Access to this full-text is provided by Frontiers.
Content available from Frontiers in Psychology
This content is subject to copyright.
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 1
REVIEW
published: 14 August 2020
doi: 10.3389/fpsyg.2020.01949
Edited by:
Changiz Mohiyeddini,
Oakland University William Beaumont
School of Medicine, United States
Reviewed by:
David Leiva,
University of Barcelona, Spain
Alexander Lithopoulos,
University of Victoria, Canada
*Correspondence:
Viktor Schønning
Viktor.Schonning@fhi.no
Specialty section:
This article was submitted to
Health Psychology,
a section of the journal
Frontiers in Psychology
Received: 11 March 2020
Accepted: 14 July 2020
Published: 14 August 2020
Citation:
Schønning V, Hjetland GJ,
Aarø LE and Skogen JC (2020) Social
Media Use and Mental Health
and Well-Being Among Adolescents –
A Scoping Review.
Front. Psychol. 11:1949.
doi: 10.3389/fpsyg.2020.01949
Social Media Use and Mental Health
and Well-Being Among Adolescents
– A Scoping Review
Viktor Schønning1*, Gunnhild Johnsen Hjetland1, Leif Edvard Aarø1and
Jens Christoffer Skogen1,2,3
1Department of Health Promotion, Norwegian Institute of Public Health, Bergen, Norway, 2Alcohol and Drug Research
Western Norway, Stavanger University Hospital, Stavanger, Norway, 3Faculty of Health Sciences, University of Stavanger,
Stavanger, Norway
Introduction: Social media has become an integrated part of daily life, with an
estimated 3 billion social media users worldwide. Adolescents and young adults are
the most active users of social media. Research on social media has grown rapidly, with
the potential association of social media use and mental health and well-being becoming
a polarized and much-studied subject. The current body of knowledge on this theme
is complex and difficult-to-follow. The current paper presents a scoping review of the
published literature in the research field of social media use and its association with
mental health and well-being among adolescents.
Methods and Analysis: First, relevant databases were searched for eligible studies
with a vast range of relevant search terms for social media use and mental health and
well-being over the past five years. Identified studies were screened thoroughly and
included or excluded based on prior established criteria. Data from the included studies
were extracted and summarized according to the previously published study protocol.
Results: Among the 79 studies that met our inclusion criteria, the vast majority (94%)
were quantitative, with a cross-sectional design (57%) being the most common study
design. Several studies focused on different aspects of mental health, with depression
(29%) being the most studied aspect. Almost half of the included studies focused on
use of non-specified social network sites (43%). Of specified social media, Facebook
(39%) was the most studied social network site. The most used approach to measuring
social media use was frequency and duration (56%). Participants of both genders were
included in most studies (92%) but seldom examined as an explanatory variable. 77%
of the included studies had social media use as the independent variable.
Conclusion: The findings from the current scoping review revealed that about 3/4 of the
included studies focused on social media and some aspect of pathology. Focus on the
potential association between social media use and positive outcomes seems to be rarer
in the current literature. Amongst the included studies, few separated between different
forms of (inter)actions on social media, which are likely to be differentially associated
with mental health and well-being outcomes.
Keywords: scoping review, social media, mental health, adolescence, well-being
Frontiers in Psychology | www.frontiersin.org 1August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 2
Schønning et al. Social Media Use and Mental
BACKGROUND
In just a few decades, the use of social media have permeated
most areas of our society. For adolescents, social media play a
particularly large part in their lives as indicated by their extensive
use of several different social media platforms (Ofcom, 2018).
Furthermore, the use of social media and types of platforms
offered have increased at such a speed that there is reason to
believe that scientific knowledge about social media in relation
to adolescents’ health and well-being is scattered and incomplete
(Orben, 2020). Nevertheless, research findings indicating the
potential negative effects of social media on mental health
and well-being are frequently reported in traditional media
(newspapers, radio, TV) (Bell et al., 2015). Within the scientific
community, however, there are ongoing debates regarding the
impact and relevance of social media in relation to mental
health and well-being. For instance, Twenge and Campbell (2019)
stated that use of digital technology and social media have
a negative impact on well-being, while Orben and Przybylski
(2019) argued that the association between digital technology
use and adolescent well-being is so small that it is more or less
inconsequential. Research on social media use is a new focus
area, and it is therefore important to get an overview of the
studies performed to date, and describe the subject matter studies
have investigated in relation to the effect of social media use on
adolescents mental health and well-being. Also, research gaps
in this emerging research field is important to highlight as it
may guide future research in new and meritorious directions.
A scoping review is therefore deemed necessary to provide a
foundation for further research, which in time will provide a
knowledge base for policymaking and service delivery.
This scoping review will help provide an overall
understanding of the main foci of research within the field
of social media and mental health and well-being among
adolescents, as well as the type of data sources and research
instruments used so far. Furthermore, we aim to highlight
potential gaps in the research literature (Arksey and O’Malley,
2005). Even though a large number of studies on social media
use and mental health with different vantage points has been
conducted over the last decade, we are not aware of any
broad-sweeping scoping review covering this area.
AIMS
This scoping review aims to give an overview of the main research
questions that have been focused on with regard to use of social
media among adolescents in relation to mental health and well-
being. Both quantitative and qualitative studies are of interest.
Three specific secondary research questions will be addressed and
together with the main research question serve as a template for
organizing the results:
•Which aspects of mental health and well-being have been
the focus or foci of research so far?
•Has the research focused on different research aims
across gender, ethnicity, socio-economic status, geographic
location? What kind of findings are reported across these
groups?
•Organize and describe the main sources of evidence related
to social media that have been used in the studies identified.
DEFINING ADOLESCENCE AND SOCIAL
MEDIA
In the present review, adolescents are defined as those between
13 and 19 years of age. We chose the mean age of 13 as our
lower limit as nearly all social media services require users to be
at least 13 years of age to access and use their services (Childnet
International, 2018). All pertinent studies which present results
relevant for this age range is within the scope of this review.
For social media we used the following definition by Kietzmann
et al. (2011, p. 1): “Social media employ mobile and web-
based technologies to create highly interactive platforms via
which individuals and communities share, co-create, discuss, and
modify user-generated content.” We also employed the typology
described by Kaplan and Haenlein’s classification scheme across
two axes: level of self-presentation and social presence/media
richness (Kaplan and Haenlein, 2010). The current scoping
review adheres to guidelines and recommendations stated
by Tricco et al. (2018).
See protocol for further details about the definitions used
(Schønning et al., 2020).
DATA SOURCES AND SEARCH
STRATEGY
A literature search was performed in OVID Medline, OVID
Embase, OVID PsycINFO, Sociological Abstracts (proquest),
Social Services Abstracts (proquest), ERIC (proquest), and
CINAHL. The search strategy combined search terms for
adolescents, social media and mental health or wellbeing. The
database-controlled vocabulary was used for searching subject
headings, and a large spectrum of synonyms with appropriate
truncations was used for searching title, abstract, and author
keywords. A filter for observational studies was applied to limit
the results. The search was also limited to publications from
2014 to current. The search strategy was translated between each
database. An example of full strategy for Embase is attached as
Supplementary Material.
STUDY SELECTION: EXCLUSION AND
INCLUSION CRITERIA
The exclusion and inclusion criteria are detailed in the protocol
(Schønning et al., 2020). Briefly, we included English language
peer-reviewed quantitative- or qualitative papers or systematic
reviews published within the last 5 years with an explicit focus
on mental health/well-being and social media. Non-empirical
studies, intervention studies, clinical studies and publications not
peer-reviewed were excluded. Intervention studies and clinical
Frontiers in Psychology | www.frontiersin.org 2August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 3
Schønning et al. Social Media Use and Mental
FIGURE 1 | Flowchart of exclusion process from unsorted results to included studies.
studies were excluded as we sought to not introduce too much
heterogeneity in design and our focus was on observational
studies. The criteria used for study selection was part of an
iterative process which was described in detail in the protocol
(Schønning et al., 2020). As per the study protocol (Schønning
et al., 2020), and in line with scoping review guidelines (Peters
et al., 2015, 2017;Tricco et al., 2018), we did not assess
methodological quality or risk of bias of the included studies.
The selection process is illustrated by a flow-chart indicating
the stages from unsorted search results to the number of included
studies (see Figure 1). Study selection was accomplished and
organized using the Rayyan QCRI software1. The inclusion
and exclusion process was performed independently by VS
and JCS. The interrater agreement was κ= 0.87, indicating
satisfactory agreement.
1https://rayyan.qcri.org/welcome
DATA EXTRACTION AND
ORGANIZATION
Details of the data extracted is described in the protocol. Three
types of information were extracted, bibliographic information,
information about study design and subject matter information.
Subject matter information included aim of study, how social
media and mental health/well-being was measured, and main
findings of the study.
Visualization of Words From the Titles of
the Included Studies
The most frequently occurring words and bigrams in the
titles of the included studies are presented in Figures 2,3.
The following procedure was used to generate Figure 1:
First, a text file containing all titles were imported into R
Frontiers in Psychology | www.frontiersin.org 3August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 4
Schønning et al. Social Media Use and Mental
FIGURE 2 | Word cloud from the titles of the included studies. Most frequent words, excluding variations of “adolescence” and “social media.” N= 113. Shades of
blue indicate word frequencies >2 and green a frequency of 2. The size of each word is indicative of its relative frequency of occurrence.
as a data frame (R Core Team, 2014). The data frame was
processed using the “tidy text”-package with required additional
packages (Silge and Robinson, 2016). Second, numbers and
commonly used words with little inherent meaning (so called
“stop words,” such as “and,” “of,” and “in”), were removed
from the data frame using the three available lexicons in the
“tidy-text”-package (Silge and Robinson, 2016). Furthermore,
variations of “adolescents” (e.g., “adolescent,” “adolescence,” and
“adolescents”) and “social media” (e.g., “social media,” “social
networking,” “online social networks”) were removed from the
data frame. Third, the resulting data frame was sorted based
on frequency of unique words, and words occurring only once
were removed. The final data frame is presented as a word
cloud in Figure 1 (N= 113). The same procedure as described
above was employed to generate commonly occurring bigrams
(two words occurring adjacent to each other), but without
removing bigrams occurring only once (N= 231). The word
clouds were generated using the “wordcloud2”-package in R
(Lang and Chien, 2018). For Figure 1, shades of blue indicate
word frequencies >2 and green a frequency of 2. For Figure 2,
shades of blue indicate bigram frequencies of >1 and green a
frequency of 1.
RESULTS
Characteristics of the Included Studies
Of 7927 unique studies, 79 (1%) met our inclusion criteria
(Aboujaoude et al., 2015;Banjanin et al., 2015;Banyai et al.,
2017;Barry et al., 2017;Best et al., 2014, 2015;Booker et al.,
2018;Bourgeois et al., 2014;Boyle et al., 2016;Brunborg et al.,
2017;Burnette et al., 2017;Colder Carras et al., 2017;Critchlow
et al., 2019;Cross et al., 2015;Curtis et al., 2018;de Lenne et al.,
2018;de Vries et al., 2016;Erfani and Abedin, 2018;Erreygers
et al., 2018;Fahy et al., 2016;Ferguson et al., 2014;Fisher et al.,
2016;Foerster and Roosli, 2017;Foody et al., 2017;Fredrick
and Demaray, 2018;Frison and Eggermont, 2016, 2017;Geusens
and Beullens, 2017, 2018;Hamm et al., 2015;Hanprathet et al.,
2015;Harbard et al., 2016;Hase et al., 2015;Holfeld and Mishna,
2019;Houghton et al., 2018;Jafarpour et al., 2017;John et al.,
2018;Kim et al., 2019;Kim, 2017;Koo et al., 2015;Lai et al.,
2018;Larm et al., 2017, 2019;Marchant et al., 2017;Marengo
et al., 2018;Marques et al., 2018;Meier and Gray, 2014;Memon
et al., 2018;Merelle et al., 2017;Neira and Barber, 2014;Nesi
et al., 2017a,b;Niu et al., 2018;Nursalam et al., 2018;Oberst
et al., 2017;O’Connor et al., 2014;O’Reilly et al., 2018;Przybylski
Frontiers in Psychology | www.frontiersin.org 4August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 5
Schønning et al. Social Media Use and Mental
FIGURE 3 | Word cloud from the titles of the included studies. Bigrams from the titles of the included studies, excluding variations of “adolescence” and “social
media.” N= 231. Shades of blue indicate bigram frequencies of >1 and green a frequency of 1. The size of each bigram is indicative of its relative frequency of
occurrence.
and Bowes, 2017;Przybylski and Weinstein, 2017;Richards
et al., 2015;Rousseau et al., 2017;Salmela-Aro et al., 2017;
Sampasa-Kanyinga and Chaput, 2016;Sampasa-Kanyinga and
Lewis, 2015;Sampasa-Kanyinga et al., 2018;Scott and Woods,
2018;Settanni et al., 2018;Spears et al., 2015;Throuvala et al.,
2019;Tiggemann and Slater, 2017;Tseng and Yang, 2015;Twenge
and Campbell, 2019;Twenge et al., 2018;van den Eijnden et al.,
2018;Wang et al., 2018;Wartberg et al., 2018;Wolke et al.,
2017;Woods and Scott, 2016;Yan et al., 2017). Among the
included studies, 74 (94%) are quantitative (Aboujaoude et al.,
2015;Banjanin et al., 2015;Banyai et al., 2017;Barry et al.,
2017;Best et al., 2014;Booker et al., 2018;Bourgeois et al., 2014;
Boyle et al., 2016;Brunborg et al., 2017;Colder Carras et al.,
2017;Critchlow et al., 2019;Cross et al., 2015;Curtis et al.,
2018;de Lenne et al., 2018;de Vries et al., 2016;Erfani and
Abedin, 2018;Erreygers et al., 2018;Fahy et al., 2016;Ferguson
et al., 2014;Fisher et al., 2016;Foerster and Roosli, 2017;Foody
et al., 2017;Fredrick and Demaray, 2018;Frison and Eggermont,
2016, 2017;Geusens and Beullens, 2017, 2018;Hamm et al.,
2015;Hanprathet et al., 2015;Harbard et al., 2016;Hase et al.,
2015;Houghton et al., 2018;Jafarpour et al., 2017;John et al.,
2018;Kim et al., 2019;Kim, 2017;Koo et al., 2015;Lai et al.,
2018;Larm et al., 2017, 2019;Marchant et al., 2017;Marengo
et al., 2018;Marques et al., 2018;Meier and Gray, 2014;Memon
Frontiers in Psychology | www.frontiersin.org 5August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 6
Schønning et al. Social Media Use and Mental
et al., 2018;Merelle et al., 2017;Neira and Barber, 2014;Nesi
et al., 2017a,b;Niu et al., 2018;Nursalam et al., 2018;Oberst
et al., 2017;O’Connor et al., 2014;Przybylski and Bowes, 2017;
Przybylski and Weinstein, 2017;Richards et al., 2015;Rousseau
et al., 2017;Salmela-Aro et al., 2017;Sampasa-Kanyinga and
Chaput, 2016;Sampasa-Kanyinga and Lewis, 2015;Sampasa-
Kanyinga et al., 2018;Scott and Woods, 2018;Settanni et al.,
2018;Spears et al., 2015;Tiggemann and Slater, 2017;Tseng and
Yang, 2015;Twenge and Campbell, 2019;Twenge et al., 2018;
van den Eijnden et al., 2018;Wang et al., 2018;Wartberg et al.,
2018;Wolke et al., 2017;Woods and Scott, 2016;Yan et al., 2017),
three are qualitative (O’Reilly et al., 2018;Burnette et al., 2017;
Throuvala et al., 2019), and two use mixed methods (Best et al.,
2015;Holfeld and Mishna, 2019) (see Supplementary Tables 1, 2
in the Supplementary Material for additional details extracted
from all included studies). In relation to study design, 45 (57%)
used a cross-sectional design (Bourgeois et al., 2014;Ferguson
et al., 2014;Meier and Gray, 2014;Neira and Barber, 2014;
O’Connor et al., 2014;Banjanin et al., 2015;Hanprathet et al.,
2015;Hase et al., 2015;Koo et al., 2015;Sampasa-Kanyinga and
Lewis, 2015;Spears et al., 2015;Tseng and Yang, 2015;Frison
and Eggermont, 2016;Sampasa-Kanyinga and Chaput, 2016;
Woods and Scott, 2016;Banyai et al., 2017;Barry et al., 2017;
Brunborg et al., 2017;Colder Carras et al., 2017;Larm et al.,
2017, 2019;Merelle et al., 2017;Oberst et al., 2017;Przybylski
and Bowes, 2017;Przybylski and Weinstein, 2017;Tiggemann
and Slater, 2017;Wolke et al., 2017;Yan et al., 2017;de Lenne
et al., 2018;Erreygers et al., 2018;Fredrick and Demaray, 2018;
Geusens and Beullens, 2018;Lai et al., 2018;Marengo et al.,
2018;Marques et al., 2018;Niu et al., 2018;Nursalam et al.,
2018;Sampasa-Kanyinga et al., 2018;Scott and Woods, 2018;
Settanni et al., 2018;Wang et al., 2018;Wartberg et al., 2018;
Critchlow et al., 2019;Kim et al., 2019;Twenge and Campbell,
2019), 17 used a longitudinal design (Cross et al., 2015;Boyle
et al., 2016;de Vries et al., 2016;Fahy et al., 2016;Frison and
Eggermont, 2016;Harbard et al., 2016;Foerster and Roosli, 2017;
Geusens and Beullens, 2017;Kim, 2017;Nesi et al., 2017a,b;
Rousseau et al., 2017;Salmela-Aro et al., 2017;Booker et al.,
2018;Houghton et al., 2018;van den Eijnden et al., 2018;Holfeld
and Mishna, 2019), seven were systematic reviews (Aboujaoude
et al., 2015;Best et al., 2015;Fisher et al., 2016;Marchant et al.,
2017;Erfani and Abedin, 2018;John et al., 2018;Memon et al.,
2018), two were meta-analyses (Foody et al., 2017:Curtis et al.,
2018), one was a causal-comparative study (Jafarpour et al., 2017),
one was a review article (Richards et al., 2015), one used a
time-lag design (Twenge et al., 2018), one was a scoping review
(Hamm et al., 2015), three used a focus-group interview design
(Burnette et al., 2017;O’Reilly et al., 2018;Throuvala et al., 2019),
and one study used a combined survey and focus-group design
(Best et al., 2014).
The most common study settings were schools [N= 42
(54%)] (Best et al., 2014;Bourgeois et al., 2014;Meier and
Gray, 2014;Neira and Barber, 2014;O’Connor et al., 2014;
Banjanin et al., 2015;Hanprathet et al., 2015;Hase et al., 2015;
Sampasa-Kanyinga and Lewis, 2015;Frison and Eggermont,
2016;Sampasa-Kanyinga and Chaput, 2016;Woods and Scott,
2016;Banyai et al., 2017;Brunborg et al., 2017;Colder Carras
et al., 2017;Foerster and Roosli, 2017;Geusens and Beullens,
2017, 2018;Kim, 2017;Larm et al., 2017, 2019;Merelle et al.,
2017;Nesi et al., 2017a,b;Przybylski and Bowes, 2017;Rousseau
et al., 2017;Salmela-Aro et al., 2017;Tiggemann and Slater, 2017;
de Lenne et al., 2018;Fredrick and Demaray, 2018;Houghton
et al., 2018;Lai et al., 2018;Marengo et al., 2018;Niu et al., 2018;
Nursalam et al., 2018;Sampasa-Kanyinga et al., 2018;Scott and
Woods, 2018;Settanni et al., 2018;van den Eijnden et al., 2018;
Wang et al., 2018;Holfeld and Mishna, 2019;Kim et al., 2019).
Fourteen of the included studies were based on data from a home
setting (Cross et al., 2015;Koo et al., 2015;Spears et al., 2015;
Boyle et al., 2016;de Vries et al., 2016;Harbard et al., 2016;
Barry et al., 2017;Frison and Eggermont, 2017;Oberst et al.,
2017;Yan et al., 2017;Booker et al., 2018;Marques et al., 2018;
Wartberg et al., 2018;Critchlow et al., 2019). Eleven publications
were reviews or meta-analyses and included primary studies
from different settings (Aboujaoude et al., 2015;Best et al., 2015;
Hamm et al., 2015;Richards et al., 2015;Fisher et al., 2016;Foody
et al., 2017;Marchant et al., 2017;Curtis et al., 2018;Erfani and
Abedin, 2018;John et al., 2018;Memon et al., 2018). One study
used both a home and school setting (Erreygers et al., 2018),
and 11 (14%) of the included studies did not mention the study
setting for data collection (Ferguson et al., 2014;Tseng and Yang,
2015;Fahy et al., 2016;Burnette et al., 2017;Jafarpour et al., 2017;
Przybylski and Weinstein, 2017;Wolke et al., 2017;O’Reilly et al.,
2018;Twenge et al., 2018;Throuvala et al., 2019;Twenge and
Campbell, 2019).
Mental Health Foci of Included Studies
For a visual overview of the mental health foci of the included
studies see Figures 2,3. Most studies had a focus on different
negative aspects of mental health, as evident from the frequently
used terms in Figures 2,3. The most studied aspect was
depression, with 23 (29%) studies examining the relationship
between social media use and depressive symptoms (Ferguson
et al., 2014;Neira and Barber, 2014;O’Connor et al., 2014;
Banjanin et al., 2015;Richards et al., 2015;Spears et al., 2015;
Tseng and Yang, 2015;Fahy et al., 2016;Frison and Eggermont,
2016, 2017;Woods and Scott, 2016;Banyai et al., 2017;Brunborg
et al., 2017;Colder Carras et al., 2017;Larm et al., 2017;Nesi
et al., 2017a;Salmela-Aro et al., 2017;Fredrick and Demaray,
2018;Houghton et al., 2018;Niu et al., 2018;Twenge et al.,
2018;Wang et al., 2018;Wartberg et al., 2018). Twenty of the
included studies focused on different aspects of good mental
health, such as well-being, happiness, or quality of life (Best et al.,
2014, 2015;Bourgeois et al., 2014;Ferguson et al., 2014;Cross
et al., 2015;Koo et al., 2015;Richards et al., 2015;Spears et al.,
2015;Fahy et al., 2016;Foerster and Roosli, 2017;Przybylski and
Bowes, 2017;Przybylski and Weinstein, 2017;Yan et al., 2017;
Booker et al., 2018;de Lenne et al., 2018;Erfani and Abedin,
2018;Erreygers et al., 2018;Lai et al., 2018;van den Eijnden
et al., 2018;Twenge and Campbell, 2019). Nineteen studies had a
more broad-stroke approach, and covered general mental health
or psychiatric problems (Aboujaoude et al., 2015;Hanprathet
et al., 2015;Hase et al., 2015;Sampasa-Kanyinga and Lewis,
2015;Spears et al., 2015;Fisher et al., 2016;Barry et al., 2017;
Jafarpour et al., 2017;Kim, 2017;Merelle et al., 2017;Oberst
Frontiers in Psychology | www.frontiersin.org 6August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 7
Schønning et al. Social Media Use and Mental
et al., 2017;Wolke et al., 2017;Marengo et al., 2018;Marques
et al., 2018;Sampasa-Kanyinga et al., 2018;Holfeld and Mishna,
2019;Kim et al., 2019;Larm et al., 2019). Eight studies examined
the link between social media use and body dissatisfaction and
eating disorder symptoms (Ferguson et al., 2014;Meier and
Gray, 2014;de Vries et al., 2016;Burnette et al., 2017;Rousseau
et al., 2017;Tiggemann and Slater, 2017;Marengo et al., 2018;
Wartberg et al., 2018). Anxiety was the focus of seven studies
(O’Connor et al., 2014;Koo et al., 2015;Spears et al., 2015;
Fahy et al., 2016;Woods and Scott, 2016;Colder Carras et al.,
2017;Yan et al., 2017), and 13 studies included a focus on the
relationship between alcohol use and social media use (O’Connor
et al., 2014;Boyle et al., 2016;Sampasa-Kanyinga and Chaput,
2016;Brunborg et al., 2017;Geusens and Beullens, 2017, 2018;
Larm et al., 2017;Merelle et al., 2017;Nesi et al., 2017b;Curtis
et al., 2018;Sampasa-Kanyinga et al., 2018;Critchlow et al.,
2019;Kim et al., 2019). Seven studies examined the effect of
social media use on sleep (Harbard et al., 2016;Woods and
Scott, 2016;Yan et al., 2017;Nursalam et al., 2018;Sampasa-
Kanyinga et al., 2018;Scott and Woods, 2018;Larm et al.,
2019). Five studies saw how drug use and social media use
affected each other (O’Connor et al., 2014;Merelle et al., 2017;
Sampasa-Kanyinga et al., 2018;Kim et al., 2019;Larm et al.,
2019). Self-harm and suicidal behavior was the focus of eleven
studies (O’Connor et al., 2014;Sampasa-Kanyinga and Lewis,
2015;Tseng and Yang, 2015;Kim, 2017;Marchant et al., 2017;
Merelle et al., 2017;Fredrick and Demaray, 2018;John et al., 2018;
Memon et al., 2018;Twenge et al., 2018;Kim et al., 2019). Other
areas of focus other than the aforementioned are loneliness, self-
esteem, fear of missing out and other non-pathological measures
(Neira and Barber, 2014;Banyai et al., 2017;Barry et al., 2017;
Colder Carras et al., 2017).
Social Media Metrics of Included Studies
The studies included in the current scoping review often focus on
specific, widely used, social media and social networking services,
such as 31 (39%) studies focusing on Facebook (Bourgeois et al.,
2014;Meier and Gray, 2014;Banjanin et al., 2015;Cross et al.,
2015;Hanprathet et al., 2015;Richards et al., 2015;Sampasa-
Kanyinga and Lewis, 2015;Spears et al., 2015;Boyle et al., 2016;
de Vries et al., 2016;Frison and Eggermont, 2016;Harbard
et al., 2016;Sampasa-Kanyinga and Chaput, 2016;Banyai et al.,
2017;Barry et al., 2017;Brunborg et al., 2017;Larm et al., 2017;
Merelle et al., 2017;Nesi et al., 2017a,b;Rousseau et al., 2017;
Tiggemann and Slater, 2017;Booker et al., 2018;de Lenne et al.,
2018;Lai et al., 2018;Marengo et al., 2018;Marques et al., 2018;
Memon et al., 2018;Sampasa-Kanyinga et al., 2018;Settanni
et al., 2018;Twenge et al., 2018), 11 on Instagram (Sampasa-
Kanyinga and Lewis, 2015;Boyle et al., 2016;Sampasa-Kanyinga
and Chaput, 2016;Barry et al., 2017;Brunborg et al., 2017;
Frison and Eggermont, 2017;Nesi et al., 2017a;Marengo et al.,
2018;Memon et al., 2018;Sampasa-Kanyinga et al., 2018), 11
including Twitter (Richards et al., 2015;Sampasa-Kanyinga and
Lewis, 2015;Spears et al., 2015;Harbard et al., 2016;Sampasa-
Kanyinga and Chaput, 2016;Barry et al., 2017;Brunborg et al.,
2017;Merelle et al., 2017;Nesi et al., 2017a;Memon et al.,
2018;Sampasa-Kanyinga et al., 2018), and five studies asking
about Snapchat (Boyle et al., 2016;Barry et al., 2017;Brunborg
et al., 2017;Nesi et al., 2017a;Marengo et al., 2018). Eight
studies mentioned Myspace (Richards et al., 2015;Sampasa-
Kanyinga and Lewis, 2015;de Vries et al., 2016;Harbard et al.,
2016;Sampasa-Kanyinga and Chaput, 2016;Larm et al., 2017;
Booker et al., 2018;Sampasa-Kanyinga et al., 2018) and two
asked about Tumblr (Barry et al., 2017;Nesi et al., 2017a).
Other media such as Skype (Merelle et al., 2017), Youtube
(Richards et al., 2015), WhatsApp (Brunborg et al., 2017), Ping
(Merelle et al., 2017), Bebo (Booker et al., 2018), Hyves (de
Vries et al., 2016), Kik (Brunborg et al., 2017), Ask (Brunborg
et al., 2017), and Qzone (Niu et al., 2018) were only included
in one study each.
Almost half (n= 34, 43%) of the included studies focus on
use of social network sites or online communication in general,
without specifying particular social media sites, leaving this up
to the study participants to decide (Best et al., 2014, 2015;
Ferguson et al., 2014;Neira and Barber, 2014;O’Connor et al.,
2014;Koo et al., 2015;Tseng and Yang, 2015;Fahy et al., 2016;
Woods and Scott, 2016;Burnette et al., 2017;Colder Carras et al.,
2017;Foerster and Roosli, 2017;Foody et al., 2017;Geusens and
Beullens, 2017, 2018;Jafarpour et al., 2017;Kim, 2017;Marchant
et al., 2017;Oberst et al., 2017;Przybylski and Weinstein, 2017;
Salmela-Aro et al., 2017;Yan et al., 2017;Curtis et al., 2018;Erfani
and Abedin, 2018;Erreygers et al., 2018;Nursalam et al., 2018;
Scott and Woods, 2018;van den Eijnden et al., 2018;Wartberg
et al., 2018;Critchlow et al., 2019;Holfeld and Mishna, 2019;
Larm et al., 2019;Throuvala et al., 2019;Twenge and Campbell,
2019). Seven of the included studies examined the relationship
between virtual game worlds or socially oriented video games and
mental health (Ferguson et al., 2014;Best et al., 2015;Spears et al.,
2015;Yan et al., 2017;van den Eijnden et al., 2018;Larm et al.,
2019;Twenge and Campbell, 2019).
In the 79 studies included in this scoping review, several
approaches to measuring social media use are utilized. The
combination of frequency and duration of social media use is
by far the most used measurement of social media use, and 44
(56%) of the included studies collected data on these parameters
(Ferguson et al., 2014;Meier and Gray, 2014;Neira and Barber,
2014;Banjanin et al., 2015;Best et al., 2015;Hanprathet et al.,
2015;Sampasa-Kanyinga and Lewis, 2015;Tseng and Yang, 2015;
Boyle et al., 2016;de Vries et al., 2016;Frison and Eggermont,
2016, 2017;Harbard et al., 2016;Sampasa-Kanyinga and Chaput,
2016;Woods and Scott, 2016;Banyai et al., 2017;Brunborg
et al., 2017;Colder Carras et al., 2017;Foerster and Roosli,
2017;Jafarpour et al., 2017;Kim, 2017;Larm et al., 2017, 2019;
Merelle et al., 2017;Nesi et al., 2017b;Oberst et al., 2017;
Rousseau et al., 2017;Tiggemann and Slater, 2017;Yan et al.,
2017;Booker et al., 2018;de Lenne et al., 2018;Erreygers et al.,
2018;Houghton et al., 2018;Lai et al., 2018;Marengo et al.,
2018;Marques et al., 2018;Niu et al., 2018;Nursalam et al.,
2018;Sampasa-Kanyinga et al., 2018;Scott and Woods, 2018;
Settanni et al., 2018;Twenge et al., 2018;van den Eijnden et al.,
2018;Twenge and Campbell, 2019). Eight studies focused on
the relationship between social media addiction or excessive use
and mental health (Banjanin et al., 2015;Tseng and Yang, 2015;
Banyai et al., 2017;Merelle et al., 2017;Nursalam et al., 2018;
Frontiers in Psychology | www.frontiersin.org 7August 2020 | Volume 11 | Article 1949
fpsyg-11-01949 August 12, 2020 Time: 19:56 # 8
Schønning et al. Social Media Use and Mental
Settanni et al., 2018;Wang et al., 2018). Bergen Social Media
Addiction Scale is a commonly used questionnaire amongst the
included studies (Hanprathet et al., 2015;Banyai et al., 2017;
Settanni et al., 2018). Seven studies asked about various specific
actions on social media, such as liking or commenting on photos,
posting something or participating in a discussion (Meier and
Gray, 2014;Koo et al., 2015;Nesi et al., 2017b;Geusens and
Beullens, 2018;Marques et al., 2018;van den Eijnden et al., 2018;
Critchlow et al., 2019).
Five studies had a specific and sole focus on the link between
social media use and alcohol, and examined how various alcohol-
related social media use affected alcohol intake (Boyle et al.,
2016;Geusens and Beullens, 2017, 2018;Nesi et al., 2017b;
Critchlow et al., 2019). Some studies had a more theory-based
focus and investigated themes such as peer comparison, social
media intrusion or pro-social behavior on social media and its
effect on mental health (Bourgeois et al., 2014;Rousseau et al.,
2017;de Lenne et al., 2018). One of the included studies looked
into night-time specific social media use (Scott and Woods, 2018)
and one looked into pre-bedtime social media behavior (Harbard
et al., 2016) to study the link between this use and sleep.
Amongst the 79 included studies, only six (8%) studies had
participants of one gender (Ferguson et al., 2014;Meier and Gray,
2014;Best et al., 2015;Burnette et al., 2017;Jafarpour et al.,
2017;Tiggemann and Slater, 2017). Sixteen studies (20%) did not
mention the gender distribution of the participants (Aboujaoude
et al., 2015;Best et al., 2015;Hamm et al., 2015;Richards et al.,
2015;Fisher et al., 2016;Woods and Scott, 2016;Foody et al.,
2017;Marchant et al., 2017;Przybylski and Weinstein, 2017;
Curtis et al., 2018;Erfani and Abedin, 2018;John et al., 2018;
Memon et al., 2018;O’Reilly et al., 2018;Twenge et al., 2018;
Twenge and Campbell, 2019). Several of these were meta-analyses
or reviews (Aboujaoude et al., 2015;Best et al., 2014;Curtis
et al., 2018;Foody et al., 2017;John et al., 2018;Erfani and
Abedin, 2018;Wallaroo, 2020). The studies that included both
genders as participants generally had a well-balanced gender
distribution with no gender below 40% of the participants. Eight
of the studies did not report gender-specific results (Harbard
et al., 2016;Nesi et al., 2017b;Curtis et al., 2018;de Lenne
et al., 2018;Niu et al., 2018;Nursalam et al., 2018;Wang et al.,
2018;Twenge and Campbell, 2019). Of the included studies,
gender was seldom examined as an explanatory variable, and
other sociodemographic variables (e.g., ethnicity, socioeconomic
status) were not included at all.
Implicit Causation Based on Direction of
Association
Sixty-one (77%) of the included studies has social media use as the
independent variable and some of the mentioned measurements
of mental health as the dependent variable (Aboujaoude et al.,
2015;Banjanin et al., 2015;Banyai et al., 2017;Barry et al.,
2017;Best et al., 2014;Booker et al., 2018;Bourgeois et