Content uploaded by Eivind Meland
Author content
All content in this area was uploaded by Eivind Meland on Dec 19, 2019
Content may be subject to copyright.
Page Proof Instructions and Queries
Thank you for choosing to publish with us. This is your final opportunity to ensure your article will be accurate at publication. Please review your proof carefully and
respond to the queries using the circled tools in the image below, which are available in Adobe Reader DC* by clicking Tools from the top menu, then clicking Comment.
Please use only the tools circled in the image, as edits via other tools/methods can be lost during file conversion. For comments, questions, or formatting requests, please
use . Please do not use comment bubbles/sticky notes .
*If you do not see these tools, please ensure you have opened this file with Adobe Reader DC, available for free at https://get.adobe.com/reader or by going to Help >
Check for Updates within other versions of Reader. For more detailed instructions, please see
https://
us.sagepub.com/ReaderXProofs.
No. Query
No queries
As you are aware there is a $100 charge per page for all pages over 5 pages. If your article is over 5pp in total
extent, please provide your institute’s VAT number (if you are in the EU) so that our Finance department can
send you the necessary invoice.
No. Query
Journal Title: Scandinavian Journal of Public Health
Article Number: 888044
Please note, only ORCID iDs validated prior to acceptance will be authorized for publication; we are unable to add or amend ORCID iDs
at this stage.
Please confirm that all author information, including names, affiliations, sequence, and contact details, is correct.
Please review the entire document for typographical errors, mathematical errors, and any other necessary corrections; check headings,
tables, and figures.
Please ensure that you have obtained and enclosed all necessary permissions for the reproduction of art works (e.g. illustrations,
photographs, charts, maps, other visual material, etc.) not owned by yourself. Please refer to your publishing agreement for further
information.
Please note that this proof represents your final opportunity to review your article prior to publication, so please do send all of your changes
now.
Please confirm that the Funding and Conflict of Interest statements are accurate.
888044SJP
https://doi.org/10.1177/1403494819888044
© Author(s) 2019
Article reuse guidelines: sagepub.com/journals-permissions
DOI: 10.1177/1403494819888044
journals.sagepub.com/home/sjp
Scandinavian Journal of Public Health, 1–9
Introduction
In Norway, approximately 40% of adolescents will
have a divorce experience (DE) by the age of 17 years,
with considerable regional differences [1]. Research
pertaining to health consequences for adolescents
exposed to marital dissolution is conflicting [2].
Earlier research focused primarily on the DE in itself,
while the focus in recent years also includes the qual-
ity of parent–child relations, custody arrangements,
inter-parental conflicts, the child’s cognitive and emo-
tional processing, concomitant changes in socio-eco-
nomic status (SES) and pre-existing psychological
vulnerability [3–6]. Without a broader perspective, it
is difficult to explain how some children fare badly
after DE, while others may benefit from moving to
less conflicting contexts or preserving relations with
both parents [2–4,6].
The view that marital conflicts are deleterious to
the health and well-being among children and ado-
lescents is widely held [1,3]. A ‘spill-over’ effect,
where the inter-parental conflict spills over and ruins
the relation between the parent and the child, may be
responsible for this association [3]. We should keep in
mind that not only overt aggression but also covert
passive aggressive control over the child is detrimen-
tal, leading especially to internalising problems [6].
Divorce and conversational difficulties with parents:
Impact on adolescent health and self-esteem
EIVIND MELAND1, HANS JOHAN BREIDABLIK2 & FRODE THUEN3
1Department of Global Public Health and Primary Care, Research Group for General Practice, Universitetet i Bergen Det
medisinsk-odontologiske fakultet, Norway, 2Department of Research and Development, District General Hospital of Førde,
Norway, and 3Center for Evidence-Based Practice, Western Norway University of Science, Norway
Abstract
Background: Divorce experience (DE) may cause health and self-conceptual problems, but these consequences might also be
caused by conflicts and lack of conversational confidence (CC) with one or both parents. We investigated how DE impacted
CC and how DE and CC impacted health complaints and self-esteem in a two-year longitudinal cohort study. Methods: The
study was performed between 2011 and 2013 among 1225 students in junior high school (aged 12 and 14 years in 2011).
We used binary logistic analyses to account for how DE impacted CC, and linear regression analyses to examine how DE
and CC impacted on subjective health and self-esteem in 2013. Results: The study revealed that former and recent DEs
impacted CC with fathers only. The impact was most evident for the more severe forms of conversational difficulties. DE
in itself predicted only self-esteem, and CC with parents mediated this association. CC with both mothers and fathers had
strong temporal causal associations with the outcomes two years later. Only CC with fathers impacted changes of the health
complaints and self-esteem in full-model residual change analyses. Conclusions: The study proves a sex-specific effect
on loss of CC between fathers and children after divorce. The impairment of CC has predictive repercussions on
the health and self-conception of adolescents in their middle teenage years. From a public-health perspective,
preserving the relation and the confidence between children and their fathers after divorce seems an important
task.
Keywords: Adolescence, parent–child relations, subjective health complaints, self-esteem, gender
Correspondence: Eivind Meland, Department of Global Public Health and Primary Care, Research Group for General Practice, Universitetet i Bergen Det
medisinsk-odontologiske fakultet, Kalfarveien 31, Bergen, 5018, Norway. E-mail: eivind.meland@uib.no
Date received 11 May 2019; reviewed 3 October 2019; accepted 7 October 2019
0010.1177/1403494819888044E. Meland et al.Divorce and conversational difficulties
research-article2019
ORIGINAL ARTICLE
2 E. Meland et al.
Several researchers modify, however, the claim that
conflict per se is detrimental for the child, maintain-
ing that impairment of health and well-being depends
on whether the child is allowed to preserve contact
with both parents and can benefit from good rela-
tions with both after divorce [4,7].
The research is unanimous that good relations
between children and their parents are promotive for
the health and well-being of the child, and that this
also builds the foundation for health later in life [6,8].
The mother–child relationship was the main focus in
earlier relational research, but during recent years,
the importance of the father–child relationship has
been emphasised [9,10].
A common measure for the quality of the relations
with parents is the self-reported ease of talking with
parents about things that are bothersome for the
child. This measure has been used in the Health
Behaviour for School-Aged Children (HBSC) stud-
ies. Young people reporting ease of communication
with their parents report positive body image [11],
higher self-rated health [12], higher life satisfaction
[13] and fewer physical and psychological complaints
[14]. Conversational confidence (CC) with fathers
has a preventive effect in maintaining emotional well-
being [15] and a positive sense of body image, par-
ticularly among girls [11]. Confidence in talking with
fathers is also associated with less aggression and vio-
lence among boys [16].
In a former cross-sectional study, two of the authors
of this paper (E.M. and H.J.B.) found that health dete-
riorated only slightly with DE, but if DE was accompa-
nied by loss of parental contact (almost exclusively
with fathers), psychological and somatic health com-
plaints were markedly exaggerated [17]. This study
from Western Norway is in accordance with most
research showing that father absence is a predictor of
detrimental psychological well-being and life adversity
for developing children and adolescents [9,10].
The review paper by East etal. admits, however,
that research on the father–child relation is inade-
quate, and they call for studies pertaining to how
divorce and the relations with each of the parents
impact health and self-conception among adoles-
cents [9]. In the present study, we were able to exam-
ine these associations in a longitudinal study, enabling
causal inferences with greater confidence.
With this background, the aims of the present
study were: to examine prospectively how DE
impacted CC with mothers and fathers; to examine
how DE and CC impacted emotional and somatic
health complaints and self-esteem; and to examine to
what extent CC with mothers and fathers mediated
or moderated the effect of DE during two years of
follow-up.
Methods
Participants
All municipalities in the county of Sogn og Fjordane
in western Norway were invited to participate in the
survey, and all except one accepted the invitation. In
2011, 68% of a total of 3075 students in grades 6 and
8 took part. In 2013, 71% of 4538 student from
grades 6, 8 and 10 responded, and 101 different
schools participated in both surveys. The total popu-
lation in Sogn og Fjordane in 2013 was 108,700
inhabitants.
The main reason for non-participation was
absence from school on the day of data collection.
Some classes dropped out because of logistical prob-
lems, but the participation across grade levels was
fairly similar: 983 in grade 6, 1050 in grade 8 and
1189 in grade 10. Students from grades 8 and 10
answered the same survey in 2011, but we only iden-
tified 1225 by person-specific codes due to insuffi-
cient coding. The coding insufficiencies were
randomly distributed between people and classes,
although insufficiencies were more prevalent among
the youngest (aged 12 years in 2011). We followed a
total of 612 boys and 613 girls: 475 from grades 6–8,
and 750 from grades 8–10 (i.e. 1225 students with
an almost identical sex distribution across the two
cohorts). This subsample comprised 48% of the orig-
inal students measured in 6th grade, and 71% of the
students measured in 8th grade in 2011. The survey
was administered late in the fall term (November/
December) at both points in time, which is consid-
ered as a stable period in the semester. Eighty-six per
cent of the participants lived in rural municipalities.
The vast majority of students completed the ques-
tionnaires. Missing data affected a maximum of 8%
concerning CC with fathers, and boys were slightly
more reluctant to answer this question than girls
were.
Measures
Most of the questions in the survey were from the
international World Health Organization–led HBSC
project, and they proved valid and reliable in former
studies [17–19]. We assessed subjective health com-
plaints (SHC) using the HBSC Symptom Checklist
extended version. Eleven questions regarding differ-
ent physical (headache, neck ache, backache, abdom-
inal pain and dizziness) and emotional health
complaints (feeling low and sad, tired and exhausted,
afraid, irritable or bad tempered, nervous, difficult
falling asleep) during the last six months constituted
the questionnaire. The answers were: ‘about every
day’, ‘more than once a week’, ‘about every week’,
Divorce and conversational difficulties 3
‘about every month’ and ‘rarely or never’. We com-
puted the mean score of somatic (five items) and
emotional health complaints (six items). The SHC
measure has been extensively used in the HBSC sur-
veys and has proved valid and reliable [18].
SES was measured with one question regarding
family finances. The pupils were asked about how
‘well off’ they considered their family to be. The
answer alternatives varied from ‘very well off’ to
‘badly off’ on a five-point Likert scale. Increasing val-
ues indicated less affluence. This question has been
used to measure SES among adolescents and has
been associated with SHC in previous studies [20].
Self-esteem (SE) was measured with Rosenberg’s
self-esteem scale [21]. The scale has 10 questions
designed to measure students’ overall evaluation of
their worthiness as human beings (e.g. ‘On the whole,
I am satisfied with myself’). The answers ranged
from ‘strongly disagree’ to ‘strongly agree’ on a four-
point Likert scale. Answers were recoded in order to
give the same direction for all answers, indicating
increased values for improved self-esteem. The
Cronbach’s alphas of the measures varied from 0.78
to 0.90, as shown in Table I.
A former study of the same population showed
that <5% of the variance in the variables was
accounted for at the level of class belongingness [22].
Therefore, we abstained from performing multilevel
analyses.
By 2011, 213 (17%) participants had experienced
divorce, and additionally, 58 (5%) had experienced the
same during the two years from 2011 to 2013, as pre-
sented in Table I. The table also shows how the students
graded their CC with mothers and fathers. Evident
from the numbers, we see that the adolescents had
greater confidence in talking about bothersome topics
with their mothers than with their fathers, although
confidence with fathers was also prevalent. These ques-
tions were well understood, with few missing answers
(maximum 5%). The measures have proved valid, with
Table I. Sex, age, experience of divorce, report of CC (the ease of talking about bothersome topics), SES, somatic and emotional health
complaints and self-esteem among 1225 students from Sogn og Fjordane surveyed at two time points (2011 and 2013).
Variables N (%) or valid
responses
Response options
(Likert scale)
Mean
(SD)
Min–max Cronbach’s
alpha
Girls 613 (50)
Boys 612 (50
8th grade in 2013 (13 years of age) 475 (39)
10th grade in 2013 (15 years of age) 750 (41)
Experienced divorce up to 2011 213 (17)
Experienced divorce 2011–2013 58 (5)
CC mother, 2011
Very easy talking with mother 487 (43)
Easy talking with mother 508 (44)
Difficult talking with mother 119 (10)
Very difficult or no contact with mother 32 (3)
CC father, 2011
Very easy talking with father 277 (25)
Easy talking with father 498 (44)
Difficult talking with father 260 (21)
Very difficult or no contact with father 94 (8)
Self-rated family affluence 1181 5 (1–5) 2.2 (0.7) 1–5
Somatic health complaints 2013 1211 1.8 (0.8) 1–5 0.78
Headache 1203 5 (1–5) 1–5
Abdominal pain 1198 5 (1–5) 1–5
Backache 1197 5 (1–5) 1–5
Dizziness 1207 5 (1–5) 1–5
Neck ache 1205 5 (1–5) 1–5
Emotional health complaints 2013 1207 2.0 (0.8) 1–5 0.82
Feeling low and sad 1190 5 (1–5) 1–5
Irritable or bad temper 1195 5 (1–5) 1–5
Nervous 1197 5 (1–5) 1–5
Difficult falling asleep 1205 5 (1–5) 1–5
Tired and exhausted 1203 5 (1–5) 1–5
Afraid 1201 5 (1–5) 1–5
Rosenberg self-esteem scale 1207 3.0 (0.6) 1–4 0.90
CC: conversational confidence; SES: socio-economic status;
4 E. Meland et al.
consistent associations with SHC and SE in former
studies [14,16,17,23]. In a former study, the prevalence
of divorce from students’ self-report conformed well
with population statistics [17].
Statistics
We presented the population with frequencies, and
we calculated the mean scores for each scale variable.
To be included in the mean score calculation,
respondents were required to answer at least 50% of
the items constituting the scale. The number of stu-
dents with valid responses on the construct measure
was accordingly higher than valid subcategory
responses. Respondents with fewer answers were not
included in the analyses. The variables were normally
distributed, except for somatic health complaints
(skewness 1.4 in 2013). Because we had a large sam-
ple and all skewness values were <2, we decided to
apply parametric statistical tests [24]. The variable
tapping self-rated family affluence had a normal dis-
tribution, with a skewness of −0.6. We entered this
variable in the regression models as a continuous
variable.
In order to judge how DE influenced the partici-
pants’ parental CC in 2013, we performed binary
logistic regressions with different cut-offs of confi-
dence as outcomes. In four separate binary logistic
regressions (two models for each parent), we entered
sex, age and self-rated family affluence (in 2011) as
adjusting variables in the analyses. Additionally, we
checked if the associations between DE and CC were
reciprocal. In the reciprocal analysis, DE during
2011–2013 among those not reporting DE in 2011
was used as the outcome, and the CC in 2011 with
mothers and fathers was the predictor in adjusted
logistic regressions controlling for sex, age and self-
reported family affluence.
CC with mothers and fathers was recoded into
four categories (‘very difficult’ and ‘no contact’ were
combined). Thereafter, we recoded the confidence
with mothers and fathers into three dummies. The
category with the most unsatisfactory confidence
(‘very difficult talking’ and ‘no contact’) served as the
reference.
To examine the predictive impact of DE and CC
with each of the parents on the outcomes, we first
performed linear regression analyses with the adjust-
ing variables entered in the analyses one by one.
Thereafter, we performed linear regressions with the
predictors one by one, with adjustments for sex, age
and family affluence. We applied temporal causal
analyses where the adjusting variables and all the sig-
nificant predictors were entered in the models.
Finally, we performed residual change analyses where
we entered the status of the outcome measures in
2011 in the linear regression models. Correlation
coefficients between the outcomes in 2011 and 2013
varied from 0.46 to 0.51, justifying the use of residual
change analyses.
The strongest correlation between the predictors
was just below 0.2, ruling out multicollinearity. We
also checked if the associations between predictors
and outcomes were different between girls and boys.
When performing stratified regression analyses
among boys and girls separately for the associations
between divorce and CC and all three main out-
comes, we found generally that the absolute values of
betas were somewhat higher for girls compared to
boys, but the confidence intervals (CIs) were vastly
overlapping. Therefore, we introduced sex as an
adjusting variable and performed the predictive anal-
yses with both sexes together.
We performed interaction analyses between DE
and CC by multiplying the dichotomous DE varia-
bles and the dummy-coded CC variables with each
other, and entered the interaction term and both
variables constituting the interaction term in linear
regression models with the health complaints and
self-esteem as outcomes.
IBM SPSS Statistics for Windows v25 (IBM
Corp., Armonk, NY) was used for the analyses. A
p-value of <0.05 was accepted as significant.
Ethics
The study was approved by the Norwegian regional
committee for ethics in medical research (approval
number 2011/510 REK vest). Pupil participation
was confidential and voluntary. Informed written
consent was obtained from the parents and pupils.
Results
We compared the participants in the study to all
study participants in 2013 from grades 8 and 10 in
order to reveal any systematic drop-out. DE and SES
were similar. The reported CC with both parents
were slightly better in the full sample compared to
participants in the present study (mean score differ-
ence=0.20 and 0.25; SD=0.83 and 0.95 for mothers
and fathers, respectively). The reason for the discrep-
ancy was that the participants were older in the lon-
gitudinal study (more drop-outs in the youngest
cohort), and the reported confidence diminished
with age.
Table II elaborates how DE before 2011 and dur-
ing 2011–2013 impacted the CC with both parents in
2013. From Table II, we see that compared to children
without DE, the CC with mothers was not affected.
Divorce and conversational difficulties 5
For fathers, we see that the conversational difficulties
were associated with DE, and that this was highly sig-
nificant, as well as for children with recent DE.
We tested if these associations were reciprocal
with CC in 2011 with mothers and fathers as predic-
tors and DE during 2011–2013 as outcome (not
shown in the tables). We adjusted for sex, age and
family affluence. The odds ratios for DE increased
from 0.7 to 1.8 for fathers and from 1.0 to 1.5 for
mothers with increasing conversational difficulties,
none of them statistically significant and with vastly
overlapping CIs.
In Table II, we also see that girls were more dis-
content with their CC with fathers compared to boys.
Boys and girls were equally satisfied with their CC
with mothers. We can also see that the oldest are less
satisfied with their CC, and that this is evident for the
most severe conversational difficulties with fathers.
The socio-economically disadvantaged reported
fewer conversational problems with fathers com-
pared to more advantaged peers.
Tables III–V report the results from the linear
regression predictor analyses. Somatic and emotional
health complaints and self-esteem were reported in
2013, while the same students reported the predic-
tors in 2011. The first full-model analysis reports the
prediction model from the adjusting variables and
the significant predictors using the status in 2013 as
outcomes (temporal causal analyses). The last two
columns report the results from the residual change
analyses (the outcome measured in 2011 entered in
the models), and give estimates on how the predic-
tors influence the change of the outcomes during the
two-year observation time.
We see from Tables III–V that DE in itself has little
influence on the self-report of health complaints and
self-esteem among the participants two years later. The
only significant association was the influence of recent
DE with self-esteem by the end of the observation in
2013. As Table V shows, this predictor lost its significant
predictive power in the full-model analysis, as it was
mediated by the other predictors in the model.
From Tables III–V, we see that CC with both par-
ents has a strong predictive power on the health com-
plaints and self-conception two years later. The
impact was apparent for every level of CC compared
to the reference category (‘very difficult’ and ‘no con-
tact’). Girls reported more health complaints than
boys did, and this was also the case when it was
examined in full-model analyses.
In all the final models, examining the residual
change associations, the CC with the fathers and not
the mothers was predictive for the change of the out-
comes. We also see from Tables III–V that the explained
variances of the models were satisfactory, document-
ing the relevance of the predictors and the models.
Interaction analyses, not shown in the tables,
revealed that interactions between DE and CC with
mothers were modest and statistically insignificant
for all interaction terms. The level of confidence with
fathers moderated, however, the impact of DE on all
the outcomes. The interaction term between DE
until 2011 and impaired CC with fathers impacted
emotional as well as somatic health complaints two
years later, with standardised coefficients of −0.09
(p=0.03) for both interactions. The interaction term
between DE during 2011–2013 and very easy CC
with fathers impacted self-esteem and somatic health
Table II. Participants’ level of CC with mothers and fathers according to their DE compared to non-divorced, expressed with ORs and 95%
CIs from binary logistic analyses with all predictors in the models.
Level of CCaMother (2013)
OR
95% CI Mother (2013)
OR
95% CI Father (2013)
OR
95% CI Father (2013)
OR
95% CI
Difficult, very
difficult and no
contact (N=1129)
Very difficult
and no contact
(N=1129)
Difficult, very
difficult and no
contact (N=1070)
Very difficult
and no contact
(N=1070)
Predictors
Divorce until 2011 1.0 0.7–1.4 1.2 0.7–2.0 2.0*** 1.5–2.8 2.7*** 1.7–4.2
Divorce during 2011–2013 1.2 0.6–2.2 0.8 0.3–2.3 1.5 0.9–2.8 3.0** 1.4–6.3
Sex (girls) 1.2 0.7–2.2 1.1 0.7–1.7 2.2*** 1.7–2.8 2.0*** 1.3–3.1
Age (15 years in 2013) 1.3* 1.0–1.8 0.9 0.8–1.7 1.5 1.2–2.0 2.7*** 1.7–4.2
Self-rated family affluence 0.9 0.7–1.1 0.9 0.7–1.2 0.8** 0.6–0.9 0.7* 0.5–1.0
Constant 2.7* 12.7*** 0.7 1.6
CC with two different cut-off levels served as outcomes in four models (two models for each parent). The models were adjusted for sex, age
(11 and 13 years in 2011) and self-reported family affluence (increasing values with less affluence). Significant ORs are shown in bold, and
increasing OR values indicate increased conversational difficulty.
*p<0.05; **p<0.01; ***p<0.001.
aThe categories in this row are merged into one level of conversational difficulty and compared tp a merged level of the remaining categories.
DE: divorce experience; OR: odds ratio; CI: confidence interval.
6 E. Meland et al.
complaints in 2013, with standardised coefficients of
−0.07 (p=0.04) and 0.10 (p=0.004), respectively.
Discussion
The present study revealed that former and recent DE
impacted CC with fathers only. The impact was most
evident for the more severe forms of conversational
difficulties. The associations between DEs and
impaired CC with fathers (or mothers) were not recip-
rocal. DEs in itself predicted health complaints and
self-esteem negligibly. The only outcome that was
associated with DE was the association between recent
DE and self-esteem at follow-up, and this association
Table III. Linear regression analyses with adjusting variables (sex, age, self-rated family affluence) and predictors (DE and CC with fathers
and mothers) of somatic health complaints in 2013 according to their status in 2011 (temporal causal analyses) expressed as standardised
regression coefficients (b).
Variables b Unadjusted
and adjusted
p-value b Full modelb
(N=1095)
p-Value b Full model residual
changeb (N=1095)
p-Value
Unadjusted analyses
Sex (boys 1, girls 2) 0.24c<0.001 0.23 <0.001 0.18 <0.001
Age 0.06c0.03
Self-rated family affluence 0.08c0.005 0.06 0.04
Adjusted for sex, age and self-rated family affluence
Divorce until 2011 0.03 0.27
Divorce 2011–2013 0.04 0.14
Difficult talking with fathera−0.17 0.001 −0.13 0.02 −0.11 0.03
Easy talking with fathera−0.33 <0.001 −0.23 <0.001 −0.18 0.001
Very easy talking with fathera−0.22 <0.001 −0.13 0.03 −0.10 0.07
Difficult talking with mothera−0.14 0.02 −0.08 0.21
Easy talking with mothera−0.42 <0.001 −0.27 0.007
Very easy talking with mothera−0.41 <0.001 −0.28 0.005
Somatic health complaints in 2011 0.41 <0.001
Explained variance for the full model analyses 0.11 0.27
The last two columns pertain to the change of somatic health complaints (residual change) from 2011 to 2013.
aAnalysed as dummy variables with ‘very difficult and no contact’ as comparison category.
bAll significant predictors in the unadjusted and adjusted analyses are entered in the full model, but only significant predictors are shown.
cBivariate analyses with the adjusting variables entered one by one.
Table IV. Linear regression analyses with adjusting variables (sex, age, self-rated family affluence) and predictors (DE and CC with fathers
and mothers) of emotional health complaints in 2013 according to their status in 2011 (temporal causal analyses) expressed as standardised
regression coefficients (b).
Variables b Unadjusted
and adjusted
p-Value b Full modelb
(N=1095)
p-Value b Full model residual
changeb (N=1095)
p-Value
Unadjusted analyses
Sex (boys 1, girls 2) 0.31c<0.001 0.28 <0.001 0.22 <0.001
Age 0.12c<0.001 0.09 0.003 0.06 0.02
Self-rated family affluence 0.09c0.001 0.07 0.02
Adjusted for sex, age and self-rated family affluence
Divorce until 2011 0.04 0.18
Divorce 2011–2013 0.04 0.17
Difficult talking with fathera−0.18 <0.001 −0.14 0.007 −0.12 0.02
Easy talking with fathera−0.37 <0.001 −0.29 <0.001 −0.22 <0.001
Very easy talking with fathera−0.30 <0.001 −0.22 <0.001 −0.15 0.005
Difficult talking with mothera−0.17 0.003 −0.11 0.08
Easy talking with mothera−0.44 <0.001 −0.25 0.007
Very easy talking with mothera−0.46 <0.001 −0.27 0.005
Emotional health complaints in 2011 0.36 <0.001
Explained variance for the full model analyses 0.17 0.29
The last two columns pertain to the change of emotional health complaints (residual change) from 2011 to 2013.
aAnalysed as dummy variables with ‘very difficult and no contact’ as comparison category.
bAll significant predictors in the unadjusted and adjusted analyses are entered in the full model, but only significant predictors are shown.
cBivariate analyses with the adjusting variables entered one by one.
Divorce and conversational difficulties 7
was mediated by the CC with parents. CC with both
mothers and fathers had strong temporal causal asso-
ciations with the outcomes two years later. Only CC
with fathers impacted changes of the health com-
plaints and self-esteem in full-model residual change
analyses, and only confidence with fathers modified
the associations between DEs and the outcomes.
The weaknesses of the study include the relatively
small number of participants experiencing recent
DE, the large proportion lost to follow-up and pos-
sibly also the context in mostly rural districts in west-
ern Norway. The great drop-out rate due to
insufficient coding, especially in the youngest age
group, represented a threat to the external validity of
the study. Selective drop-out of those with better CC
with parents may reduce the generalisability of the
study. As the differences impacted both parents, it is
unlikely that the selective importance of fathers
should be invalid. Reliance on self-reports and an
identical questionnaire at both time points introduce
the possibility of common method variance.
The strengths of the study were its longitudinal
design and the evaluation of CC with both mothers
and fathers during two-year follow-up. We examined
both health impairment and self-conceptual prob-
lems. Both mediation and moderation were exam-
ined, and we performed both temporal causal and
residual change analyses. The study also adjusted for
possible confounders, most importantly the self-
rated SES that is linked with both DE and subjective
health.
The county of Sogn og Fjordane is mostly a rural
district county, although urban areas exist. The prev-
alence of divorce is lower than the country average.
In several studies, however, two of the authors of this
study (E.M. and H.J.B.) have demonstrated that
adolescent health and health behaviour problems are
similar to national and even international findings
[17,25]. The external validity of our findings there-
fore seems safeguarded. Due to small numbers (58
experiencing divorce during 2011–2013) and lack of
statistical power, we cannot rule out reciprocal cau-
sality, although the analyses showed statistically
insignificant findings.
Our study demonstrating lack of or negligible asso-
ciations between DE and later health and self-concep-
tual problems is somewhat conflicting with earlier
research. An early meta-analysis confirmed that
divorce was associated with reduced emotional well-
being and other impairments [26], although the
authors emphasised that these associations were atten-
uated when controlling for relevant confounders. The
former study from Sogn og Fjordane demonstrated
that the health impairments were mostly confined
among adolescents losing contact with the non-custo-
dial parent [17]. The associations between single-par-
ent custody and health and health behaviour
impairments were evident in a long-term follow-up
study with several measurement points from Norway
[27], and father absence also predicted depressive
symptoms in a longitudinal study from the UK [10].
Recent reviews emphasise the importance of parental
Table V. Linear regression analyses with adjusting variables (sex, age, self-rated family affluence) and predictors (DE and CC with fathers
and mothers) of self-esteem in 2013 according to their status in 2011 (temporal causal analyses) expressed as standardised regression coef-
ficients (b).
Variables b Unadjusted
and adjusted
p-Value b Full modelb
(N=1095)
p-Value b Full model residual
changeb (N=1095)
p-Value
Unadjusted analyses
Sex (boys 1, girls 2) −0.41c<0.001 −0.38 <0.001 −0.29 <0.001
Age −0.12c<0.001 −0.07 0.006
Self-rated family affluence −0.09c0.002 −0.06 0.003
Adjusted for sex, age and self-rated family affluence
Divorce until 2011 −0.05 0.06
Divorce 2011–2013 −0.07 0.01
Difficult talking with fathera0.16 <0.001 0.11 0.02 0.09 0.04
Easy talking with fathera0.31 <0.001 0.23 <0.001 0.17 0.001
Very easy talking with fathera0.34 <0.001 0.22 <0.001 0.16 0.001
Difficult talking with mothera0.19 0.001 0.14 0.02
Easy talking with mothera0.41 <0.001 0.26 0.004
Very easy talking with mothera0.53 <0.001 0.35 <0.001
Self-esteem in 2011 0.38 <0.001
Explained variance for the full model analyses 0.24 0.35
The last two columns pertain to the change of self-esteem (residual change) from 2011 to 2013.
aAnalysed as dummy variables with ‘very difficult and no contact’ as comparison category.
bAll significant predictors in the unadjusted and adjusted analyses are entered in the full model, but only significant predictors are shown.
cBivariate analyses with the adjusting variables entered one by one.
8 E. Meland et al.
cooperation, mutual caring and shared custody after
divorce [6,9].
Our findings confirming the importance of CC
between children and their parents are consistent
with the literature [11–16]. When this confidence is
disrupted by impaired contact with the non-custo-
dial parent, the child’s self-esteem is also impacted
negatively [23]. A recent cross-sectional study docu-
ments the independent importance of the father–
child relationship [28].
We were unable to evaluate how the inter-parental
conflicts impacted our results. It may well be, how-
ever, that impaired relational confidence between
fathers and their children is a result of inter-parental
conflict – the so-called spill-over effect [29]. Many
researchers in this field will confirm that inter-paren-
tal conflicts are deleterious for children’s health and
well-being, especially when they are drawn into the
conflicts [1,3,6]. Conflicting results from a longitudi-
nal study with adjustment for many pre-divorce fac-
tors revealed beneficial effects of shared custody
compared to sole custody by the mother. Surprisingly,
the level of inter-parental conflict did not moderate
any of the outcomes [30].
With support from empirical studies, other
researchers also maintain that except for extremely
conflictual cases, shared custody seems to prevent
impaired health among children and adolescents
[4,7]. Therefore, a new paradigm in child law has
been promoted, shifting the ethical demand from par-
ents’ custody rights to mutual obligations for the
child. This paradigm aims at preserving confidence
and relations with both parents after divorce by shared
responsibilities in planning mutual parenting [31,32].
In conclusion, this study proves a sex-specific
effect on loss of CC between fathers and children
after divorce. The impairment of CC has predictive
repercussions on the health and self-conception of
adolescents in their middle teenage years. From a
public-health perspective, preserving the relationship
and the CC between children and their fathers after
divorce seems an important task.
Declaration of conflicting interests
The authors declare that there is no conflict of
interest.
Funding
The authors received no financial support for the
research, authorship and/or publication of this article.
ORCID iDs
Eivind Meland https://orcid.org/0000-0001-9161
-1680
Hans Johan Breidablik https://orcid.org/0000-
0002-0251-430X
References
[1] FHI. Folkehelserapporten – Helsetilstanden i Norge. Barn
og unges helse: oppvekst og levekår. Oslo: Folkehelsa, 2018.
[2] Kelly JB. Children’s living arrangements following sepa-
ration and divorce: insights from empirical and clinical
research. Fam Process 2007;46:35–52.
[3] Harold GT and Sellers R. Annual research review: inter-
parental conflict and youth psychopathology: an evidence
review and practice focused update. J Child Psychol Psychia-
try 2018;59:374–402.
[4] Nielsen L. Joint versus sole physical custody: outcomes for
children independent of family income or parental conflict.
J Child Custody 2018;15:35–54.
[5] Bohman H, Laftman SB, Paaren A, etal. Parental separa-
tion in childhood as a risk factor for depression in adult-
hood: a community-based study of adolescents screened for
depression and followed up after 15 years. BMC Psychiatry
2017;17:117.
[6] Lansford JE. Parental divorce and children’s adjustment.
Perspect Psychol Sci 2009;4:140–152.
[7] Fabricius WV and Suh GW. Should infants and tod-
dlers have frequent overnight parenting time with fathers?
The policy debate and new data. Psychol Public Policy Law
2017;23:68–84.
[8] Waylen A, Stallard N and Stewart-Brown S. Parenting and
health in mid-childhood: a longitudinal study. Eur J Public
Health 2008;18:300–305.
[9] East L, Jackson D and O’Brien L. Father absence and
adolescent development: a review of the literature. J Child
Health Care 2006;10:283–295.
[10] Culpin I, Heron J, Araya R, etal. Father absence and depres-
sive symptoms in adolescence: findings from a UK cohort.
Psychol Med 2013;43:2615–2626.
[11] Fenton C, Brooks F, Spencer NH, etal. Sustaining a positive
body image in adolescence: an assets-based analysis. Health
Soc Care Community 2010;18:189–198.
[12] Pedersen M, Carmen M, Alcón G, etal. Family and health.
In: Currie C, Roberts C, Morgan A, etal. (eds) Young peo-
ple’s health in context Health Behaviour in School-aged Children
(HBSC) study: international report from the 2001/2002 sur vey.
Copenhagen: World Health Organization, 2004, pp.173–
177.
[13] Levin KA and Currie C. Adolescent toothbrushing and the
home environment: sociodemographic factors, family rela-
tionships and mealtime routines and disorganisation. Com-
munity Dent Oral Epidemiol 2010;38:10–18.
[14] Moreno C, Sanchez-Queija I, Munoz-Tinoco V, etal. Cross-
national associations between parent and peer commu-
nication and psychological complaints. Int J Public Health
2009;54:235–242.
[15] Sheeber LB, Davis B, Leve C, etal. Adolescents’ relation-
ships with their mothers and fathers: associations with
depressive disorder and subdiagnostic symptomatology. J
Abnorm Psychol 2007;116:144–154.
[16] Lambert SF and Cashwell CS. Preteens talking to parents:
perceived communication and school-based aggression.
Fam J 2004;12:122–128.
[17] Reiter SF, Hjorleifsson S, Breidablik HJ, etal. Impact of
divorce and loss of parental contact on health complaints
among adolescents. J Public Health 2013;35:278–285.
[18] Haugland S and Wold B. Subjective health complaints in
adolescence – reliability and validity of survey methods. J
Adolesc 2001;24:611–624.
[19] Aanesen F, Meland E and Torp S. Gender differences in
subjective health complaints in adolescence: the roles of
Divorce and conversational difficulties 9
self-esteem, stress from schoolwork and body dissatisfac-
tion. Scand J Public Health 2017;45:389–396.
[20] Quon EC and McGrath JJ. Subjective socioeconomic sta-
tus and adolescent health: a meta-analysis. Health Psychol
2014;33:433–447.
[21] McMullen T and Resnick B. Self-esteem among nursing
assistants: reliability and validity of the Rosenberg Self-
Esteem Scale. J Nurs Meas 2013;21:335–344.
[22] Diseth A, Breidablik HJ and Meland E. Longitudinal rela-
tions between perceived autonomy support and basic
need satisfaction in two student cohorts. Educ Psychol
2018;38:99–115.
[23] Turunen J, Fransson E and Bergstrom M. Self-esteem in
children in joint physical custody and other living arrange-
ments. Public Health 2017;149:106–112.
[24] Terrell SR. Statistics translated. A step-by-step guide to ana-
lyzing and interpreting data. New York: The Guilford Press,
2012.
[25] Meland E, Rydning JH, Lobben S, etal. Emotional, self-
conceptual, and relational characteristics of bullies and the
bullied. Scand J Public Health 2010;38:359–367.
[26] Amato PR and Keith B. Parental divorce and the well-
being of children: a meta-analysis. Psychol Bull 1991;110:
26–46.
[27] Thuen F, Breivik K, Wold B, et al. Growing up with one
or both parents: the effects on physical health and health-
related behavior through adolescence and into early adult-
hood. J Divorce Remarriage 2015;56:451–474.
[28] Suh GW, Fabricius WV, Stevenson MM, etal. Effects of the
interparental relationship on adolescents’ emotional security
and adjustment: the important role of fathers. Dev Psychol
2016;52:1666–1678.
[29] Rohner RP. The parental ‘acceptance–rejection syndrome’:
universal correlates of perceived rejection. Am Psychol
2004;59:830–840.
[30] Gunnoe ML and Braver SL. The effects of joint legal cus-
tody on mothers, fathers, and children controlling for fac-
tors that predispose a sole maternal versus joint legal award.
Law Hum Behav 2001;25:25–43.
[31] Marcus P. Parental responsibilities: reformulating the para-
digm for parent–child relationships. Part 1: What is wrong
with the ways in which we deal with the children of sepa-
rated parents and how to put them right. J Child Custody
2017;14:83–105.
[32] Marcus P. Parental responsibilities: reformulating the para-
digm for parent–child relationships. Part 2: Who has respon-
sibilities to children and what are these responsibilities? J
Child Custody 2017;14:106–133.