	NA	0.52	self-esteem	Rosenberg	SNSs	Milošević-Đorđević and Žeželj (2014)	−0.35
Mitra and Rangaswamy (2019)	India	264	21.56	0.62	depression	Clinically Useful Depression Outcome Scale	SM	SMDS	0.40
Ndasauka et al. (2016)	UK	256	21.40	0.53	loneliness	UCLA	Twitter	Microblog Excessive Use Scale	0.22
Omar and Subramanian (2013)	Malaysia	400	19.50	NA	loneliness	UCLA	Facebook	Kim and Karidakis (2009)	0.25
Ponnusamy et al. (2020)	Malaysia	364	22.50	0.51	life satisfaction, loneliness	SWLS, UCLA	Instagram	Bergen	0.59, 0.19
Pontes (2017)	Portugal	495	13.02	0.47	depression, anxiety	DASS-21	Facebook	Bergen	0.33, 0.31
Pontes et al. (2018)	UK	511	27.50	0.65	distress	Symptom Checklist-6	SM	Bergen	0.44
Przepiórka and Błachnio (2020)	Poland	426	14.67	0.49	depression	CES-D	Facebook	FIQ	0.35
Rajesh and Rangaiah (2020)	India	114	24.00	0.32	loneliness	UCLA	Facebook	Bergen	0.38
Raudsepp (2019)	Estonia	249	15.30	0.53	depression	CES-D	SM	Bergen	0.34
Robinson (2019)	US	504	20.40	0.82	depression	PHQ-9	SM	Bergen	0.20
Şahin (2017)	Turkey	612	20.34	0.62	life satisfaction	SWLS	SM	Şahin and Yağcı (2017)	−0.31
Saleem et al. (2014)	Pakistan	600	19.50	0.50	loneliness	UCLA	Facebook	IAT	0.23
Satici (2019)	Turkey	280	21.04	0.58	positive affect, negative affect, life satisfaction, loneliness	PANAS, SWLS, UCLA	Facebook	Bergen	−0.16, 0.16, −0.19 0.15
Satici and Uysal (2015)	Turkey	311	20.86	0.58	life satisfaction, happiness	SWLS, SHS	Facebook	Bergen	−0.32, −0.32
Savci et al. (2018)	Turkey	187	NA	0.47	positive affect, negative affect	PANAS	SM	SMDS	−0.23, 0.36
Sheldon et al. (2020)	US	337	23.35	.58	life satisfaction	Life Position Scale	Facebook, Instagram,Snapchat	Bergen	−0.13, −0.15, −0.04
Shettar et al. (2017)	India	100	27.55	0.46	loneliness	UCLA	Facebook	Bergen	0.24
Soraci et al. (2020)	Italy	217	32.14	0.64	anxiety, depression, self-esteem	Adult PROMIS Emotioanl Distress/Anxiety-Short Form, Adult PROMIS Emotioanl Distress/Depression-Short Form, Rosenberg	Facebook	Bergen	0.47, 0.48, 0.23
Spraggins (2009)	US	350	20.00	0.81	social anxiety, life satisfaction, loneliness,depression, self-esteem, happiness	Social Avoidance and Distress Scale, SWLS,UCLA, CES-D, Rosenberg, Oxford Happiness Questionnaire	SNSs	GPIUS	0.22, −0.13, 0.28, 0.28, −0.30, −0.28
Steggink (2015)	Netherlands	315	28.74	0.43	Social loneliness, depression, social anxiety	SELSA-S, DASS-21,Self-Consciousness Scale; Screen for Child Anxiety Related Emotional Disorders	Facebook	Bergen	0.09, 0.10, 0.10
Stockdale and Coyne (2020)	US	385	20.01	0.53	life satisfaction, anxiety, depression	Blais et al. (1989) & SWLS, Spence Child Anxiety Inventory,CES-DC	SNSs	Merlo et al. (2013)	0.11, 0.24, 0.28
Tesi (2018)	Italy	580	32.00	0.62	loneliness	UCLA	SM	Bergen	0.19
Turel et al. (2018)	Israel	215	26.99	0.73	well-being	WHO-5	Facebook	Bergen	−0.57
Turel and Qahri-Saremi (2016), #1	US	341	23.00	0.52	self-esteem	Rosenberg	Facebook	Turel and Bechara (2016)	−0.05
Turel and Qahri-Saremi (2016), #2	US	60	19.50	NA	self-esteem	Rosenberg	Facebook	Turel and Bechara (2016)	0.01
Uysal et al. (2013)	Turkey	297	20.10	0.53	happiness	SHS	Facebook	Bergen	−0.37
van den Eijnden et al. (2016), #1	Netherlands	724	14.36	0.54	self-esteem, depression	Rosenberg, Kutcher Adolescent Depression Scale	SM	SMDS	−0.19, 0.37
van den Eijnden et al. (2016), #2	Netherlands	873	14.28	0.48	depression, loneliness	Kutcher Adolescent Depression Scale, UCLA	SM	SMDS	0.29, 0.24
van den Eijnden et al. (2018), #1	Netherlands	275	13.90	1	life satisfaction	SWLS	SM	SMDS	−0.48
van den Eijnden et al. (2018), #2	Netherlands	263	13.90	0	life satisfaction	SWLS	SM	SMDS	−0.33
van Rooij et al. (2017)	Netherlands	3945	13.04	.50	life satisfaction, depression, self-esteem, loneliness, social anxiety	Students’ Life Satisfaction Scale, DML, Rosenberg, \|UCLA, SASA	SM	Meerkerk (2007)	−0.30, 0.45, −0.29, 0.32, 0.20
Vangeel et al. (2016)	Belgium	1002	15.21	0.51	self-esteem, loneliness, depression	Rosenberg, Rasch-Type Loneliness Scale, DML	SNSs	Bergen	−0.23, 0.25, 0.33
Vernon et al. (2017)	Australia	874	14.40	0.59	depression	Michigan Study of Adolescent Life Transitions	SNSs	IAT	0.31
Walburg et al. (2016), #1	France	115	16.61	0	depression, suicidal ideation	CES-D, Garrison et al. (1999)	Facebook	IAT	0.37, 0.23
Walburg et al. (2016), #2	France	171	16.43	1	depression, suicidal ideation	CES-D, Garrison et al. (1999)	Facebook	IAT	0.10, 0.20
Wan (2009)	China	335	23.50	0.56	loneliness	UCLA	Xianonei	IAT	0.39
Wang et al. (2016), #1	China	380	19.93	0.75	life satisfaction	SWLS	Weibo	Bergen	−0.11
Wang et al. (2016), #2	China	535	19.83	0.78	life satisfaction	SWLS	Weibo	Bergen	0.05
Wang et al. (2018)	China	365	15.96	0.52	depression, Self-esteem	CES-D, Rosenberg	SNSs	FIQ	0.18, −0.07
Wegmann et al. (2015)	Germany	334	19.27	0.72	depression	BSI	SNSs	IAT	0.48
Wong et al. (2020)	Hong Kong	300	20.89	0.59	depression, anxiety	DASS-21	SM	Bergen	0.34, 0.34
Wood et al. (2016)	US	209	20.23	0.73	depression, anxiety	DASS-21	SM	FIQ	0.18, 0.21
Worsley et al. (2018)	UK	1029	19.80	0.75	depression	PHQ-9	SM	Bergen	0.27
Yam et al. (2019)	Hong Kong	307	21.64	.68	anxiety, depression	HADS	SM	Bergen	0.19, 0.18
Young et al. (2020)	US	124	30.58	0.80	depression, life satisfaction	PHQ-9, SWLS	SM	GPIUS	0.40, −0.20
Yu et al. (2016)	Macao	395	19.05	0.63	loneliness	UCLA	SNSs	Bergen	0.24
Yurdagül et al. (2019)	Turkey	491	15.92	0.59	loneliness, depression, anxiety, social anxiety	UCLA, SDHS, STAI, SASA	Instagram	Bergen	0.14, 0.25, 0.22, 0.28
Zaffar et al. (2015)	Pakistan	150	NA	0.50	depression, anxiety, loneliness	PHQ-9, Severity Measure for Generalized Anxiety Disorder, UCLA	Facebook	Bergen	0.39, 0.51, 0.07

NA = not available; FM = proportion of females in the sample; MH = mental health indicator; MHMeas = measures of mental health; SM = type of social media measured; SMmeas = measures of social media; ES = effect size.

Effect sizes were obtained from the authors.

Mean correlation between problematic SM use and mental health

The same sample analyzed in Andreassen et al. (2016) and Andreassen et al. (2017) had over 23,000 participants, and that in Boer et al. (2020) had 154,981 participants. As the inverse variance (N-3) was used as a weight to compute the mean correlation, these two studies would receive extremely large weights. However, the common winsorization method, which recodes the sample size at two or three standard deviations above the mean (Lipsey & Wilson, 2001) was not appropriate, as the standard deviation shown in Table 1 was also extremely large (13,536.57). The sample sizes of these two samples were set at 3 times the mean (5,519).

Table 3 lists the weighted mean correlations between problematic SM use and mental health indicators. Eighty-five effect sizes were related to well-being. Of these, 4 were for happiness, 30 for life satisfaction, 3 for positive affect, 2 for mental health, 42 for self-esteem, 3 for overall well-being and 1 for psychiatric well-being. As expected, the correlations between problematic SM use and well-being indicators were negative, ranging from −.11 to −.30. The mean correlations of problematic SM use with happiness, life satisfaction, positive affect and self-esteem were significantly different from 0. The homogeneity test for the correlation between problematic SM use and self-esteem was significant, indicating significant between-study variation.

Table 3.

Summary of mean correlations between problematic social media use and mental health.

	k	$\bar{r}$	95% CI		Q	p
			upper	lower
Well-being^a	85	−0.16	−0.20	−0.13	117.21	0.01**
Happiness	4	−0.30	−0.38	−0.21	2.96	0.40
Life satisfaction	30	−0.11	−0.18	−0.03	35.94	0.18
Positive affect	3	−0.18	−0.34	−0.01	0.73	0.69
Mental Health	2	−0.29	−0.71	0.28	0.07	0.80
Self-esteem	42	−0.17	−0.22	−0.13	57.26	0.05*
Overall well-being	3	−0.29	−0.65	0.17	4.54	0.10
Distress	136	0.27	0.25	0.29	150.37	0.17
Anxiety	17	0.30	0.25	0.35	17.00	0.39
Depression	59	0.31	0.28	0.33	46.69	0.86
Overall distress	4	0.27	0.01	0.49	3.36	0.34
Loneliness	29	0.21	0.17	0.25	25.34	0.61
Negative affect	4	0.08	−0.29	0.44	4.55	0.21
Social Anxiety	17	0.30	0.24	0.35	19.65	0.24
Social loneliness	2	0.19	−0.76	0.88	1.00	0.32
Suicidal ideation	3	0.18	0.02	0.34	0.53	0.77

One correlation between problematic use of social media and psychiatric well-being was reported and thus mean correlation was not computed.

p < 0.05; ^** p < 0.01.

Many studies examined the correlations between problematic SM use and distress indicators, and all mean correlations were positive. The relation between problematic SM use and depression attracted most research attention. The mean correlation was moderate at $\bar{r}$ = .31, indicating possible moderate detrimental impact of problematic SM use. Except for the correlations of problematic SM use with negative affect and social loneliness, other mean correlations were significantly different from 0.

Moderator analyses of the relation between problematic SM use and self-esteem

Due to insufficient numbers of effect sizes, moderator analyses were conducted for correlations of problematic SM use with self-esteem (k = 42), life satisfaction (k = 30), depression (k = 59), and loneliness (k = 29). The mean correlations were computed for categories of moderators with at least 4 effect sizes. Table 4 lists the categorical moderator effects on the relation between problematic SM use and mental health. Regarding publication outlet, 1 correlation between problematic SM use and self-esteem was reported in a book chapter, 1 in a conference, 1 in a Master’s thesis, 4 in Doctoral dissertations and 35 in journals. The effect of publication outlet was not significant with Q_B = 3.43, indicating the absence of a file-drawer problem. For the country effect, 17 countries reported the relation between problematic SM use and self-esteem, and 3 of them had at least 4 effect sizes. Again, Q_B = 1.32 was not significant, indicating that the mean correlations of US, Poland, and Turkey were comparable.

Table 4.

Categorical moderator effects on the relations between problematic social media use and mental health.

Moderator	k	$\bar{r}$	95% CI		Q_B
Moderator	k	$\bar{r}$	lower	upper	Q_B
r (PUSM, SE)
Publication Status					3.43
Journal	35	−0.19	−0.24	−0.15
Dissertation	4	−0.06	−0.27	0.17
Country					1.32
US	10	−0.08	−0.18	0.02
Poland	4	−0.11	−0.31	0.10
Turkey	6	−0.16	−0.29	−0.03
SE Measures					0.94
SISES	7	−0.13	−0.25	−0.00
Rosenberg	33	−0.18	−0.23	−0.13
PUSM Measures					0.05
Bergen	12	−0.13	−0.24	−0.03
FIQ	8	−0.12	−0.25	0.02
Platform					2.03
Social media	22	−0.19	−0.25	−0.14
Facebook	18	−0.13	−0.20	−0.06
r (PUSM, LS)
Country					13.57^***
Turkey	4	−0.31	−0.45	−0.14
US	11	−0.08	−0.16	−0.01
PUSM Measures					0.15
Bergen	13	−0.09	−0.21	0.04
FIQ	7	−0.05	−0.24	0.15
Platform					0.08
Social media	12	−0.16	−0.25	−0.07
Facebook	12	−0.14	−0.24	−0.05
r (PUSM, DEP)
Country					2.21
Turkey	9	0.26	0.16	0.36
Germany	5	0.36	0.21	0.50
US	9	0.30	0.21	0.40
Netherlands	4	0.32	0.12	0.49
DEP Measures					0.58
PHQ-9	8	0.28	0.18	0.37
CES-D	10	0.31	0.23	0.39
DASS-21	11	0.32	0.24	0.39
SDHS	6	0.29	0.17	0.40
PUSM Measures					0.81
Bergen	30	0.30	0.26	0.34
FIQ	5	0.30	0.16	0.42
IAT	5	0.32	0.19	0.45
Xanidas	4	0.29	0.12	0.44
SMDS	4	0.34	0.17	0.48
Platform					0.66
Social media	35	0.30	0.26	0.34
Facebook	23	0.32	0.27	0.37
r (PUSM, LON)
Platform					1.92
Social media	10	0.24	0.18	0.30
Facebook	12	0.19	0.13	0.25

k = total number of correlations included in the analysis; Q_B = between-group homogeneity statistic; r= correlation coefficient; PUSM = problematic use of social media; SE = self-esteem; LS = life satisfaction; DEP = depression; LON = loneliness; SE Measures = self-esteem measures, SISES = Single Item Self-Esteem Scale, Rosenberg = Rosenberg Self-Esteem Scale; Platform = platform users addicted to; PUSM Measures = measures of problematic use of social media, Bergen = Bergen Facebook Addiction Scale, Bergen Social Media addiction Scale and their adaptations, FIQ = Facebook Intrusion Questionnaire and its adaptations; IAT = adaptations of Internet Addiction Test, Xanidas = Xanidas & Brignell (2016), SMDS = Social Media Disorder Scale, DEP Measures = depression measures, PHQ-9 = Patient Health Questionnaire, CES-D = Center for Epidemiologic Studies-depression scale, DASS-21 = Depression Anxiety and Stress Scales-21, SDHS = Short Depression-Happiness Scale.

***

p < 0.001.

The majority of studies used the Rosenberg Self-Esteem Scale (Rosenberg, 1965) to assess global self-worth, and the mean correlation for these studies was $\bar{r}$ = −.18. Seven studies used the Single Item Self-Esteem Scale (Robins et al., 2001) and these had a mean correlation of $\bar{r}$ = −.13. The correlation between problematic SM use and self-esteem did not vary with self-esteem measure. The included studies used 18 measures of problematic SM use. Twelve samples used the Bergen Social Media Addiction Scale (Andreassen et al., 2016), Bergen Facebook Addiction Scale (Andreassen et al., 2012) or their adaptations, and 8 samples used the Facebook Intrusion Questionnaire (Elphinston & Noller, 2011) and its adaptations. The mean correlations of these measures were quite comparable.

Twenty-two effect sizes were related to problematic use in general; 18 were for problematic Facebook use; 1 for problematic Twitter use, and 1 for problematic WhatsApp use. The platform to which users were addicted did not exert a significant effect on the relation between problematic SM use and self-esteem.

Table 5 presents the effects of mean age and gender composition of the sample on the correlations between problematic SM use and mental health indicators. The mean age was reported in 38 samples, and the age effect was not significant. The effect of gender composition was significant, and the meta-regression model was $\bar{r}$ = −.38 + .36 × proportion of females in the sample. Thus, the expected correlation was $\bar{r}$ = −.38 for all-male samples, and $\bar{r}$ = −.02 for all-female samples.

Table 5.

Effects of continuous moderators.

correlation	Mean age			females
correlation	k	a	b	k	a	b
r (PUSM, SE)	38	−0.23	0.00	40	−0.38	0.36^*
r (PUSM, LS)	28	−0.32	0.01	30	−0.01	−0.16
r (PUSM, DEP)	57	0.29	0.00	58	0.39	−0.12
r (PUSM, LON)	28	0.18	0.00	26	0.29	−0.14

r = correlation coefficient; PUSM = problematic use of social media; SE = self-esteem; LS = life satisfaction; DEP = depression; LON = loneliness; female = proportion of females in the sample.

p < 0.05.

Moderator analyses the relation between problematic SM use and life satisfaction

The effect of publication status was not examined, because journal articles were the only outlet having more than 3 effect sizes. For the country effect, Turkey and US are the only two countries with more than 3 effect sizes and the mean correlations were moderate and small, respectively ( $\bar{r}$ = −.31 and $\bar{r}$ = −.08, respectively). The country difference was significant with Q_B = 13.57. The mean correlations were not compared among different life satisfaction measures, as the Satisfaction with Life Scale (Diener et al., 1985) was the only measure having more than 3 effect sizes. The effect of measure of problematic SM use was not significant, and the means for studies using the Bergen and FIQ were both small. Similarly, the effects of platform and of two continuous moderators were not significant.

Moderator analyses the relation between problematic SM use and depression

The effect of publication status cannot be examined, because 56 effect sizes were reported in journal articles, and 1 effect size was each reported in a book chapter, thesis, and dissertation. As shown in Table 4, four countries, namely Turkey, Germany, US and Netherlands, had more than 3 effect sizes, and the mean correlations for these countries were about moderate. The effect of study country was not significant.

The included studies used 19 depression measures, of which 4 had more than 3 effect sizes. The effect of depression measure was not significant, and mean correlations for these measures were comparable. Regarding measures of problematic SM use, most studies used the Bergen Social Media Addiction Scale, Bergen Facebook Addiction Scale and their adaptations. The effect of measure of problematic SM use was minimal as the mean correlations were similar.

Most studies assessed global problematic use, and Facebook was the only specific platform attracted many researchers’ attention. Again, the platform did not moderate the relation between problematic SM use and depression. The effect of mean age was minimal and not significant, as the regression coefficient was close to 0. Similarly, gender composition was not related to the magnitude of the correlation between problematic SM use and depression.

Moderator analyses the relation between problematic SM use and loneliness

Journal articles had 24 effect sizes, and other publication outlets had less than 4 effect sizes. Therefore, the effect of publication status was not examined. Similarly, the effects of study country and measures of loneliness and problematic SM use were not examined, as only one category of moderators had sufficient effect sizes. As shown in Table 4, the effect of platform was not related to the magnitude of the correlation between problematic SM use and loneliness. As shown in Table 5, the mean age and gender composition of the sample did not exert a significant effect.

Discussion

The psychological and social impact of problematic SM use has drawn widespread research attention. Although previous meta-analyses (Marino et al., 2018a, 2018b) were conducted to quantitatively synthesize these empirical findings, they did not address some research gaps. Those meta-analyses were narrow in scope, as they focused on problematic Facebook use. This study broadened the research scope by including studies examining problematic use of all platforms. This undertaking leaded to the number of included studies is more than 5 times of that in Marino et al. (2018a) and more than 15 times that in Marino et al. (2018b). As this study has a large number of effect sizes, it has more stable estimates of mean correlations than previous research. Furthermore, as the number of effect size was small in previous meta-analyses and thus they examined a limited number of moderator effects. For a complete assessment of moderators, moderators including measures of problematic SM use and mental health, and platform to which users were addicted were incorporated to understand whether research contexts can moderate the correlation. Finally, a small number of mental health indicators were included in the previous meta-analyses. For example, Marino et al. (2018a) examined only 5 indictors, namely psychological distress, depression, anxiety, general well-being and life satisfaction. The mean correlations of 14 indicators were computed in this study. Thus, the current meta-analysis has significant contributions beyond previous research.

The findings of this meta-analysis support a detrimental effect of problematic SM use. Specifically, the correlations between problematic SM use and various well-being indicators were negative, ranging from small to moderate (Cohen, 1992). The correlations between problematic SM use and various distress indicators were positive and also ranged from small to moderate (Cohen, 1992). The correlations of problematic SM use with anxiety, depression, and distress were comparable to those in Marino et al. (2018a).

The correlations of problematic SM use with self-esteem, life satisfaction, loneliness and depression had sufficiently large numbers of effect sizes, so their moderator effects were examined. The magnitude of correlation between problematic SM use and self-esteem was stronger for samples with more male participants. This finding indicates the negative effect of problematic SM use on self-esteem is relatively severe for males, and thus they are especially in need of help. Previous research (Pluhar et al., 2019; Stevens et al., 2019; Winkler et al., 2013) shows cognitive behavioral therapy, dialectical behavior therapy, group therapy and multidimensional treatment were effective in treatment of Internet addiction and online gaming addiction. These treatments can be considered and modified to treat male users at high risk of SM addiction.

The study country was related to the correlation between problematic SM use and life satisfaction. The correlation was moderate for Turkey, and small for US. Although the country effect on the relation between problematic Facebook use and psychological distress was significant, Marino et al. (2018a) found that Western countries had larger effect sizes than Asian countries. The inconsistent findings can be caused by the composition of both the Western and Asian country categories in Marino et al. (2018a). In Marino et al. (2018a), US only contributed one effect size in the Western country category, and Asian countries included Turkey, Taiwan and Thailand.

Implications, limitations, and future directions

This meta-analysis has important implications for practice and research. As the magnitudes of correlations varied by mental health indicator, future practitioners and researchers should use multiple indicators for a comprehensive assessment of the effect of problematic SM use on well-being and distress. The magnitudes of these correlations did not vary with publication status, instruments, SM platforms or mean age. The strength of correlation between problematic SM use and mental health was similar in most research conditions. Thus, the magnitude of these correlations can generalize across most moderator conditions, supporting the stability of correlations between problematic SM use and mental health indicators in most research conditions. Moreover, the moderating effect of study country was investigated to examine possible cultural differences. Turkish studies revealed a relatively strong correlation between SM use and life satisfactions. Since some countries had few studies, more international research should be conducted to examine the possible country effect.

This study searched articles published in English. Different languages or countries might have different dominant SM platforms. Few included studies investigated main SM platforms other than Facebook from different countries (e.g. WeChat in China). Therefore, future research should examine whether the moderating effect of SM platform is different across cultures. As each SM platform has distinctive features, the country effect may be confounded with platform effect.

One limitation of this study was that it treated multiple effect sizes from a single sample independent. For example, correlations between multiple measures of problematic SM use and multiple mental health indicators were all coded. Moreover, some moderator effects could not be examined, because those moderators did not have enough effect sizes to warrant stable estimates of mean correlations. Lastly, systematic reviews tend to use multiple reviewers or readers (Buscemi et al., 2005). Only one reader performed data extraction and coding, possibly leading to errors.

Footnotes

Appendix

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding for this study was provided by Ministry of Science and Technology (MOST) of the Republic of China, Taiwan Grant No. 108-2511-H-018-026.

ORCID iD

Chiungjung Huang

References

Aladwani

A. M.

Almarzouq

(2016). Understanding compulsive social media use: The premise of complementing self-conceptions mismatch with technology. Computers in Human Behavior, 60, 575–581. https://doi.org/10.1016/j.chb.2016.02.098

American Psychiatric Association (1994). Diagnostic and statistical manual of mental disorders, DSM-IV (4th ed.). APA.

Andreassen

C. S.

(2015). Online social network site addiction: A comprehensive review. Current Addiction Reports, 2, 175–184. https://doi.org/10.1007/s40429-015-0056-9

Andreassen

C. S.

Billieux

Griffiths

M. D.

Kuss

D. J.

Demetrovics

Mazzoni

Pallesen

(2016). The relationship between addictive use of social media and video games and symptoms of psychiatric disorders: A large-scale cross-sectional study. Psychology of Addictive Behaviors, 30, 252–262. https://doi.org/10.1037/adb0000160

Andreassen

C. S.

Pallesen

Griffiths

M. D.

(2017). The relationship between addictive use of social media, narcissism, and self-esteem: Findings from a large national survey. Addictive Behaviors, 64, 287–293. https://doi.org/10.1016/j.addbeh.2016.03.006

Andreassen

C. S.

Torsheim

Brunborg

G. S.

Pallesen

(2012). Development of a Facebook addiction scale. Psychological Reports, 110, 501–517. https://doi.org/10.2466/02.09.18.PR0.110.2.501-517

Bagby

R. M.

Ryder

A. G.

Schuller

D.R.

Marshall

M. B.

(2004). The Hamilton Depression Rating Scale: Has the gold standard become a lead weight? American Journal of Psychiatry, 161, 2163–77. https://doi.org/10.1176/appi.ajp.161.12.2163

Baturay

M. H.

Toker

(2017). Self-esteem shapes the impact of GPA and general health on Facebook addiction: A mediation analysis. Social Science Computer Review, 35, 555–575. https://doi.org/10.1177/08944393166566

Beck

A. T.

Steer

R. A.

Brown

G. K.

(1996). Manual for the Beck Depression Inventory–II. San Antonio, TX: Psychological Corporation.

10.

Beck

A. T.

Steer

R. A.

(1987). Beck Depression Inventory manual. New York: Psychological Corporation.

11.

Biolcati

Mancini

Pupi

Mugheddu

(2018). Facebook Addiction: Onset Predictors. Journal of Clinical Medicine, 7, 118. https://doi.org/10.3390/jcm7060118

12.

Błachnio

Przepiórka

Pantic

(2015). Internet use, Facebook intrusion, and depression: Results of a cross-sectional study. European Psychiatry, 30, 681–684. https://doi.org/10.1016/j.eurpsy.2015.04.00

13.

Brailovskaia

Teismann

Margraf

(2018). Physical activity mediates the association between daily stress and Facebook addiction disorder (FAD)—A longitudinal approach among German students. Computers in Human Behavior, 86, 199–204. https://doi.org/10.1016/j.chb.2018.04.045

14.

Buscemi

Hartling

Vandermeer

Tjosvold

Klassen

T. P.

(2005). Single data extraction generated more errors than double data extraction in systematic reviews. Journal of Clinical Epidemiology, 59, 697–703. https://doi.org/10.1016/j.jclinepi.2005.11.010

15.

Caldiroli

Serati

Buoli

(2018). Is Internet addiction a clinical symptom or a psychiatric disorder? A comparison with bipolar disorder. The Journal of Nervous and Mental Disease, 206, 644–656. https://doi.org/10.1097/NMD.0000000000000861.

16.

Caplan

S. E.

(2010). Theory and measurement of generalized problematic Internet use: A two-step approach. Computers in Human Behavior, 26, 1089–1097. https://doi.org/10.1016/j.chb.2010.03.012

17.

Card

N. A.

(2012). Applied meta-analysis for social science research. New York: Guilford Press.

18.

Carr

C. T.

Hayes

R. A.

(2015). Social media: Defining, developing, and divining. Atlantic Journal of Communication, 23, 46–65. https://doi.org/10.1080/15456870.2015.972282

19.

Chabrol

Laconi

Delfour

Moreau

(2017). Contributions of psychopathological and interpersonal variables to problematic Facebook use in adolescents and young adults. International Journal of High Risk Behavioral Addiction, 6, e32773. https://doi.org/10.5812/ijhrba.32773.

20.

Chamberlain

Lochner

Stein

Goudriaan

Holst

Zohar

Grant

J. E.

(2016). Behavioural addiction-a rising tide? European Neuropsychopharmacology, 26, 841–855. https://doi.org/10.1016/j.euroneuro.2015.08.013

21.

Choi

S. B.

Lim

M. S.

(2016). Effects of social and technology overload on psychological well-being in young South Korean adults: The mediatory role of social network service addiction. Computers in Human Behavior, 61, 245–254. https://doi.org/10.1016/j.chb.2016.03.032

22.

Cohen

(1992). A power primer. Psychological Bulletin, 112, 155–159. https://doi.org/10.1037/0033-2909.112.1.155

23.

De Cock

Vangeel

Klein

Minotte

Rosas

Meerkerk

. (2014). Compulsive use of social networking sites in Belgium: Prevalence, profile, and the role of attitude toward work and school. Cyberpsychology, Behavior, and Social Networking, 17, 166–171. https://doi.org/10.1089/cyber.2013.0029

24.

Diener

E. D.

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The Satisfaction With Life Scale. Journal of Personality Assessment, 49, 71–75. https://doi.org/10.1207/s15327752jpa4901_13

25.

Durak

H. Y.

(2018). Modeling of variables related to problematic internet usage and problematic social media usage in adolescents. Current Psychology, 39(4), 1375–1387. https://doi.org/10.1007/s12144-018-9840-8

26.

Dussault

Fernet

Austin

Leroux

(2009). Revisiting the factorial validity of the Revised UCLA Loneliness Scale: A test of cCompeting models in a sample of teachers. Psychological Reports, 105, 849–856. https://doi.org/10.2466/PR0.105.3.849-856

27.

Elphinston

R. A.

Noller

(2011). Time to face it! Facebook intrusion and the implications for romantic jealousy and relationship satisfaction. Cyberpsychology, Behavior, and Social Networking, 14, 631–635. https://doi.org/10.1089/cyber.2010.0318.

28.

Faber

R. J.

O’Guinn

T. C.

(1992). A clinical screener for compulsive buying. Journal of Consumer Research 1992, 19(3), 459–469. https://doi.org/10.1086/209315

29.

Ferris

Hollenbaugh

. (2018). A Uses and Gratifications Approach to Exploring Antecedents to Facebook Dependency. Journal of Broadcasting & Electronic Media, 62, 51–70. https://doi.org/10.1080/08838151.2017.1375501.

30.

Frost

R. L.

Rickwood

D. J.

(2017). A systematic review of the mental health outcomes associated with Facebook use. Computers in Human Behavior, 76, 576–600. https://doi.org/10.1016/j.chb.2017.08.001

31.

Giota

K. G.

Kleftaras

(2013). The role of personality and depression in problematic use of social networking sites in Greece. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 7(3), article 1. https://doi.org/10.5817/CP201336

32.

Griffiths

(2000). Does Internet and computer “addiction” exist? Some case study evidence. CyberPsychology & Behavior, 3, 211–218. https://doi.org/10.1089/109493100316067.

33.

Hamilton

(1960). A rating scale for depression. Journal of Neurology, Neurosurgery, and Psychiatry, 23, 56–62. https://doi.org/10.1136/jnnp.23.1.56

34.

Hamilton

(1967). Development of a rating scale for primary depressive illness. British Journal of Social and Clinical Psychology, 6, 278–296. https://doi.org/10.1136/jnnp.23.1.56

35.

Hartshore

T. S.

(1993). Psychometric properties and confirmatory analysis of the UCLA Loneliness Scale. Journal of Personality Assessment, 61, 182–195. https://doi.org/10.1207/s15327752jpa6101_14

36.

Hawi

N. S.

Samaha

(2018): Identifying commonalities and differences in personality characteristics of Internet and social media addiction profiles: traits, self-esteem, and self-construal. Behaviour & Information Technology, 38(2), 110–119. https://doi.org/10.1080/0144929X.2018.1515984

37.

Heo

Murphy

C. F.

Meyers

B. S.

(2007). Relationship between the Hamilton Depression Rating Scale and the Montgomery- Åsberg Depression Rating Scale in depressed elderly: A m-analysis. American Journal of Geriatric Psychiatry, 15, 899–905.

38.

Hong

Huang

Lin

Chiu

(2014). Analysis of the psychological traits, Facebook usage, and Facebook addiction model of Taiwanese university students. Telematics and Informatics, 31, 597–606. https://doi.org/10.1016/j.tele.2014.01.001

39.

Jasso-Medrano

López-Rosales

(2018). Measuring the relationship between social media use and addictive behavior and depression and suicide ideation among university students. Computers in Human Behavior, 87, 183–191. https://doi.org/10.1016/j.chb.2018.05.003

40.

Joseph

Linley

P. A.

Harwood

Lewis

C. A.

McCollam

(2004). Rapid assessment of well-being: The Short Depression-Happiness Scale (SDHS). Psychology and Psychotherapy: Theory, Research and Practice, 77, 463–478. https://doi.org/10.1348/1476083042555406.

41.

Kanat-Maymon

Almog

Cohen

Amichai-Hamburger

(2018). Contingent self-worth and Facebook addiction. Computers in Human Behavior, 88, 227–235. https://doi.org/10.1016/j.chb.2018.07.011

42.

Kang

Y. S.

(2007). Effect of the body satisfaction and self respect for the job selection of the university students. Gyeongbuk, Korea: Graduate School of Daegu Haanny University.

43.

Keles

McCrae

Grealish

(2019). A systematic review: The influence of social media on depression, anxiety and psychological distress in adolescents. International Journal of Adolescence and Youth, 25(1), 79–93. https://doi.org/10.1080/02673843.2019.15908

44.

Ketharanathan

Hanwella

Weerasundera

de Silva

V. A.

(2016). Diagnostic validity and factor analysis of Montgomery-Asberg Depression Rating Scale in Parkinson disease population. Journal of Geriatric Psychiatry and Neurology, 29, 115–119. https://doi.org/10.1177/0891988715606232

45.

Kircaburun

(2016). Self-esteem, daily Internet use and social media addiction as predictors of depression among Turkish adolescents. Journal of Education and Practice, 7, 64–72.

46.

Kircaburun

Demetrovics

Tosuntaş

Ş. B.

(2019). Analyzing the links between problematic social media use, dark triad traits, and self-esteem. International Journal of Mental Health and Addiction, 17, 1496–1507. https://doi.org/10.1007/s11469-018-9900-1

47.

Kim

Park

(2015). Factors affecting Internet gaming addiction: SNS addiction tendencies, self-esteem, and interpersonal relationships among male middle school students. Indian Journal of Science and Technology, 8(S8), 212–218. https://doi.org/10.17485/ijst/2015/v8iS8/70509

48.

Koc

Gulyagci

(2013). Facebook addiction among Turkish college students: The role of psychological health, demographic, and usage characteristics. CyberPsychology, Behavior, and Social Networking, 16, 279–284. https://doi.org/10.1089/cyber.2012.0249

49.

Kroenke

Spitzer

R. L.

Williams

J. B.

(2001). The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine, 16, 606–613. https://doi.org/10.1046/j.1525-1497.2001.016009606.x

50.

Laconi

Verseillié

Chabrol

(2018). Exploration of the problematic Twitter and Facebook uses and their relationships with psychopathological symptoms among Facebook users. International Journal of High Risk Behaviors & Addiction, 7, e61775. https://doi.org/10.5812/ijhrba.61775.

51.

LaRose

Mastro

Eastin

M. S.

(2001). Understanding Internet usage: A social-cognitive approach to uses and gratifications. Social Science Computer Review, 19, 395–413. https://doi.org/10.1177/089443930101900401

52.

Lipsey

M. W.

Wilson

D. B.

(2001). Practical meta-analysis. Sage.

53.

Lovibond

S. H.

Lovibond

P. F.

(1995). Manual for the depression anxiety stress scales (2nd. Ed.). Psychology Foundation.

54.

Lukat

Margraf

Lutz

van der Veld

W. M.

Becker

E. S.

(2016). Psychometric properties of the Positive Mental Health Scale (PMH-scale). BMC Psychology, 4(1), 8. https://doi.org/10.1186/s40359-016-0111-x

55.

Marino

Gini

Vieno

Spada

M. M.

(2018a). The associations between problematic Facebook use, psychological distress and well-being among adolescents and young adults: A systematic review and meta-analysis. Journal of Affective Disorders, 226, 274–281. https://doi.org/10.1016/j.jad.2017.10.007

56.

Marino

Gini

Vieno

Spada

M. M.

(2018b). A comprehensive meta-analysis on problematic Facebook use. Computers in Human Behavior, 83, 262–277. https://doi.org/10.1016/j.chb.2018.02.009

57.

McShane

B. B.

Böckenholt

Hansen

K. T.

(2016). Adjusting for publication bias in meta-analysis: an evaluation of selection methods and some cautionary notes. Perspectives on Psychological Science, 11, 730–749. https://doi.org/10.1177/1745691616662243

58.

Meerkerk

(2007). Pwned by the Internet. Explorative research into the causes and consequences of compulsive internet use. Rotterdam, Netherland: Erasmus Universiteit Rotterdam.

59.

Miele

G. M.

Tilly

S. M.

First

Frances

(1990). The definition of dependence and behavioural addictions. British Journal of Addiction, 85(11), 1421–1423. https://doi.org/10.1111/j.1360-0443.1990.tb01623.x

60.

Montgomery

S. A.

Åsberg

(1979). A new depression scale designed to be sensitive to change. British Journal of Psychiatry, 134, 382–389. https://doi.org/10.1192/bjp.134.4.382

61.

Moqbel

Kock

(2018). Unveiling the dark side of social networking sites: Personal and work-related consequences of social networking site addiction. Information & Management, 55, 109–119. https://doi.org/10.1016/j.im.2017.05.001

62.

Pavot

Diener

(1993). Review of the satisfaction with life scale. Psychological Assessment, 5, 164–172. https://doi.org/10.1037/1040-3590.5.2.164

63.

Pluhar

Kavanaugh

J. R.

Levinson

J. A.

Rich

(2019). Problematic interactive media use in teens: comorbidities, assessment, and treatment. Psychology Research and Behavior Management, 12, 447–455. https://doi.org/10.2147/prbm.s208968

64.

Pontes

H. M.

(2017). Investigating the differential effects of social networking site addiction and internet gaming disorder on psychological health. Journal of Behavioral Addictions, 6, 601–610. https://doi.org/10.1556/2006.6.2017.075

65.

Radloff

L. S.

(1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385–401. https://doi.org/10.1177/014662167700100306

66.

Robins

R.W.

Hendin

H. M.

Trzesniewski

K. H.

(2001). Measuring global self-esteem: construct validation of a single-item measure and the Rosenberg self-esteem scale. Personality and Social Psychology Bulletin, 27, 151–161. https://doi.org/10.1177/0146167201272002

67.

Rosenberg

(1965). Society and the adolescent self-image. Princeton University Press.

68.

Russell

Peplau

L. A.

Cutrona

C. E.

(1980). The revised UCLA Loneliness Scale: Concurrent and discriminant validity evidence. Journal of Personality and Social Psychology, 39, 472–480. https://doi.org/10.1037/0022-3514.39.3.472

69.

Ryan

Chester

Reece

Xenos

(2014). The uses and abuses of Facebook: A review of Facebook addiction. Journal of Behavioral Addictions, 3, 133–148. https://doi.org/10.1556/JBA.3.2014.016

70.

Salokangas

R. K. R.

Poutanen

Stengård

(1995). Screening for depression in primary care: Development and validation of the depression scale, a screening instrument for depression. Acta Psychiatrica Scandinavica, 92(1), 10–16. https://doi.org/10.1111/j.1600-0447.1995.tb09536.x

71.

Santen

Gomeni

Danhof

Pasqua

O. D.

(2008). Sensitivity of the individual items of the Hamilton depression rating scale to response and its consequences for the assessment of efficacy. Journal of Psychiatric Research, 42, 1000–1009. https://doi.org/10.1016/j.jpsychires.2007.11.004

72.

Satici

S. A.

(2018). Facebook addiction and subjective well-being: A study of the mediating role of shyness and loneliness. International Journal of Mental Health and Addiction, 17(1), 41–55. https://doi.org/10.1007/s11469-017-9862-8

73.

Satici

S. A.

Uysal

(2015). Well-being and problematic Facebook use. Computers in Human Behavior, 49, 185–190. https://doi.org/10.1016/j.chb.2015.03.005

74.

Schmitt David

Allik

(2005). Simultaneous administration of the Rosenberg self-esteem scale in 53 nations: Exploring the universal and culture-specific features of global self-esteem. Journal of Personality and Social Psychology, 89, 623–642. https://doi.org/10.1037/0022-3514.89.4.623

75.

Stevens

M. W. R.

King

D. L.

Dorstyn

Delfabbro

P. H.

(2019). Cognitive-behavioral therapy for internet gaming disorder: a systematic review and meta-analysis. Clinical Psychology and Psychotherapy, 26, 191–203. https://doi.org/10.1002/cpp.2341

76.

Szabó

(2010). The short version of the Depression Anxiety Stress Scales (DASS-21): Factor structure in a young adolescent sample. Journal of Adolescence, 33, 1–8. https://doi.org/10.1016/j.adolescence.2009.05.014

77.

Tang

C. S.

Koh

Y. W.

Gan

(2017). Addiction to Internet use, online gaming, and online social networking among young adults in China, Singapore, and the United States. Asia-Pacific Journal of Public Health, 29, 673–682. https://doi.org/10.1177/1010539517739558

78.

Turel

Qahri-Saremi

(2016) Problematic use of social networking sites: antecedents and consequence from a dual system theory perspective. Journal of Management Information Systems, 33, 1087–1116. https://doi.org/10.1080/07421222.2016.1267529

79.

Uysal

Satici

S. A.

Akin

(2013). Mediating effect of Facebook addiction on the relationship between subjective vitality and subjective happiness. Psychological Reports, 113, 948–953. https://doi.org/10.2466/02.09.18.PR0.113x32z3

80.

van den Eijnden

Koning

Doornwaard

van Gurp

Bogt

T. T

. (2018). The impact of heavy and disordered use of games and social media on adolescents’ psychological, social, and school functioning. Journal of Behavioral Addiction, 7, 697–706. https://doi.org/10.1556/2006.7.2018.65

81.

van den Eijnden

Lemmens

J. S.

Valkenburg

P. M

. (2016). The social media disorder scale. Computers in Human Behavior, 61, 478–487. https://doi.org/10.1016/j.chb.2016.03.038

82.

Vangeel

De Cock

Klein

Minotte

Rosas

Meerkerk

(2016). Compulsive use of social networking sites among secondary school adolescents in Belgium. In Walrave

Ponnet

Vanderhoven

Haers

Segaert

(Eds.). Youth 2.0: Social media and adolescence (pp. 179–191). Springer International Publishing.

83.

Verma

Kumari

(2016). A study on addiction to social networking sites and psychological well being among working adults. International Journal of Humanities and Social Sciences, 5, 153–161.

84.

Wan

(2009). Gratifications and loneliness as predictors of campus-SNS websites addiction and usage pattern among Chinese college students. (Unpublished Master’s Thesis) Chinese University of Hong Kong, Hong Kong.

85.

Winkler

Dorsing

Rief

Shen

Glombiewski

J. A.

(2013). Treatment of internet addiction: a meta-analysis. Clinical Psychology Review, 33, 317–329. https://doi.org/10.1016/j.cpr.2012.12.005

86.

Worsley

J.D.

McIntyre

J. C.

Bentall

R. P.

Corcoran

. (2018). Childhood maltreatment and problematic social media use: The role of attachment and depression. Psychiatry Research, 267, 88–93. https://doi.org/10.1016/j.psychres.2018.05.023.

87.

Young

(1998). Internet addiction: The emergence of a new clinical disorder. CyberPsychology & Behavior, 1, 237–244. https://doi.org/10.1089/cpb.1998.1.237