International Journal of Psychology

International Journal of Psychology, 2022

Vol. 57, No. 5, 606–612, DOI: 10.1002/ijop.12840

Alexithymia, narcissism and social anxiety in relation to

social media and internet addiction symptoms
Michael Lyvers1 , Alessandra Salviani1 , Stephanie Costan1 ,
and Fred Arne Thorberg2
1 School of Psychology, Bond University, Gold Coast, Qld, Australia
2 School of Psychology, University of Queensland, Brisbane, Qld, Australia

T here has been considerable interest in trait correlates of excessive or problematic use of the internet, known as internet
addiction. However, the concept of internet addiction has been criticised as too broad. Specific forms of excessive
internet use, for example, social media, may have different trait correlates compared to internet use in general. The present
study compared levels of internet and social media addiction symptoms in relation to three traits previously linked to one
or the other form of excessive behaviour: alexithymia, narcissism and social anxiety. There were 217 young adult social
media- and internet-using participants aged 18–35 years recruited from two university campuses in southeast Queensland,
Australia. They completed an online questionnaire battery that included a demographics questionnaire and widely used,
validated measures of narcissism, alexithymia and social anxiety. Hierarchical regressions indicated that after controlling
for demographic variables, internet addiction symptoms were predicted by social anxiety, narcissism and alexithymia,
whereas social media addiction symptoms were predicted only by social anxiety and narcissism. Results suggest that
the association of alexithymia with internet addiction symptoms does not encompass excessive use of social media, and
support the contention that the concept of internet addiction may be too broad.

Keywords: Social media addiction; Internet addiction; Narcissism; Alexithymia; Social anxiety.

Alexithymia, a personality trait characterised by difficul- features (Lyvers et al., 2020b). The authors reported that
ties identifying and describing feelings and an externally although personality traits of alexithymia, narcissism and
oriented thinking style (Bagby et al., 2020), has been impulsivity were significant predictors of excessive or
implicated as a risk factor for compulsive behaviours problematic alcohol use in their young adult sample, only
including alcohol use disorders (Thorberg et al., 2009) narcissism and impulsivity predicted excessive or prob-
and eating disorders (Harrison et al., 2010). Alexithymia lematic social media use; alexithymia did not approach
has also been linked to excessive or problematic use significance. A subsequent study in a similar sample
of the internet, often termed internet addiction when (Lyvers et al., 2020a) did however find that alexithymia
addiction-related features are present such as craving, predicted greater self-disclosure on Facebook, which the
adverse impacts on various life domains, and impair- authors attributed to lower proneness to embarrassment
ment of control (Lyvers et al., 2016; Mahapatra & in alexithymic individuals.
Sharma, 2018). Various interpretations of the role of Trait narcissism has been linked to excessive or prob-
alexithymia as a risk factor for addictions have been lematic social media use in various studies (Andreassen
proposed, including self-medication of negative mood et al., 2017; Eskisu et al., 2017; McCain & Camp-
(Thorberg et al., 2009) and compensation for inadequate bell, 2018). Presumably narcissistic individuals find
early maternal bonding (Lyvers et al., 2019). in social media a readily available tool to project and
In contrast to the well-established association between reinforce their feelings of superiority to others. Other
alexithymia and internet addiction, a recent study research has indicated that social anxiety may also
found no association of alexithymia with excessive or predispose to both internet and social media addic-
problematic use of social media, known as social media tion (Erwin et al., 2004), as those with high levels of
addiction when similarly defined by addiction-related social anxiety often prefer to spend time engaging in

Correspondence should be addressed to Michael Lyvers, School of Psychology, Bond University, Gold Coast, Qld 4229, Australia. (E-mail:
mlyvers@[Link])
Funding: The authors have no funding to disclose.

© 2022 International Union of Psychological Science.

 SOCIAL MEDIA AND INTERNET ADDICTION 607

activities alone (including surfing the internet; Weinstein more inclusion criteria, followed by four cases that were
et al., 2015) and prefer to interact with others online as multivariate outliers, the final sample of 217 participants
opposed to in person (Desjarlais & Willoughby, 2010). with no missing data was obtained. This was considerably
Like traits of alexithymia and narcissism, social anxiety more than the 97 indicated by Soper’s online a pri-
appears to have an early developmental onset and is ori sample size calculator ([Link]
typically well-established and stable by young adulthood statcalc/[Link]?id=1) assuming a small effect
(American Psychiatric Association [APA], 2013), thus size of 0.15, power of 0.80 and six predictors in mul-
social anxiety can be regarded as a trait-like characteristic. tiple regression, however a broad range of social media
Despite the reported high prevalence of internet addic- and internet use was sought. Participants in the final sam-
tion symptoms worldwide (Cheng & Li, 2014; Durkee ple had an age range of 18–35 years (M = 22.33 years,
et al., 2012; Moreno et al., 2011), the construct of internet SD = 4.27), and included 167 females (77%), 44 males
addiction itself has been criticised as too broad, with the (20.3%), and 6 indicating “other” (2.8%). All participants
recommendation that specific forms of excessive or prob- were current students. A majority of participants (133;
lematic online activity be evaluated instead (Starcevic & 61%) indicated that they used social media for personal
Aboujaoude, 2017). Types of excessive online activity reasons five or more times per day, whereas 67 (31%) did
that reflect specific needs, such as internet gaming, use of so more than once per day, 14 (6%) once per day, 2 (1%)
internet porn, or acquiring and maintaining social rela- once per week, and 1 (0.004%) once per month.
tionships via social media use, may show different trait
correlates as compared to other types of internet use, or Materials
to excessive use of the internet in general. The rationale
for the present study was to test this idea by comparing Demographics
the roles of trait narcissism, alexithymia and social anx-
iety as predictors of social media and internet addiction This questionnaire asked for participants’ age, gender,
symptoms in regression models using a sample of regular frequency of use of social media (providing examples
young adult users of the internet and social media. The of popular platforms such as Facebook and Whatsapp),
goal was to determine if there were any substantial differ- highest education level achieved to date, English fluency
ences in the relationships of these traits, particularly alex- and whether they had ever suffered a serious brain injury
ithymia and narcissism (given the evidence cited earlier), or were taking medications for a psychiatric disorder (for
to symptoms of excessive or problematic social media screening purposes).
use versus the more general and more widely studied con-
struct of internet addiction. Based on previous research Narcissistic Personality Inventory 16
findings and interpretations cited earlier, social anxiety
was expected to predict both internet and social media Narcissistic Personality Inventory 16 (NPI-16; Ames
addiction symptoms, whereas alexithymia was only et al., 2006). The NPI-16 is a 16-item self-report assess-
expected to predict internet addiction symptoms, and nar- ment of trait narcissism derived from the widely used
cissism was only expected to predict social media addic- NPI-40 developed by Raskin and Terry (1988). Each
tion symptoms, in the corresponding regression models. item consists of a pair of statements where par-
ticipants endorse the statement that best represents
METHOD them. Narcissism-consistent responses are scored as 1
(e.g., “I like to be the centre of attention”) whereas
Participants narcissism-inconsistent responses are scored as 0
(e.g., “I prefer to blend in with the crowd”). Total
Approvals were obtained from the ethics committees of scores can be calculated as either the sum or mean of
two universities in southeast Queensland, Australia, prior narcissism-consistent responses; higher scores indicate
to participant recruitment. The study was advertised by higher levels of trait narcissism. The present study used
campus-wide email at one university and by posting on the sum of narcissism-consistent responses, which can
an electronic noticeboard at the other. The announce- range from 0 to 16. The Cronbach alpha reliability
ment asked for participants for research on the associa- coefficient in the current study was acceptable at .71.
tion between personality and online activities. Inclusion
criteria specified that participants had to be between 18 Toronto Alexithymia Scale 20
and 35 years old, regular users of the internet and social
media, and fluent in English, with no history of trau- Toronto Alexithymia Scale 20 (TAS-20; Bagby
matic head injury or current use of psychiatric medica- et al., 1994). The TAS-20 is a 20-item self-report mea-
tion. A link to an online questionnaire battery, set up via sure of the key facets of alexithymia: difficulty identifying
the online survey hosting platform Qualtrics, was pro- feelings (DIF; e.g., “I am often confused about what emo-
vided. After removal of 39 cases that did not meet one or tion I am feeling”), difficulty describing feelings (DDF;

© 2022 International Union of Psychological Science.

608 LYVERS ET AL.

e.g., “I find it hard to describe how I feel about people”), Cronbach alpha reliability coefficient in the current study
and externally oriented thinking (EOT; e.g., “I prefer was excellent at .90.
talking to people about their daily activities rather than
their feelings”). Participants rate their agreement with
each statement on a five-point Likert scale ranging from Procedure
1 (Strongly Disagree) to 5 (Strongly Agree). The sum
After clicking the link to the online survey and pro-
of responses to items provides subscale scores and a
viding informed consent, participants were presented
total score; the latter can range from 20 to 100, with
standardised instructions to complete the instruments
higher scores indicating higher levels of alexithymia.
online at one point in time. Order of questionnaires
In the present sample, the Cronbach’s alpha reliability
was uniquely randomised per participant following the
coefficient of the total TAS-20 was high at .86.
demographics questionnaire. At one university, partici-
pants (comprising most of the sample) were not offered
Social Interaction Anxiety Scale any incentive, whereas at the other university they were
offered research participation credit for an undergraduate
Social Interaction Anxiety Scale (SIAS; Mattick & psychology subject. After completion of data collection,
Clarke, 1998). The SIAS is a 20-item measure designed statistical analyses were conducted using IBM Statistical
to assess anxiety when interacting with others. Items Package for the Social Sciences (SPSS) version 26.
(e.g., “I am tense mixing in a group”) are rated on a Gender was coded as 1 = male, 2 = female, 3 = other.
five-point Likert scale with values ranging from 0 (“not Highest education level was coded as 1 = before grade
at all characteristic or true of me”) to 4 (“extremely 12, 2 = grade 12, 3 = undergraduate or trade school
characteristic or true of me”). Scores can range from 0 degree, 4 = postgraduate degree.
to 80, with higher scores indicating higher levels of social
anxiety. The Cronbach alpha reliability coefficient in the
current study was excellent at .89. RESULTS

The desired broad range of scores on the measures was

Social Media Disorder Scale
achieved in the current sample. Scores ranged from 25
Social Media Disorder Scale (SMDS; Van den to 100 (M = 51.84, SD = 13.59) on the IAT, 0 to 9
Eijenden et al., 2016). The SMDS consists of nine items (M = 2.23, SD = 1.96) on the SMDs, 22 to 87 (M = 53.19,
assessing symptoms of social media addiction: Preoc- SD = 13.90) on the TAS-20, 0 to 14 (M = 3.19, SD = 2.75)
cupation, tolerance, withdrawal, displacement, escape, on the NPI-16, and 21 to 80 on the SIAS (M = 52.65,
problems, deception, displacement, and conflict. Items SD = 13.66); skewness and kurtosis values were within
are based on the Diagnostic and Statistical Manual of the acceptable range for all variables. Harman test indi-
Psychiatric Disorders 5th Edition (DSM-5; APA, 2013) cated that total variance for a single factor was well less
criteria for internet gaming disorder; for example, “Dur- than 50%, suggesting that common method bias did not
ing the past year, have you tried to spend less time on influence the data. Assumption testing indicated that all
social media, but failed?” Respondents indicate whether criteria for regression were met. Correlations amongst
each statement applied to them in the past year by ticking variables are shown in Table 1. Correlations were in
“yes” (1) or “no” (0). The sum of item responses yields expected directions except for the NPI-16 measure of nar-
a total score that can range from 0 to 9; higher scores cissism, which was uncorrelated with internet and social
indicate more social media addiction symptoms. In the media addiction symptoms but was significantly nega-
present sample, the Cronbach alpha index of internal tively correlated with both alexithymia and social anxiety.
reliability was acceptable at .73. Narcissism was nevertheless used in the planned regres-
sion models reported below.
Internet Addiction Test
TABLE 1
Internet Addiction Test (IAT; Young, 2017). The Intercorrelations amongst variables (N = 217)
IAT is a 20-item scale based on the DSM criteria for
pathological gambling and other identified tendencies in Variable 1 2 3 4 5
addictive behaviours. The IAT requires respondents to 1. Internet addiction -
∗∗
rate items on a five-point Likert scale from 0 to 5 as to 2. Social media addiction 0.59 -
∗∗
the degree to which their internet use affects their daily 3. Alexithymia 0.34 0.10 -
∗∗ ∗∗ ∗∗
4. Social anxiety 0.46 0.22 0.49 -
routine, social life, productivity, sleeping pattern, and ∗∗ ∗∗
5. Narcissism −0.01 0.07 −0.25 −0.31 -
feelings. Scores can range from 2 to 100, with higher
scores indicating more problems due to internet use. The **p < .001.

© 2022 International Union of Psychological Science.

 SOCIAL MEDIA AND INTERNET ADDICTION 609

TABLE 2
Hierarchical multiple regression predicting internet addiction symptoms

Predictor 𝛥R2 𝛽 B SE B 95% CI for B

Step 1 .02
Age −.13 −.42 .26 [−.93, .08]
Gender .05 1.51 2.07 [−2.56, 5.59]
Education .01 .21 1.62 [−2.98, 3.41]
Step 2 .24∗∗∗
Age −.08 −.27 .22 [−.71, .18]
Gender .03 .82 1.83 [−2.79, 4.42]
Education .06 1.14 1.43 [−1.68, 3.96]
∗∗
Narcissism .17 .85 .32 [.23, 1.47]
∗
Alexithymia .18 .18 .07 [.04, .32]
∗∗∗
Social anxiety .42 .41 .07 [.28, .55]

Note: CI = confidence interval; SE B = standard error of unstandardized coefficient. *p < .05, **p < .01, ***p < .001.

TABLE 3
Hierarchical multiple regression predicting social media addiction symptoms

Predictor 𝛥R2 𝛽 B SE B 95% CI for B

Step 1 .02
Age −.10 −.05 .04 [−.12, .03]
Gender .05 .21 .30 [−.38, .80]
Education −.04 −.10 .04 [−.57, .36]
∗∗
Step 2 .06
Age −.08 −.04 .04 [−.11, .03]
Gender .04 .17 .29 [−.42, .75]
Education −.01 −.03 .23 [−.48, .43]
∗
Narcissism .15 .11 .05 [.01, .21]
Alexithymia .01 .00 .01 [−.02, .02]
∗∗
Social anxiety .25 .04 .01 [.01, .06]

Note: CI = confidence interval; SE B = standard error of unstandardized coefficient. *p < .05, **p < .01.

Regression on internet addiction symptoms 6% of additional variance and was significant, Fchange(3,
210) = 4.60, p = .004. As shown in Table 3, social anxiety
A hierarchical linear regression was conducted on IAT was the strongest predictor, followed by narcissism; alex-
scores, with demographic covariates of age, gender, and ithymia’s contribution was virtually zero. The final model
education level entered at step 1, followed by total scores explained only 8% of variance in excessive social media
on the NPI-16, TAS-20, and SIAS at step 2. Step 1 use, F(6, 210) = 3.01, p = .008.
accounted for only 2% of variance and was not significant,
F(3, 213) = 1.47, p = .22. Step 2 accounted for 24%
Exploratory regression and mediation test
of additional variance and was significant, Fchange(3,
210) = 22.43, p < .001. As can be seen in Table 2, all three An exploratory third hierarchical regression analysis was
traits were significant; social anxiety was the strongest conducted on internet addiction symptoms with the same
predictor, followed by alexithymia and narcissism. The two steps as before, but with the addition of SMDS scores
final model accounted for 26% of variance in internet at a third step to see if the variance in internet addiction
addiction symptoms, F(6, 210) = 12.17, p < .001. symptoms accounted for by narcissism in regression 1
might be attributed to the excessive social media use
Regression on social media addiction previously reported to be associated with narcissism (as
symptoms described earlier) and which was supported by the results
of regression 2. In step 3, entry of SMDS scores explained
A hierarchical linear regression was conducted on SMDS 24% of additional variance in IAT scores, Fchange(1,
scores, with demographic covariates of age, gender, and 209) = 97.15, p < .001, 𝛽 = .51; social anxiety (𝛽 = .29,
education level entered at step 1, followed by total scores p < .001) and alexithymia (𝛽 = .17, p = .003) remained
on the NPI-16, TAS-20 and SIAS at step 2. Step 1 highly significant, however the contribution of narcissism
accounted for only 1.9% of variance and was not signif- became nonsignificant (𝛽 = .10, p = .07), suggesting
icant, F(3, 213) = 1.36, p = .26. Step 2 accounted for mediation by social media addiction scores. Mediation

© 2022 International Union of Psychological Science.

610 LYVERS ET AL.

was then tested via JASP 0.14.1. Path analysis using Watson et al., 2013, for how to interpret suppressor vari-
bootstrapping with 1000 bias-corrected replications, con- ables in psychology research using self-report scales).
trolling for alexithymia and social anxiety as covariates, The most interesting finding was that in regression
suggested full mediation of the relationship between models, traits predicting symptoms of social media addic-
narcissism and internet addiction by excessive social tion were not the same as those predicting symptoms of
media use, p = .03. The 95% confidence interval for internet addiction. In the present study, alexithymia inde-
the indirect effect did not include zero (.01, .15) and the pendently predicted the latter but not the former, whereas
standardised estimate for the direct effect of narcissism narcissism and social anxiety were similarly predictive
of .09 was not significant, p = .09. of both (although exploratory path analysis suggested
that the contribution of narcissism to variance in internet
addiction symptoms could be explained by social media
DISCUSSION use). The present results thus support the contention of
Starcevic and Aboujaoude (2017) that the concept of
Regression analyses indicated that after controlling for internet addiction may be too broad, as different forms
demographic covariates, narcissism and social anxiety of excessive online behaviour likely reflect different pro-
were significant predictors of both social media and inter- cesses and predispositions given the many different uses
net addiction symptoms, whereas alexithymia was a sig- of the internet. In the present context, trait alexithymia
nificant predictor of internet addiction symptoms only. was linked to symptoms of addictive use of the internet
This result is in line with previous work indicating a in general but not social media in particular. The asso-
link between alexithymia and internet addiction but not ciation of alexithymia with excessive internet use is thus
excessive social media use in young adults (Lyvers et al., likely attributable to use of the internet for purposes other
2020b). People with high alexithymia are reported to suf- than pursuit of social interactions via social media. The
fer from persistent loneliness and interpersonal difficul- specific modalities of internet use associated with alex-
ties (Qualter et al., 2009; Spitzer et al., 2005) as well ithymia remain to be determined.
as avoidance of, or dissatisfaction with, close personal The present study had several limitations. First, the
relationships (Humphreys et al., 2009); the present fail- proportion of variance in internet addiction symptoms
ure to find a positive association between alexithymia and explained by the first regression model (26%) was sub-
social media use seems congruent with such evidence. On stantially greater than for social media addiction symp-
the other hand, a robust association of alexithymia with toms (8%), where the second model accounted for only
internet addiction symptoms was replicated in the present a small, though significant, amount of variance. Other
study. Present results are also consistent with the notion trait factors not assessed in the present investigation have
that people with social anxiety tend to find online inter- been reported to show significant relationships with levels
actions preferable to in-person ones, and are also more of social media use and/or internet addiction, including
likely to report symptoms of internet addiction. Present extraversion, neuroticism, reward sensitivity and impul-
findings are further consistent with reports cited earlier of sivity (Correa et al., 2010; Lyvers et al., 2016, 2020b;
a link between trait narcissism and excessive or problem- Marino et al., 2018). The predictor variables in the present
atic social media use; indeed, exploratory path analysis study thus offered a limited picture of the potential roles of
indicated that the association of narcissism with inter- personality traits as risk factors for excessive use of social
net addiction symptoms in the present sample could be media or of the internet in general. Further, although
explained by social media use. Interestingly, although nar- the traits examined in the present study are reported
cissism was not significantly correlated with either inter- to have an early onset and tend to be stable over time
net addiction or social media addiction scores, it was (APA, 2013; Hiirola et al., 2017; Lyvers et al., 2019),
a significant positive predictor of both in the first two the cross-sectional design of the present investigation pre-
(planned) regression models. This suggested a suppres- cludes causal interpretation of the observed relationships.
sor effect of one or both of the other two variables, as Longitudinal research is recommended in future work to
narcissism showed significant negative correlations with clarify the nature of the observed links between trait fac-
both social anxiety and alexithymia. Post hoc exploratory tors and various excessive online behaviours as reported
hierarchical regressions showed that when social anxiety in this and other studies cited earlier.
was added to regressions of narcissism on social media Another potential limitation concerns the measure-
addiction or internet addiction scores, narcissism became ment of constructs of interest via self-report instruments,
significant; by contrast, adding alexithymia to the regres- although this approach is perhaps unavoidable for those
sions did not have this effect. The suppressor effect of constructs, and the instruments were all well-known,
social anxiety may reflect that fact that the NPI-16 narcis- widely used, reliable and validated. A positive aspect of
sism scale used in the present study contains some items the current study was that there was no indication of com-
that effectively constitute an inverse measure of social mon method bias according to Harman single factor test
anxiety (e.g., “I like to be the centre of attention;” see in the present dataset. Finally, the present study used a

© 2022 International Union of Psychological Science.

 SOCIAL MEDIA AND INTERNET ADDICTION 611

nonclinical sample, hence the findings cannot be assumed Cheng, C., & Li, A. Y. (2014). Internet addiction prevalence
to generalise to clinical samples of individuals exhibit- and quality of (real) life: A meta-analysis of 31 nations
ing compulsive online behaviour, or who have been diag- across seven world regions. Cyberpsychology, Behavior and
nosed with conditions such as social anxiety disorder Social Networking, 17, 755–760. [Link]
cyber.2014.0317
or narcissistic personality disorder. Nevertheless, given
Correa, T., Hinsley, A. W., & Gil de Zúñiga, H. (2010). Who
current conceptions of addictive disorders and personal-
interacts on the web? The intersection of users’ personality
ity traits as distributed along continua in the population and social media use. Computers in Human Behavior, 26,
(e.g., APA, 2013; Substance Abuse and Mental Health 247–253. [Link]
Services Administration, 2016), the present study’s find- Desjarlais, M., & Willoughby, T. (2010). A longitudinal study
ings provide support for the recommendation that future of the relation between adolescent boys and girls’ computer
investigations of traits linked to excessive or problem- use with friends and friendship quality: Support for the social
atic online behaviour should be conducted in relation to compensation or the rich-get-richer hypothesis? Computers
specific forms of internet use, rather than internet use as in Human Behavior, 26, 895–905. [Link]
conceived broadly (i.e., internet addiction), in order to .chb.2010.02.004
advance our understanding of potential risk factors. Durkee, T., Kaess, M., Carli, V., Parzer, P., Wasserman, C.,
Floderus, B., Apter, A., Balazs, J., Barzilay, S., Bobes, J.,
Brunner, R., Corcoran, P., Cosman, D., Cotter, P., Despalins,
COMPLIANCE WITH ETHICAL STANDARDS R., Graber, N., Guillemin, F., Haring, C., Kahn, J. P., …
Wasserman, D. (2012). Prevalence of pathological internet
All procedures performed in studies involving human par- use among adolescents in Europe: Demographic and social
ticipants were in accordance with the ethical standards factors. Addiction, 107, 2210–2222. [Link]
j.1360-0443.2012.03946.x
of the Bond University Human Research Ethics Commit-
Erwin, B. A., Turk, C. L., Heimberg, R. G., Fresco, D. M., &
tee and with the 1964 Helsinki Declaration and its later Hantula, D. A. (2004). The internet: Home to a severe pop-
amendments or comparable ethical standards. ulation of individuals with social anxiety disorder? Journal
of Anxiety Disorders, 18, 629–646. [Link]
INFORMED CONSENT .janxdis.2003.08.002
Eskisu, M., Hosoglu, R., & Rasmussen, K. (2017). An
investigation of the relationship between Facebook usage,
Informed consent was obtained from all individual adult
Big Five, self-esteem and narcissism. Computers in Human
participants included in the study. Behavior, 69, 294–301. [Link]
.12.036
Manuscript received February 2021
Harrison, A., Sullivan, S., Tchanturia, K., & Treasure, J. (2010).
Revised manuscript accepted February 2022
Emotional functioning in eating disorders: Attentional bias,
First published online March 2022
emotion recognition, and emotion regulation. Psycholog-
ical Medicine, 40, 1887–1897. [Link]
REFERENCES S0033291710000036
Hiirola, A., Pirkola, S., Karukivi, M., Markkula, N., Bagby,
American Psychiatric Association (APA). (2013). Diagnos- R., Joukamaa, M., Jula, A., Kronholm, E., Saarijärvi, S.,
tic and statistical manual of mental disorders: DSM-5™ Salminen, J. K., Suvisaari, J., Taylor, G., & Mattila, A. K.
(5th ed.). American Psychiatric Publishing. [Link] (2017). An evaluation of the absolute and relative stability
10.1176/[Link].9780890425596 of alexithymia over 11 years in a Finnish general population.
Ames, D. R., Rose, P., & Anderson, C. P. (2006). The NPI-16 Journal of Psychosomatic Research, 95, 81–87. [Link]
as a short measure of narcissism. Journal of Research in .org/10.1016/[Link].2017.02.007
Personality, 40, 440–450. [Link] Humphreys, T. P., Wood, L. M., & Parker, J. D. A. (2009). Alex-
.03.002 ithymia and satisfaction in intimate relationships. Personal-
Andreassen, C. S., Pallesen, S., & Griffiths, M. D. (2017). The ity and Individual Differences, 46, 43–47. [Link]
relationship between addictive use of social media, narcis- .1016/[Link].2008.09.002
sism, and self-esteem: Findings from a large national survey. Lyvers, M., Cutinho, D., & Thorberg, F. A. (2020a). Alex-
Addictive Behaviors, 64, 287–293. [Link] ithymia, impulsivity, disordered social media use, mood and
.addbeh.2016.03.006 alcohol use in relation to Facebook self-disclosure. Com-
Bagby, R. M., Parker, J. D., & Taylor, G. J. (1994). The puters in Human Behavior, 103, 174–180. [Link]
twenty-item Toronto Alexithymia Scale - item selection and .1016/[Link].2019.09.004
cross-validation of the factor structure. Journal of Psycho- Lyvers, M., Karantonis, J., Edwards, M. S., & Thorberg,
somatic Research, 38, 23–32. [Link] F. A. (2016). Traits associated with internet addiction in
3999(94)90005-1 young adults: Potential risk factors. Addictive Behaviors
Bagby, R. M., Parker, J. D. A., & Taylor, G. J. (2020) Reports, 3, 56–60. [Link]
Twenty-five years with the 20-item Toronto Alexithymia .001
Scale. Journal of Psychosomatic Research, 131, 109940. Lyvers, M., Mayer, K., Needham, K., & Thorberg, F. A. (2019).
[Link] Parental bonding, adult attachment, and theory of mind:

© 2022 International Union of Psychological Science.

612 LYVERS ET AL.

A developmental model of alexithymia and alcohol-related evidence of its construct validity. Journal of Personality and
risk. Journal of Clinical Psychology, 75, 1288–1304. https:// Social Psychology, 54(890), 902. [Link]
[Link]/10.1002/jclp.22772 0022-3514.54.5.890
Lyvers, M., Narayanan, S. S., & Thorberg, F. T. (2020b). Spitzer, C., Siebel-Jurges, U., Barnow, S., Grabe, H., & Frey-
Disordered social media use and risky drinking in young berger, H. (2005). Alexithymia and interpersonal problems.
adults: Differential associations with addiction-linked traits. Psychotherapy and Psychosomatics, 74, 240–246. https://
Australian Journal of Psychology, 71, 223–231. [Link] [Link]/10.1159/000085148
.org/10.1111/ajpy.12236 Starcevic, V., & Aboujaoude, E. (2017). Internet addic-
Mahapatra, A., & Sharma, P. (2018). Association of inter- tion: Reappraisal of an increasingly inadequate concept.
net addiction and alexithymia: A scoping review. Addictive CNS Spectrums, 22, 7–13. [Link]
Behaviors, 81, 175–182. [Link] S1092852915000863
.2018.02.004 Substance Abuse and Mental Health Services Administration;
Marino, C., Gini, G., Vieno, A., & Spada, M. M. (2018). A Office of the Surgeon General. (2016). Facing addiction in
comprehensive meta-analysis on problematic Facebook use. America: The surgeon General’s report on alcohol, drugs,
Computers in Human Behavior, 83, 262–277. [Link] and health. US Department of Health and Human Services.
.org/10.1016/[Link].2018.02.009 [Link]
Mattick, R. P., & Clarke, J. C. (1998). Development and vali- ch4.f1/
dation of measures of social phobia scrutiny fear and social Thorberg, F. A., Young, R. M., Sullivan, K. A., & Lyvers, M.
interaction anxiety. Behavior Research and Therapy, 36, (2009). Alexithymia and alcohol use disorders: A critical
455–470. [Link] review. Addictive Behaviors, 34, 237–245. [Link]
McCain, J. L., & Campbell, W. K. (2018). Narcissism and .1016/[Link].2008.10.016
social media use: A meta-analytic review. Psychology of Van den Eijenden, R. J., Lemmens, J. S., & Valkenburg,
Popular Media Culture, 7, 308–327. [Link] P. M. (2016). The social media disorder scale. Computers
ppm0000137 in Human Behavior, 61, 478–487. [Link]
Moreno, M. A., Jelenchick, L., Cox, E., Young, H., & Chris- [Link].2016.03.038
takis, D. A. (2011). Problematic internet use among US Watson, D., Clark, L. A., Chmielewski, M., & Kotov, R. (2013).
youth: A systematic review. Archives of Pediatric and Ado- The value of suppressor effects in explicating the construct
lescent Medicine, 165, 797–805. [Link] validity of symptom measures. Psychological Assessment,
archpediatrics.2011.58 25, 929–941. [Link]
Qualter, P., Quinton, S. J., Wagner, H., & Brown, S. (2009). Weinstein, A., Dorani, D., Elhadif, R., Bukovza, Y., Yarmulnik,
Loneliness, interpersonal distrust, and alexithymia in uni- A., & Dannon, P. (2015). Internet addiction is associated with
versity students. Journal of Applied Social Psychology, social anxiety in young adults. Annals of Clinical Psychiatry,
39, 1461–1479. [Link] 27, 4–9 PMID: 25696775.
.00491.x Young, K. S. (2017). IAT manual. Center for Internet Addic-
Raskin, R., & Terry, H. (1988) A principal-components anal- tion Recovery. [Link]/wp-content/uploads/2015/
ysis of the Narcissistic Personality Inventory and further 11/[Link]

© 2022 International Union of Psychological Science.

Copyright of International Journal of Psychology is the property of John Wiley & Sons, Inc.
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.