Journal of Affective Disorders 228 (2018) 60–67

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Parental bonding and adolescents’ depressive and anxious symptoms in T

Pakistan
⁎
Amna Khalida,b, , Farah Qadirc, Stella W.Y. Chand, Matthias Schwannauere
a
School of Health in Social Science, University of Edinburgh, Edinburgh EH8 9AG, United Kingdom
b
Department of Behavioral Sciences, Fatima Jinnah Women University, The Mall, Rawalpindi, Pakistan
c
Quaid-e-Azam International Hospital, Rawalpindi, Pakistan
d
Section of Clinical Psychology, School of Health in Social Science, University of Edinburgh, Edinburgh EH8 9AG, United Kingdom
e
Head of Clinical & Health Psychology, School of Health in Social Sciences, University of Edinburgh, Edinburgh, United Kingdom

A R T I C L E I N F O A B S T R A C T

Keywords: Introduction: A quantitative cross sectional study was carried out to investigate the role of parental bonding in
Parental bonding relation to depressive and anxious symptoms among secondary school adolescents in Pakistan. The study also
Pakistan aimed to investigate the construct validity of the parental bonding inventory in the cultural context of Pakistan.
Adolescents Methods: The sample consisted of 1124 adolescents recruited from eight secondary schools in Rawalpindi,
Depression
Pakistan. Urdu translated versions of Parental Bonding Instrument (Qadir et al., 2005) and Hospital Anxiety and
Anxiety
Well-being
Depression Scales (Mumford et al., 1991) were administered in classroom settings. Confirmatory factor analysis
Structural Equation Modeling (CFA) and Structural Equation Modeling (SEM) were used to analyze the data.
Results: Results from the CFA of the Urdu version of the Parental Bonding Instrument supported a modified
version of the three factor model proposed by Kendler (1996) consisting of warmth, protectiveness and au-
thoritarianism. Based on SEM, we found a significant relationship between low parental warmth and depression
symptoms among adolescents, whereas maternal protectiveness was a significant predictor for anxiety. Parental
authoritarianism was not a significant predictor for depression or anxiety.
Discussion: Findings of this study indicate that parental bonding is a central construct within a developmental
framework in the cultural context of Pakistan, and is important when considering long-term psychosocial
functioning of individuals. It should be explored further in clinical populations of Pakistani adolescents to as-
certain significance of these constructs for interventions.

1. Introduction 1997; Murphy et al., 1997). Kendler (1996) also proposed a three factor
solution for the 16 item version of PBI comprising of warmth, protec-
Parent-child bonding is crucial in adolescents as they strive towards tiveness and authoritarianism which has demonstrated a good fit in three
autonomy (Moretti and Peled, 2004). This bonding is an individual's subsequent confirmatory factor analytic studies (Sato et al., 1999; Terra
interpretation of their relationship with their parents (Cassidy, 2008). et al., 2009; Uji et al., 2006). In a Japanese sample, Uji et al. (2006)
There is robust evidence suggesting that parental bonding plays a sig- found support for a three factor model using exploratory factor analysis
nificant role in predicting adolescents’ psychosocial functioning (Doyle but preferred a four factor structure consisting of care, indifference,
et al., 2000, 2002; Moretti and Holland, 2003). overprotection and autonomy based on the superior fit of the four factor
The Parental Bonding Instrument (PBI) is a widely used self report model in confirmatory factor analysis. These factor analyses have
measure of parental bonding developed by Parker et al. (1979) to shown inconclusive results in the factor structure of PBI across and
capture the parenting behaviors in two dimensions consisting of care within cultures, particularly for the control dimension. This suggests
(warmth, emotional closeness, and empathy) and overprotection (con- that perceptions of parenting behaviors differ in different populations
trol, intrusion and the level of independence and autonomy). However, and may have differential effect on mental health of adolescents. As-
the two factor structure of the instrument showed inconsistent findings suming the factor structure to be applicable across different cultures
across different studies demonstrating a superior fit of a three-factor can significantly affect the authenticity of the study outcomes therefore,
structure (Cubis et al., 1989; Gomez-Beneyto et al., 1993; Kendler et al., the empirical question on the applicability of a construct needs to be

⁎
Corresponding author at: Department of Behavioral Sciences, Fatima Jinnah Women University, The Mall, Rawalpindi, Pakistan
E-mail addresses: khalid.a@fjwu.edu.pk (A. Khalid), drfarahqadir@yahoo.co.uk (F. Qadir), stella.chan@ed.ac.uk (S.W.Y. Chan), m.schwannauer@ed.ac.uk (M. Schwannauer).

https://doi.org/10.1016/j.jad.2017.11.050
Received 23 February 2017; Received in revised form 24 August 2017; Accepted 11 November 2017
Available online 21 November 2017
0165-0327/ © 2017 Elsevier B.V. All rights reserved.
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

investigated rather than assumed (Cheung and Leung, 1998). Therefore 2.2. Measures
in order to explore the effect of parental bonding on mental health it is
pivotal to examine the validity of the construct in the culture. 2.2.1. Parental bonding instrument
Irrespective of the lack of consensus regarding the dimensions of (PBI 16 item; Kendler, 1996) is a self-report measure of perceived
parental bonding there is a general consensus in the literature that optimal parental behaviors adapted from the 25 item original version (Parker
parenting lowers the risk of developing psychopathology; Adam et al., et al., 1979). It has three proposed dimensions; warmth, protectiveness
1994; Canetti, Bachar, Galili-Weisstub et al., 1997; Rey, 1995; Dale et al., and authoritarianism. Items are rated on a four point likert scale ran-
2010; Rikhye et al., 2008). Warmth from parents as measured by PBI is ging between 0 (very like) to 3 (very unlike) with higher score in-
associated with lower levels whereas parental control is associated with dicating the parenting attribute to be stronger. PBI is considered to be
higher levels of depressive and anxious symptoms among adolescents in the most reliable measure of parent-child bond in both clinical and non-
cross-sectional (Reitz et al., 2006) and longitudinal studies (Raudino et al., clinical settings (Enns et al., 2002). The 25 item PBI has been translated
2013). There are however, some limitations of the current literature. and administered among adult Pakistani female sample (Qadir et al.,
Firstly, previous studies mainly focused on maternal bonding. Secondly, 2005), the present study used the 16 items of the same translated
most of the studies employed samples that were predominantly from version. Perceptions of bonding with mothers and fathers were mea-
Western cultures. Relatively fewer studies have examined the relationship sured separately. The Cronbach's alpha for Warmth was 0.57for the
between perceptions of parenting and adolescent psychopathology across father and 0.61 for the mother, for Protectiveness the alpha was 0.48
different ethnic and cultural groups. Distinct differences have been ob- and 0.50 and for Authoritarianism alpha for the father was 0.67 while
served between collectivist and individualist cultures (Papps et al., 1995). for the mother it was 0.75.
Studies indicated the collectivist cultures to report non-optimal parental
bonding as measured by Western instruments (Dinh et al., 1994). In some 2.2.2. Hospital anxiety and depression scale
of these cultures, parental control did not predict depression, whereas (HADS; Zigmond and Snaith, 1983). It is a screening tool indicating
separateness from parents did (Aydin and Oztutuncu, 2001) in others likelihood of depression and anxiety comprising of 7 items for mea-
strong parental control and warmth seemed to coexist (Dekovic et al., suring each. Although the scale was initially designed for use with
2004). Studies examining differences in parenting across Caucasian adults but has also shown good psychometric properties with adoles-
Americans and Asian American culture show that Caucasian American cents (White et al., 1999). The scale is scored on a four point likert scale
parents emphasize on their child's ability to build a “sense of self” while ranging between 0 and 3. White et al. (1999) proposed a score of 7–9
Asian American parents stress the importance of developing a sense of for adolescents' possible depression and a score of 10 and above in-
connectedness with their families (Wang and Leichtman, 2000). Asian dicating probable depression. For anxiety sub-scale they proposed a
culture has been reported to support over-protection and strictness score of 9 through 11 indicating the probable presence of anxiety
(Chung, 1997) which could have a positive function within the culture. whereas a score of 12 and above suggested the presence of anxiety.
These studies highlight the need to explore these aspects further to en- Since HADS is not a diagnostic tool therefore, the scores obtained on
hance out understanding of parenting and mental health. this measure only reflect depressive and anxious symptomatology. The
Islamic Republic of Pakistan is an Eastern collectivist society present study used the Urdu translated version of HADS by Mumford
(Routamaa and Hautala, 2008). Individual and socio-political lives of et al. (1991). The Cronbach's alpha reliability estimate for depression
the people of Pakistan are heavily directed by the traditional Islamic subscale was 0.54 and for anxiety it was 0.67.
values. Pakistani society values parental reverence, emphasis is placed
on conforming to the rules and societal norms to sustain authority of 2.3. Procedure
the parents (Fuligni et al., 1999). Considering these socio-familial dy-
namics it is inevitable that many aspects of an adolescent's life in Pa- Out of a list of secondary schools acquired from the Educational
kistan would differ from their counter parts in the West. However, Directorate Islamabad, eight schools were selected using the Active
much of our understanding of psychosocial development in adolescents Data Software. The questionnaires were administered in class during
and their mental health outcomes are based on Western theorization of school times. Prior to administration information sheets and consent
these concepts. It is unclear whether the literature developed through forms were distributed to the participants. Participants were informed
theory and research in the Western cultures would also be applicable that the information provided will be kept confidential and that they
for the Pakistani adolescent population. have the right to withdraw at any point.
To explore this line of research further the aims of the present study
are two-fold: First, it aimed to examine the construct of parental 2.4. Data analysis
bonding among Pakistani adolescents by carrying out Confirmatory
Factor Analysis of the 16 item version of PBI; Second, it aimed to use Preliminary analysis was carried out in the Statistical Package for
Structural Equation Modeling (SEM) to test the hypothesis that warmth Social Sciences version 19 (SPSS). Mplus Version 7 (Muthen and
from mother and father will be negatively associated with symptoms of Muthen, 1998–2012) was used for conducting the Confirmatory Factor
depression and anxiety in Pakistani adolescents whereas protectiveness Analysis CFA of the scale as well as building and testing the SEM model.
and authoritarianism will be positively correlated with these symptoms. Some of the variables on PBI were non-normally distributed there-
fore non-parametric tests are reported. Robust maximum likelihood
2. Methods (MLM) estimator developed for non-normally distributed data for SEM
was used. The following indices were selected to assess goodness-of-fit
2.1. Participants of the models: The Root-Mean-Square-Error-of-Approximation (RMSEA;
best if less than 0.06), the Comparative Fit Index (CFI: best if close to
Participants were recruited from 8 secondary schools in Rawalpindi, 0.95 or greater), Tucker-Lewis Fit Index (TLI; Tucker and Lewis, 1973)
Pakistan. 1124 adolescents studying in grades 6th through 12th were re- (best if 0.90 or above) and the Standardized-Root-Mean-Square-Re-
cruited. The sample included 621 (55.25%) boys and 503 (44.75%) girls sidual (SRMR; best if less than 0.08). A non-significant value of χ2 is
between the ages of 11 and 18 (mean = 14.2, SD = 1.75). Participation required as an indicator of the goodness of fit of the model. The χ2 is
was voluntary and written permission was obtained from the local gov- however highly affected by the sample size such that larger sample size
ernment educational representatives, school representatives and partici- would increaseχ2 leading to an erroneous rejection of the model (Kline,
pants. Both University of Edinburgh, UK and Fatima Jinnah Women 2005). Therefore, the present study used additional fit indices and Sa-
University, Pakistan granted ethical approval for this study. torra-Bentler's Maximum Likelihood Mean Adjusted Chi Square χ2SB

61
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

Table 1
Intercorrelations between PBI mother/father sub-scales, depression and anxiety scores and comparison of PBI across gender (n = 1124).

1 2 3 4 5 6 7 8

1. Father warmth 1.00 − 0.27** − 0.08** 0.47** − 0.17** − 0.08** − 0.16** − 0.25**
2. Father authoritarianism 1.00 − 0.09** − 0.13** 0.73** − 0.14** 0.11** 0.10**
3. Father protectiveness 1.00 − 0.12** − 0.08** 0.64** 0.10** 0.01
4. Mother warmth 1.00 − 0.15** − 0.14** − 0.23** − 0.29**
5. Mother authoritarianism 1.00 − 0.12** 0.10** 0.08**
6. Mother protectiveness 1.00 0.15** 0.02
7. Anxiety 1.00 0.33**
8. Depression 1.00

Mean Rank
Boys 547.68 551.16 540.49 557.48 561.20 530.85 – –
Girls 580.80 576.50 589.68 568.69 564.10 601.57 – –
U 146975.5 149140.5 142511.5** 153066 155375.5 136529**
r − 0.05 − 0.04 − 0.07 − 0.01 − 0.004 0.11

** Correlation significant at 0.01.

(1988). Each indicator was fixed to load on the factor it was hypothe- 3.3. Structure equation model of parental bonding, depression and anxiety
sized to measure. In Mplus, residual terms for all indicators are not
correlated by default, in this study they were not tampered with. The The SEM Model (Fig. 3) tested direct effects of maternal and pa-
modifications to the models were made following the guidelines pro- ternal bonding on depressive and anxious symptoms. Depression and
posed by Byrne (2012). The modifications were made based on the anxiety scores were added as measured variables. Based on the CFA
modification indices only when they were theoretically justifiable. models, the modified three factor models for both mother and father
were analyzed as predictors of depression and anxiety. The SEM Model
had 6 latent variables and 32 dependent variables. PBI was used to
3. Results
assess parenting perceptions for both mother and father therefore, it
was hypothesized that the residuals will correlate across the indicators
3.1. Preliminary findings
of mother and father bonding. Depression and anxiety were also mod-
eled to correlate with each other.
The mean anxiety and depression scores of the participants was 7.91
The SEM Model had 139 free parameters. The hypothesized model
(S.D = 4.0) and 6.05 (S.D = 3.31) respectively. Participants' mean
demonstrated a good fit of the data, MLM χ2 (421, N= 1124) =
scores on warmth were higher than protective or authoritarian. Warmth
876.75*, p < 0.001; CFI = 0.95; RMSEA (90% CI) = 0.031
was negatively associated with protectiveness and authoritarianism for
(0.028–0.034), SRMR= 0.049. Parameter estimates of the measure-
both parents (see Table 1). There was a strong correlation between
ment model were all significant; p < .001.
perceptions of warmth, protectiveness and authoritarianism from father
Beta coefficients show that paternal and maternal warmth predict a
and mother. Adolescents scoring high on warmth were less anxious and
decrease in depressive symptomatology. No effect was observed for
depressed. Participants with high authoritarianism scores reported
parental authoritarianism and protectiveness on depression. Maternal
more anxious and depressive symptoms. Whereas, participants’ scores
protectiveness predicted increase in the participant's anxiety scores.
on protectiveness were positively associated with anxiety (see Table 1).
Whereas warmth, authoritarianism and paternal protectiveness were
Table 1 also reports results of differences in perceptions of parenting
not significantly associated with the participants anxiety scores. For
behaviors between boys and girls. Mann Whitney U test was carried
standardized beta coefficients see Fig. 3.
out. The results show that the perceptions of parenting behaviors do not
differ across gender except for protectiveness from mother and father
4. Discussion
where girls scored higher than boys, however, the effect size for this
association was very small.
To our knowledge this is the first study exploring the factor struc-
ture of the 16 item Urdu version of PBI in Pakistani adolescents. The
3.2. Confirmatory factor analysis of parental bonding instrument results from our confirmatory factor analysis corroborate a modified
version of Kendler's (1996) three-factor model. Whereas Parker's two
Three models were tested for mother and father forms separately: factor model (1979) did not adequately capture the Pakistani ado-
two factors Parker's model (Parker et al., 1979), three factors Kendler's lescent's perceptions of parenting. This supports previous literature
model (Kendler, 1996) and four factors Uji's model (Uji et al., 2006). Fit proposing that overprotection dimension should be further divided into
indices of these models are presented in Table 2 two factors (Cubis et al., 1989; Gomez-Beneyto et al., 1993; Kendler,
The three factor Kendler's model was modified based on the mod- 1996; Kendler et al., 1997; Murphy et al., 1997; Sato et al., 1999). It is
ification indices to generate an optimal fit. During model modification not surprising that none of the a priori models fitted the data without
item 23 (Is overprotective of me) was loaded on warmth instead of pro- modification as parenting is a culturally loaded phenomenon. PBI was
tectiveness and item 18 (Does not talk with me very much) was cross- originally developed to capture the parenting practices of Western
loaded on both warmth and protectiveness. Item 4 (Seemed emotionally cultures and therefore reflects characteristics and values of those cul-
cold to me) was dropped from the model due to insignificant R2. The tures. The differences in Western and Eastern cultures prevail, this
final model for mother with significant standardized coefficients is perhaps is a strong reason why variation in factor structures across
provided in Fig. 1. Same process was repeated for the father form (See studies and cultures has been reported. Uji et al. (2006) proposed a four
Fig. 2). Chi square for model comparison using χ2SB shows that the factor model which came out to be superior in an Eastern culture. The
modified model has a significantly better fit to the data as compared to four factor model therefore, should be explored further in the Pakistani
the previously proposed models (See Table 2). Items retained in the adolescent sample using the full version of PBI. In this study, the four
final models are provided in Table A1 in the Appendix A. factor model was rejected on the basis that there were only two items in

62
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

Table 2
Fit indices from CFA of PBI mother and father scale.

Parker model Kendler model Uji modela Modified model

Mother Father Mother Father Mother Father Mother Father

χ2 839.411 782.497 627.164 586.408 408.337 517.898 260.446 330.958

df 103 103 101 101 98 98 86 86
p < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
CFI 0.83 0.80 0.83 0.80 0.93 0.89 0.94 0.90
TLI 0.80 0.77 0.80 0.77 0.92 0.85 0.93 0.88
SRMR 0.083 0.073 0.083 0.073 0.052 0.049 0.049 0.050
RMSEA 0.068 0.065 0.068 0.065 0.043 0.053 0.042 0.050
90% CI 0.063–0.073 0.060–0.071 0.063–0.073 0.060–0.071 0.038–0.049 0.048–0.058 0.037–0.048 0.045–0.056
χ2SB (df, p) 670.087(17, < 0.001) 485.026 411.558(15, < 0.001) 269.204(15, < 0.001) 160.266(3, < 0.001) 192.393(12, < 0.001) – –
(17, < 0.001)

Note: Corresponding items in short form for: Parker model: Care 1, 4, 5, 11, 12, 17, 18; Overprotection factor 7, 8, 9, 13, 15, 19, 21, 23, 25. Uji model: Care 1, 5, 11, 12, 17; Indifference 4,
18, 24; Overprotection 8, 9, 13, 19, 23; Autonomy 7, 15, 21, 25.
a
Uji's model was the best fitting model. However, for the indifference dimension item 18 (Does not talk with me very much) showed a negative residual variance and was deleted from
the model. That left only item 4 (Seemed emotionally cold to me) in the indifference factor. Building a model using a single indicator to measure latent variable of indifference would
require to fulfill assumptions of 1). no measurement error and 2). an assumption that item 4 alone can explain the latent construct of indifference (Sagan and Pawełek, 2014). Therefore,
four factor model was not an appropriate choice for the short version of PBI in the current sample.

indifference factor out of which factor loading of one item was insig- We therefore, recommend generalizing the results of the study with
nificant. Therefore, it is possible that when the full version of PBI is caution. The reason for low alpha could also be the cultural sensitivity
used the four factor model could be a better fit. of the items and the construct itself. The linguistic and translation
In the present study, reliability estimates of PBI for both mother and glitches due to cultural uniqueness has been highlighted in previous
father were lower than the ones reported in the previous literature studies which used the same instrument (Qadir et al., 2005; Chao,
(Martin et al., 2004). A plausible explanation for this could be the 1994). The results of this study highlight the significance of con-
variation in loadings of some of the items or the fact that there were ceptualizing Pakistani adolescents’ experience of their bond with their
fewer number of items in each subscale (Tavakol and Dennick, 2011). parents and identify discreet parenting attributes. Future research

Fig. 1. Confirmatory factor model modified for PBI (mother) with

standardised robust maximum likelihood parameter estimates N
= 1124. All coefficients are statistically significant, *p < 0.001.

63
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

Fig. 2. Confirmatory factor model modified for PBI (father) with

standardised robust maximum likelihood parameter estimates N
= 1124. *p < 0.001.

Fig. 3. SEM for testing direct effect of parental bonding on de-

pression and anxiety with standardised robust maximum like-
lihood parameter estimates N = 1124. Coefficients which were
statistically significant are shown as *p < 0.01.

needs to explore the culture specific interpretation of this construct et al., 2009). The results reflect the cultural milieu of Pakistan where
particularly with reference to children and adolescents. females are to be sheltered and guarded to the extent of overprotection.
With the exception of protectiveness from mother and father by the The effects of protectiveness on anxiety and depression could be con-
female participants no gender differential was found in the present founded by the fact that girls perceive their parents to be more pro-
study. Previous research has indicated similar findings where percep- tective as compared to boys. This can affect the association between
tions of protectiveness have varied across gender with females re- protectiveness and mental health in boys. Therefore, in future the ef-
porting more protectiveness from both mother and father (Seganfredo fects of gender and other potential demographic factors should be

64
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

ideally controlled for when examining the relationship between par- style. Research proves that PBI has been established as a measure of
enting and mental health. parenting style rather than a measure of anxiety or depressive symptoms
In this study, depressive symptoms were only significantly predicted of the parents which is established by the fact that parental protectiveness
by parental warmth but not by parental authoritarianism and protec- is seen in both anxious and healthy parents (Parker, 1979). It is however
tiveness. In previous research depression has been associated more seen that parents with mental disorder show less care to their children
strongly with parental warmth than parental control (Greenberger and (e.g. Arrindell et al., 1983, 1989; Parker, 1979). In case of overprotection,
Chen, 1996; Rapee, 1997; Rohner and Britner, 2002; Mcleod et al., results are inconsistent where some studies report that parents with an-
2007). This vulnerability to depression may develop through the child's xiety disorders are more over protective of their children as compared to
dysfunctional cognitions which they develop when they do not receive healthy controls (Wiborg and Dahl, 1997) while other studies do not
optimal parental care (Zafiropoulou and Avagianou, 2014). provide support to this argument (Arrindell et al., 1983, 1989; Parker,
A robust association has been established in Western (Cunha et al., 1979). Since there was no control group in this study and parents of the
2008; Raudino et al., 2013) as well as in Asian (Peng et al., 2011) adolescents were not screened for mental illnesses; it cannot be ruled out
samples between adolescents’ high perception of protectiveness, au- that the maternal over protection reported by the participants is not be-
thoritarianism and lower perceptions of parental warmth predicting cause of the mother's anxiety. This aspect needs to be further examined in
anxiety. The SEM model in the present study indicates that protec- Pakistani sample.
tiveness from mother was a significant predictor of anxiety among Overall the results from this Pakistani adolescent sample provides
Pakistani adolescents. A meta-analysis by McLeod et al. (2007) found support to the proposal that parental warmth is a key factor in de-
that parental control dimension has a stronger association with anxiety signing prevention and intervention programs for adolescents’ depres-
as compared to the warmth dimension. Furthermore, Wood et al. sion and the control dimension is more relevant to anxiety disorders.
(2003) in their meta-analysis also found inconsistent support for asso-
ciation between parenting characterized by warmth and adolescents’ 4.1. Limitations
anxiety. Research suggests that when parents exert excessive control
over children, children may not develop the essential sense of self The present study has used a screening tool for measuring depres-
sufficiency which induces feelings of vulnerability to threat and anxiety sive and anxious symptomatology and does not claim to diagnose
(Chorpita and Barlow, 1998; Wood, 2006). The results that paternal clinical depression and anxiety among Pakistani adolescents therefore;
parenting had no effects on anxiety sheds light upon the much debated the generalizability to clinical samples is limited. Furthermore, par-
similarities and differences between maternal and paternal parenting enting behaviors have been measured using a self-report measure
roles in Eastern samples. Childrearing experience may appear to be the completed by the adolescents. PBI is recognized as a robust tool for
same across cultures, but it may have different impact on children's measuring unbiased parental behaviors reported retrospectively by the
development because of the distinct meanings associated with these offspring and has shown stability in results across multiple informants
behaviors in different cultures (Bornstein, 1995). Generally, the father (Murphy et al., 2010). However, using parental reports to cross-validate
is considered to be more stern and controlling as compared to the the information received could have increased the reliability and va-
mother who are expected to be kind and loving. This is distinction is lidity of the findings. Finally, the cross-sectional design of this study
more stark in the Pakistani culture. This could be a possible reason for does not establish causality.
participants feeling more anxious if they perceived their mother to be
controlling since it is not considered a cultural parenting norm. Fathers 5. Conclusion
are also not expected to take the lead in parenting, instead the mother is
responsible to look after the children. Studies from Western cultures The overall findings of the present study provide support to the
exploring paternal parenting in relation to adolescents’ mental health proposal that parenting perceptions should be an integral part of as-
outcomes are rare. Few studies show paternal control as a significant sessment and intervention for adolescents’ health and well-being in
predictor of child anxiety (van der Bruggen et al., 2008) but the effect Pakistan. This study provides support for the three factor model of PBI.
sizes are small; others did not find a significant effect of paternal control It provides a baseline data for future research in the field of adolescent's
(Hudson and Rapee, 2002). Thus, there is evidence that maternal par- mental health and the significance of the parent child dyad. A better
enting is a stronger predictor of child anxiety, but this evidence is scarce understanding of parental bonding in different cultures can help un-
and inconclusive. Although we found significant association between derstand, address and prevent potential psychopathology not just in this
parental warmth and depression and maternal protectiveness and an- relationship but also in other interpersonal contacts. Educating parents
xiety, the effect sizes of these associations were small. This was in line about culturally appropriate optimal parenting practices will help en-
with a meta-analysis, which found that parenting accounts for only 4% hance children's future wellbeing.
variance in anxiety and care accounts for only 8% of variance in de-
pression among children and adolescents (Mcleod et al., 2007). This Acknowledgements
seems to contest the assumption that parenting is critical for psycho-
logical well-being of children, however, it should be taken into con- We are grateful to participating schools and pupils.
sideration that there are a number of factors which could have mod-
erated the effect of parenting on the adolescents’ mental health such as Author's statement
the use of self-report measures and single informant design. It should
also be noted that these effect sizes reflect the direct effects parenting Ethical considerations were made during this project by getting
has on mental health it does not take into account the indirect effects written consent from educational authorities, schools and respondents.
parenting would have through other significant factors such as per- Ethical approval was granted by the University of Edinburgh and
ceived social support, emotion regulation etc (Caspi et al., 2003; Fatima Jinnah Women University. For details kindly refer to the
Wilhelm et al., 2006). This implies that mental health problems in manuscript.
children and adolescents can be a result of a complex set of interactions
between different factors. Furthermore, the effects of parenting may Funding
seem small but it could potentially play a catalytic role among a sub-
group of children who are vulnerable to depression and anxiety. Higher Education Commission of Pakistan sponsored AK's Phd
One can debate that low warmth and high protectiveness could be a which has resulted into this manuscript.
manifestation of parental depression or anxiety rather than their parenting University of Edinburgh provided token money for data collection.

65
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

Appendix A

See Table A1.

Table A1
Items retained in the final model for mother and father forms.

Dimension Items

Warmth 1. Speaks to me in a warm and friendly voice

5. Appears to understand my problems and worries
11. Enjoys talking things over with me
12. Frequently smiles at me
17. Can make me feel better when I am upset
18.a Does not talk with me very much
23. Is overprotective of me

Protectiveness 8. Does not want me to grow up

9. Tries to control everything I do
13. Tends to baby me
19. Tries to make me feel dependent on her
18.a Does not talk with me very much

Authoritarianism 7. Likes me to make my own decisions

15. Let me decide things for myself
21. Gives me as much freedom as I want
25. Let me dress in any am I please

a
Item 18 was cross-loaded on both warmth and protectiveness.

References CAT no. 032ss. H5219-00CYHS. Health Canada, Child and Family Division, Ottawa.
Enns, M.W., Cox, B.J., Clara, I., 2002. Parental bonding and adult psychopathology: re-
sults from the US National Comorbidity Survey. Psychol. Med. 32 (6), 997–1008.
Adam, K.S., Keller, A., West, M., Larose, S., Goszer, L.B., 1994. Parental representation in Fuligni, A.J., Tseng, V., Lam, M., 1999. Attitudes toward family obligations among
suicidal adolescents: a controlled study. Aust. N.Z. J. Psychiatry 28, 418–425. American adolescents with Asian, Latin American, and European backgrounds. Child
Arrindell, W.A., Emmelkamp, P.M., Monsma, A., Brilman, E., 1983. The role of perceived Dev. 70 (4), 1030–1044.
parental rearing practices in the aetiology of phobic disorders: a controlled study. Br. Gomez-Beneyto, M., Pedros, A., Tomas, A., Aguilar, K., Leal, C., 1993. Psychometric
J. Psychiatry 143 (2), 183–187. properties of the parental bonding instrument in a Spanish sample. Soc. Psychiatry
Arrindell, W.A., Kwee, M.G., Methorst, G.J., Van der Ende, J., Pol, E., Moritz, B.J., 1989. Psychiatr. Epidemiol. 28 (5), 252–255.
Perceived parental rearing styles of agoraphobic and socially phobic in-patients. Br. Greenberger, E., Chen, C., 1996. Perceived family relationships and depressed mood in
J. Psychiatry 155 (4), 526–535. early and late adolescence: a comparison of European and Asian Americans. Dev.
Aydin, B., Oztutuncu, F., 2001. Examination of adolescents' negative thoughts, depressive Psychol. 32 (4), 707–716.
mood, and family environment. Adolescence 36 (141), 77–83. Hudson, J.L., Rapee, R.M., 2002. Parent-child interactions in clinically anxious children
Bornstein, M.H., 1995. Form and function: implications for studies of culture and human and their siblings. J. Clin. Child Adolesc. Psychol. 31 (4), 548–555.
development. Cult. Psychol. 1 (1), 123–137. Kendler, K.S., 1996. Parenting: a genetic-epidemiologic perspective. Am. J. Psychiatry
Byrne, B.M., 2012. A Primer of LISREL: Basic Applications and Programming for 153 (1), 11–20.
Confirmatory Factor Analytic Models. Springer Science & Business Media. Kendler, K.S., Sham, P.C., MacLean, C.J., 1997. The determinants of parenting: an epi-
Canetti, L., Bachar, E., Galili-Weisstub, E., De-Nour, A.K., Shalev, A.Y., 1997. Parental demiological, multi-informant, retrospective study. Psychol. Med. 27 (03), 549–563.
bonding and mental health in adolescence. Adolescence 32, 381–394. Martin, G., Bergen, H.A., Roeger, L., Allison, S., 2004. Depression in young adolescents:
Caspi, A., Sugden, K., Moffitt, T.E., 2003. Influence of life stress on depression: mod- investigations using 2 and 3 factor versions of the Parental Bonding Instrument. J.
eration by a polymorphism in the 5-HTT gene. Science 301, 386–389. Nerv. Ment. Dis. 192 (10), 650–657.
Cassidy, J., 2008. The nature of the child's ties. In: Cassidy, J., Shaver, P.R. (Eds.), McLeod, B.D., Wood, J.J., Weisz, J.R., 2007. Examining the association between par-
Handbook of Attachment: Theory, Research, and Clinical Applications. Guilford enting and childhood anxiety: a meta-analysis. Clin. Psychol. Rev. 27 (2), 155–172.
Press, New York, pp. 3–22. Moretti, M.M., Peled, M., 2004. Adolescent-parent attachment: bonds that support
Chao, R.K., 1994. Beyond parental control and authoritarian parenting style: under- healthy development. Paediatr. Child Health 9 (8), 551–555.
standing Chinese parenting through the cultural notion of training. Child Dev. 65 (4), Moretti, M.M., Holland, R., 2003. Navigating the journey of adolescence: Parental at-
1111–1119. tachment and the self from a systemic perspective. In: Clinical Applications of
Cheung, F.M., Leung, K., 1998. Indigenous personality measures chinese examples. J. Attachment Theory. Guildford, New York, pp. 41–56.
Cross-Cult. Psychol. 29 (1), 233–248. Mumford, D.B., Tareen, I.A.K., Bajwa, M.A.Z., Bhatti, M.R., Karim, R., 1991. The trans-
Chorpita, B.F., Barlow, D.H., 1998. The development of anxiety: the role of control in the lation and evaluation of an Urdu version of the Hospital Anxiety and Depression
early environment. Psychol. Bull. 124 (1), 3. Scale. Acta Psychiatr. Scand. 83 (2), 81–85.
Chung, W., 1997. Asian American children. In: Lee, E. (Ed.), Working with Asian Murphy, Eleanor, Wickramaratne, Priya, Weissman, Myrna, 2010. The stability of par-
Americans: A Guide for Clinicians. Guilford, New York, pp. 165–174. ental bonding reports: a 20-year follow-up. J. Affect. Disord. 125 (1), 307–315.
Cubis, J., Lewin, T., Dawes, F., 1989. Australian adolescents' perceptions of their parents. Murphy, E., Brewin, C.R., Silka, L., 1997b. The assessment of parenting using the Parental
Aust. N.Z. J. Psychiatry 23, 35–47. Bonding Instrument: two or three factors? Psychol. Med. 27 (02), 333–342.
Cunha, M., Soares, I., Pinto-Gouveia, J., 2008. The role of individual temperament, family Muthen, L.K., Muthen, B.O., 1998. Mplus User's Guide, Seventh edition. Muthen &
and peers in social anxiety disorder: a controlled study. Int. J. Clin. Health Psychol. 8 Muthen, Los Angeles, CA.
(3), 631–655. Papps, F., Walker, M., Trimboli, A., Trimboli, C., 1995. Parental discipline in Anglo,
Dale, R., Power, K., Kane, S., Stewart, A.M., Murray, L., 2010. The role of parental Greek, Lebanese, and Vietnamese cultures. J. Cross-Cult. Psychol. 26 (1), 49–64.
bonding and early maladaptive schemas in the risk of suicidal behavior repetition. Parker, G., 1979. Reported parental characteristics of agoraphobics and social phobics.
Arch. Suicide Res. 14 (4), 311–328. Br. J. Psychiatry 135 (6), 555–560.
Dekovic, M., Wissink, I., Meijer, A.M., 2004. The role of family and peer relations in Parker, G., Tupling, H., Brown, L.B., 1979. A parental bonding instrument. Br. J. Med.
adolescent antisocial behaviour: comparison of four ethnic groups. J. Adolesc. 27 (5), Psychol. 52 (1), 1–10.
497–514. Peng, Z.W., Lam, L.T., Jin, J., 2011. Factors associated with social interaction anxiety
Dinh, K.T., Sarason, B.R., Sarason, I.G., 1994. Parent-child relationships in Vietnamese among Chinese adolescents. East Asian Arch. Psychiatry.: Off. J. Hong Kong Coll.
immigrant families. J. Fam. Psychol. 8 (4), 471–488. Psychiatr. 21 (4), 135.
Doyle, A.B., Moretti, M.M., Canada. Health Promotion and Programs Branch. Childhood Qadir, F., Stewart, R., Khan, M., Prince, M., 2005. The validity of the Parental Bonding
and Youth Division, 2000. Attachment to Parents and Adjustment in Adolescence: Instrument as a measure of maternal bonding among young Pakistani women. Soc.
Literature Review and Policy Implications. Childhood and Youth Division, Health Psychiatry Psychiatr. Epidemiol. 40, 276–282.
Canada. Rapee, R.M., 1997. Potential role of childrearing practices in the development of anxiety
Doyle, A.B., Moretti, M.M., Brendgen, M., Bukowski, W., 2002. Parent Child Relationships and depression. Clin. Psychol. Rev. 17 (1), 47–67.
and Adjustment in Adolescence: Findings from the HSBC and NLSCY Cycle 2 Studies. Raudino, A., Fergusson, D.M., Horwood, L.J., 2013. The quality of parent/child

66
A. Khalid et al. Journal of Affective Disorders 228 (2018) 60–67

relationships in adolescence is associated with poor adult psychosocial adjustment. J. analysis. Psychometrika 38, 1–10.
Adolesc. 36, 331–340. Uji, M., Tanaka, N., Shono, M., Kitamura, T., 2006. Factorial structure of the parental
Reitz, E., Dekovic, M., Meijer, A.M., 2006. Relations between parenting and externalizing bonding instrument (PBI) in Japan: a study of cultural, developmental, and gender
and internalizing problem behaviour in early adolescence: child behaviour as mod- influences. Child Psychiatry Human. Dev. 37 (2), 115–132.
erator and predictor. J. Adolesc. 29 (3), 419–436. Tavakol, Mohsen, Dennick, Reg, 2011. Making sense of Cronbach's alpha. Int. J. Méd.
Rey, J.M., 1995. Perceptions of poor maternal-care are associated with adolescent de- Education 2, 53.
pression. J. Affect. Disord. 34, 95–100. van der Bruggen, C.O., Stams, G.J.J.M., Bögels, S.M., 2008. Research review: the relation
Rikhye, K., Tyrka, A.R., Kelly, M.M., Gagne, G.G., Mello, A.F., Mello, M.F., Carpenter, between child and parent anxiety and parental control: a meta-analytic review. J.
L.L., 2008. Interplay between childhood maltreatment, parental bonding, and gender Child Psychol. Psychiatry 49, 1257–1269.
effects: impact on quality of life. Child Abus. Negl. 32 (1), 19–34. Wang, Q., Leichtman, M.D., 2000. Same beginnings, different stories: a comparison of
Rohner, R.P., Britner, P.A., 2002. Worldwide mental health correlates of parental ac- American and Chinese children's narratives. Child Dev. 71 (5), 1329–1346.
ceptance-rejection: review of cross-cultural and intracultural evidence. Cross-Cult. White, D., Leach, C., Sims, R., Atkinson, M., Cottrell, D., 1999. Validation of the Hospital
Res. 36, 16–47. Anxiety and Depression Scale for use with adolescents. Br. J. Psychiatry 175 (5),
Routamaa, V., Hautala, T.M., 2008. Understanding cultural differences-the values in a 452–454.
cross-cultural context. Int. Rev. Bus. Res. Pap. 4 (5), 129–137. Wiborg, I.M., Dahl, A.A., 1997. The recollection of parental rearing styles in patients with
Sagan, A., Pawełek, B., 2014. Single-indicator SEM with measurement error: case of Klein panic disorder. Acta Psychiatr. Scand. 96 (1), 58–63.
I Model. In: Advances in Latent Variables. Springer International Publishing, pp. Wilhelm, K., Mitchell, P.B., Niven, H., 2006. Life events, first depression onset and the
171–183. serotonin transporter gene. Br. J. Psychiatry 188, 210–215.
Sato, T., Narita, T., Hirano, S., Kusunoki, K., Sakado, K., Uehara, T., 1999. Confirmatory Wood, J.J., 2006. Parental instrusiveness and children's separation anxiety in a clinical
factor analysis of the Parental Bonding Instrument in a Japanese population. Psychol. sample. Child Psychiatry Human. Dev. 37, 73–87.
Med. 29 (1), 127–133. Wood, J.J., McLeod, B.D., Sigman, M., Hwang, W.C., Chu, B.C., 2003. Parenting and
Seganfredo, A.C.G., Torres, M., Salum, G.A., Blaya, C., Acosta, J., Eizirik, C., Manfro, G.G., childhood anxiety: theory, empirical findings, and future directions. J. Child Psychol.
2009. Gender differences in the associations between childhood trauma and parental Psychiatry 44 (1), 134–151.
bonding in panic disorder. Rev. Bras. Psiquiatr. 31 (4), 314–321. Zafiropoulou, Maria, Avagianou, Penelope-Alexia, 2014. Parental bonding and early
Terra, L., Hauck, S., Fillipon, A.P., Sanchez, P., Hirakata, V., Schestatsky, S., Ceitlin, L.H., maladaptive schemas. J. Psychol. Abnorm.
2009. Confirmatory factor analysis of the Parental Bonding Instrument in a Brazilian Zigmond, A.S., Snaith, R.P., 1983. The hospital anxiety and depression scale. Acta
female population. Australas. Psychiatry 43 (4), 348–354. Psychiatr. Scand. 67 (6), 361–370.
Tucker, L.R., Lewis, C., 1973. A reliability coefficient for maximum likelihood factor

67