Journal of Youth and Adolescence, Vol. 34, No. 2, April 2005, pp.

97–110 (

C 2005)

DOI: 10.1007/s10964-005-3209-7

Parenting, Marital Conflict and Adjustment From

Early- to Mid-Adolescence: Mediated by Adolescent
Attachment Style?

Anna Beth Doyle1 and Dorothy Markiewicz2

Received August 12, 2003; revised March 4, 2004; accepted April 6, 2004

Contributions of 3 dimensions of parenting (psychological control, warmth, and behavioural con-

trol), marital conflict, and attachment style (anxiety and avoidance) to adjustment from early to
middle adolescence were assessed. Mediation of marital conflict effects by parenting, and of par-
enting effects by attachment were examined. Adolescents (n = 175) initially age 13 years reported
parenting practices, attachment styles, school grades, self-esteem, and internalizing and externalizing
problems twice (T1, T2) 2 years apart. T1 marital conflict was associated with lower self-esteem,
more externalizing symptoms, and lower academic achievement at T2, all but the latter mediated by
parental warmth. T1 parental psychological control was associated with increases in internalizing
symptoms over time, an effect not mediated by attachment insecurity, which contributed indepen-
dently. T1 parental warmth was associated with decreases in externalizing symptoms and increases
in self-esteem over time, the latter mediated by attachment security.

KEY WORDS: parenting; adolescence; adjustment; marital discord; attachment.

Numerous studies have established that parenting family, that is, marital discord and adolescent–parent at-
practices have significant influence on adolescent adjust- tachment, to affect adjustment.
ment. Recently, the importance of the distinction between With respect to parental control, Steinberg and oth-
psychological and behavioural forms of control has been ers highlight the conceptual and empirical differences in
highlighted, and the need for more careful examination children’s psychological processes associated with be-
of the effects of parental use of psychological control havioural and psychological control, as well in the par-
(Barber and Harmon, 2002). In this paper, the influence ents’ conscious or unconscious motives driving their
of parental psychological control, and 2 other common child-rearing behaviour (Steinberg et al., 1989; Steinberg,
dimensions of parenting, warmth, and behavioural con- 1990; Silk et al., 2003). While behavioral control (e.g.,
trol, in the development of internalizing and externalizing monitoring) is aimed at influencing the child’s actions
problems during early adolescence are examined. As well, with the goal of assisting the child to acquire normative,
pathways are assessed by which these parenting dimen- socially appropriate behavior patterns, psychological con-
sions interact with other important influences within the trol is directed at manipulating the child’s thoughts and
feelings. The latter is hypothesized to interfere with the
1 Professor of Psychology, Concordia University, Montreal, QC, Canada,
child’s development of autonomy and to be the conse-
H4B 1R6. Received PhD in Developmental Psychology from Stanford
quence of the parents’ attempts to satisfy their own needs
University. Research interests include parenting, attachment, and ad- to maintain their status and power through keeping the
justment in adolescence. To whom correspondence should be addressed child emotionally dependent on them. Psychological con-
at Concordia University, Montreal, QC, Canada, H4B 1R6; e-mail: trol is intrusive and involves efforts to manipulate the child
abdoyle@vax2.concordia.ca. through the use of techniques such as inducing guilt, with-
2 Professor of Psychology and Applied Human Sciences, Concordia Uni-

versity, Montreal, QC, Canada, H3G 1M8. Received PhD in Social Psy-
drawing love, provoking shame, and arousing anxiety.
chology from Ohio State University. Research interests include close In a recent thorough review of research find-
personal relationships and adjustment. ings concerning the child and adolescent correlates of

97
0047-2891/05/0400-0097/0

C 2005 Springer Science+Business Media, Inc.
98 Doyle and Markiewicz

psychological control, Barber and Harmon (2002) noted ination of potential mediating effects than the concurrent
that psychological control is consistently linked to a data used in the Stone et al. study.
wide variety of indices of child adjustment across varied There have been several calls for a broader look at the
methodologies and samples. More use of psychological interrelations between parenting and attachment (Belsky,
control is associated with lower self concept/esteem, more 1999; Cummings and Cummings, 2002; De Wolff and
internalizing problems, poorer academic performance, van Ijzendoorn, 1997). The importance of both marital
and, less consistently, more externalizing problems. conflict and psychological control for youth problem be-
Much literature supports the link between parental havior suggests that child attachment security might be a
warmth and behavioural control and adolescent adjust- mediator of both effects. That is, both covert conflict tac-
ment. Parental warmth and behavioural control are as- tics and psychological control are “anxiety-driven” (Stone
sociated with greater academic achievement and self- et al., 2002). In this study it was hypothesized that psy-
esteem, as well as with less depression and delinquency chological control is a pattern of insensitive parenting
(e.g., Allen and Hauser, 1996; Lamborn et al., 1991; which, like marital conflict, undermines the child’s sense
Parish and McCluskey, 1992; Steinberg et al., 1992). Al- of self as valuable and the parent as sensitive and respon-
though psychological control has been demonstrated to sive to his/her needs (Davies and Cummings, 1994; Pettit
be an important correlate of child adjustment, further re- and Laird, 2002). Thus, psychological control was ex-
search is needed to explore its influence in the context pected to be associated with more insecure child–parent
of other important parenting behaviours (i.e., warmth and attachment, specifically with anxious-preoccupied attach-
behavioural control) and family factors such as marital ment. This anxious attachment would then result in poor
conflict and child attachment security. There is an ab- child adjustment. Moreover, although parental warmth has
sence of study of the unique contributions of parental been shown to be associated with secure child attachment
psychological control, warmth and behavioural control in (e.g., Karavasilis et al., 2003), it remains to be established
a family-wide context and of mechanisms mediating their whether the effects of parental warmth on adjustment are
effects. mediated through security of attachment in adolescence.
With respect to influences in a family-wide context, Thus, the role of child–parent attachment insecurity as a
existing research demonstrates that marital conflict is as- mediating mechanism was assessed in this study.
sociated with child adjustment difficulties and that it tends Although adolescence is a time of expansion of so-
to disrupt parenting (Erel and Burman, 1995; Stone et al., cial roles, both mother and father continue to be important
2002). Most focus has been on the negative effects of mar- attachment figures (Hazan and Zeifman, 1994; Fraley and
ital conflict on parental warmth and behavioural control. Davis, 1997; Stone et al., 2002). Evidence is accumu-
With respect to psychological control, in two samples of lating that fathers and mothers play different as well as
children between 10- and 15-years-old Stone et al. found similar roles as attachment figures, and that adolescence is
support for both independent-additive and indirect effect a particularly important time for the father–adolescent re-
models of inter-parental conflict and parental psycholog- lationship (e.g., Grossmann et al., 1999; Lieberman et al.,
ical control for youth problem behaviour. However, the 1999). Many studies of attachment in adolescence have
indirect association between inter-parental conflict and utilized an index combining relationships with mother and
behavior problems held only for covert forms of conflict father, whether based on interview (e.g., Van-Ijzendoorn
but not for overt forms. Similar to Stone et al. in this study and Bakermans-Kranenburg, 1996) or questionnaire (e.g.,
the possible mediating role of psychological control in the Paterson et al., 1995). Thus, in this study, indices of at-
relationship of marital conflict with adjustment is exam- tachment security based on relationships with both mother
ined. However, this present study addresses two of the and father were utilized.
issues raised by Stone et al. First, the roles of parental In most research on the associations of parenting
warmth and behavioural control, as well as psychological behavior with child adjustment, concurrent data is used
control, are examined. The meaning of particular parent- (Pettit and Laird, 2002). One difficulty with this design
ing behaviors might differ depending on other charac- is its inability to tease out the causal directions among
teristics (e.g., warmth) of the parent–child relationship. the relationships found (Patterson and Fisher, 2002). For
Such contextual effects might account for the conflicting example, a significant correlation between the use of psy-
results of Fauber et al., 1990, who, using other measures chological control and depression is consistent with the
of parenting as well as psychological control, found these view that the former causes the latter, but is also consistent
to fully mediate the association of marital conflict with with depressed children eliciting a greater variety of par-
child adjustment. Second, this study used a short-term enting efforts, including psychological methods, as well
longitudinal design which permits a more effective exam- as with some other factor (e.g., marital conflict) resulting
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 99

in both child depression and parents’ use of psychologi- and employment of the parent(s). Scores on the Socioe-
cal control. In this study, using a longitudinal design and conomic Index of Occupations in Canada (Blishen et al.,
a cross-lag statistical approach, evidence for causal rela- 1987), on the basis of income and education weighted
tions among variables is examined. That is, associations equally, range from 19 to a high of 101. Mean SES was
of marital conflict, parenting, and attachment style to level 43.60 (SD = 13.49) for mothers and 43.08 (SD = 13.28)
of adjustment 2 years later were examined. Moreover, the for fathers, characteristic of teachers, social work-
ability of these measures to predict changes in adjustment ers, personnel clerks and sales occupations. Participants
over time was assessed. As well, the converse ability, endorsing 1 ethnic background (75%) indicated either En-
of initial level of child adjustment to predict changes in glish Canadian (34%); French Canadian (4%); European
parenting and attachment over time, was examined. On (17%); Latin American, African, or Asian (1.5% each);
the basis of the literature, parental psychological con- or other (3%). Participants endorsing 2 (28.8%) or 3
trol was expected to be an important predictor of inter- (6.0%) ethnic backgrounds primarily indicated English
nalizing problems and self-esteem. Parental warmth and or French Canadian, and/or European. Results did not
behavioural control were also expected to predict adjust- differ by family structure, SES, or ethnicity.
ment, in particular externalizing problems. On the other Adolescents who continued to participate at T2
hand, child adjustment was expected to have little influ- (n = 175) and those who dropped out (n = 69) did not
ence on changes in parenting and attachment over time. differ significantly on T1 measures of parental behavioural
Pettit et al. (2001) identified both child gender and control, warmth, psychological control, marital conflict,
prior level of adjustment as important moderators of the attachment styles, depressive symptoms, or most demo-
link between psychological control and adjustment out- graphics. However, T2 participants reported significantly
comes. For example, higher levels of psychological con- higher grades (M = 1.93, SD = 0.62, versus M = 1.63,
trol were associated with more anxiety and depression for SD = 0.60), t(117) = 3.38, p < 0.001, and less variety of
girls and for those adolescents who were high in preado- delinquent acts (M = 3.58, SD = 3.65, versus M = 5.76,
lescent anxiety and depression. This study explored both SD = 4.84), t(96) = 3.38, p < 0.001. T2 participants
child gender and prior level of adjustment as potential also tended more often to be girls, to come from 2-parent
moderators of the role of parenting in the adjustment of homes, and to report higher self-esteem, relative to non-
young adolescents. participants, p < .10.
In sum, in this study, the role of psychological con-
trol, warmth and behavioural control, marital conflict, and Procedure
attachment security in the adjustment of young adoles-
cents is examined across a 2-year period. Hypotheses con- Written consent for participation at T1, required from
cerning causal direction of effect and mediating relations both adolescent and parent, was 54% (refusals 20% and
among these variables are explored. no response 26%). At T2, 2 years later, written consent,
required from the adolescent alone, was 93%. Attrition
METHOD from T1 to T2 was 29%, consisting of .5% refusal and
6.5% no response at T2, 2% refusals and 11% no response
Participants at an interim testing session, 6% moves from school, and
4% repeated absences at testing.
Participants were 175 adolescents (110 girls) At both T1 and T2, participating students completed
drawn from an original sample of 246 adolescents (145 questionnaires on family relationships (including parent-
girls) who had participated in the first year of a 3-year ing style, marital conflict, and attachment styles to mother
longitudinal study of family factors and adolescent and father), and adjustment (depressive symptoms, self-
adjustment. Participants were 13-years-old at Time 1 (T1, esteem, delinquent acts, and school grades). Data collec-
SD = 0.80), enrolled in grades 7 and 8 (n = 87 in each) tion was conducted in groups of about 20 at school, either
of a large, public, English-language high school in an in the classroom if numbers warranted or in a separate
older suburb of a large Canadian city. From information room. At T1, parental behavior, marital conflict, attach-
provided by the adolescents on a demographic question- ment styles, and grades were assessed during 1 class pe-
naire, most (79%) were from 2-parent homes, of which riod; the remaining adjustment measures and a measure
90% were intact and 10% reconstituted. For adolescents of parent demographics were completed during a second
in single parent families, 92% lived with their mother and class period approximately 1.5 months later. At T2, ques-
8% with their father. Socioeconomic status (SES) was tionnaires assessing parenting style, marital conflict, and
derived from information on the occupation, job activities most measures of adjustment were completed during the
100 Doyle and Markiewicz

first data collection; attachment styles were assessed in This questionnaire provides continuous self-report ratings
the second session, along with measures not relevant to of four attachment styles: Secure, Dismissing, Preoccu-
this study. pied, and Fearful. Participants were asked to rate the ex-
tent to which four paragraphs, each describing a different
Measures: Predictors of adjustment attachment style, pertained to their relationship with a par-
ent. For example, participants were asked to indicate on
Parenting Style a 7-point scale (1 = not at all, 7 = very much) the extent
to which the following paragraph (secure) typified their
At both T1 and T2, dimensions of parenting were relationship with their father: “It is easy for me to become
assessed by items rated on a 4-point scale ranging from emotionally close to my father. I am comfortable depend-
“never” to “always.” Parental Psychological Control was ing on my father and having my father depend on me. I
assessed by 3 items (e.g. “Do your parents punish you by don’t worry about being alone or having my father not ac-
saying you cause them distress?,” M = 2.67, SD = 0.52, cept me.” Adolescents completed two versions of the RQ,
α = 0.62) selected from Haapasalo and Tremblay (1994). one assessing the relationship with the mother, and the
Behavioral Control was assessed by 6 items (e.g. “In the second the relationship with the father (Bartholomew, per-
course of a day, how often do your parent actually know sonal communication, 1996; Scharfe and Bartholomew,
where you are?,” M = 1.95, SD = 0.53, α = 0.64) se- 1994). Students with only 1 parent were instructed to skip
lected from Smith and Krohn (1995). Warmth was as- the questionnaire about the other parent; students in re-
sessed by 8 items (e.g. “I can count on my Mom/Dad to constituted families chose whether to rate their biological
help me out, if I have some kind of problem,” M = 3.00, parent or step-parent.
SD = 0.62, α = 0.80) selected from Fletcher et al. (1995). In this study, the four ratings of attachment styles
The punishment scale from which the Psychological Con- were combined to yield dimensions of attachment anxiety
trol items were taken distinguished parents of aggressive and avoidance (Bartholomew and Shaver, 1998; Griffin
boys from parents of boys who were not aggressive (Haa- and Bartholomew, 1994). For each parent, the Anxiety
pasalo and Tremblay, 1994); higher levels of Behavioral dimension was obtained by summing the ratings of the
Control have been associated with lower rates of delin- two attachment styles with negative self-models (Preoc-
quency (Smith and Krohn, 1995); higher Warmth distin- cupied and Fearful) and subtracting the ratings of the two
guishes authoritative from non-authoritative parents, and styles with positive models of self (Secure and Dismiss-
children of authoritative parents are higher in social com- ing). Scores for Attachment Anxiety could therefore range
petence, and lower in delinquency and distress (Fletcher from −12 to +12, with higher scores representing greater
et al., 1995). anxiety about the worthiness and lovableness of the self.
The Avoidance dimension was obtained by summing the
Marital Conflict ratings of the attachment styles with negative models of
the parent (Dismissing and Fearful) and subtracting the
A short version of the Children’s Perception of In- ratings of the styles with positive models of the parent
terparental Conflict Scale (CPIC; Grych et al., 1992) was (Secure and Preoccupied). Scores for Attachment Avoid-
constructed to assess adolescents’ perceptions of the fre- ance could also range from −12 to +12, with higher
quency, intensity, and resolution of their parents’ marital scores representing greater tendency to avoid the parent
conflict. At both T1 and T2, adolescents rated each of and to view the parent as unavailable. Attachment Anxiety
12 items on a 3-point scale, (e.g., “I often see my parents scores for mother and father were correlated .31, and At-
arguing” as “false,” “sort of true,” or “true”). Participants tachment Avoidance scores .35. Thus, scores for mother
described the parents they lived with, or, if living with only and father were averaged (Anxiety M = 5.56, SD = 1.86;
1 parent, their parents when they were together. Higher Avoidance M = 4.60, SD = 2.06).
scores represent more conflict (M = 1.61, SD = 0.47; The RQ ratings correlate moderately with attach-
α = 0.88). The CPIC has been validated with parent ment styles determined by interview (Bartholomew and
reports of conflict and with child adjustment measures Horowitz, 1991), and provide a rapid assessment of at-
(Grych et al., 1992). tachment quality. Test–retest reliability is reported to av-
erage .51 across 8 months (Scharfe and Bartholomew,
Attachment to Mother and Father 1994). Validity data indicate that attachment styles as-
sessed using the RQ correlate as expected with mea-
The Relationship Questionnaire (RQ), adapted from sures of self-concept; interpersonal functioning; rep-
Bartholomew and Horowitz (1991), was used at both T1 resentations of family relationships as determined by
and T2 to assess the quality of attachment to each parent. interview; and family, peer, and self-ratings (Bartholomew
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 101

and Horowitz, 1991; Horowitz et al., 1993; Griffin and and delinquent behaviours. Participants indicate whether
Bartholomew, 1994). RQ ratings also are consistent with a series of 112 internalizing symptoms and externalizing
Hazan and Shaver’s (1987) traditional three-category symptoms are “not true,” “sometimes true,” or “very true”
model (Brennan et al., 1991). of them (scored 0, 1, and 2, respectively). Internalizing
Problems are measured by summing scores on the With-
Measures of Adjustment drawn (8 items, e.g., “I feel lonely”), Somatic Complaints
(10 items, e.g., “I feel dizzy”), and Anxious/Depressed
Symptoms of Anxiety/Depression and (13 items, e.g., “I cry a lot,” “I worry a lot”) subscales,
Delinquent Behaviours: Time 1 while Externalizing Problems are assessed by summing
scores on the Rule-Breaking Behaviour (15 items, e.g.,
At T1 a 53-item version of the 70-item Multiple Ad- “I cut classes or skip school”, “I hang around kids who
jective Attribute Checklist- Revised (MAACL; Zuckerman get in trouble”) and Aggressive Behaviour (17 items,
and Lubin, 1985) was used to assess dysphoric feel- e.g., “I have a hot temper”) subscales. Higher scores re-
ings. The adapted mood checklist consisted of the 10 flect higher problem behaviours. For the current sample,
items of the anxiety scale, 9 of the 12 items of the both second-order scales showed good internal consis-
depression scale, 13 of 15 items of the hostility scale, tency (α = .92, M = 12.60, SD = 9.97, and α = 0.91,
and the 21 items of the positive affect scale. The 5 M = 15.18, SD = 10.14, for Internalizing and External-
items omitted from these subscales were those judged izing, respectively). The YSR is a commonly used mea-
in pilot testing as unclear for young adolescents (e.g., sure of children’s behaviour problems with extensive ev-
“cross,” “forlorn,” “incensed,” “sunk” and “tormented”). idence of its reliability and validity (Achenbach, 1999).
Participants were asked to indicate the words which de-
scribe how they generally feel. The Anxiety, Depression General Self-Esteem
and Hostility subscales are significantly and moderately
correlated (i.e., between .4 and .6); thus, the second- At both T1 and T2, a 26-item version of the Self-
order Dysphoria scale was used (M = 5.59, SD = 5.42, Description Questionnaire II (SDQII, Marsh, 1988), was
α = 0.79). used to assess self-concept with respect to general self-
At T1, an adaptation of the Self-Report Delinquency esteem (GSE), same- and opposite-sex peer relations, and
Scale (SRDS, Elliott et al., 1985) was used to assess the academics, as well as pro-social behaviour. Only data from
past-year frequency of 39 major and minor delinquent the GSE subscale (5 items; e.g., “Overall I have a lot
acts, including status and property offences, and violent to be proud of,” M = 4.22, SD = 0.66; α = 0.81) were
offences against persons. Items range in severity from examined in this report. Students rated each item on a
theft under $5, purchasing alcohol as a minor, and van- 6-point scale (1 = false, 6 = true). The SDQ correlates
dalism to breaking and entering and assault. For exam- well with the Perceived Self-Competence Scales (Harter,
ple, participants respond “yes” or “no” to items such as 1982) and has the advantage of better discriminant validity
“Have you ever purposely damaged or destroyed property between the subscales (Marsh and Gouvernet, 1989).
(includes vandalism/graffiti) belonging to your school or
employer?,” and “Have you ever taken a vehicle for a ride
or drive without the owner’s permission?” If the partici- School Grades
pants respond “yes,” they also indicated how many times
in the past year on a 15-point scale, which ranged from At T1 students indicated their average letter grade
0 to 13-or-more times a week. The variety of the 19 delin- (F = 0, D or C = 1, B = 2, A = 3; M = 1.93, SD =
quent acts committed most frequently in the past year 0.62). At T2, students’ reports of grades (failing = 0,
(M = 3.58, SD = 3.65, α = 0.87) was computed as the below average = 1, average = 2, above average = 3) in
index of delinquency. The original SRDS correlates well each of 4 school subjects (English, history, mathematics
with official delinquency rates and with teacher and par- and science) were averaged (M = 2.00, SD = 0.52).
ent reports of delinquent behaviour (Elliott and Ageton,
1980; Elliott et al., 1985). Marlowe-Crown Social Desirability Scale (SD)

Symptoms of Anxiety/Depression and To control for defensive responding, a short version

Delinquent Behaviours: Time 2 of the Marlowe-Crown Social Desirability Scale (Stra-
han and Gerbasi, 1972) was given to the participants at
At T2 the Youth Self-Report (YSR, Achenbach, 2001) both T1 (7 items, α = 0.49) and T2 (15 items, M = 0.46,
was used to assess both symptoms of anxiety/depression SD = 0.20; α = 0.68). Participants were asked to indicate
102 Doyle and Markiewicz

Table I. Intercorrelations of Time 1 Predictors and Control Variables Controlling for Social Desirability Responding

N 1 2 3 4 5 6 7 8 9 10 11 12a

1. Marital conflict 166 — −.43∗∗ −.35∗∗ .28∗∗ .24∗∗ .36∗∗ .16∗ .17∗ −.18∗ −.20∗ −0.10 −.21∗∗
2. Parental warmth 173 — .42∗∗ −.24∗∗ −.37∗∗ −.54∗∗ −.20∗∗ −.23∗∗ .45∗∗ .28∗∗ .17∗ .18∗
3. Behavioral control 172 — −0.02 −.16∗ −.33∗∗ −0.01 −.37∗∗ .15∗ .18∗ −0.11 0.12
4. Psychological control 172 — .20∗ −0.01 .35∗∗ 0.14 −.23∗∗ −0.09 −0.09 −.31∗∗
5. Attachment anxiety w. parents 175 — .34 ∗∗ 0.11 .17∗ −.24∗∗ −.16∗ −.23∗∗ 0.12
6. Attachment avoidance w. parents 175 — 0.11 .19∗ −.33∗∗ −.26∗∗ −0.05 −0.13
7. Dysphoria 175 — 0.02 −.37∗∗ −0.08 −0.01 −.25∗∗
8. Delinquency 175 — −0.14 .32∗∗ .21∗∗ .26∗∗
9. Self-esteem 175 — −.17∗ 0.03 −.39∗∗
10. Grades 166 — 0.00 −0.04
11. Socioeconomic status 156 — −.21∗∗
12. Social desirabilitya 175 —

a Zero-order correlations.
∗p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001.

“True” or “False” for each item (e.g., “No matter who I’m relations with parental Warmth, Anxious Attachment, and
talking to, I’m always a good listener”). The abbreviated Self-Esteem were significant but low.
forms have been found to correlate highly with the orig- Intercorrelations of the 4 major T2 adjustment vari-
inal scale (r = 0.90, Strahan and Gerbasi, 1972), which ables are presented in Table II. Correlations with T2 SD
has been established as assessing the tendency to respond were low to moderate, highest for Externalizing Problems.
with social defensiveness (e.g., Lobel and Teiber, 1994). Controlling for SD, correlations were non-significant to
moderate in magnitude. Socio-economic status was not
associated significantly with any outcome variables and
RESULTS was thus dropped from further analysis.

Preliminary Analyses
T1 Marital Conflict, Parenting and Attachment as
Intercorrelations of potential T1 predictor variables Predictors of T2 Adjustment
are presented in Table I. Because of missing data due
to single-parent families, and/or absences at either T1 or Hypotheses concerning predictions from Marital
T2, final sample size ranged from 152 (analyses involv- Discord, parenting behaviours and attachment to adjust-
ing marital discord) to 175. Correlations with T1 social ment were tested via hierarchical multiple regression.
desirability (SD) were low to moderate, highest for T1 Gender and T2 Social Desirability responding were en-
Psychological Control and Delinquent behaviours. Con- tered first as control variables. The importance of parent-
trolling for SD, parental Psychological Control was in- ing for level of T2 adjustment was assessed by predicting
dependent of Behavioural Control and negatively associ- from the T1 parenting behaviours as a block to each of
ated with Warmth. Family relationships (Marital Conflict, the four T2 adjustment variables. Because of the inter-
attachment quality), parenting and T1 adjustment were est in parenting as a mediator of Marital Conflict effects
associated in expected ways. Socio-economic status cor- on adjustment, and in attachment security as a mediator

Table II. Intercorrelations of Time 2 Measures Controlling for Social Desirability Responding

N 1 2 3 4 5a

1. Internalizing behaviour 176 — .18∗ −.56∗∗ −0.08 −.40∗∗

2. Externalizing behaviour 176 −.24∗∗ −.27∗∗ −.60∗∗
3. Self-esteem 163 .24∗∗ .26∗∗
4. Grades 174 0.14
5. Social desirabilitya 168 —

a Zero-order correlations.
∗p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001.
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 103

Table III. Hierarchical Multiple Linear Regression: Prediction of T2 Internalizing Problems

T2 internalizing Change in internalizing T1 to T2

Step Predictor R 2 βa s2r R 2 βa s2r

1 Sex .20∗∗∗ −.16∗ .03 .26∗∗∗ −.15∗ .02

Dysphoria T1 — — .19∗ .03
Social desirability T2 −.39∗∗∗ .10 −.35∗∗∗ .10
2/− Marital conflict .02t .02 .00 — — —
3/2 Behavioural control .06∗∗ .05 .02 .03 .04 .00
Warmth .02 .00 .06 .00
Psychological control .23∗∗ .04 .16∗ .02
4/3 Attachment anxiety .04∗∗ .19∗ .03 .04∗ .19∗ .03
Attachment avoidance .07 .00 .09 .00
R2 = .31, F(8, 144) = 8.11∗∗∗ R2 = .33, F(8, 151) = 9.22∗∗∗

a Beta on last step.

tp < 0.10; ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < .001.

of parenting effects on adjustment, Marital Conflict was p < .01). In addition, when entered on Step 3, parental
entered as a predictor before parenting, and Anxious and Warmth uniquely predicted lower Externalizing Problems
Avoidant Attachment after. To conduct a strong test of the (ß = −.17, p < .05) and higher Self-Esteem (ß = .29,
importance of Marital Conflict, parenting and attachment p < .01). Thus, consistent with much previous research,
for adjustment, their associations with change in adjust- Marital Conflict and parenting behaviours were associ-
ment from T1 to T2 were assessed by repeating these anal- ated with later adolescent adjustment. Different aspects
yses controlling for level of T1 adjustment on the first step. of parenting were related uniquely to the different out-
Regression analyses of level of T2 adjustment in- comes: Psychological Control was particularly significant
dicated that perceived Marital Conflict was associated for higher internalizing behaviours; Warmth was partic-
with a tendency to more Internalizing Problems, and sig- ularly significant for lower Externalizing Problems and
nificantly more Externalizing Problems, lower T2 Self- higher Self-Esteem.
Esteem and lower school grades (see Tables III–VI). With respect to attachment insecurity, the 2 predic-
Moreover, T1 parenting behaviours as a block added tors as a block added significantly to the prediction of
significantly to the prediction of T2 Internalizing Prob- Internalizing Problems, Self-Esteem, and School Grades
lems, Externalizing Problems, and Self-Esteem; but only (see Tables III, V and VI). As shown, Attachment Anxiety
a trend for School Grades. When entered on Step 3 (not contributed uniquely to all 3 of these outcomes, and At-
shown in Tables III–VI), parental Psychological Con- tachment Avoidance contributed uniquely and negatively
trol uniquely predicted Internalizing Problems (ß = .23, to Self-Esteem.

Table IV. Hierarchical Multiple Linear Regression: Prediction of T2 Externalizing Problems

T2 externalizing Change in externalizing T1 to T2

Step Predictor R 2 βa s2r R 2 βa s2r

1 Sex .40∗∗∗ .18∗∗ .03 .51∗∗∗ .15∗ .02

Delinquency (D)T1 — — .46∗∗∗ .08
Gender × D — — −.24∗∗ .03
Social desirability T2 −.53∗∗∗ .25 −.46∗∗∗ .17
2/− Marital conflict .02∗ .01 .00 — — —-
3/2 Behavioural control .04∗ −.09 .01 .03∗ −.03 .00
Warmth −.13 .01 −.12 .01
Psychological control (PC) .01 .00 −.07 .00
4/3 Attachment anxiety .01 .01 .00 .00 −.01 .00
Attachment avoidance .08 .00 .06 .00
R 2 = .43, F (8, 144) = 15.01∗∗∗ R 2 = .51, F (9, 150) = 19.41∗∗∗

a Beta on last step.

tp < 0.10; ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < .001.
104 Doyle and Markiewicz

Table V. Hierarchical Multiple Linear Regression: Prediction of T2 Self-Esteem

T2 self-esteem Change in self-esteem T1 to T2

Step Predictor R 2 βa s2r R 2 βa s2r

1 Sex 0.07∗∗ 0.11 0.01 0.26∗∗∗ 0.08 0.01

Self-esteem T1 — — 0.29∗∗∗ 0.06
Social desirability T2 0.18∗ 0.03 0.14∗ 0.02
2/− Marital conflict 0.09∗∗∗ −0.09 0.01 — — —
3/2 Behavioural control 0.09∗∗∗ −0.02 0.00 0.05∗ 0.02 0.00
Warmth 0.14 0.01 0.07 0.00
Psychological control (PC) −0.14t 0.01 −0.03 0.00
4/3 Attachment anxiety 0.07∗∗∗ −0.20∗ 0.03 0.06∗∗∗ −0.17∗ 0.02
Attachment avoidance −0.19∗ 0.02 −0.19∗ 0.02
R 2 = 0.28, F (8, 144) = 8.38∗∗∗ R 2 = 0.33, F (8, 151) = 10.79∗∗∗

a Beta on last step.

tp < 0.10; ∗ p < 0.05; ∗∗ p < 0.01; ∗∗∗ p < 0.001.

With respect to predicting change in adjustment from p < .01). In addition, parenting behaviour tended to pre-
T1 to T2, these regression analyses were conducted again, dict change in School Grades (see Table VI), with the
controlling for the appropriate T1 adjustment measure. In interaction of Gender with Behavioral Control predicting
contrast to the analyses of level of T2 adjustment, per- uniquely (sr2 = .03, p < .05). Parental Behavioral Con-
ceived Marital Conflict was never a significant predictor trol predicted increases in School Grades over time for
of changes in adjustment over time, and thus was dropped boys (ß = .16, p < .05), but not for girls (ß = −.06, ns).
from further analyses. With respect to the parenting vari- Similar to analyses of level of T2 adjustment, At-
ables as predictors, as shown in Tables III–V, results were tachment Anxiety and Avoidance were significant predic-
quite similar to the previous analyses for changes in In- tors of increases in Internalizing Problems and decreases
ternalizing Problems, Externalizing Problems and Self- in Self-Esteem over time (see Tables III and V). For
Esteem over time. For Internalizing Problems, however, changes in School Grades, although attachment insecurity
parenting variables were no longer significant predictors as a block was no longer a significant predictor, Attach-
when entered as a block on the second step. For External- ment Anxiety entered alone was (ß = −.16, R 2 = .06,
izing Problems, the interaction of Gender with T1 Delin- p < .05, not shown in Table VI).
quency was also significant. T1 Delinquent Behaviour Thus, as hypothesized, parenting, in particular
was associated more strongly with T2 Externalizing Prob- parental Warmth and Psychological Control were sig-
lems for girls (ß = .46, p < .001) than for boys (ß = .22, nificant predictors of changes in adjustment over time.

Table VI. Hierarchical Multiple Linear Regression: Prediction of T2 Reported School Grades

T2 School grades Change in school grades T1 to T2

Step Predictor R 2 βa s2r R 2 βa s2r

1 Gender (G) .03 −.09 .01 .21∗∗∗ .00 .00

Grades T1 — — .40∗∗∗ .18
Social desirability T2 .06 .03 .08 .01
2/− Marital conflict .07∗∗ −.19∗ .02 — — —
3/2 Behavioural control (BC) .05t −.06 .00 .05t −.09 .00
Warmth .05 .00 .03 .00
Psychological control (PC) .05 .00 .02 .00
G × BC .22∗ .02 .26∗∗ .03
4/3 Attachment anxiety .04∗ −.20∗ .03 .02 −.15t .02
Attachment avoidance −.05 .00 −.04 .00
R 2 = .13, F (9, 142) = 3.41∗∗∗ R 2 = .23, F (9, 142) = 6.03∗∗∗

a Beta on last step.

tp < .10; ∗ p < .05; ∗∗ p < .01; ∗∗∗ p < .001.
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 105

T1 parental Psychological Control was associated posi- level of T1 attachment were entered on the first step, fol-
tively with increases in Internalizing Behaviour. Parental lowed by T1 measures of adjustment for which change
Warmth predicted decreases in Externalizing Problems, over time had been predicted by attachment on the second
and increases in Self-Esteem. The only effect of Behav- step. For predicting Attachment Anxiety, these were T1
ioral Control was for boys on School Grades. Attachment Dysphoria, Self-Esteem and School Grades. For Attach-
insecurity, particularly Attachment Anxiety also signifi- ment Avoidance, this was Self-Esteem. Though attach-
cantly predicted increased maladjustment over time. ment at T1 was a strong predictor of T2 attachment in both
cases (partial rs = 0.27 and 0.54, p < 0.001), level of T1
adjustment did not predict changes in Attachment Anxiety
Cross-lag Correlations Assessing Direction of Effects or Avoidance over time. Thus, attachment insecurity, in
particular Attachment Anxiety appears to causally affect
The previous analyses of changes in adjustment over Internalizing Problems, Self-Esteem, and School Grades.
time could be interpreted as indicating that parenting and
attachment contribute to changes in adjustment over time.
Parenting as a Mediator of the Effects of Marital
To strengthen this argument, we examined evidence for
Conflict on T2 Adjustment
the reverse possibility, that level of early adjustment de-
termines changes in parenting and attachment over time.
The potential mediating role of parenting behaviour
With respect to predicting changes in parenting, 3
in the associations between Marital Conflict and level of
regression analyses were conducted in which the 3 T2 par-
T2 adjustment was examined according to the criteria of
enting measures were predicted in separate analyses from
Baron and Kenny (1986). As noted in Table I, Marital
Gender, T2 Social Desirability, T1 Psychological Control
Conflict was significantly correlated with parenting be-
entered on the first step, and the relevant T1 adjustment
haviours. Moreover, as noted in the analyses of level of
measures entered on Step 2. For parental Psychological
T2 adjustment, Psychological Control was a significant
Control the T1 adjustment measures were Dysphoria and
correlate of Internalizing Problems and Warmth was a
Delinquency; for Warmth these were Delinquency and
significant correlate of Externalizing Problems and Self-
Self-Esteem, and for Behavioral Control it was School
Esteem. In addition, Warmth and Behavioral Control were
Grades. T1 parenting variables were highly significant
significantly correlated with T2 School Grades, control-
predictors of T2 parenting in all 3 analyses (p < 0.001).
ling for Gender and Social Desirability Responding (par-
Neither T1 Dysphoria nor Delinquency, however, signif-
tial rs = 0.25 and 0.21, p < 0.01); thus, the precondi-
icantly predicted changes in parental Psychological Con-
tions for testing mediation were fulfilled for the 4 main
trol. Nor did level of T1 School Grades predict changes in
indices of adjustment (Baron and Kenny, 1986). When
Behavioral Control. However, change in parental Warmth
parenting behaviours were included in the predictions of
over time was predicted from T1 Self-Esteem (R 2 =
adjustment, Marital Conflict was no longer a significant
0.03, p < 0.05; ß = 0.19, sr2 = 0.03, p < 0.01).
predictor of T2 Internalizing Problems (ß = 0.05, ns),
Thus, evidence suggests that parental Psychological
Externalizing Problems (ß = 0.03, ns), or Self-Esteem
Control, and to a lesser extent Behavioural Control, con-
(ß = −0.14, ns). Thus, parenting fully mediated the link
tribute to changes in adolescent adjustment over time, but
between Marital Conflict and level of T2 Self-Esteem,
there was no evidence of the converse, that adolescent ad-
and Internalizing and Externalizing Problems. However,
justment contributed to changes in parental Psychological
Marital Conflict remained a significant negative predictor
Control or Behavioral Control over time. Moreover, there
of School Grades (ß = −0.21, p < 0.05) and thus was
was evidence that parental Warmth contributed to changes
not mediated by parenting.
in Externalizing Problems over time, but no evidence of
the converse, that adolescent Delinquency contributed to
decreases in Warmth over time. However, evidence of Attachment Insecurity as Mediating Parenting
transactional effects emerged for Self-Esteem, in find- Effects on T2 Adjustment
ing both that parental Warmth contributed to increases in
Self-Esteem over time and that adolescent Self-Esteem The role of attachment as a mediator of the asso-
contributed to increases in parental Warmth over time. ciations between parenting and T2 adjustment was also
With respect to adjustment as predicting changes in examined according to the criteria of Baron and Kenny
attachment over time, for each of the 2 T2 attachment (1986). With respect to the preconditions for mediation
measures, separate regression analyses were conducted between parenting and level of T2 adjustment, as shown in
in which Gender, T2 Social Desirability responding, and Table I, Warmth and Behavioral Control were significantly
106 Doyle and Markiewicz

correlated with both Attachment Anxiety and Avoidance, Potential Moderating Effects of Gender, Initial Level
and Psychological Control was significantly correlated of Adjustment, and Parenting
with Attachment Avoidance. Attachment Avoidance was
also significantly correlated with T2 Externalizing Prob- On the basis of the literature, potential moderating
lems, controlling for Gender and Social Desirability Re- effects of gender, one parenting variable on another, and
sponding (partial r = 0.19, p < 0.01, thus fulfilling all initial level of adjustment were tested by examining the ef-
preconditions for mediation for all 4 main adjustment fects of 2-way interactions of these variables. In addition,
outcomes (Baron and Kenny, 1986). To test for media- the potential moderation of 1 dimension of attachment
tion of the joint contributions of the parenting variables insecurity on the other was examined, given that secure
as a block, the regressions described in Tables III – VI attachment is characterized as low Anxiety and low Avoid-
were rerun with Attachment Anxiety and Avoidance en- ance. As recommended by Jaccard et al. (1990), the con-
tered as predictors prior to the entry of the parenting mea- tinuous independent variables were z-standardized prior
sures (not shown). For level of T2 Internalizing Problems, to creating the interaction terms, which were then entered
both attachment (R 2 = 0.05, p < 0.01) and parenting on the last step of the regression, and the significance
continued to significantly predict outcome (R 2 = 0.05, level of each interaction term was inspected separately.
p < 0.05), with Psychological Control still contribut- The hierarchical regression analyses were then repeated
ing uniquely (ß = 0.23, sr2 = 0.04, p < 0.01). For T2 on the z-standardized dependent variable, including only
Self-Esteem, where parental Warmth had previously con- the significant interaction terms to increase test power
tributed uniquely, attachment insecurity predicted sig- to obviate the difficulty of detecting interactions in field
nificantly (R 2 = .09, p < .001) and parenting only studies (McClelland and Judd, 1993). The only signifi-
tended to predict (R 2 = 0.03, p < 0.10), with Psycho- cant interaction effects were those between Gender and
logical Control tending to predict uniquely (ß = −0.14, initial level of Delinquency, and between Gender and Be-
sr2 = 0.00, p < 0.10). Moreover, for level of T2 External- havioural Control, as described in previous sections.
izing Problems, where Warmth had again previously con-
tributed uniquely, even though attachment insecurity only
tended to predict significantly (R 2 = 0.02, p < 0.10), DISCUSSION
parenting no longer predicted outcome (R 2 = 0.02,
ns.). With respect to level of T2 School Grades, predic- The principal hypotheses of the study were con-
tions were examined without Marital Conflict because in firmed. That is, parenting, including psychological con-
that case parenting predicted significantly (R 2 = 0.08, trol, and attachment security had important implications
p < .01), with Warmth predicting uniquely (ß = 0.19, for both the level of adjustment and changes in adjust-
sr2 = 0.02, p < 0.05). When attachment insecurity was ment over time. Analyses of direction of effect confirmed
included first, parenting was no longer a significant predic- that parenting and attachment affect adolescent adjust-
tor of School Grades (R 2 = 0.04, ns). Thus, attachment ment more than the reverse. Marital conflict was also
insecurity mediated the association between parenting, in associated with level of maladjustment, and these asso-
particular Warmth, and level of T2 Externalizing Prob- ciations were meditated by parenting. In addition, attach-
lems and School Grades. However, Psychological Con- ment security mediated the effects of parental warmth on
trol effects on level of T2 Internalizing Problems and adjustment, although not the effects of psychological con-
Self-Esteem (trend) were not mediated by attachment. trol. Each of these findings is discussed more fully in the
With respect to attachment insecurity mediating par- following sections.
enting effects on changes in adjustment from T1 to T2,
results were similar to the previous analyses for Internal- Associations of Parenting with
izing problems, Externalizing Problems and Self-Esteem. Adolescent Adjustment
For School Grades, parenting still tended to be signif-
icant (R 2 = 0.05, p < .10), with the interaction of As hypothesized, greater reported parental psycho-
gender and behavioural control predicting uniquely (see logical control was associated both with more internal-
Table VI). Thus, attachment insecurity mediated the ef- izing problems 2 years later and with increases in in-
fects of parenting, in particular Warmth, on changes in Ex- ternalizing problems over time. Moreover, psychological
ternalizing Problems and Self-Esteem over time, but not control was the only aspect of parenting to make a unique
the effects of Psychological Control on changes in Inter- contribution to internalizing problems, and accounted for
nalizing Problems, nor the effects of Behavioral Control most of the variance in outcome attributable to parenting.
for boys on changes in school grades. That is, psychological control was the most important
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 107

parenting influence on adolescent internalizing problems. ure to replicate previous findings may again reflect the
The results for internalizing problems are congruent with current smaller sample size or the fact that grades were
the extensive extant literature (summarized by Barber and self-reported. Nevertheless, it is noteworthy that increases
Harmon, 2002). However, this study extends those find- in reported school grades was the only outcome predicted
ings, which focus almost exclusively on concurrent asso- by behavioural control, and that the prediction was moder-
ciations with the exception of a few longitudinal studies ated by gender such that the effect was for boys only. The
(Conger et al., 1997; Pettit et al., 2001). The fact that failure to find stronger associations between behavioural
psychological control effects were found for internaliz- control and outcome is not inconsistent with the litera-
ing problems is consistent with the view that this parental ture (Barber and Harmon, 2002) and likely reflects poor
technique induces shame, anxiety, and guilt in particular. discrimination by the present behavioral control items,
The cross-lag correlations examining directions of which were taken from a study of a less normative sample
influence indicated that initial levels of internalizing did (Smith and Krohn, 1995).
not predict changes in psychological control over time.
Taken together, the present findings imply rather strongly
Marital Conflict and Adjustment:
that parental psychological control is not a response to
Mediated by Parenting?
child behaviour but is rather a contribution by the parent
that has significant negative influence on child internaliz-
Marital conflict showed consistent associations with
ing problems, consistent with the interpretation of Barber
level of adjustment 2 years later; 3 associations were sig-
and Harmon (2002).
nificant, and for internalizing problems there was a trend.
However, except for a tendency to predict self-
Moreover, parenting, in particular psychological control
esteem, psychological control did not significantly predict
and warmth, mediated the effects of marital conflict on
the other 3 adjustment measures. The failure to find sig-
level of internalizing and externalizing problems, and self-
nificant effects on self-esteem as well as on internalizing
esteem. These finding are consistent with the findings of
problems is puzzling, perhaps reflecting that the negative
Fauber et al. (1990) in a cross-sectional study of inter-
effects of psychological control on the self-concept are
parental conflict, psychological control, rejection and lax
less conscious and/or more difficult to assess by a mea-
control. The present findings contrast with those of Stone
sure focussing primarily on positive self-statements.
et al. (2002), who conducted a cross-sectional study which
The lack of association between parental psycho-
examined only psychological control as a measure of par-
logical control and externalizing symptoms is not un-
enting, and who also assessed covert conflict. Interest-
expected, given the inconsistency of findings in the lit-
ingly, in the present longitudinal study, marital conflict
erature. It is noteworthy, however, that parental warmth
was not associated with change in adjustment over time
significantly predicted decreases in externalizing prob-
whereas parenting and attachment were. Taken together
lems and increases in self-esteem. These results are par-
with the significant concurrent associations between mar-
tially consistent with those of Pettit and Laird (2002), who
ital conflict and parenting, these results suggest that mar-
found effects of both psychological control and parental
ital conflict affects internalizing and externalizing prob-
warmth in a larger, younger sample. The present cross-lag
lems and self-esteem indirectly, through parenting. That
analyses strengthen the causal role attributed to parental
is, marital conflict affects parenting negatively, and how
warmth in reducing externalizing problems, in that ini-
parents directly interact with their children relates causally
tial adolescent delinquent behaviour was not associated
to changes in these indices of adjustment. However, with
with changes in parental warmth over time. Nevertheless,
respect to reported school grades, marital conflict had di-
similar analyses found evidence of bidirectional effects
rect effects over time, independent of parenting. The stress
between parental warmth and in self-esteem over time.
and anxiety caused by interparental conflict per se may
These findings not only attest to the positive influence
have particular and direct implications for children’s mo-
of parental warmth on adolescent self-esteem but to the
tivation and ability to concentrate on schoolwork both at
likelihood that parents find it easier to respond warmly to
home and at school, apart from effects on parenting.
self-confident adolescents.
In this study, there was no association between
parental psychological control and adolescent school Attachment Insecurity and Adolescent Adjustment
achievement, in contrast to other findings (Barber and
Shagle, 1992; Steinberg et al., 1989). Instead, in this study As expected, there was considerable evidence in
behavioural control tended to positively predict increases this study that attachment insecurity, particularly anxiety
in school achievement for boys but not for girls. This fail- about the self as worthy of parental love, was associated
108 Doyle and Markiewicz

with maladjustment. Attachment anxiety was significantly worthy, which in turn is associated in adolescence with
associated with increases in internalizing problems and higher school grades. However, for boys, greater parental
decreases in self-esteem over time, and with level of re- behavioural control also tends to directly affect school
ported school achievement. Attachment avoidance was performance.
also uniquely associated with decreases in self-esteem In contrast to findings concerning parental warmth
over time. The links between attachment and adolescent and adjustment, in contrast to expectation, there was no
adjustment are consistent with the literature (e.g., Cooper indication that attachment insecurity mediated the effects
et al., 1998; Nada-Raja et al., 1992). The present longitu- of psychological control on internalizing problems. In
dinal study goes beyond the previous correlational litera- fact, psychological control was only weakly associated
ture, however, to strengthen the argument that attachment with attachment anxiety and non-significantly with at-
security contributes causally to internalizing problems and tachment avoidance. It is puzzling that the expected asso-
self-esteem. Moreover, this study highlights that in com- ciations were not found, given the similarity between the
parison with attachment avoidance, attachment anxiety is parenting behaviours defined as psychologically control-
more implicated in the development of internalizing prob- ling and those identified as leading to anxious attachment.
lems, consistent with the central role of anxiety in both. The weak associations found may reflect limitations in
In contrast, both styles of insecurity, anxiety and avoid- the measures of psychological control used in this study,
ance, were implicated in decreasing self-esteem, despite which relied on very few items. Few studies, if any, have
the theory that the anxiety dimension reflects attributions previously examined associations between psychological
to the self more strongly than the avoidance dimension. control and attachment security. Thus, speculations that
The implications of attachment anxiety and avoidance for parents’ use of psychological control undermines child
attributions about self and other merit more study. attachment security must await further research.
In the current study, moderation of the associations of
psychological control with internalizing and externalizing
Attachment Insecurity as a Mediator of Parenting behaviours by gender and initial level of adjustment were
Effects on Adjustment explored, similar to the approach taken by Pettit and Laird
(2002). In that study, the effect of parental psychological
Given the importance of security of attachment for control on delinquency depended on the level of parent
adolescent adjustment, the role of attachment anxiety involvement (warmth); in the current study, however, the
and avoidance in mediating the effects of parenting on effects of psychological control were not moderated by
adjustment was examined. As hypothesized, attachment either the level of parental warmth or behavioural control.
insecurity mediated the effects of parenting, in particu-
lar warmth, on changes in externalizing problems and
self-esteem. Interestingly, attachment anxiety and avoid- Limitations
ance both appeared to mediate warmth; for self-esteem
both contributed uniquely, however, and for externalizing The sample size of the current study was relatively
problems, neither contributed uniquely. This finding im- small, precluding the use of structural equation modeling.
plies that lower parental warmth affects both adolescents’ Moreover, as noted previously, the index of psychologi-
view of themselves as worthy and their view of parents as cal control consisted of only 3 items assessing parental
available. Attachment theory would suggest that effects punishment style: using derision (calling the adolescent
on anxiety might be greater for some adolescents and names), guilt- and anxiety-induction (saying the adoles-
avoidance greater for others, depending on other aspects cent causes them distress), and coercion/hostility (argu-
of parent–child interaction, but such speculation remains ing). The fact that significant effects, consistent with the
to be verified by future research. In this study, there was literature, were found with this limited measure attests to
no indication that one aspect of parenting was moderated the robustness and importance of the construct. In addi-
by another. tion, more associations between psychological control and
Attachment insecurity, in particular attachment anx- attachment security/anxiety might be found with a more
iety, did not appear to mediate the effect of behavioural differentiated and reliable measure of attachment than the
control for boys on changes in reported school achieve- single-item paragraphs of the Relationship Questionnaire.
ment over time. This finding, taken together with the find- A significant limitation is the use of adolescent self-
ing that attachment anxiety was associated with lower report for all measures. Nevertheless, adolescent self-
school achievement, suggests that optimal parenting has report is well established as an important source of infor-
implications for the adolescent’s sense of self as love- mation on both parenting and on adolescent internalizing
Parenting, Marital Conflict and Adjustment From Early- to Mid-Adolescence 109

behaviours and self-esteem. The use in this study of both Bartholomew, K., and Shaver, P. R. (1998). Methods of assessing adult
social desirability responding and initial level of adjust- attachment: Do they converge? In Simpson, J. A., and Rholes, W.
S. (eds.), Attachment Theory and Close Relationships. Guilford,
ment as controls helps ameliorate this limitation. New York, pp. 25–45.
Belsky, J. (1999). Interactional and contextual determinants of attach-
ment security. In Cassidy, J., and Shaver, P. R. (eds.), Handbook of
Conclusions
attachment: Theory, Research, and Clinical Applications. Guilford,
New York, pp. 249–264.
This study substantiates the importance of psycho- Blishen, B. R., Carroll, W. K., and Moore, C. (1987). The 1981 socioeco-
nomic index for occupations in Canada. Can. Rev. Soc. Anthropol.
logical control as a risk factor for the adjustment of young
24: 465–488.
adolescents over time, in particular for the exacerbation Brennan, K. A., Shaver, P. R., and Tobey, A. E. (1991). Attachment styles,
of internalizing problems. Psychological control appears gender and parental problem drinking. J. Soc. Pers. Relationships
8: 451–466.
to mediate associations between marital discord and in-
Conger, K. J., Conger, R. D., and Scaramella, L. V. (1997). Parents,
ternalizing problems, but to be independent and additive siblings, psychological control, and adolescent adjustment. J. Adol.
to effects of attachment anxiety on that outcome. Thus, Res. 12: 113–138.
Cooper, M. L., Shaver, P. R., and Collins, N. L. (1998). Attachment
the particular influence of parental psychological control
styles, emotion regulation, and adjustment in adolescence. J. Pers.
on adolescent adjustment merits continued investigation. Soc. Psychol. 74: 1380–1397.
Warmth is confirmed as an important aspect of parent- Cummings, E. M., and Cummings, J. S. (2002). Parenting and attach-
ing for self-esteem and externalizing problems, mediat- ment. In Bornstein, M. H. (ed.), (2002). Handbook of Parenting:
Vol. 5: Practical Issues in Parenting (2nd edn.). Erlbaum, Mahwah,
ing associations of marital conflict with these aspects of NJ.
adjustment. Importantly, the effects of warmth on these Davies, P., and Cummings, E. M. (1994). Marital conflict and child
outcomes are mediated by attachment insecurity. adjustment: An emotional security hypothesis. Psychologic. Bull.
116: 387–411.
De Wolff, M., and van Ijzendoorn, M. H. (1997). Sensitivity and attach-
ment: A meta-analysis on parental antecedents of infant attachment.
ACKNOWLEDGMENTS Child Dev. 68: 571–591.
Elliott, D. S., and Ageton, S. (1980). Reconciling differences in estimates
This research was completed with assistance from of delinquency. Am. Sociologic. Rev. 45: 95–110.
Elliott, D. S., Huizinga, D., and Ageton, S. (1985). Explaining Delin-
the Social Sciences and Humanities Research Council of quency and Drug Use. Sage Beverly Hills, CA.
Canada and the Quebec Fonds FCAR. Portions of this Erel, O., and Burman, B. (1995). Interrelatedness of marital relations
paper were presented at the Society for Research in Child and parent–child relations: A meta-analytic review. Psychologic.
Bull. 118: 108–132.
Development, Tampa, April, 2003. Thanks to the partic- Fauber, R., Forehand, R., Thomas, A. M., and Wierson, M. (1990). A
ipating staff, teachers and students of Lasalle Catholic mediational model of the impact of marital conflict on adolescent
Comprehensive School, Lester B. Pearson School Board. adjustment in intact and divorced families: The role of disrupted
parenting. Child Dev. 61: 1112–1123.
Fletcher, A. C., Darling, N. E., Steinberg, L., and Dornbusch, S. (1995).
REFERENCES The company they keep: Relation of adolescents’ adjustment and
behavior to their friends’ perceptions of authoritative parenting in
the social network. Dev. Psychol. 31: 300–310.
Achenbach, T. M. (1999). The child behavior checklist and related in- Fraley, R. C., and Davis, K. E. (1997). Attachment formation and transfer
struments. In Maruish, M. E. (ed.), The Use of Psychological Test- in young adults’ close friendships and romantic relationships. Pers.
ing for Treatment Planning and Outcomes Assessment (2nd edn.). Relationships 4: 131–144.
Erlbaum, Mahwah, NJ. Griffin, D. W., and Bartholomew, K. (1994). Models of the self and other:
Achenbach, T. M. (2001). Achenbach System of Empirically Based Fundamental dimensions underlying measures of adult attachment.
Assessment—Youth Self-Report Syndrome Scales. University of J. Pers. Soc. Psychol. 67: 433–445.
Vermont, Burlington, VT. Grossmann, K. E., Grossmann, K., and Zimmermann, P. (1999). A wider
Allen, J. P., and Hauser, S. T. (1996). Autonomy and relatedness in view of attachment and exploration: Stability and change during the
adolescent-family interactions as predictors of young adults’ states years of immaturity. In Cassidy, J., and Shaver, P. R. (eds.), Hand-
of mind regarding attachment. Dev. Psychopathol. 8: 793–809. book of Attachment: Theory, Research, and Clinical Applications.
Barber, B. K., and Harmon, E. L. (2002). Violating the self: Parental The Guilford Press, New York, pp. 760–786.
psychological control of children and adolescents. In Barber, B. K. Grych, J. H., Seid, M., and Fincham, F. D. (1992). Assessing mar-
(ed.), Intrusive Parenting: How Psychological Control Affects Chil- ital conflict from the child’s perspective: The children’s per-
dren and Adolescents. American Psychological Association, Wash- ception of interparental conflict scale. Child Dev. 63: 558–
ington, DC, pp. 15–52. 572.
Barber, B. K., and Shagle, S. (1992). Adolescent problem behaviors: A Haapasalo, J., and Tremblay, R. E. (1994). Physically aggressive boys
sociological analysis. Fam. Perspect. 26: 493–515. from ages 6 to 12: Family background, parenting behavior, and
Baron, R. M., and Kenny, D. A. (1986). The moderator-mediator variable prediction of delinquency. J. Consult. Clin. Psychol. 62: 1044–
distinction in social psychological research: Conceptual, strate- 1052.
gic, and statistical considerations. J. Pers. Soc. Psychol. 51: 1173– Harter, S. (1982). The perceived competence scale for children. Child
1182. Dev. 53: 87–97.
Bartholomew, K., and Horowitz, L. M. (1991). Attachment styles among Hazan, C., and Shaver, P. (1987). Romantic love conceptualized
young adults: A test of a four-category model. J. Pers. Soc. Psychol. as an attachment process. J. Pers. Soc. Psychol. 52: 511–
61: 226–244. 524.
110 Doyle and Markiewicz

Hazan, C., and Zeifman, D. (1994). Sex and the psychological Paterson, J., Pryor, J., and Field, J. (1995). Adolescent attachment to
tether. In Bartholomew, K., and Pearlman, D. (eds.), Advances parents and friends in relation to aspects of self-esteem. J. Youth
in Personal Relationships. London: Jessica Kingsley pp. 151– Adolesc. 24: 365–376.
177. Pettit, G. S., and Laird, R. D. (2002) Psychological control and moni-
Horowitz, L. M., Rosenberg, S. E., and Bartholomew, K. (1993). In- toring in early adolescence: The role of parental involvement and
terpersonal problems, attachment styles, and outcome in brief earlier child adjustment. In Barber, B. K. (ed.), Intrusive Parent-
dynamic psychotherapy. J. Consult. Clin. Psychol. 61: 549– ing: How Psychological Control Affects Children and Adolescents.
560. American Psychological Association, Washington, DC, pp. 97–
Jaccard, J., Turrisi, R., and Wan, C. K. (1990). Interaction Effects in 123.
Multiple Regression. Sage, London. Pettit, G. S., Laird, R. D., Dodge, K. A., Bates, J. E., and Criss, M. M.
Karavasilis, L., Doyle, A. B., and Markiewicz, D. (2003). Associations (2001). Antecedents and behavior-problem outcomes of parental
between parenting style and attachment to mother in middle child- monitoring and psychological control in early adolescence. Child
hood and adolescence. Int. J. Behav. Dev. 27: 153–164. Dev. 72: 583–598.
Lamborn, S. D., Mounts, N. S., Steinberg, L., and Dornbusch, S. (1991). Scharfe, E., and Bartholomew, K. (1994). Reliability and stability of
Patterns of competence and adjustment among adolescents from adult attachment patterns. Pers. Relationships 1, 23–43.
authoritative, authoritarian, indulgent and neglectful families. Child Silk, J. S., Morris, A. S., Kanaya, T., and Steinberg, L. (2003). Psycholog-
Dev. 62: 1049–1065. ical control and autonomy granting: Opposite ends of a continuum
Lieberman, M., Doyle, A. B., and Markiewicz, D. (1999). Developmen- or distinct constructs? J. Res. Adolesc. 3: 113–128.
tal patterns in security of attachment to mother and father in late Smith, C., and Krohn, M. D. (1995). Delinquency and family life among
childhood and early adolescence: Associations with peer relations. male adolescents: The role of ethnicity. J. Youth Adolesc. 24: 69–93.
Child Dev. 70: 202–213. Steinberg, L. (1990). Autonomy, conflict, and harmony in the family
Lobel, T. E., and Teiber, A. (1994). Effects of self-esteem and need for relationship. In Feldman, S. S., and Elliott, G. R. (eds.), At the
approval on affective and cognitive reactions: Defensive and true threshold: The developing adolescent. Harvard University Press,
self-esteem. Pers. Individual Diff. 16: 315–321. Cambridge, MA, pp. 255–276.
Marsh, H. W. (1988). Self-Description Questionnaire I and II. Psycho- Steinberg, L., Elman, J. S., and Mounts, N. S. (1989). Authoritative par-
logical Corporation, San Antonio, TX. enting, psychosocial maturity, and academic success among ado-
Marsh, H. W., and Gouvernet, P. J. (1989). Multidimensional self- lescents. Child Dev. 60: 1424–1436.
concepts and perceptions of control: Construct validation of re- Steinberg, L., Lamborn, S. D., Dornbusch, S. M., and Darling, N. (1992).
sponses by children. J. Educ. Psychol. 81: 57–69. Impact of parenting practices on adolescent achievement: Author-
McClelland, G. H., and Judd, C. M. (1993). Statistical difficulties of itative parenting, school involvement, and encouragement to suc-
detecting interactions and moderators effects. Psychol. Bull. 114: ceed. Child Dev. 63: 1266–1281.
376–390. Stone, G., Buehler, C., and Barber, B. K. (2002). Interparental con-
Nada-Raja, S., McGee, R., and Stanton, W. R. (1992). Perceived at- flict, parental psychological control, and youth problem behavior.
tachments to parents and peers and psychological well-being in In Barber, B. K. (ed.), Intrusive Parenting: How Psychological
adolescence. J. Youth Adolesc. 21(4): 471–485. Control Affects Children and Adolescents. American Psychologi-
Parish, T. S., and McCluskey, J. J. (1992). The relationship between cal Association, Washington, DC, pp. 53–95.
parenting styles and young adults’ self-concepts and evaluations of Van-Ijzendoorn, M. H., and Bakermans-Kranenburg, M. J. (1996). At-
parents. Adolescence 27: 915–918. tachment representations in mothers, fathers, adolescents, and clin-
Patterson, G. R., and Fisher, P. A. (2002). Recent developments in our ical groups: A meta-analytic search for normative data. J. Consult.
understanding of parenting: Bidirectional effects, causal models, Clin. Psychol. 64: 8–21.
and the search for parsimony. In Bornstein, M. H. (ed.), (2002). Zuckerman, M., and Lubin, B. (1985). Manual for the Multiple Affect
Handbook of Parenting: Vol. 5: Practical Issues in Parenting (2nd Adjective Checklist Revised. Educational and Industrial Testing Ser-
edn.). Erlbaum, Mahwah, NJ. vice, San Diego.