Scandinavian Journal of Psychology, 2016 DOI: 10.1111/sjop.

12343

Personality and Social Psychology

Social and emotional loneliness and self-reported difficulty initiating and
maintaining sleep (DIMS) in a sample of Norwegian university students
AMIE C. HAYLEY,1 LUKE A. DOWNEY,1 CON STOUGH,1 BØRGE SIVERTSEN,2,3,4 MARIT KNAPSTAD2,5
and SIMON ØVERLAND2,6
1
Centre for Human Psychopharmacology, Swinburne University of Technology, Hawthorn, Australia
2
Department of Public Mental Health, Domain for Mental and Physical Health, Norwegian Institute of Public Health, Bergen, Norway
3
Regional Centre for Child and Youth Mental Health and Child Welfare, Uni Research Health, Bergen, Norway
4
Department of Psychiatry, Helse Fonna HF, Haugesund, Norway
5
Department of Clinical Psychology, Faculty of Psychology, University of Bergen, Bergen, Norway
6
Department of Psychosocial Science, Faculty of Psychology, University of Bergen, Bergen, Norway

Hayley, A. C., Downey, L. A., Stough, C., Sivertsen, B., Knapstad, M. & Øverland, S. (2016). Social and emotional loneliness and self-reported difficulty
initiating and maintaining sleep (DIMS) in a sample of Norwegian university students. Scandinavian Journal of Psychology.

Social and emotional loneliness negatively impact several areas of health, including sleep. However, few comprehensive population-based studies have
evaluated this relationship. Over 12,000 students aged 21–35 years who participated in the student survey for higher education in Norway (the SHoT
study) were assessed. Loneliness was assessed using the Social and Emotional Loneliness Scale. Difficulty initiating and maintaining sleep (DIMS) was
assessed by a single-item subjective response on the depression scale of the Hopkins Symptoms Checklist (HSCL-25). Social loneliness was associated
with more serious DIMS (unadjusted proportional odds-ratio [OR] = 2.69, 95% CI = 2.46–2.95). This association was attenuated following adjustment for
anxiety (adjusted OR = 1.92, 95% CI = 1.75–2.10) and depression (adjusted OR = 1.48, 95% CI = 1.34–1.63), however was not substantially altered
when all demographics and psychological distress were accounted for (fully adjusted OR = 1.46, 95% CI = 1.30–1.63). Emotional loneliness was also
associated with more serious DIMS (unadjusted proportional OR = 2.33, 95% CI = 2.12–2.57). Adjustment for anxiety (adjusted OR = 1.96, 95%
CI = 1.78–2.15) and depression (adjusted OR = 1.64, 95% CI = 1.48–1.80) attenuated, but did not extinguish this relationship in the fully adjusted model
(adjusted OR = 1.22, 95% CI = 1.09–1.31). Mediation analyses revealed that the social loneliness-DIMS association was fully attributed to psychological
distress, while the emotional loneliness-DIMS association was only partially mediated, and a direct association was still observed. Associations between
social and emotional loneliness and subjective DIMS were embedded in a larger pattern of psychological distress. Mitigating underlying feelings of
loneliness may reduce potentially deleterious effects on sleep health and psychological wellbeing in young adults.
Key words: Emotional, loneliness, psychopathology, social, sleep disturbance, student, adult.
Amie C. Hayley, Centre for Human Psychopharmacology, Faculty of Health, Arts and Design, Swinburne University of Technology, Hawthorn, VIC 3122,
Australia. Tel. +61 3 92145585; e-mail: [email protected]

INTRODUCTION emotional factors are similarly implicated in overall health and

wellbeing outcomes, including psychiatric and sleep health
Social engagement is a fundamental component of the human (Heinrich & Gullone, 2006).
condition (Baumeister & Leary, 1995). Loneliness is a complex Extant cross-sectional and longitudinal studies have noted a
psychosocial construct which typically occurs when there is a strong and independent bidirectional relationship between
discrepancy between the quality and quantity of an individuals’ loneliness and a wide spectrum of psychopathologies (Wilson,
desired and actual social and emotional relationships (Peplau & Krueger, Arnold et al., 2007). Loneliness has been shown to be
Perlman, 1982). Theoretically, loneliness can be examined as a correlated with interpersonal sensitivity (low self-esteem) (Jackson
singular concept, or can be divided into separate components & Cochran, 1991), depression and anxiety (Cacioppo, Hughes,
pertaining to the degree of perceived quantitative (social) or Waite, Hawkley & Thisted, 2006), and several manifestations of
qualitative (emotional) deficiencies (Heinrich & Gullone, 2006; suicidal ideation (Stravynski & Boyer, 2001). Moreover,
Peplau & Perlman, 1982). Indeed, social loneliness has previously depressed individuals typically report higher rates of loneliness
been linked to subjective deficiencies in social structure (having a than non-depressed individuals (Hsu, Hailey & Range, 1987), and
network and someone to rely on) and emotional loneliness is mediatory relationships have been noted with other, peripherally
considered to reflect the subjective experience of being lonely related health outcomes, including sleep (Segrin & Burke, 2015;
(despite the possible existence of social networks) (Russell, Zawadzki, Graham & Gerin, 2013). The reciprocal association
Cutrona, Rose & Yurko, 1984). Formation and maintenance of between psychiatric symptomology and sleep disturbance is well
social relationships is considered to be particularly salient during established in medicine (Ford & Kamerow, 1989). Considerable
periods of early adulthood (Heinrich & Gullone, 2006). Emerging clinical and epidemiological research has demonstrated that
research has indicated the existence of a nonlinear, U-shaped psychiatric disturbance can precede the experience of poorer sleep
prevalence distribution for loneliness, with as many as 21% of quality (Hayley et al., 2015), and psychopathology often
individuals aged 20–34 years old frequently report these contributes to poorer sleep outcomes (Baglioni, Battagliese, Feige
experiences (Victor & Yang, 2012). Many of these social and et al., 2011; Ford & Kamerow, 1989). These associations are

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
2 A. C. Hayley et al. Scand J Psychol (2016)

similarly considered to exist on a dose-response curve of was obtained from the Norwegian Data Inspectorate for Social Sciences,
increasing symptomatic severity (Ohayon & Roth, 2003), and and informed written consent was received from all participants prior to
data collection. The SHoT study was initiated by the 10 largest student
these shared pathologies are more pronounced during periods of
welfare organisations in Norway (located in Oslo, Bergen and
early adulthood (Breslau, Roth, Rosenthal & Andreski, 1996; Trondheim). The study surveyed higher education students aged less than
Buysse, Angst, Gamma, Ajdacic, Eich & R€ossler, 2008). 35 years, representing 71% of all students enrolled in Norwegian colleges
Loneliness is also both directly and peripherally associated with and universities. Initial invitations to participate in the research were
sleep disruption (Cacioppo, Hawkley, Crawford, et al., 2002), and emailed to students, which contained an explanation of the study, and an
active hyperlink to an anonymous online questionnaire. This invitation
indirect pathways to shared psychiatric health outcomes are often
was sent to 47,514 randomly selected students, and was stratified by study
observed (Segrin & Burke, 2015; Zawadzki et al., 2013). Lonely institutions, faculties and departments across higher education institutions
individuals typically display greater objectively measured sleep in the selected regions. The overall response rate for the survey was
disruption, such as poorer sleep efficiency, greater wake time after 25.3%, and the current study included a total sample of 12,043 students.
sleep onset (Cacioppo, Hawkley, Berntson et al., 2002) and Written informed consent was obtained from all participants prior to data
collection, and approval for conducting the SHoT study was granted by
increased sleep fragmentation (Kurina, Knutson, Hawkley,
the Data Protection Officer for research at the Norwegian Social Science
Cacioppo, Lauderdale & Ober, 2011). Young adults who report Data Services.
higher levels of loneliness similarly report poorer subjective sleep
quality, including longer sleep latency, longer perceived sleep
duration, and greater daytime dysfunction (Cacioppo, Hawkley, Measures
Crawford, 2002; Hawkley & Cacioppo, 2003). Difficulty initiating Social and emotional loneliness. Loneliness was assessed using the Social
and maintain sleep (DIMS) is regarded as the most prevalent self- and Emotional Loneliness Scale (Wittenberg, 1986), which is an extension
reported sleep complaint experienced during early adulthood of the tool originally developed by Russell et al. (1984). The scale
(Hysing, Pallesen, Stormark, Lundervold & Sivertsen, 2013), consists of two five-item measures which assess concepts of social
(“I don’t get much satisfaction from the group I participate in,” “there are
affecting between 9 (Janson, Gislason, De Backer et al., 1995) and
good people around me who understand my views and beliefs,” “I belong
42% of young adults (Farnill & Robertson, 1990). Self-reported to a network of friends,” “There are people I can count on for
DIMS has been shown to be negatively associated with several companionship,” “Mostly, everyone around me seems like a stranger) and
health and wellbeing indices later in life, including poorer quality emotional loneliness (“There is no one I have felt close to for a very long
of life (Baldwin, Ervin, Mays et al., 2010) and greater instances of time,” “I have a romantic partner who gives me support and
encouragement,” “I don’t have one specific relationship in which I feel
psychopathology (Neckelmann, Mykletun & Dahl, 2007), and thus
understood,” “I am an important part of the emotional well-being of
close examination of these association during these critical age another person,” “I don’t have a special love relationship”). Participants
periods is imperative to improve later health outcomes. Although are required to indicate how often they feel the way depicted in the
the association between symptoms of loneliness and differing questions, and responses are presented on a five-point Likert-scale
aspects of subjective sleep problems has been established to some response matrix (1 = Never to 5 = Very often). Total scores on the Social
and Emotional Loneliness Scale range from 5–25 for each concept,
effect, as yet, no studies have directly examined the association
whereby higher scores reflect more pronounced feelings of loneliness. For
between loneliness and self-reported DIMS, particularly among the purpose of the current study, scores were used both as a continuous
cohorts of young adults (Jensen, Dehlin, Hagberg, Samuelsson & factor, and the sum score was converted to quartiles, with the highest
Svensson, 1998), and no research has yet explored the possible quartile being an indication of highest levels of reported loneliness.
mediating effect of psychiatric symptoms in the expression of this Previous psychometric assessments of the coefficient alphas for both
constructs of the scale have yielded acceptable values (0.78 and 0.76 for
important relationship.
emotional and social loneliness, respectively) (Russell et al., 1984). When
Loneliness is a pervasive subjective experience, and is both compared to other measures of loneliness (such as the UCLA loneliness
directly and peripherally associated with negative psychiatric, sleep scale; Russell, Peplau & Ferguson, 1978), evaluations of convergent
and wellbeing outcomes. Self-reported DIMS is common among validity for the Wittenberg (1986) scale yielded acceptable correlation
young adults, and is associated with considerable health burden. coefficients of 0.81 for social loneliness and 0.59 for emotional loneliness
(Robinson, Shaver & Wrightsman, 1991).
Despite this, no research has yet explicitly evaluated the direct and
indirect role of loneliness and psychiatric illness in the expression of
these associations. Systematic evaluations of the association between Psychological distress. Psychological distress was assessed using The
Hopkins Symptoms Checklist (HSCL-25). The HSCL-25 is a 25-item a
loneliness and self-reported DIMS are critical for characterizing the
self-report symptom inventory which measures five underlying dimensions
effect on health during critical periods of early adulthood. Therefore, of psychological distress, including anxiety and depression (Derogatis,
the present study aimed to assess the association between social and Lipman, Rickels, Uhlenhuth & Covi, 1974). The scale is divided into two
emotional feelings of loneliness and subjective symptoms of DIMS, main subscales; a ten-item anxiety symptom scale, and a fifteen-item
whilst systematically evaluating the relative impact of several depressive symptom scale. The response scale for each question includes
four categories (“Not at all,” “A little,” “Quite a bit,” “Extremely,”).
peripherally related lifestyle, health and psychiatric factors among a
Scores on the HSCL-25 can be interpreted both as a dichotomous and
group of college-age Norwegian students. continuous outcome. The HSCL-25 has been utilized in several cross-
cultural settings among groups of immigrants (Hoffmann, McFarland,
Kinzie et al., 2006), minority groups (Mouanoutoua, 1995), psychiatric
METHODS populations (Veijola, Jokelainen, J., L€aksy et al., 2003), and in general-
population samples (Nyman, Miettunen, Freimer et al., 2011). Independent
psychometric assessments of the HSCL-25 have indicated adequate
Sample
reliability and sensitivity of the sub-scales (Deane, Leathern & Spicer,
Data for this study were extracted from the SHoT study, a large national 1992). When evaluating HSCL depression subscale, we omitted the DIMS-
student survey for higher education students in Norway. Ethical approval specific item, and replaced it with the mean of the remaining 14 items,

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Scand J Psychol (2016) Loneliness and sleep in university students 3

thereby preserving the recommended total score to enable comparison with sex, by entering the product of these variables in separate blocks.
other studies. However, as no significant interaction effects were observed, the results
are presented as a whole group. Effect sizes are presented as odds-ratios
Difficulty Initiating and Maintaining Sleep (DIMS). Difficulties initiating (OR) with 95% confidence intervals. The mediation analyses (using the
and/or maintaining sleep (DIMS) is measured on one item of the continuous variables) were conducted using the PROCESS macro by Hays
depression subscale of the Hopkins Symptoms Checklist (HSCL-25), and (2013) in SPSS. A significant mediation effect was said to occur if the
this item was used in the current study as a proxy indicator for sleep confidence intervals (CI) for the indirect effect did not include zero. The
difficulties. For the current study, subjective DIMS was indicated as both a CIs for the indirect effect was BCa bootstrapped CI based on 1,000
categorical (all four ordinal levels) and dichotomous outcomes (combining samples.
“Quite a bit,” “Extremely” as indicating the presence of DIMS).

Demographic and lifestyle information. Demographic information was RESULTS

obtained via self-report, and included assessment of age, marital/cohabitant
status (married or cohabitant, partner/ lives alone, or single), care for Differences in social and emotional loneliness for the whole
children status (yes/no), income (in thousand Norwegian Kroner, NOK) group according to demographical and clinical characteristics are
and self-reported financial difficulties (never/seldom, sometimes, often). presented in Table 1. Overall, 8.4% (n = 1,112) of individuals
Body mass index (BMI) was calculated based on self-reported weight in
reported symptoms of Social Loneliness, and 9.1% (n = 1,161)
kilograms divided by the square of height in meters. BMI was then
calculated and split into four categories: underweight, normal weight, reported experiencing Emotional Loneliness. Women, those aged
overweight and obese, based on recommended cut-offs (National Heart, 29–34 year age-group, those who have care for children, those
Lung and Blood Institute, 1998). Data was collected on smoking status who have economic problems, have a one-year study program,
and was classified as current or previous smoker or never smoked. Weekly who are classified as physically inactive and are a current smoker,
level of exercise (0 days, 1–3 days, or 4+ days per week) was also
have a normal AUDIT classification, who are obese, report higher
obtained. Alcohol use was assessed by the Alcohol Use Disorders
Identification Test (AUDIT). The AUDIT was originally developed by the rates of anxiety and depression and those who self-report DIMS
World Health Organization to assess alcohol consumption, drinking “quite a bit” and “extremely” reported significantly higher mean
behaviors, and alcohol-related problems (Saunders, Aasland, Babor, de la scores on social loneliness. For emotional loneliness, the
Fuente & Grant, 1993), and includes 10 items examining drinking significant demographics were largely comparable to those of
behaviors. An overall cut-off of 8 was applied for the current study,
social loneliness; with the exception of males, those who had
indicating a hazardous drinking pattern. This tool has been validated
among diverse groups of adults, and is considered appropriate for use in hazardous AUDIT scores and those who have no care for children
the general population (Allen, Litten, Fertig & Babor, 1997). Psychometric reporting higher mean emotional loneliness scores.
evaluation of the AUDIT at this cut-off has revealed good sensitivity and Table 2 and Fig. 1 shows the association between the quartiles
specificity (0.96 and 0.85, respectively) in general population samples of social loneliness and the odds of self-reporting DIMS. In the
(Tsai, Tsai, Chen & Liu, 2005).
crude model, social loneliness was significantly associated with
increased odds of reporting DIMS in a dose-response manner. For
example, those in the highest, 3rd and 2nd quartile of loneliness
Statistical analysis
had 2.69, 1.67 and 1.26-fold increased odd, respectively, of
All analyses were performed using both the SPSS statistical software reporting more serious self-reported DIMS compared to those
package, version 23 (SPSS Inc., Chicago, IL, USA) and STATA/SE 14.1 without social loneliness (all ps < 0.05). In the highest quartile
statistical package, and all tests were two-tailed with conventional
p < 0.05 as significance threshold. Independent samples t-tests and
of social loneliness, subsequent independent adjustment for
analysis of variance (ANOVA) were used to examine differences in social demographic and lifestyle (model 2) (adjusted OR = 2.90, 95%
and emotional loneliness according to demographical and lifestyle CI = 2.63–3.19), study-related (model 3) (adjusted OR= 2.64,
variables. Ordered logistic regression analyses were conducted to examine 95% CI = 2.41–2.89) and social factors (model 5) (adjusted
the predictive effect of social and emotional loneliness on subjective OR = 2.68, 95% CI = 2.43–2.96) did not substantially attenuate
DIMS. Ordered logistic regression is an extension of logistic regression
accounting for ordered categorical outcomes (here DIMS). The model is
the association. To examine the contribution of co-existing mental
based on proportional odds, which in the current case means that the odds health problems in the relationship between social loneliness and
for “extremely” DIMS versus all the lower categories of DIMS in self-reported DIMS, further adjustment was made for
combination, was similar to the odds for “quite a bit” DIMS versus psychological distress symptoms. These analyses showed that
“a little bit” and “not at all” DIMS in combination, etc. This assumption adjustment for anxiety and depression (adjusted OR = 1.40, 95%
of “proportional odds” was met in the models concerning emotional
loneliness, but not in the social loneliness models. Thus, to show the odds
CI = 1.27–1.54) did reduce the size of the OR, however, the
pertaining to each DIMS category, we additionally performed multinomial association was not substantially altered when all demographics
logistic regression analyses (setting “not at all” DIMS as base). Six models and psychological distress factors were simultaneously accounted
were examined: (1) model 1: crude; (2) model 2: model 1 + adjusting for for (fully adjusted model OR = 1.46, 95% CI = 1.30–1.63).
age, gender, income, physical exercise, smoking, BMI and AUDIT As depicted in Table 3 and Fig. 1, a similar dose-response
classification; (3) model 3: model 1 + adjusting for study length program,
semester; (4) model 4: model 1 + adjusting for active in social activities,
association was noted in the crude model between emotional
number of friends, cohabitant status, care for children; (5) model 5: model loneliness and increased odds of self-reporting DIMS.
1 + additional adjustment for Anxiety (HSCL Anxiety subscale) and Specifically, those in the highest, 3rd and 2nd quartile of social
additional adjustment for Depression (HSCL Depression subscale); and loneliness were at 2.33, 1.89 and 1.47-fold increased odd,
(6) model 6: Fully adjusted model. All analyses were completed using the respectively, of reporting more serious subjective DIMS
unweighted data. We tested for multicollinearity by inspecting tolerance
values and Variance Inflation Factors (VIFs) using the ‘collin’ command
compared to those without emotional loneliness (all ps < 0.05). In
in STATA, and all values were well within the recommended limits the highest quartile of emotional loneliness, separate adjustment
(O’Brien, 2007). We also tested for interactions between loneliness and for demographic and lifestyle factors (model 2) (adjusted

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
4 A. C. Hayley et al. Scand J Psychol (2016)

Table 1. Mean score on the social and emotional loneliness subscales by demographical and clinical factors.

Social loneliness Emotional loneliness

n % Mean (SD) P-value* Mean (SD) P-value*

Gender
Males 4,581 33.5 11.19 3.76 0.001 13.45 3.27 <0.001
Females 9,082 66.5 10.74 3.61 12.74 3.04
Age group
18–20 years 1,767 12.9 10.92 3.56 <0.001 13.34 3.22 <0.001
21–22 years 3,678 26.9 10.72 3.63 13.21 3.13
23–25 years 4,887 35.8 10.70 3.58 12.85 3.10
26–28 years 2,006 14.7 11.07 3.68 12.67 3.12
29–34 years 1,325 9.7 11.80 4.04 12.86 3.14
Has care for children
Yes 863 6.3 11.30 3.79 0.003 12.09 2.67 <0.001
No 12,762 93.4 10.86 3.66 13.04 3.16
Economic problems
Never/seldom 8,065 59.0 10.69 3.62 <0.001 12.88 3.17 <0.001
Sometimes 3,971 29.1 11.00 3.64 13.02 3.06
Often 1,602 11.7 11.64 3.84 13.34 3.19
Study program
One year 488 3.6 11.41 3.85 <0.001 13.17 3.15 <0.001
Lower grade 8,311 60.8 11.11 3.72 13.22 3.18
Higher grade 4,836 35.4 10.47 3.50 12.55 3.03
Physical exercise (days/week)
0 days 3,980 29.1 11.60 3.84 <0.001 13.20 3.20 <0.001
1–3 days 8,540 62.5 10.58 3.53 12.84 3.08
4+ days 1,122 8.2 10.71 3.77 13.27 3.32
Smoking
Current smoker 1,194 8.7 11.61 3.84 <0.001 13.49 3.20 <0.001
Previous smoker 1,075 7.9 11.55 3.75 12.87 3.01
Never smoked 11,025 80.7 10.75 3.62 12.93 3.14
AUDIT classification
Normal 7,870 57.6 11.11 3.77 <0.001 12.82 3.19 <0.001
Hazardous 5,445 39.9 10.49 3.46 13.11 3.05
BMI classification
Underweight 550 4.0 11.64 4.16 <0.001 13.49 3.38 <0.001
Normal weight 9,337 68.3 10.74 3.57 12.83 3.06
Overweight 2,614 19.1 11.00 3.75 13.17 3.19
Obese 775 5.7 11.64 3.86 13.61 3.39
HSCL-25 Anxiety
Under cut-off 9,877 72.3 10.29 3.40 <0.001 12.70 3.06 <0.001
Over cut-off 3,648 26.7 12.51 3.88 13.72 3.22
HSCL-25 Depression
Under cut-off 8,952 65.5 9.84 3.10 <0.001 12.44 2.93 <0.001
Over cut-off 4,573 33.5 12.95 3.81 14.01 3.26
DIMS
Not at all 5,065 37.1 10.16 3.46 <0.001 12.41 3.02 <0.001
A little 4,929 36.1 10.83 3.48 13.00 3.09
Quite a bit 2,340 17.1 11.65 3.74 13.62 3.15
Extremely 1,191 8.7 12.77 4.11 14.03 3.30

Notes: *p values refer to outcome of Independent samples t-tests and Analysis of Variance (ANOVA) for differences in social and emotional loneliness
according to demographical and lifestyle variables.

OR = 2.36, 95% CI = 2.13–2.61) and study-related factors extinguished, when all demographics and psychological distress
(model 4) (adjusted OR= 2.28, 95% CI = 2.07–2.50) did not factors were simultaneously accounted for (adjusted OR = 1.22,
substantially attenuate the association, however further adjustment 95% CI = 1.09–1.31).
for social factors (model 5), reduced the strength of the Mediation analyses were conducted to examine whether
association somewhat (adjusted OR = 1.92, 95% CI = 1.72– symptoms of psychological distress mediated the relationships
2.15). Further adjustment for psychological distress symptoms between social loneliness and self-reported DIMS. Figure 2
revealed a slight increase in the OR when anxiety and depression presents the results of a mediation analyses based on 1000
were entered into the model (adjusted OR = 1.62, 95% bootstrapped samples and adjusted for age and sex. The
CI = 1.47–1.79). This association was further attenuated, but not standardized regression coefficient between social loneliness and

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Scand J Psychol (2016) Loneliness and sleep in university students 5

Table 2. Social loneliness (by quartiles) associated with difficulties initiating and maintaining sleep (DIMS), ordered logistic regression models

Social loneliness quartile

Highest quartile 3rd quartile 2nd quartile

Adjustment variables OR 95% CI OR 95% CI OR 95% CI

Model 1: social loneliness only 2.69 2.46–2.95 1.67 1.53–1.81 1.26 1.15–1.38
Model 2: model 1 + demographical and lifestyle 2.90 2.63–3.19 1.73 1.58–1.90 1.31 1.19–1.44
factors (age, gender, income, physical exercise,
smoking, BMI, and AUDIT classification)
Model 3: model 1 + study related factors (study 2.64 2.41–2.89 1.65 1.52–1.80 1.25 1.14–1.37
length program, semester)
Model 4: model 1 + social factors (active in social 2.68 2.43–2.96 1.63 1.49–1.78 1.27 1.16–1.39
activities, number of friends, cohabitant status, care for children)
Model 5: model 1 + anxiety (HSCL Anxiety subscale) 1.40 1.27–1.54 1.21 1.11–1.33 1.14 1.04–1.25
and depression (HSCL Depression subscale)
Model 6: fully adjusted model 1.46 1.30–1.63 1.25 1.13–1.38 1.18 1.07–1.30

Fig. 1. Multinomial regression analyses of the association between social and emotional loneliness (exposure) and difficulties initiating and/or maintaining
sleep (DIMS) (outcome). Bars represent odds-ratios, and error bars represent 95% confidence intervals (the Y-axis has a logarithmic scale). Reference
categories: “Not at all DIMS” and lowest quartiles loneliness. # Adjusted for demographical factors (age, gender cohabitant status, income, care for
children), lifestyle factors (physical exercise, smoking, BMI, and AUDIT classification), study related factors (study length program, semester) and social
factors (active in social activities, number of friends).

Table 3. Emotional loneliness (by quartiles) associated with difficulties initiating and maintaining sleep (DIMS)

Emotional loneliness quartile

Highest quartile 3rd quartile 2nd quartile

Adjustment variables OR 95% CI OR 95% CI OR 95% CI

Model 1: emotional loneliness only 2.33 2.12–2.57 1.89 1.71–2.07 1.47 1.35–1.60
Model 2: model 1 + demographical and lifestyle factors 2.36 2.13–2.61 1.89 1.71–2.10 1.47 1.34–1.60
(age, gender, income, physical exercise, smoking,
BMI, and AUDIT classification)
Model 3: model 1 + study related factors (study length program, semester) 2.28 2.07–2.50 1.85 1.68–2.04 1.46 1.34–1.58
Model 4: model 1 + social factors (active in social 1.92 1.72–2.15 1.66 1.49–1.84 1.39 1.27–1.52
activities, number of friends, cohabitant status, care for children)
Model 5: model 1 + anxiety (HSCL Anxiety subscale) 1.62 1.47–1.79 1.51 1.37–1.67 1.30 1.19–1.41
and depression (HSCL Depression subscale)
Model 6: fully adjusted model 1.22 1.09–1.31 1.27 1.13–1.42 1.20 1.09–1.31

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
6 A. C. Hayley et al. Scand J Psychol (2016)

Fig. 2. Model of social loneliness as a predictor of difficulties initiating and/or maintaining sleep (DIMS), mediated by symptoms of anxiety and
depression, adjusted for age and sex. The confidence interval for the indirect effect is bias-corrected and accelerated (BCa) bootstrapped CI based on 1,000
samples.

anxiety was statistically significant (indirect effect b = 0.042, following separate adjustment for psychological distress (anxiety
p < 0.001), as was the standardised regression coefficient between and depression). Mediational analyses revealed that depression
and anxiety and DIMS (indirect effect, b = 0.371, p < 0.001). and anxiety largely mediated the association between social
This was similarly observed for the association between social loneliness and self-reported DIMS; however, these symptoms
loneliness and depression (indirect effect b = 0.074, p < 0.001) only partially mediated the relationship between emotional
and depression and self-reported DIMS (indirect effect b = 0.556, loneliness and self-reported DIMS, where a direct association was
p < 0.001), however, the direct effect between social loneliness retained.
and self-reported DIMS was not retained following simultaneous Loneliness is a distressing and pervasive subjective experience
adjustment for psychological distress (b = 0.0001, p > 0.05). which can have manifest deleterious implications for many areas
These results indicate that the association between social of health, including sleep (Hawkley & Cacioppo, 2010). Clinical
loneliness and self-reported DIMS was largely mediated by manipulation studies and in-home assessments have indicated that
symptoms of psychological distress. loneliness is associated with chronic sleep disruption, including
Mediation analyses were conducted to examine whether poorer objective sleep efficiency and greater time spent awake
symptoms of psychological distress mediated the relationships after sleep onset (Cacioppo, Hawkley, Berntson et al., 2002;
between emotional loneliness and self-reported DIMS. Figure 3 Doane & Thurston, 2014). This observation is mirrored among
presents the results of a mediation analyses based on 1,000 most (Kurina et al., 2011; Mahon, 1994), but not all (Hawkley,
bootstrapped samples and adjusted for age and sex. The Preacher & Cacioppo, 2010) epidemiological cohort studies, with
standardised regression coefficient between emotional loneliness lonely individuals typically reporting greater objectively disrupted
and anxiety (indirect b = 0.028) and anxiety and self-reported sleep, including greater sleep fragmentation (Kurina et al., 2011),
DIMS (indirect b = 0.378, both p < 0.001) were found to be as well as higher rates of subjective feelings of poor sleep
significant. This association was similarly observed between efficiency (Mahon, 1994), longer perceived sleep latency and
emotional loneliness and depression (b = 0.056) and depression sleep duration (Cacioppo, Hawkley, Berntson et al., 2002) and
and DIMS (b = 0.517, both p < 0.001), and direct effect between greater daytime impairment (Hawkley et al., 2010). Loneliness
emotional loneliness and self-reported DIMS was sustained has also been shown as a significant moderator of the associations
following simultaneous adjustment for psychological distress between daily stress and objectively monitored sleep duration and
(b = 0.015, p < 0.001). These results indicate that the association latency (Doane & Thurston, 2014). In the current study, those
between emotional loneliness and self-reported DIMS was who experienced social and emotional loneliness were
partially mediated by psychological distress, and a direct link significantly more likely to concurrently report symptoms of
from emotional loneliness to self-reported DIMS was retained. subjective DIMS, with approximately one fifth of individuals
reporting notable subjective sleep difficulty (DIMS responses –
quite a bit and extremely). We similarly observed an independent,
DISCUSSION dose-response association between the highest quartile of
The current study sought to investigate the association between loneliness and subjective DIMS in both unadjusted and fully
social and emotional feelings of loneliness and subjective adjusted regression models; however, this relationship was found
symptoms of DIMS whilst systematically evaluating the to be marginally stronger for social loneliness following
mediational effects of psychological distress in this relationship simultaneous adjustment for confounding factors.
among a group of college-age Norwegian students. Following Mechanistically, social loneliness likely directly impact
adjustment for a number of demographic, lifestyle and social nocturnal sleep (and vice versa) via reciprocal biological
factors, a significant dose-response association was found between pathways which mimic states of autonomic arousal and stress.
the highest quartile of social and emotional loneliness and self- Indeed, these negative bio-psycho-social feedback loops are
reported DIMS. These associations were however attenuated recognised to have significant mediatory effects in the instigation

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Scand J Psychol (2016) Loneliness and sleep in university students 7

Fig. 3. Model of emotional loneliness as a predictor of difficulties initiating and/or maintaining sleep (DIMS), mediated by symptoms of anxiety and
depression, adjusted for age and sex. The confidence interval for the indirect effect is BCa bootstrapped CI based on 1,000 samples.

and maintenance of persistent sleep and psychological pathology, likely that these underlying mechanistic profiles also remain true
and are considered to contribute to the transition of a transient for this sub-category. Conversely, the direct association between
somatic complaint to that of a chronic, treatment-resistant illness emotional loneliness and subjective DIMS was maintained
(Lamont, James, Boivin & Cermakian, 2007). In turn, these following adjustment for psychological symptomology, and thus,
factors are likely prospectively maintained by external it is possible that the discrepancy between satisfying romantic
psychosocial factors and intrinsic psychiatric cues, such as actual relationships and/or the absence of close emotional attachments
and perceived levels of social engagement and support. The (emotional loneliness) have more salient impact on sleep health
association between social loneliness and subjective DIMS may than for perceived lack of satisfying friendship relationships or a
be more closely linked to intrinsic or underlying internal factors, lack of access to social networks (social loneliness), per se.
such as feelings of psychological distress, given the conceptually Clinically, these observations may assist with interventional
close relationship considered to exist between these factors approaches aimed at reducing distress and improving health
(Cunningham, Wheaton & Giles, 2015). Indeed, symptoms of outcomes (including sleep) among affected or vulnerable
psychological distress negatively impact sleep (Cunningham individuals, particularly those with current or lifetime history of
et al., 2015), and a large portion of the variance reported in the psychopathology.
current study was found to be attributable to underlying Interpretation of the findings presented in the current study
psychopathology, indicating the possible mediatory role of must be considered in light of some methodological limitations.
psychological distress in the expression of this association. First, we acknowledge that isolating a single item from the HSCL
Extant clinical, epidemiological and meta-analytic research has scale as a proxy indicator for sleep difficulty (subjective DIMS)
demonstrated the existence of an independent, reciprocal and may introduce a design-driven factor to our analyses. Despite this,
mediational association between psychopathology and sleep the value of single-item sleep variable has previously been noted
health (Alvaro, Roberts & Harris, 2013). Emergent research has in population-based samples (Mallon, Broman & Hetta, 2002),
similarly implicated loneliness as a risk factor for poorer sleep and this methodology is considered to have utility for screening
outcomes in non-clinical samples (Kurina et al., 2011) and the purposes in survey-based research (Kravitz, Ganz, Bromberger,
potential marginal role of psychopathology (Cacioppo et al., Powell, Sutton-Tyrrell & Meyer, 2003; Kuppermann, Lubeck,
2006). Considerably less is known about both the direct and Mazonson et al., 1995). Second, the use of a lesser-known scale
indirect effect of these factors and the possible mediational effect for classifying loneliness may somewhat restrict the generalisation
of underlying psychological symptoms on sleep outcomes. of findings presented within the current study. Despite this, the
Mediational analyses performed in the current study indicated a use of the adapted and extended Social and Emotional Loneliness
significant indirect’ but not direct, relationship between social Scale (Wittenberg, 1986) is advantageous as it differentiates
loneliness and subjective DIMS. Conceptually, it is possible that between two conceptually distinct components of loneliness
the indirect associations observed stemming from social loneliness (Heinrich & Gullone, 2006), and the measure is considered to
to psychological distress (both depression and anxiety) is driven have comparable validity to other well-known loneliness scales
by reciprocal patterns of heightened arousal and distress. In turn, (Robinson et al., 1991). Indeed, loneliness is often assessed as a
the association running from psychological symptomology to unidimensional concept, particularly when assessed in the context
subjective DIMS may be maintained by compromised of sleep outcomes (see, Cacioppo, Hawkley, Berntson et al.,
recuperative processes linked to heightened states of autonomic 2002; Hawkley et al., 2010). Thus, the use of this particular scale
arousal, coupled with reduced metabolic, endothelial and physical provides greater insight and specificity into the degree of
functioning (Meerlo, Sgoifo & Suchecki, 2008). These factors association between each of these individual subscales, and
likely culminate in poorer perceptions of subjective sleep indices. provides greater information as to what drives the observed
Comparably significant indirect mediational pathways were associations between poor sleep and loneliness. Moreover, sleep
observed when emotional loneliness was assessed; and thus it is data was obtained using one joint, subjective variable of DIMS,

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
8 A. C. Hayley et al. Scand J Psychol (2016)

and we were thus unable to examine these singular sleep items Breslau, N., Roth, T., Rosenthal, L. & Andreski, P. (1996). Sleep
separately. Although such self-reported sleep parameters typically disturbance and psychiatric disorders: A longitudinal epidemiological
study of young Adults. Biological Psychiatry, 39, 411–418.
differ from those obtained from objective assessments
Buysse, D. J., Angst, J., Gamma, A., Ajdacic, V., Eich, D. & R€ ossler, W.
(Lauderdale, Knutson, Yan, Liu & Rathouz, 2008), recent studies (2008). Prevalence, course, and comorbidity of insomnia and
have showed that self-reported sleep assessments can be depression in young adults. Sleep, 31, 473–480.
recommended for the characterization of sleep parameters in both Cacioppo, J. T., Hawkley, L. C., Berntson, G. G., Ernst, J. M., Gibbs, A.
clinical and population-based research (Zinkhan, Berger, Hense C., Stickgold, R. & Hobson, J. A. (2002). Do lonely days invade the
nights? Potential social modulation of sleep efficiency. Psychological
et al., 2014). Lastly, the study response rate for this study was
Science, 13, 384–387.
notably low (25.3%), and thus it is possible that the observed Cacioppo, J. T., Hawkley, L. C., Crawford, L. E., Ernst, J. M., Burleson,
associations are somewhat underrepresented, and the observed M. H., Kowalewski, R. B. et al. (2002). Loneliness and health:
associations may, to some effect, be distorted and/or inflated due Potential mechanisms. Psychosomatic Medicine, 64, 407–417.
to the discrepancies in response frequencies between groups. It is Cacioppo, J. T., Hughes, M. E., Waite, L. J., Hawkley, L. C. & Thisted,
R. A. (2006). Loneliness as a specific risk factor for depressive
possible that the use of a single, email-based survey approach
symptoms: Cross-sectional and longitudinal analyses. Psychology and
contributed to this low response-rate, as web-based platforms Aging, 21, 140–151.
typically yield lower overall participation rates when compared to Cunningham, T. J., Wheaton, A. G. & Giles, W. H. (2015). The
traditional mail approaches, such as mail survey or face-to face association between psychological distress and self-reported sleep
interviews (Sheehan, 2001). duration in a population-based sample of women and men. Sleep
Disorders, 8, 1–8. doi:10.1155/2015/172064.
Deane, F. P., Leathern, J. & Spicer, J. (1992). Clinical norms, reliability
and validity for the hopkins symptom checklist–21. Australian Journal
CONCLUSION of Psychology, 44, 21–25.
In summary, the findings of the current study suggest that Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H. & Covi, L.
feelings of social and emotional loneliness are associated with (1974). The Hopkins Symptom Checklist (HSCL): A self-report
symptom inventory. Behavioral Science, 19, 1–15.
increased risk for self-reporting DIMS among Norwegian
Doane, L. D. & Thurston, E. C. (2014). Associations among sleep, daily
college-age students, independent of a number of health, lifestyle experiences, and loneliness in adolescence: Evidence of moderating
and social factors. Examination of the role of underlying and bidirectional pathways. Journal of Adolescence, 37, 145–154.
psychological distress was found to explain this association for Farnill, D. & Robertson, M. F. (1990). Sleep disturbance, tertiary-
social loneliness, however for emotional loneliness, a direct transition stress, and psychological symptoms among young first-year
Australian college students. Australian Psychologist, 25, 178–188.
association was retained. Recognition of the importance of
Ford, D. E. & Kamerow, D. B. (1989). Epidemiologic study of sleep
loneliness as a risk factor for poor subjective sleep parameters disturbances and psychiatric disorders: An opportunity for prevention?
may be important in mitigating the potentially deleterious JAMA, 262, 1479–1484.
peripheral effects on quality of life and overall wellbeing in Hawkley, L. C. & Cacioppo, J. T. (2003). Loneliness and pathways to
young adults. disease. Brain, Behavior, and Immunity, 17, 98–105.
Hawkley, L. C. & Cacioppo, J. T. (2010). Loneliness matters: A
theoretical and empirical review of consequences and mechanisms.
ACH, BS, MK and SO were involved in the development and design of
Annals of Behavioral Medicine, 40, 218–227.
the study. BS, MK and SO collated the data and performed data analyses.
Hawkley, L. C., Preacher, K. J. & Cacioppo, J. T. (2010). Loneliness
ACH, LD, CS, BS, MK and SO interpreted the data and wrote the
impairs daytime functioning but not sleep duration. Health psychology,
manuscript. ACH, LD, CS, BS, MK and SO were involved in drafting,
29, 124–129.
editing and critical appraisal of the manuscript. All authors have approved
Hayley, A. C., Skogen, J. C., Sivertsen, B., Wold, B., Berk, M., Pasco, J.
the manuscript for submission. We would like to thank the members of
A. & Øverland, S. (2015). Symptoms of depression and difficulty
the SHoT steering committee; Hege R akil (SiB), Espen Munkvik (SiT)
initiating sleep from early adolescence to early adulthood: A
and Kari Jussie Lønning (SiO) who initiated and designed the SHoT-
longitudinal study. Sleep, 38, 1599–1606.
survey, and therefore made this research possible.
Hays, A. F. (2013). Introduction to mediation, moderation, and
conditional process analysis: A regression-based approach. New
York, NY: Guilford Publications.
REFERENCES Heinrich, L. M. & Gullone, E. (2006). The clinical significance of
Allen, J. P., Litten, R. Z., Fertig, J. B. & Babor, T. (1997). A review of loneliness: A literature review. Clinical Psychology Review, 26, 695–
research on the alcohol use disorders identification test (AUDIT). 718.
Alcoholism: Clinical and Experimental Research, 21, 613–619. Hoffmann, C., McFarland, B. H., Kinzie, J. D., Bresler, L., Rakhlin, D.,
Alvaro, P. K., Roberts, R. M. & Harris, J. K. (2013). A systematic review Wolf, S. & Kovas, A. E. (2006). Psychological distress among recent
assessing bidirectionality between sleep disturbances, anxiety, and Russian immigrants in the United States. International Journal of
depression. Sleep, 36, 1059–1068. Social Psychiatry, 52, 29–40.
Baglioni, C., Battagliese, G., Feige, B., Spiegelhalder, K., Nissen, C., Hsu, L. R., Hailey, B. J. & Range, L. M. (1987). Cultural and emotional
Voderholzer, U. et al. (2011). Insomnia as a predictor of depression: A components of loneliness and depression. The Journal of Psychology,
meta-analytic evaluation of longitudinal epidemiological studies. 121, 61–70.
Journal of Affective Disorders, 135, 10–19. Hysing, M., Pallesen, S., Stormark, K. M., Lundervold, A. J. & Sivertsen,
Baldwin, C. M., Ervin, A.-M., Mays, M. Z., Robbins, J., Shafazand, S., B. (2013). Sleep patterns and insomnia among adolescents: A
Walsleben, J. & Weaver, T. (2010). Sleep disturbances, quality of life, population-based study. Journal of Sleep Research, 22, 549–556.
and ethnicity: The Sleep Heart Health Study. Journal of Clinical Sleep Jackson, J. & Cochran, S. D. (1991). Loneliness and psychological
Medicine: JCSM, 6, 176–183. distress. The Journal of Psychology, 125, 257–262.
Baumeister, R. F. & Leary, M. R. (1995). The need to belong: Desire for Janson, C., Gislason, T., De Backer, W., Plaschke, P., Bj€ ornsson, E.,
interpersonal attachments as a fundamental human motivation. Hetta, J. et al. (1995). Prevalence of sleep disturbances among young
Psychological Bulletin, 117, 497–529. adults in three European countries. Sleep, 18, 589–597.

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd
Scand J Psychol (2016) Loneliness and sleep in university students 9

Jensen, E., Dehlin, O., Hagberg, B., Samuelsson, G. & Svensson, T. Robinson, J. P., Shaver, P. R. & Wrightsman, L. S. (1991). Measures of
(1998). Insomnia in an 80-year-old population: Relationship to personality and social psychological attitudes. San Diego, CA:
medical, psychological and social factors. Journal of Sleep Research, Academic Press.
7, 183–189. Russell, D., Cutrona, C. E., Rose, J. & Yurko, K. (1984). Social and
Kravitz, H. M., Ganz, P. A., Bromberger, J., Powell, L. H., Sutton-Tyrrell, emotional loneliness: An examination of Weiss’s typology of
K. & Meyer, P. M. (2003). Sleep difficulty in women at midlife: A loneliness. Journal of Personality and Social Psychology, 46, 1313–
community survey of sleep and the menopausal transition. Menopause, 1321.
10, 19–28. Russell, D., Peplau, L. A. & Ferguson, M. L. (1978). Developing a
Kuppermann, M., Lubeck, D. P., Mazonson, P. D., Patrick, D. L., Stewart, measure of loneliness. Journal of Personality Assessment, 42, 290–
A. L., Buesching, D. P. & Filer, S. K. (1995). Sleep problems and 294.
their correlates in a working population. Journal of General Internal Saunders, J. B., Aasland, O. G., Babor, T. F., de la Fuente, J. R. & Grant,
Medicine, 10, 25–32. M. (1993). Development of the Alcohol Use Disorders Identification
Kurina, L. M., Knutson, K. L., Hawkley, L. C., Cacioppo, J. T., Test (AUDIT): WHO Collaborative Project on Early Detection of
Lauderdale, D. S. & Ober, C. (2011). Loneliness is associated with Persons with Harmful Alcohol Consumption–II. Addiction, 88, 791–
sleep fragmentation in a communal society. Sleep, 34, 1519–1526. 804.
Lamont, E. W., James, F. O., Boivin, D. B. & Cermakian, N. (2007). Segrin, C. & Burke, T. J. (2015). Loneliness and sleep quality:
From circadian clock gene expression to pathologies. Sleep Medicine, Dyadic effects and stress effects. Behavioral Sleep Medicine, 13, 241–
8, 547–556. 254.
Lauderdale, D. S., Knutson, K. L., Yan, L. L., Liu, K. & Rathouz, P. J. Sheehan, K. B. (2001). E-mail survey response rates: A review. Journal of
(2008). Self-reported and measured sleep duration: How similar are Computer-Mediated Communication, 6, 2.
they? Epidemiology (Cambridge, Mass.), 19, 838–845. Stravynski, A. & Boyer, R. (2001). Loneliness in relation to suicide
Mahon, N. E. (1994). Loneliness and sleep during adolescence. Perceptual ideation and parasuicide: A population-wide study. Suicide and Life-
and Motor Skills, 78, 227–231. Threatening Behavior, 31, 32–40.
Mallon, L., Broman, J. E. & Hetta, J. (2002). Sleep complaints predict Tsai, M.-C., Tsai, Y.-F., Chen, C.-Y. & Liu, C.-Y. (2005). Alcohol Use
coronary artery disease mortality in males: a 12–year follow-up study Disorders Identification Test (AUDIT): Establishment of cut-off scores
of a middle-aged Swedish population. Journal of Internal Medicine, in a hospitalized Chinese population. Alcoholism: Clinical and
251, 207–216. Experimental Research, 29, 53–57.
Meerlo, P., Sgoifo, A. & Suchecki, D. (2008). Restricted and disrupted Veijola, J., Jokelainen, J., L€aksy, K., Kantoj€arvi, L., Kokkonen, P.,
sleep: Effects on autonomic function, neuroendocrine stress systems J€arvelin, M.-R. & Joukamaa, M. (2003). The Hopkins Symptom
and stress responsivity. Sleep Medicine Reviews, 12, 197–210. Checklist-25 in screening DSM–III–R axis–I disorders. Nordic Journal
Mouanoutoua, V. L. L. G. (1995). Hopkins Symptom Checklist-25, of Psychiatry, 57, 119–123.
Hmong Version: A screening instrument for psychological distress. Victor, C. R. & Yang, K. (2012). The prevalence of loneliness among
Journal of Personality Assessment, 64, 376–383. adults: A case study of the United Kingdom. The Journal of
National Heart, Lung and Blood Institute (1998). Clinical guidelines on Psychology, 146, 85–104.
the identification, evaluation, and treatment of overweight and obesity Wilson, R. S., Krueger, K. R., Arnold, S. E., Schneider, J. A., Kelly, J. F.,
in adults: The evidence report. Obesity Reviews, Suppl 2, 51S–209S. Barnes, L. L. et al. (2007). Loneliness and risk of alzheimer disease.
Neckelmann, D., Mykletun, A. & Dahl, A. A. (2007). Chronic insomnia Archives of General Psychiatry, 64, 234–240.
as a risk factor for developing anxiety and depression. Sleep, 30, 873– Wittenberg, M. (1986). Emotional and social loneliness: An examination
880. of social skills, attributions, sex role, and object relations perspectives.
Nyman, E., Miettunen, J., Freimer, N., Joukamaa, M., M€aki, P. & (Doctoral dissertation), University of Rochester.
Ekelund, J. et al. (2011). Impact of temperament on depression and Zawadzki, M. J., Graham, J. E. & Gerin, W. (2013). Rumination and
anxiety symptoms and depressive disorder in a population–based birth anxiety mediate the effect of loneliness on depressed mood and sleep
cohort. Journal of Affective Disorders, 131, 393–397. quality in college students. Health Psychology, 32, 212–222.
O’Brien, R. M. (2007). A caution regarding rules of thumb for variance Zinkhan, M., Berger, K., Hense, S., Nagel, M., Obst, A., Koch, B. et al.
inflation factors. Quality and Quantity, 41, 637–690. (2014). Agreement of different methods for assessing sleep
Ohayon, M. M. & Roth, T. (2003). Place of chronic insomnia in the characteristics: a comparison of two actigraphs, wrist and hip
course of depressive and anxiety disorders. Journal of Psychiatric placement, and self-report with polysomnography. Sleep Medicine, 15,
Research, 37, 9–15. 1107–1114.
Peplau, L. A. & Perlman, D. (1982). Perspectives on loneliness. New
York: Wiley. Received 7 August 2016, accepted 11 October 2016

© 2016 Scandinavian Psychological Associations and John Wiley & Sons Ltd