Parent-Child Relationships and Self-Concepts of Adolescents

dinnerworkableΠολεοδομικά Έργα

16 Νοε 2013 (πριν από 3 χρόνια και 8 μήνες)

137 εμφανίσεις


1


DARMSTADT

UNIVERSITY

OF TECHNOLOGY

INSTITUTE OF PSYCHOL
OGY

Division of Personality

and Developmental Psychology


What Means and Difference Scores Don’t Tell Us:

Parent
-
Child Relationships and Self
-
Concepts of Adolescents
Du
r
ing School
-
to
-
Work Transitio
n


J. Gowert Masche


Paper presented as part of the symposium “Heterogeneity of dyadic relationships: Are all
frien
d
ships, antipathies, sibling, and parent
-
child relationships the same?” at the 17th Biennial
Meeting of the International Society for the Stu
dy of Behavioural Development (ISSBD),

August 2

6, 2002 at Ottawa, Ontario, Canada.


1 Introduction

This symposium investigates several dyadic relationships involving chil
dren and adolescents.
All papers share one methodological approach, namely calcula
ting mean and discrep
ancy meas
-
ures. Thus, instead of individuals, properties of the dyad are subject to the analyses. This
sounds as a good idea at first sight: People (generally) are not lonesome riders but form systems
of dyads, families, groups, etc
. So, why not analysing data on the system level? How
ever, I was
doubtful how useful this way of doing it would be. So, my contribution presents a test of this
approach. First, I’ll employ this methodological approach by using data on parent
-
ado
lesce
nt
relationships. As you will see, there were corre
lations between adoles
cents’ self
-
es
teem on the
one hand and on the other hand mean dyadic satisfaction and discrepancy in satis
faction between
adoles
cents and mothers or fathers, respectively. Give
n that the statistical ap
proach is explicitly
at the dyad level, I will also adopt a dyadic interpretation of the results.

Second, I will run another analysis on the same data set. This time, I shall not aggregate
the data into dyadic composites but in
stead will leave them as they are. The results also will be
inter
preted. We might see how the two interpretations will fit or whether one should be pre
-
ferred over the other. Third, I will take advantage of the fact that my data set is longitudinal and

will show you how things developed over time. More specifically, my sample of families with
adolescents was drawn at tenth and thus final grade of German middle
-
school (
Realschule
). Stu
-
dents gradu
ating from middle
-
school have the choice either to cont
inue school education, for
example at a college
-
bound school track, or they can enter a professional training. I assume that
before graduating from middle
-
school, students going to continue school were comparable in
terms of similar correlations between f
amily variables and self
-
esteem to those students going to
leave school after graduation. If this as
sumption holds, then I can define the group of students
leaving school as the “experimental group”, the group of students continuing school as the “con
-
tr
ol group”. I will call the first group the
apprentices
, the second the
students
. Both groups were

2

investigated first three months before leaving school, second at the time of beginning the new
education that was either an apprentice
ship or further schoo
l education, and third three more
months later.

According to individuation theory of parent
-
adolescent relationships (Youniss & Ketter
-
linus, 1987; Youniss & Smollar, 1985, 1989), adolescents’ identity development is strongly in
-
fluenced by the parent
-
ad
olescent relationships. This should also apply to self
-
esteem that may
be regarded as one of the central issues of identity. However, Roberts and Bengtson (1993) sug
-
gested that the importance of parents for adolescents’ self
-
esteem may vanish when the r
ole of a
child in a family loses its salience because of the acquisition of new roles as worker, spouse, or
parent. In addition, when I planned this study, I had the idea that the status transition from
school to work may alter the role expectations withi
n the family. Possibly, the adolescent claims
a more adult role, which also is assigned to her/him by the parents. So, I hypothesise that in the
course of the school
-
to
-
work transition, the correlation between the quality of the parent
-
adoles
-
cent dyad a
nd adolescents’ self
-
esteem will decline. For the control group, no change is ex
-
pected, so that after the transition, the correlations between the dyad quality and self
-
esteem
should be higher in stu
dents than in apprentices.

2 Methods

The total numbers

of non
-
divorced families participating in waves 1, 2, and 3 of data collections
were 52, 45, and 41. The numbers of families with apprentices were 22, 21, and 20 at the three
times, and the according numbers of families with students were 30, 24, and 21.

In each family,
the target adolescent (mostly age 15

17), the mother, and the father filled in several question
-
naires and were interviewed orally. The sample was German
-
speaking and predominantly mid
-
dle
-
class.

The adolescents separately assessed their

relationships to their mothers and fathers, and
mothers and fathers assessed their relationships to their adolescent children. Each assessment
consisted of four measures.

First, the
Relationship Satisfaction Scale

(Masche, 2000, derived from Family Adapt
abil
-
ity Measure; Skinner, Steinhauer, & Santa
-
Barbara, 1983; and Dependability and Emotional
Hesitancy Scales; Kreppner, 1995) is a 15
-
item scale that contains items such as “She is always
there for me if I want to talk to her”, “If I am in a bad way, I c
an always rely on his help”. In a
representative sample (Masche & B
arber, 2001), this scale reached an α reliability of .87 for the
adolescent
-
mother and of .90 for the adolescent
-
father relationship.

Second, the family members had to describe each other and their relationships towards
each other during five minutes eac
h in free
Five
-
Minutes Speech Samples

(FMSS, Fricke et al.,
1993). Three measures were extracted from these oral statements, (a) the apparent
global satis
-
faction

(rated from
dissatisfied

0 to
satisfied

3), (b) the
amount of critique

expressed, and (c) th
e
number of positive statements
. The amount of critique was reverse coded so that for each meas
-
ure, higher numbers indicated a higher level of satisfaction.

As preliminary principal component analyses showed, these measures for each of the four
assessed
relation
ships (adolescent
-
mother, mother
-
adolescent, adolescent
-
father, father
-
adoles
-
cent) loaded highly on a single factor.
1

Thus, the
z

scores of these measures were summed up to



1

The scales d
escribed and some other scales not mentioned in this paper due to time constraints were factor
analyzed (PCA) at each time of measurement separately. The factor structures were equivalent between the times.
The scales described in this paper belonged to
the same factor, but different dyadic relationships loaded in part on
different factors. See Masche (submitted) for details.


3

one global measure for each relationship. The mean dyadic quality was ca
lculated by summing
up the respective recipro
cal assessments, divided by two. Two discrepancy measures were cal
-
culated. The first measure was the adoles
cent’s assessment minus the respective parent’s as
-
sessment, e. g., the adolescent’s satisfaction w
ith her mother minus the mother’s satisfaction with
her adolescent child. The second measure was the absolute value (without sign) of the first one.
This makes the latter measure independent of the direc
tion of the discrepancy, i. e. independent
of whet
her for example the mother or the adolescent was more satisfied.

The adolescents’ self
-
esteem was self
-
assessed, i. e. rated by the adolescents, and sepa
-
rately rated by mothers and fa
thers. Each of these family members was asked how satisfied the
adoles
cent currently was about her/himself. The respondents marked their assessment on a scale
from 0 to 10 with
dissatisfied

and
very satisfied

as anchors.

3 First results: Effects of levels and discrepancies of dyadic relationships on self
-
esteem

Tables 1
and 2 show the correlations between adolescents’ self
-
esteem and the levels and di
s
-
crepancies of the parent
-
adolescent dyadic relationships. As expected, the mean levels of rel
a
-
tionship quality were positively related to the self
-
esteem assessments of t
he adolescents the
m
-
selves as well as of their mothers and their fathers (mother
-
adolescent dyad only).

Table 1:

Correlations between adolescents’ self
-
esteem and mother
-
adolescent dyadic relationships

Self
-
esteem
as rated by…

Relationship quality

Level:


Mean quality

Discrepancy

(adol.


mo.)

Discrepancy
|adol.


mo.|

Adolescents’
assessment

Mothers’
a
s
sessment

Adolescents

.42**

-
.10

-
.21

.30*

.41**

Mothers

.43**

-
.45***

-
.05

.13

.62***

Fathers

.42**

-
.13

-
.18

.28*

.43**



p

< .10, *
p

< .05, **
p

<
.01, ***
p

< .001


Table 2:

Correlations between adolescents’ self
-
esteem and father
-
adolescent dyadic relationships

Self
-
esteem as
rated by…

Relationship quality

Level:

Mean quality

Discrepancy
(adol.


fa.)

Discrepancy
|adol.


fa.|

Adolescents’
assess
ment

Fathers’
a
s
sessment

Adolescents

.39**

-
.14

-
.18

.25


.39**

Mothers

.35*

-
.37**

-
.35*

.08

.49***

Fathers

.19

-
.36*

-
.34*

-
.05

.35*



p

< .10, *
p

< .05, **
p

< .01, ***
p

< .001

For the mother
-
child dyad, the two discrepancy measures were different
ly related to ad
o
-
lescents’ self
-
esteem: While the difference between adolescents’ and mothers’ relationship a
s
-
sessments was negatively related to self
-
esteem as assessed by mothers, the absolute difference
was not. That is, the size of the discrepancy h
ad no effect, but the more the adolescents ex
-
ceeded their mothers’ relationship assessments, the lower was their self
-
esteem.


4

For the father
-
child dyad, both discrepancy measures were negatively related to both par
-
ents’ self
-
esteem assessments. The larg
er the absolute discrepancy between fathers and adoles
-
cents and the more the adolescents exceeded their fathers’ relationship assessments, the lower
were the parental self
-
esteem reports.

4 First discussion: High levels are beneficial, high discrepancies
are dysfunctional

Given these results from the first wave of data collections, their interpretation seems quite obvi
-
ous. First, as expected, the level of dyad functioning is related to the individual’s emotional
well
-
being. Second, discrepancies may be
interpreted as divergent needs and wishes in the rela
-
tionship or as a lack of mutual communication. These disturbances of the dyads were negatively
related to the adolescents’ well
-
being. However, the direction of the discrepancies turned out to
be infl
uential. For the mother
-
adolescent dyad, it was not the absolute size of the discrepancy.
But the degree to which adolescents exceeded their mothers in relationship satisfaction was
nega
tively related to adolescents’ self
-
esteem. One may speculate that

a dyad that fulfils the
wishes of the adolescent only, may fail to reach an age
-
appropriate balance of wishes and thus
may be dysfunctional for the adolescent developmental tasks of defining a positive self. Inter
-
estingly, mothers seemed to be more sens
itive to this outcome than the adolescents themselves.

For fathers, the same effect occurred and in addition also a correlation between the abso
-
lute degree of discrepancy and low adolescents’ self
-
esteem. These two effects seem somewhat
contradictory b
ecause if the difference between fathers’ and adolescents’ relationship assess
-
ments is negative, the two correlation coefficients lead to opposite self
-
esteem estimates. Actu
-
ally, subsequent analyses showed that the negative correlation between absolute

relationship dis
-
crepancy and self
-
esteem was only due to those cases where the difference between adolescents
and fathers was positive and thus the two discrepancy measures were equivalent (
r
s for discrep
-
ancy/maternal self
-
esteem rating:
-
.56**/.15 and
for discrepancy/paternal self
-
esteem rating:
-
.62***/
-
.04 for adolescent


father


0 and adolescent


father < 0, resp.). Thus, for both par
-
ents, not the absolute degree of discrepancy but the extent of adolescents exceeding their parents’
relationship
assessments was negatively related to adolescents’ levels of self
-
esteem.

Given this way of analysing the data on a dyad level, I preferred a dyadic interpretation:

If the dyad complied to the adolescents’ wishes


making her or him satisfied


at the exp
ense of
parental wishes


making them less satisfied

, the dyad failed to achieve an age
-
appropriate
symmetry and thus was assumed to be dysfunctional for the devel
opmental task of defining a
positive self. However, I want to challenge this interpretati
on. First, although the analysis was
done on a dyad level, I want to add an individual interpretation: Maybe, if adolescents value their
relationships to their parents highly and higher than their par
ents do, this may indicate an ideali
-
sation of the par
ents that may be accompanied by a bad self
-
image as in Bartholomew’s preoc
-
cupied attachment style (Bartholomew & Horowitz, 1991). Second, as we will see, a data analy
-
sis on an individual level provides still another and as I think more compelling interp
retation.

5 A closer look: Data analysis on an individual level

Tables 1 and 2 also provide the correlations between adolescents’ and mothers’ or fathers’, re
-
spectively, relationship assessments and adolescents’ self
-
esteem. It can be seen that adoles
-
ce
nts’ satisfactions with the relationships towards their mothers and tentatively towards their fa
-
thers were positively related to self
-
esteem as assessed by the adolescents themselves. In addi
-
tion, adolescents’ satisfaction with the mother relationship w
as also positively correlated to fa
-
thers’ self
-
esteem ratings. However, the paternal relationship assessments were by far more

5

closely related to all self
-
esteem ratings, and all these correlation coefficients were statistically
significant (
p

< .05) or
even highly (
p

< .001) significant. The more satisfied the parents were in
their relationships with their adolescent children, the higher they rated their self
-
esteem. The
adolescents’ relationship assessments were in most cases only related to their own

self
-
esteem
ratings but not their parents’ ones.

6 Discussion of the results of the individual
-
level analysis

If one compares the columns in Tables 1 and 2, it is obvious that the effects of the discrepancy
measures mentioned in Section 3 can be explain
ed by the parental relationship assessments
alone. In fact, as the correlations with the parental data were generally higher than those with the
dyad measures, it can be concluded that the dyad measures combined relevant parental and less
relevant adolesc
ent data. Thus, interpretations based on this mixture of relevant and irrelevant
information are misleading. In fact, as we now see, the first two interpretations given in Section
4 were wrong because both implied a negative relationship between adolesce
nts’ relationship
assessments and self
-
esteem. But in contrary, adolescents were the more satisfied with them
-
selves the higher they and their parents assessed the relationship quality.

The parental self
-
esteem ratings were correlated only to the parent
al relationship satisfac
-
tion. Obviously, parents who were satisfied with their children assumed the same satisfaction in
them. So, this result says more about consistency in paternal cognitions than about adolescents’
self
-
esteem. But still, also the a
dolescents’ own self
-
esteem ratings were correlated with the pa
-
rental relationship satisfaction. This is not due to shared variance in relationship measures. The
partial correlations of self
-
assessed adolescent self
-
esteem and maternal and paternal rela
tionship
satisfaction controlling for the adolescents’ satisfaction with both parents were .33 each for
mothers and fathers (
p

< .05). Thus, independent of the own relationship satisfaction, the ado
-
lescents felt better with themselves when the parents al
so were satisfied with them. As was sug
-
gested 100 years ago by Cooley (1983), the experience of the own image in others may influence
the self
-
image.

So, what lesson can be learned from this? Aggregating individual data into dyad compos
-
ites or, more gen
erally spoken, system composites may result in misinformation. This is because
systems such as families or parent
-
child dyads are not static but dynamic. These dynamics are
lost when aggregating data over persons. As Malloy and Kenny (1986, p. 207, emph
asis added)
put it, the aggregation of data such as averaging over persons “allows analyses only at the dyad
or group level.
The study of

individual level processes and
social interaction processes

(e.g.,
reciprocity and compensation)
is necessarily precl
uded
.” And, as Stern (1994) suggested, the
acting entity is the individual although it may be influenced by others. Therefore, the analysis of
systems requires the inclusion of data analyses at the individual level.

The Social Relations Model (SRM, Mall
oy & Kenny, 1986) claims to include data on in
-
dividual as well as group levels and to be well
-
suited for social interaction processes between
persons. It employs relational data such as behaviours from one person to another or


as in my
paper


relation
ship assessments from one person with regard to another one. Each person is
actor as well as object. For example, the mother assesses the relationships to the other family
members and also the relationships to her are assessed by the others. This so
-
cal
led round robin
design by and by seems to become ubiquitous in family research. The Social Relations Model
estimates the dependent behaviour variable
X
ijk

as dependent from the grand mean
m
k

plus ef
-
fects of actors

a
i
, relationship partners

b
j
, actor
-
part
ner interaction

g
ij
, and measurement error

e
ijk
:

X
ijk

=
m
k

+
a
i

+
b
j

+
g
ij

+
e
ijk

(Malloy & Kenny, 1986, p. 208)


6


Although reciprocal data are analysed in the model, reciprocity is obviously lost in these
computations. The interaction term denotes a stati
stical interaction as in an ANOVA. The So
-
cial Relations Model sees behaviour as determined from some combination of person and (so
-
cial) environment, but the relationship is nonreciprocal as is obvious from the model equation.
Thus, according to the dif
ferentiation by Buss (1977), this “interaction” is of a different type than
the reciprocal or bi
-
directional interaction in a systems perspective. In fact, what is called “inter
-
action” in the Social Relations Model is equivalent to Mischel’s and Shoda’s
(1995) intra
-
indi
-
vidual patterns of behaviour variability which they interpret as signatures of personality. The
inadequacy of the Social Relations Model becomes fully apparent when one considers two as
-
sumptions of the model: First, the interaction part
ners are seen as random factors, second, the
effects are estimated under the assumption of an infinite number of interaction partners. Even in
modern times, the assumption of an infinite number of random fathers does not seem plausible…

However, other app
lications of the model have included bi
-
directional interactions. E. g.,
Stevenson, Leavitt, Thompson, and Roach (1988) calculated so
-
called dyadic reciprocity coeffi
-
cients that seem to be the mean covariance of
g
ij

with
g
ji

that is the covariance of the

unique ad
-
justments between each pair of two subjects. But from a social learning perspective, also the
actor effect
a
i

may reflect results of experiences in interactions with family members. The actor
effect then includes those aspects of (learned) beh
aviour that are exerted on all family members.
Thus, the statement that the actor effect is “a person’s
natural tendency

to behave in a particular
way” (Stevenson et al., 1988, p. 101, emphasis added) may fit to ideas of the 18th century (e. g.,
Rousseau)

but not to contemporary psychological approaches.

Thus, if one would like to see what is going on in a family, to me it seems necessary to
investi
gate the individuals and their interplay. Only if there are data from different perspectives
at the individ
ual level, such as the parent and adolescent data in the present study, then there is a
chance to find dynamics between family members. Often, dynamics become clearer in a longi
-
tudinal study. Thus, finally I would like to present two sets of longitudina
l findings, first, how
the corre
lations between adolescents’ self
-
assessed self
-
esteem and family variables developed
during the school
-
to
-
work transition, and second, in how far the parents were able to assess their
children’s self
-
esteem.

7 Longitudinal

findings

As Table 3 shows, the correlations between adolescents’ self
-
esteem and parent
-
adolescent rela
-
tionship satisfaction changed over time. For apprentices, only at time 2 when the apprenticeship
began, their satisfaction with their fathers was stro
ngly related to their self
-
esteem level. This
peak was significantly different from the insignificant correlations at times 1 and 3. All other
correlations between dyadic relationship measures and self
-
esteem became considerably smaller
after time 1 so t
hat prior significant relationships became insignificant. For mothers’ assessment
of the relationship quality, this change approached significance.

For students, the correlations between their own relationship assessments and self
-
esteem
were stable over
time or even showed a tentative increase so that their satisfaction with mothers
at time 3 was significantly and significantly more strongly related to self
-
esteem than the satis
-
faction of apprentices with their mothers. Second, like for apprentices, als
o for students, the cor
-
relations between parental relationship estimates and self
-
esteem dropped to insignificance by
the start of the new school. This drop reached significance for the paternal relationship assess
-
ments. However, unlike the apprentices
, the correlations recovered from this drop by time 3.
That is, the loss of significance of the parental relationship assessments was only temporary.


7

Table 3:

Correlations between adolescents’ self
-
assessed self
-
esteem and parent
-
adolescent relationship
m
easures before leaving middle
-
school, when starting an apprenticeship/changing schools, and
three months later.

Relationship
Satisfaction

Apprentices

Students

Time 1

Time 2

Time 3

Time 1

Time 2

Time 3

Adol.


Mother

.41


-
.01

-
.16
a

.19

.23

.56
a
*

Mother



Adol.
d

.50*

-
.17

-
.02
b

.34


.02

.47
b


Adol.


Father
e

.21

.70***

-
.07

.26

.32

.40

Father


Adol.

.34

.03

.11

.43
c
*

.01
c

.35

Note
. Numbers with subscripts differ significantly.
a:
z

= 2.08,
p

< .05.
b:
z

= 1.42,
p

< .10, c:
z

= 1.36,
p

< .10
(each
probability one
-
tailed).

d

The small sample size allows longitudinal comparisons of correlations of dependent data for students between time
1 and time 2 only. Thus, less powerful comparisons for independent data were employed for apprentices and for
time
s 1 to 3. For apprentices, correlations of mother
-
adolescent relationsh
ip quality with self
-
esteem changed over
time (χ
2
(2) = 5.00,
p

< .10).

e

Change of correlations over time for apprentices: χ
2
(2) = 7.38,
p

< .05



p

< .10, *
p

< .05, ***
p

< .001

Table 4 shows the concordance of the parental ratings with adolescents’ own
self
-
esteem
ratings. The mothers of both apprentices and students were by far less concordant in their esti
-
mates after the end of middle
-
school. Because of the larger size of the student subsample and
because of a higher correlation at time 1, this chang
e reached signifi
cance for students only.
Second, the fathers of both apprentices and students were more concordant
in
their estimates
after time 1 so that the father
-
student correlation approached significance at times 2 and 3.

Table 4:

Correlations bet
ween adolescents’ self
-
assessed self
-
esteem and parental self
-
esteem
assessments before leaving middle
-
school, when starting an apprenticeship/changing schools,
and three months later.

Parent

Apprentices

Students

Time 1

Time 2

Time 3

Time 1

Time 2

Time 3

Mother

.34

-
.09

.16

.50
a,b
**

.06
a

.05
b

Father

.17

.30

.34

.29

.39


.41


Note
. Numbers with subscripts differ significantly (in both instances
z

= 1.86, one
-
tailed
p

< .05).



p

< .10, **
p

< .01

8 Discussion of the longitudinal findings

Before finish
ing middle
-
school, adolescents’ self
-
assessed self
-
esteem correlated highly with
their parents’ satisfaction with them. Obviously, both later apprentices and students esteem
them
selves as they experience their parents evaluating them. Apart from an insi
gnificant tempo
-
rary disturbance at time 2, as expected, the correlations for students were stable over time. How
-
ever, for apprentices, parents lost their significance for their self
-
esteem. This confirms the as
-

8

sumption that the status transition from
school to work reduces the salience of parents for self
-
esteem (Roberts & Bengtson, 1993). This result is an extension of Youniss’s and Smollar’s (e.
g., 1985; Youniss & Ketterlinus, 1987) individuation approach. Youniss and Smollar had not yet
assumed a

decline in the parental importance for the adolescents’ identity formation.

An additional result was the peak of paternal relevance for the apprentices’ self
-
esteem
just after beginning the apprenticeship. Presumably, fathers are in the role of experts

for this
transition. If one can communicate well with one’s father at this time and can profit out of his
expertise, then one can feel well
-
prepared for this new life
-
stage.

It is interesting that after graduation from middle
-
school, mothers were hardly
able to as
-
sess their offspring’s self
-
esteem. This may be the result of a decline in communication fre
-
quency or in self
-
disclosure. Building own private “territories” has been described as an aspect
of separation during adolescence (Youniss & Smollar,
1985). However, although the raise of the
correlation coefficients was not significant, fathers seemed to take over some part of the maternal
closeness as they were better able than before to assess the adolescents’ levels of self
-
esteem.

In a nutshell, a
s Malloy and Kenny (1986) have mentioned before, aggregating individual
data in means or discrepancy scores proved to be inadequate for the analysis of dyadic system
processes. On the contrary, by analyses on the individual level, adolescents’ “looking
-
gl
ass self”
became transparent: Influences of parental (and not averaged or discrepant dyadic) satisfaction
was important for adolescents’ self
-
satisfaction. Furthermore, it could be shown how the family
dynamics were transformed during school
-
to
-
work trans
ition. The special importance of fathers
during this period of time was revealed. Insights such as those can only be obtained by employ
-
ing a longitudinal approach and by investigating the relationships from the viewpoints of adoles
-
cents, mothers, and fa
thers.

References

Bartholomew, K. & Horowitz, L. M. (1991). Attachment styles among young adults: A test of a four
-
category model.
Journal of Personality and Social Psychology
,
61
, 226

244.

Buss, A. R. (1977).
The trait
-
situation controversy and the conce
pt of interaction.
Personality and Social
Psychology Bulletin
,
3
, 196

201.

Cooley, C. H. (1983).
Human nature and the social order
. New Brunswick, NJ: Transaction Publ. (Re
-
print of the 1964 ed. First published in 1902)

Fricke, T., Plagge, S., Hahlweg, K.,

Thurmaier, F., Engl, J. & Eckert, V. (1993).
Freie Partnerbeurtei
lung:
Ein neuer Weg in der Diagnostik von Paarbeziehungen

[Open partner assessment: A new way in
the diagnostic of couple relationships].
System Familie
,
6
, 171

180.

Kreppner, K. (1995). D
ifferential experiences within the family during adolescence: Consistencies of re
-
lationship assessments and concrete communication behaviors over time. In J. J. Hox, B. F. van
der Meulen, J. M. A. M. Janssens, J. J. F. ter Laak, & L. W. C. Taveccio (Eds.)
,
Advances in fam
-
ily re
search

(pp. 103

122). Amsterdam: Thesis Publishers.

Malloy, T. E. & Kenny, D. A. (1986). The Social Relations Model: An integrative method for personality
research.
Journal of Personality
,
54
, 199

225.

Masche, J. G. (2000).
Does a
happy marriage make positive parent
-
adolescent relationships
and self
-
satisfied children?

Poster paper presented at the Eighth SRA Biennial Meeting, Chicago, IL,
USA.
Published in

Resources in Education
, ERIC document number ED 443 576.

Masche, J. G. (subm
itted).
Abgrenzung einmal anders: Eltern
-
Kind
-
Beziehungen und Selbstkonzept Ju
-
gendlicher nach Realschulabschluß

[Another kind of separation: Parent
-
child relationships and
adolescent self
-
esteem after finishing middle school]. In J. G. Masche & S. Walper
(Eds.),
Facet
-
ten der Individuation: Was hilft, was hilft nicht?

[special issue].
Zeitschrift für Familienfor
-
schung
.


9

Masche, J. G. & Barber, B. K. (2001).
Connectedness and separation in parent
-
adolescent rela
tionships:
Indicators of a successful identit
y development?

Poster paper presented at the Bien
nial Meeting of
the Society for Research in Child Development, Minneapolis, MN, USA. Pub
lished in
Resources
in Education
, ERIC document number ED 451 962.

Mischel, W. & Shoda, Y. (1995).
A cognitive
-
affect
ive system theory of personality: Reconceptualizing
situations, dispositions, dynamics, and invariance in personality structure.
Psychological Review
,
102
, 246

268.

Roberts, R. E. L. & Bengtson, V. L. (1993). Relationship with parents, self
-
esteem, and psy
chological
well
-
being in young adulthood.
Social Psychology Quarterly
,
56
, 263

277.

Skinner, H. A., Steinhauer, P. D., & Santa
-
Barbara, J. (1983). The Family Assessment Measure.
Ca
na
-
dian Journal of Community Mental Health, 2,
91

105.

Stern, W. (1994).
Di
e Differentielle Psychologie in ihren methodischen Grundlagen

[The Differential
Psychology in its methodical foundations] (reprint of the 2nd ed., 1911). Berne, Switzerland:
Hans Huber.

Stevenson, M. B., Leavitt, L. A., Thompson, R. H., & Roach, M. A. (198
8). A Social Relations Model
analysis of parent and child play.
Developmental Psychology
,
24
, 101

108.

Youniss, J. & Ketterlinus, R. D. (1987). Communication and connectedness in mother
-

and father
-
adoles
-
cent relationships.
Journal of Youth and Adolescenc
e
,
16
, 265

280.

Youniss, J. & Smollar, J. (1985).
Adolescent relations with mothers, fathers, and friends
. Chicago: Uni
-
versity of Chicago Press.

Youniss, J. & Smollar, J. (1989). Adolescents’ interpersonal relationships in social context. In T. J.
Berndt

& G. W. Ladd (Eds.),
Peer relationships in child development

(pp. 300

316). New York:
John Wiley & Sons.


I thank PD Dr. Kurt Kreppner and Prof. Dr. Hans
-
Georg W. Voß for helpful comments
on an earlier version of this manu
script.

Please address correspon
dence to: Dr. J. Gowert Masche, Darmstadt University of Tech
-
nol
ogy, Institute of Psychology, Steubenplatz 12, D
-
64293 Darmstadt, Germany. E
-
mail:
masche@psychologie.tu
-
darmstadt.de. Internet: http://www.tu
-
darmstadt.de/fb/fb3/psy/diffpsy