Вы находитесь на странице: 1из 18

Associations between Parental Psychological

and Behavioral Control and Youth Internalized


and Externalized Behaviors

Brian K. Barber and Joseph E. Olsen


Brigham Young University

Shohha C. Shagle
Northwestern University

BARBER, BWAN K.; OLSEN, JOSEPH E.; and SHAGLE, SHOBHA C . Associations between Parental
Psychological and Behavioral Control and Youth Internalized and Externalized Behaviors.
CHILD DEVELOPMENT, 1994, 65, 1120-1136. In this study, psychological control of children was
conceptually and empirically distinguished trom behavioral control. Further, it was demon-
strated as hypothesized that psychological control was more predictive of adolescent internalized
problems, and that behavioral control was more predictive of extemalized problems. Subjects
were 473 fifth-, eighth-, and tenth-grade males and females from a Southem suburb. Control was
measured by the Child Report of Parent Behavior Inventory and the Colorado Self-Report of
Family Functioning Inventory. Problem behaviors were measured with the Child Behavior
Checklist. First- and second-order factor analyses discriminated psychological and behavioral
control, and structural equation analyses demonstrated the differential prediction of internalized
and externalized problems. These last analyses were conducted using youth-reported data and
validated using a subsample of 227 mother-youth pairs.

Types and patterns of parental control- the construct. It is therefore not surprising
ling behavior and their effects on children's that the empirical evidence for the effects of
development have been the focus of con- parental control on children has often been
siderable research over the past several inconsistent or equivocal (Barber, 1992;
decades. Although specific labels vary, Rollins & Thomas, 1979). There have been
researchers have consistently identified some attempts to integrate and synthesize
control in their conceptualizations of salient the various conceptualizations of control in
parental behaviors. Some examples include order to bring clarity to the construct and
restrictiveness (Baldwin, 1955; Baumrind & its effects on children's development. These
Black, 1967; Becker, 1964); firm control and include Rollins and Thomas's (1979) differ-
psychological control (Burger & Armentrout, entiation between coercive, inductive, and
1971; Schaefer, 1965); demandingness (Roe unspecified control attempts, and Maccoby
& Siegelman, 1963; Siegelman, 1965); au- and Martin's (1983) twofold classification
thoritative, authoritarian, and permissive scheme of responsiveness and demanding-
control (Baumrind, 1967, 1971); directive/ ness that built heavily on Baumrind's work,
conventional, assertive, and supportive con- .^^^ ^^^^ ^^^^^^ .^ ^.^ ^ ^ j ^ ^^.
trol Baumnnd, 1991a, 1991b); overprotec- ^^^.^^^ ^^^ approach to
tion (Parker, Tuphng, & Brown, 1979); etc. -r • i l i_ l J -^ a: i. u
/c r> II- s T-u nn-rm r i ..• specifymg parental control and its effects by
(See Rollins & Thomas [19791 for a listing ^^^^^.^ ^^^^ difference between psycho-
"^''°"' conceptualizations of logical and behavioral control. The evolu-
tion of research on parental control has not
The several ways in which parental con- included careful attention to the locus of pa-
trol of children has been conceptualized and rental control attempts, that is, whether the
operationalized testifies to the complexity of particular parental behavior tries to exercise
The data for this project were collected in collaboration with the Section on Social and
Emotional Development, National Institute of Child Health and Human Development. We thank
Samuel E. Bratton, Jr., Coordinator, Research and Evaluation, Knox County Department of Pub-
lic Instruction; the numerous principals and teachers; and the many families from Knoxville,
Tennessee, for participating in the study. Reprint requests should be sent to Brian K. Barber,
Brigham Young University, Center for Studies ofthe Family, 930 SWKT, Provo, UT 84602.
[Child Development, 1994,65,1120-1136. © 1994 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/94/6504-0002$01.00]
Barber, Olsen, and Shagle 1121
control over the child's psychological world logical control is defined as patterns of fam-
or is an attempt to regulate the child's be- ily interaction that intrude upon or impede
havior. This distinction between parental the child's individuation process, or the rela-
psychological and behavioral control has tive degree of psychological distance a child
been implicit in some work (e.g., Schaefer's experiences from his or her parents and fam-
[1965] psychological control vs. firm con- ily (Sabatelli & Mazor, 1985). Individuation
trol). Also, Steinberg (1990) has recently em- is a common term in developmental theories
phasized the importance of the distinction of personality, particularly for adolescents
and noted that findings from the child devel- (e.g.. Bios, 1979), and it has been viewed as
opment literature can be interpreted to sup- an important correlate of identity formation,
port it. However, systematic attempts to the- a central task of adolescence (Cooper, Gro-
oretically justify the distinction and track tevant, & Condon, 1982, 1983; Erickson,
the effects of the two types of control are 1968). Despite the fact that it is parents and
lacking. family from which adolescents are con-
ceived to separate or individuate, the role of
Two fundamental presuppositions family interaction in the individuation pro-
about human development justify the dis- cess has received relatively little theoretical
tinction between psychological control and or empirical attention (Sabatelli & Mazor,
behavioral control. They are that developing 1985). These authors note the conceptual
children require; (1) an adequate degree of similarity between individuation as a per-
psychological autonomy, that is, that they sonal process and differentiation as a family
learn through their social interactions that systems process, and suggest that the two are
they are effective, competent individuals integrally connected. In poorly differenti-
with a clear sense of personal identity; and ated families, distances are regulated to ob-
(2) sufficient regulation of behavior to en- struct psychological separation and au-
able them to learn that social interaction is tonomy experienced by family members,
governed by rules and structures that must whereas in well-differentiated families, in-
be recognized and adhered to in order to be dividual members experience adequate con-
a competent member of society. nectedness while at the same time main-
taining their individuality.
Developmental theories and empirical
findings converge in highlighting autonomy Researchers using clinical samples have
and regulation as particularly critical issues observed that some parents in disturbed
in development. Autonomy has been a cen- families resist their adolescents' differentia-
tral construct in classic theories of develop- tion through binding or constraining interac-
ment (e.g., Erikson, 1968), with concepts of tions that interfere with the autonomous
separation (A. Freud, 1958), individuation functioning of the adolescents (Hauser,
(Bios, 1979), and identity (Erikson, 1968; 1991; Hauser et al., 1984; Stierlin, 1974).
Marcia, 1980) assuming principal impor- Similarly, children who consistently per-
tance during adolescence. At the same time, ceive that their parents are always trying to
it is evident that healthy development re- change them, or who experience parental
quires adequate structure. Social learning manipulative behavior that threatens a dis-
theory has long identified the importance of ruption or discontinuance of the emotional
discipline in socializing dependency and bond between parent and child (e.g., love
conformity (Bandura, 1964; Bandura & Wal- withdrawal), will likely have difficulty rec-
ters, 1959). Social psychological theories, ognizing their own uniqueness or adequacy
such as Lewin's (1935) field theory, have or will be unwilling to trust their own ideas
also argued for structure and guidance to en- or individuality for fear of losing key connec-
hance favorable personality development. tions with caregivers.
As Maccoby and Martin (1983) note, chil-
dren must learn to both inhibit disruptive In contrast to psychological control,
behavior and to engage in socially approved where the presence of control poses risk for
behavior, and adult caregivers are looked to adolescent development, we focus on the
to provide the stmcture and the predictable absence of sufficient behavioral control as
contingencies for self-regulatory mecha- defining risk. For the purposes of this study,
nisms to develop. Particularly during adoles- we define behavioral control as family inter-
cence, parental management practices are action that is disengaged and provides insuf-
critical in fostering socially competent be- ficient parental regulation of the child's be-
havior among youth (Patterson, 1982). havior, as in excessive behavioral autonomy,
lack of rules and restrictions, and/or lack of
For the purposes of this study, psycho- knowledge of a child's day-to-day behavior.
1122 Child Development
As was the case with psychological control, in rebellion against excessive behavioral re-
dyadic interactions indicative of insufficient strictions, our concern is with insufficient
behavioral control refiect more general fam- regulation (lack of demandingness, in Mac-
ily system properties. Families in which par- coby & Martin's terms), because ofthe theo-
ents are permissive and employ inadequate retical points described above, and because
management strategies appear to be charac- of the weight of empirical evidence docu-
terized at the system level by relationships menting an association between low behav-
that are disengaged (i.e., low levels of emo- ioral control and extemalized problems,
tional bonding between family members such as impulsivity, aggression, delin-
[Olson et al., 1983; Patterson 1976, 1982; quency, drug use, and sexual precocity
Pulkinnen, 1982]). (Baumrind, 1971, 1991a; Dishion & Loeber,
1985; Dombusch et al., 1985; Loeber & Dis-
The relevance of distinguishing be- hion, 1984; Maccoby & Martin, 1983;
tween psychological and behavioral control McCord, 1979, 1990; Miller, McCoy, Olson,
extends beyond just a conceptual clarifica- & Wallace, 1986; Olweus, 1980; Patterson,
tion. A fundamental purpose of this study is Capaldi, & Bank, 1989; Patterson & Stout-
to demonstrate also that deficiencies in these hamer-Loeber, 1985; Pulkinnen, 1982; Volk,
two areas of socialization lead to different Edwards, Lewis, & Sprenkle, 1989). The
outcomes in children. Specifically, we pro- risk for the development of extemalized
pose that children experiencing psycho- problems among adolescents from femilies
logical control are at risk for internalized exhibiting inadequate behavioral control ap-
problems, whereas children experiencing pears to be substantial. Theoretical explana-
inadequate behavioral regulation are at tions for this association include the failure
greater risk for the development of external- of such an environment to socialize self-
ized problems. The link between psycholog- regulation, conformity, and/or dependency
ically controlling environments and inter- favorable to personality development. Poor
nalized chairacteristics in children can be supervision and monitoring also appear to
understood through the crippling effect such encourage association with deviant peers
control has on the development of psycho- and expose adolescents to risks and tempta-
logical autonomy. Such parenting is unre- tions without resources to resist. It is also
sponsive to the psychological needs of chil- possible that adolescents act out in order to
dren (Maccoby & Martin, 1983), who are determine the limits of acceptable behavior
likely to exhibit passive, inhibited, and over- or to find meaningful relationships lacking
controlled characteristics (Beavers, 1982). in their disengaged families.
Children do so partly because they fail to
learn to be aware of or to express personal
initiative or self-reliance (White, 1989), and Three final points are important. The
partly because they have learned that they first is that because one of the control con-
have little power or control in their interac- structs we are positing represents excessive
tions with others, particularly parents (e.g., control (psychological control) and the other
learned helplessness [Seligman & Peterson, insufficient control (behavioral control), it is
1986]). Children who are thus unequipped tempting to simply consider them at oppo-
with awareness or confidence in their own site ends of a single control continuum. Our
worth and identity or who have learned that interest, however, is in demonstrating that
the expression of psychological autonomy is the domain over which control is exercised,
fruitless or unacceptable are likely to turn not the magnitude of control, is pivotal in
inward or withdraw to fend for themselves specifying the effects of control. We view
psychologically as they encounter the control over psychological processes and be-
stresses and pressures of social interaction. havior to be different phenomena that have
These predictions are consistent with Baum- contrasting effects on children. Psychologi-
rind's (1991b) correlational findings that pa- cal control and behavioral control, as we de-
rental officious behavior (i.e., intrusive, fine them, are linear constructs ranging firom
meddling, overcontrolling) is associated sig- zero-points to increasingly higher levels of
nificantly with young adolescents' inter- psychological and behavioral control. The
nalizing problems. constructs are conceptually orthogonal in
that the level of one is not necessarily a func-
Inadequate behavioral control, on the tion ofthe level ofthe other. The association
other hand, appears to be associated with the between psychological and behavioral con-
development of externalized problems in trol may vary substantially between families.
adolescents. Although it is also possible that There may be families that impose high lev-
adolescents engage in delinquent behavior els of both psychological and behavioral
Barber, Olsen, and Shagle 1123
control, but there may also be families emotionally and psychologically from par-
whose intrusion into the psychological au- ents may not be able to distinguish control
tonomy of the child may not extend to strict behaviors that intmde on their psychological
rules and regulations on behavior, and who autonomy from more behaviorally oriented
are, in fact, disengaged or quite permissive control. Also, the increased freedom and as-
behaviorally. sociation with people outside ofthe family
that occurs at adolescence makes supervi-
The second point relates to the role of sion and monitoring particularly salient at
supportive relationships within the family. this time period. Thus, it may not be until
Historically, as often as control has been children begin to make these moves toward
identified as an important socialization be- psychological autonomy and more behav-
havior, so has a dimension of nurturant rela- ioral freedom that the distinctions between
tionships between parents and children psychological and behavioral control and
(Maccoby & Martin, 1983; Rollins & their effects become evident. Because our
Thomas, 1979). Family or parental support data set includes some preadolescents (fifth
has not occupied a central role in our theo- grade), we will be able to test the viability
rizing in this article because our purpose is of the model at this age relative to the ado-
to specify meaningful dimensions of control. lescent subjects (eighth and tenth grade).
Also, because of its consistent association
with both internalized and externalized be-
haviors (Maccoby & Martin, 1983; Rohner, Method
1986; Rollins & Thomas, 1979), parental Sample.—Data for this study come from
supportive behavior does not appear to be the Tennessee Adolescents in Families Proj-
useful in discriminating types of youth prob- ect (TAIFS), a school-based survey study of
lems, another purpose of the present study. pre-, early-, and middle-adolescent students
Although efforts to document an interaction from the Knox County, Tennessee, school
between support and control in predicting system. The data were collected in the
child outcomes have generally not been suc- Spring of 1990. Extensive surveys were ad-
cessful (Rollins & Thomas, 1979; Thomas, ministered to participating students in class-
Cecas, Weigert, & Rooney, 1974), we room or larger groups. The assessment took
thought it important to assure that the ob- approximately IV2 hours, after which stu-
tained effects between our dimensions of dents were treated to a pizza party. Students
control and adolescent difficulty are not in- carried home assessment packets to parents
fiuenced by any associations between con- containing survey instruments on the family
trol and support, especially because some environment and child behavior and person-
theorists have conceptualized enmeshed ality. These took approximately 45 min to
family relationships (high psychological complete. The packet included instruments
control) to be a component of cohesive or for both mother and father with separate
supporting family relationships (e.g., Olson business reply envelopes for their return di-
et al., 1983). Therefore, our analyses will in- rectly to the principal investigator. A check
clude a test of the model that incorporates a for $5 was sent to each parent who mailed
measure of parental supportive behavior. back a completed questionnaire. Funds
were not available to follow-up nonrespon-
The third point relates to the develop- dents. Approximately 50% of mothers and
mental implications of the model. There is 25% of fathers responded.
ample theory to defend the importance of
autonomy and regulation as critical elements Findings reported in this article come
of development at virtually any phase of the from analyses of the suburban subsample,
life course from childhood through adult- the largest subset of the data. The entire en-
hood. However, it is likely that the proposed rollment ofthe fifth, eighth, and tenth grades
pattems of control and problem behaviors of a typical suburban community served as
will be most evident during adolescence. the sampling frame. Students carried home
This is so because of the substantial devel- letters describing the study and requesting
opmental changes that occur during adoles- signed parental consent for their children's
cence that relate directly to the psychologi- participation. An average of 65% of the par-
cal and behavioral processes at issue in the ents across the three grades consented. All
model. The individuation process, though children who received parental consent par-
evident throughout the life course, has ac- ticipated in the project, resulting in a sample
celerated prominence during adolescence of 524 students. Because this sample had so
(Josselson, 1980; Sabatelli & Mazor, 1985). few non-white and low-income families, we
Children who have not yet begun to separate have limited these initial analyses to a sub-
1124 Child Development
sample of white and middle-income families nalizing items depict specific behaviors
(defined as students receiving no lunch sub- related to misconduct.
sidy). The resulting sample size was 473, Because of the theoretical convergence
made up of 45% males, 55% females, and between the social-psychological and family
broken down by grade as follows: fifth grade systems conceptualizations of family control,
N = 123, eighth grade N = 188, tenth grade our interest was to measure the family con-
N = 162. The average age for the sample structs at both the dyadic (parent-child) and
was 13.7. Sixty-nine percent ofthe students the system levels. Accordingly, we selected
of this subsample reported living with both the latest revision (Schludermann & Schlud-
parents and had fathers (88%) and mothers ermann, personal communication, 1988) of
(56%) who worked full time. Schaefer's (1965) Child Report of Parent Be-
Measures.—Youth problem behaviors havior Inventory (CRPBI) to measure family
were measured with the Child Behavior control at the dyadic level. This is a 30-item
Checklist—Youth Self Report (Achenbach &c version which, like its precedent versions,
Edelbrock, 1987). The main analytical pro- measures three dimensions of parenting,
cedure used to test the theoretical model two of which are control variables: psycho-
was structural equation analysis, which calls logical control versus psychological au-
for the use of observed variables as indica- tonomy and firm versus lax control. The
tors of hypothesized latent constructs. Colorado Self-Report of Family Functioning
Therefore, we did not use scale scores from Inventory (CSRFFI) was employed to mea-
the CBC, but instead employed selected raw sure control at the family level (Bloom,
items measuring internalized and external- 1985). This is a 60-item instrument that re-
ized problem behaviors. The selection of sulted from a series of factor analyses of four
these items was based on a series of factor commonly used family functioning invento-
analyses of the Depressed and Delinquent ries. It includes dimensions relevant to this
subscales of the CBC, conducted separately study such as enmeshment, disengagement,
by sex and grade of respondent. Our intent etc. '
was to derive subsets of items that were ap-
plicable to both males and females across Results
the age range represented by the sample.
The selected items clustered together con- The first analyses focused on the youth-
sistently in all analyses. reported information because these were the
most complete data and large samples sizes
The three items measuring internalized were necessary to test the measurement
problems were: "I feel lonely," "I feel con- model. Next, analyses were conducted using
fused or in a fog," and "I am unhappy, sad, both youth and mother data.
or depressed" (Cronbach's alpha = .78). The The first task was to select items from
three items representing externalized prob- the two family instruments to measure the
lems were: "I use alcohol or drugs for non- control constructs. Both instruments were
medical purposes," "I cut classes or skip examined, and specific items were selected
school," and "I swear or use dirty language" based on face validity in measuring the fam-
(Cronbach's alpha = .64). The reliabilities ily control constructs as they have been the-
for these measures are typical for scales with orized. This process resulted in the a priori
as few as three items. Each of the selected selection of 20 items to measure psychologi-
items have been commonly used in past re- cal control. These included the 10 items
search to measure internalized and external- from the CRPBI psychological control scale,
ized problems (Cicchetti & Toth, 1991), but which indexes control of psychological au-
it should be noted that they represent rela- tonomy and love withdrawal. Ten additional
tively mild levels of problem behaviors. items were selected from various scales of
This results from the constraint that the the CSRFFI describing family properties
items be equally valid for males and fe- that appear to inhibit individuation.
males, and for preadolescents as well as mid-
adolescents. We caution against generaliz- Ten items were selected from the two
ing results to more severe forms of problem instruments to measure behavioral control.
behavior. Of most importance to this study Four of these came from the firm control
was to identify variables that were distin- scale ofthe CRPBI, which indexes excessive
guished in terms of their internalizing and behavioral autonomy. Six were selected
externalizing characteristics. The internaliz- from various scales of the CSRFFI that ap-
ing items describe feelings of isolation, con- peared to measure disengaged family rela-
fusion, and low affect. In contrast, the exter- tionships. In addition, we included five
Barber, Olsen, and Shagle 1125
TABLE 1
ITEMS SELECTED TO MEASURE PSYCHOLOGICAL CONTROL

CSRFFI
YA37 It is difficult for family members to take time away from the family.
YA61 Family members feel guilty if they want to spend some time alone.
YA25 Family members find it hard to get away from each other.
YA49 Family members feel pressured to spend most free time together.
YA69 The children in our family have little influence on anything of real
importance.
YA46 In our family, parents do not check with the children before making
important decisions.
YA21 In our family, we know where all family members are at all times.
YA24 Parents make all the important decisions in our family.
YA36 There is strict punishment for breaking rules in our family.
YA72 It seems like there is never any place to be alone in our house.
(1 = very untrue for my family, 2 = fairly untrue for my family,
3 = fairly true for my family, 4 = very true for my family)

CRPBI
YB56 My mother insists that I must do exactly as I am told.
YB58 My mother is always telling me how I should behave.
YB55 My mother says, if I really cared for her, I would not do things that
cause her to worry.
YB59 My mother is very strict with me.
YB64 My mother wants to control whatever I do.
YB67 My mother is always trying to change me.
YB76 My mother will avoid looking at me when I have disappointed her.
YB79 My mother, if I have hurt her feelings, stops talking to me until I
please her.
YB73 My mother is less fnendly with me if I do not see things her way.
YB61 My mother would like to be able to tell me what to do all the time.
(1 = not like her, 2 = somewhat like her, 3 = a lot like her)

items to measure parental monitoring taken from both psychological and behavioral
from the work of Dombusch and Steinberg items, the other six factors had theoretical
(as in Dombusch, Ritter, Liederman, Rob- integrity in that each was comprised entirely
erts, & Fraleigh, 1987). The two sets of items of either psychological or behavioral control
with corresponding response scales are items. Items measuring psychological con-
listed in Tables 1 and 2. trol loaded on three separate factors: PSY-
DYADl, PSYDYAD2, PSYFAM; items
Two questions were important in test- chosen to measure behavioral control also
ing the theoretical model: (1) Can the theo- loaded on three separate factors: BEHD-
rized concepts of psychological and behav- YAD, MONITOR, BEHFAM. This analysis
ioral control be adequately discriminated as provided initial evidence for the distinction
independent constructs? and (2) Do the con- between psychological control and behav-
cepts distinguish between internalized and ioral control. If these constructs were simply
extemalized problem behaviors? opposite ends of the same continuum, we
Distinguishing psychological and be- would have expected items measuring each
havioral control.—A series of exploratory to load together with opposite valences. This
and confirmatory factor analyses were con- did not occur. These analyses were also con-
ducted to distinguish the two control con- ducted separately on boys and girls, and no
structs. First, factor analyses with oblimin significant differences in factor structure
rotation were conducted on the entire set of were noted.
35 items chosen to measure both constructs.
Four of the items were eventually removed Items loading on the PSYDYADl factor
because they had high secondary loadings. came from the psychological control scale of
Analysis ofthe remaining 31 items revealed the CRPBI. We labeled this factor Authori-
a seven-factor solution explaining 54.5% of tarianlIntrusive because the items appear to
the variance (Table 3). With the exception of represent extreme levels of parental control
the factor labeled MIX, which had loadings that intrude on individual development.
1126 Child Developnient
TABLE 2
ITEMS SELECTED TO MEASURE BEHAVIORAL CONTROL

CSRFFI
YA35 Family members are not punished or reprimanded when they do some-
thing wrong.
YA60 There are very few rules in our family.
YA23 Members of our family can get away with almost anything.
YA45 Each family member does as he or she wishes without concern about
the others.
YA62 Family members seem to avoid contact with each other when at home.
Members of our family generally go their own way.
(1 = very untrue, 2 = fairly untrue for my family; 3 = fairly true for
my family; 4 = very true for my family)

CRPBI
YB71 My mother gives me as much freedom as I want.
YB80 My mother lets me do anything I like to do.
YB77 My mother lets me go out any evening I want.
YB74 My mother lets me go any place I please without asking.
(1 = not like her, 2 = somewhat like her, 3 = a lot like her)

Monitoring
YB172 How much do your parents really know where you go at night?
YB175 How much do your parents really know where you are most after-
noons after school?
YB173 How much do your parents really know how you spend your money?
YB174 How much do your parents really know what you do with your free
time?
YB171 How much do your parents really know who your friends are?
(1 = don't know, 2 = know a little, 3 = know a lot)

This labeling is consistent with Baumrind's degree from the permissiveness in the fam-
(1991a, 1991b) recent work in which she ily in general that is exhibited in the BEH-
conceptualized the presence of intru- FAM factor. This factor was labeled Laissez-
siveness to distinguish importantly between faire to be consistent with the CSRFFI
types of directive parenting. PSYDYAD2 subscale from which two of the items origi-
was also comprised of items from the psy- nated. Both factors were independent of the
chological control scale of the CRPBI and Monitoring factor, which indexes parental
has been labeled Love Withdrawal. Three awareness of children's activities.
of the four items describe parental behavior
that manipulates the affectional bond be- The fact that there were multiple di-
tween parent and child by making parental mensions of each construct is not surprising
acceptance contingent on the child pleasing since the items came from existing instru-
the parent. This type of behavior has high ments with demonstrated reliabilities. Also,
affective intensity and appears to be power- because these instruments have different re-
ful in generating anxiety in children (Mac- sponse patterns, it is possible that the ob-
coby & Martin, 1983). The PSYFAM factor tained factor structure was infiuenced by
contains items from the CSRFFI that repre- method variance. It was therefore necessary
sent Enmeshed family relationships. These to conduct a series of second-order factor
relationships are characterized by over- analyses to determine whether there was an
identification, where loyalty and consensus overarching two-dimensional factor struc-
among family members prevent individua- ture, that is, that the three dimensions of
tion (Minuchin, 1974; Olson et al., 1983). psychological control were actually subdi-
mensions of the more general construct of
The BEHDYAD factor refiects what we psychological control, and the same for be-
have labeled Unrestricted Autonomy Grant- havioral control. Structural equation analysis
ing from parents to adolescents. This ap- (LISREL VII; Joreskog & Sorbom, 1989)
pears to be an extreme form of laxness (items was used to conduct these analyses.
originated in the firm vs. lax control scale
of the CRPBI) that differs perhaps only in Figure 1 depicts the second-order
Barber, Olsen, and Shagle 1127
TABLE 3
FACTOR ANALYSIS (Oblimin Rotation) OF ITEMS SELECTED A PRIORI TO MEASURE
PSYCHOLOGICAL AND BEHAVIORAL CONTROL

Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6 Factor 7


PSYDYADl (Authoritarian/Intmsive):
YB56 J4 .04 .11 .01 -.02 .07 .15
YB58 Jl .05 .02 .03 -.04 -.02 .01
YB55 M --17 -.01 -.10 -.04 -.07 -.10
YB59 33 .08 -.04 .11 .02 -.11 27
YB64 ^ .08 -.14 .05 -.10 -.31 -.13
YB67 M .08 -.18 .07 -.13 -.31 -.05
BEHDYAD (Unrestricted Autonomy Granting);
YB71 05 -M 06 .05 .09 .16 .08
YB80 -.01 -J2 -.04 -.06 -.04 .00 -.12
YB77 01 -71 -.03 -.04 .01 -.12 -.14
YB74 -.11 -M -16 .09 -.09 -.22 -.00
MONITOR (Monitoring):
YB172 02 .07 M .06 -.00 .04 .05
YB175 06 .17 79 .06 .00 .08 -.06
YB173 -.02 -.04 79 .02 .02 -.10 .10
YB174 -.01 -.09 78 -.02 -.01 -.08 -.04
YB171 -.00 .00 .68 -.15 -.06 -.01 -.18
MIX:
YA69 -.03 -.03 .06 .79 -.05 .03 .13
YA46 -.04 .04 .04 -.05 -.10 .07
YA62 -.01 .08 -.11 M -.10 .04 -.02
YA68 16 -.04 -.03 .58 .26 .10 -.32
YA45 02 -.16 -.08 ^ -.01 -.03 -.20
PSYFAM (Enmeshed):
YA37 20 -.13 .08 -.04 -M .29 .11
YA61 -.22 .06 .04 .03 -35 -.27 .04
YA25 10 .06 -.07 .03 -32 .12 -.10
YA49 05 -.01 .03 .15 -=52 -.15 -.11
PSYDAD2 (Love Withdrawal):
YB76 05 -.08 .01 -.06 -.02 -77 -.00
YB79 12 -.22 .07 .07 -.00 -M .21
YB73 17 .14 -.02 .09 .03 -.57 -.08
YB61 43 .13 -.08 .07 .04 -A5 -.13
BEHFAM (Laissez-faire):
YA35 01 .02 .04 -.04 -.06 .05 -Jl
YA60 -.10 -.15 .11 .02 .06 .01 -J6
YA23 -.06 -.11 -.12 .17 -.10 -.08 -50
Eigenvalues 5.20 3.58 2.44 1.88 1.48 1.20 1.15
Cum% Var 16.8 28.3 36.2 42.3 47.0 50.9 54.5
NOTE.—N = 524.

model that best represents the theory. Two the fit of an altemate model with a single
second-order latent constructs, psychologi- second-order factor. This second model did
cal control and behavioral control, are indi- not differentiate between psychological and
cated by each of the three latent constructs behavioral control but hypothesized instead
representing the factors revealed in the ex- a single global dimension of undifferenti-
ploratory factor analysis discussed above. ated control with the seven primary factors
Each of these six variables is, in turn, indi- as subdimensions.
cated by its respective raw items. Since the
MIX factor was made up of items measuring The chi-square for the first model
both general constructs, this variable, with (933.21, df = 425) was significantly lower
its respective raw items, was allowed to indi- than that for the second model (1031.68, df
cate both of the second-order constructs. = 427; x^ difference = 98.47, p < .001), in-
The fit of this model was then compared to dicating that the model positing the exis-
1128 Child Development

FIG. 1.—Second-order factor model of Psychological Control and Behavioral Control

tence of two second-order factors fit the data involvement in externalized problem behav-
better than the model that considers all iors. The theory was tested using structural
seven factors as subdimensions of a general, equation models (LISREL VII; Joreskog &
unspecified control factor. These results pro- Sorbom, 1989) in which the two parental
vided further empirical support for the inde- control variables were used simultaneously
pendence of the psychological control and to predict measures of internalized problems
behavioral control constructs. Further, they and externalized problems.
demonstrated that the dyadic and system
measures are compatible in defining these In order to reduce the number of param-
constructs. It is not often that both levels of eter estimates in the model, and thereby en-
family functioning are addressed in a given hance the reliability ofthe coefficients, only
study, and it is encouraging to see that the a subset of items was used to measure the
theorized constructs can be discerned both family control constructs. The dyadic mea-
in specific interactions between family sure of psychological control (10 items from
members as well as in the systemic proper- the CRPBI) was used to measure psycholog-
ties ofthe family. ical control, and the monitoring variables
were used to measure behavioral control.
Discriminating between problem be- These factors were selected because they
haviors.—Having provided evidence for the appeared to be the strongest measures of
empirical independence of the two control their respective constructs in the second-
variables, the next step was to test the con- order factor analysis. The dependent vari-
struct validity of the variables by relating ables were measured as described above.
them to the outcome variables. A central part
of the theory was that children from psycho- Figure 2 presents the results of this
logically controlling environments would be analysis. The theory was supported with re-
at greater risk for internalized problem be- gard to the role of the parental control vari-
haviors, and that children exposed to behav- ables in discriminating between the depen-
iorally undercontrolling family environ- dent variables. Psychological control was
ments would have a greater likelihood for significantly related to internalized prob-
Barber, Olsen, and Shagle 1129

N=422 •p<.05
-233.31 *'p<.01
p=.003 *"p<.001
AdiGF -.94

FIG. 2.—Family control and adolescent problem behaviors using all youth-reported data

lems (.20) but not to externalized problems has not typically begun. Therefore, the in-
(.01). Monitoring (behavioral control) also trusive and inhibiting effects of psychologi-
had a significant association with internal- cal control may not be apparent at this ear-
ized problems (-.16), but it had a much lier age.
larger relationship with externalized prob-
lems ( — .50), as predicted. The fit of this Replication with multiple data sources.—
model was improved by allowing some of Recently, concern has arisen regarding
the error terms to be correlated for each of structural equation modeling of single-
the control constructs. source data (Bank, Dishion, Skinner, & Pat-
terson, 1990; Lorenz, Conger, Simons, &
The model was also validated separately Whitbeck, 1991), as was done in the analyses
by sex and grade of child. Models were run just described. The concern is essentially
separately for male and female youth, and that when a single source of data is used,
there was no significant difference in the structural parameters become systematically
functioning of the model by sex (x^ differ- infiated because of shared method variance.
ence = 1.19, df = 4). Models were also run This is an important issue and is addressed
separately by grade. There was no differ- in the next analysis, but it should be noted
ence between the eighth- and tenth-grade that the focus of the previous analysis was
models (x^ difference = 3.43, df = 4). The not on the absolute magnitude of path coef-
fifth-grade model did differ from the other ficients, but on the relative strength of the
grades in that the control constructs did not theorized paths. It was of concern to see if
differentiate between internalized and ex- the family control variables distinguished
ternalized problems. This difference will between internalized and externalized prob-
have to be validated on a larger sample (the lems. The analyses showed that both types
fifth-grade sample was the smallest subsam- of family control had significantly different
ple of the data), but to the extent that this predictive power for the two outcomes.
finding is meaningful, it highlights the de-
velopmental implications of the theory. For The next step in the analysis tested the
preadolescents, the individuation process model using multiple sources of informa-
characteristic of the adolescent transition tion. Not enough fathers responded to per-
1130 Child Development
mit the use of both parents' data, so these control, and disengagement as behavioral
analyses were conducted using youth and control. Youth-reported data were used for
mother responses. The use of multiple the outcome variables because it can be ar-
sources of data has important implications gued theoretically that a child is the best in-
for two types of validity. First, it strengthens formant of his or her own feelings and be-
construct validity because constructs are de- havior away from home.
fined by converging measures reported by
different agents. This minimizes the likeli- The results of this analysis (Fig. 3) are
hood that a measured construct merely re- very similar to the previous analysis that
fiects the particular perspective of a single used only youth-reported data. Once again,
informant. In the present case, if the distinc- psychological control was predictive of in-
tion between psychological control and be- ternalized problems (.29) but not external-
havioral control holds when measuring each ized problems (.07), and behavioral control
construct from both youth and mothers, then was more strongly associated with external-
there is more reason for confidence that ized problems ( - .39) than with internalized
these constructs are actually different and problems ( — .23). The major difference be-
represent a shared reality in the family tween the two models was a substantial in-
rather than just the child's viewpoint. This crease in the variance accounted for in inter-
multi-agent modeling also provides a more nalized problems (.08 to .20) when both
accurate picture of predictive validity by youth and mother data were used to measure
reducing the problem of infiated correla- family control. This was largely attributable
tions between constructs due to informant- to a stronger association between psycholog-
specific method variance shared by the con- ical control and internalized problems in the
structs. In this way, more confidence is mother-youth model. There was also a less
gained in both the absolute and relative dramatic decrease in the variance accounted
magnitude ofthe path coefficients. for in externalized problems (.25 to .18), due
to a somewhat weaker relationship with the
For this analysis we selected items from measure of behavioral control used in this
the original a priori list of psychological and model.
behavioral control items that had adequate
correspondence between mother and child With the exception of the first psycho-
reports. We used only two items (each item logical control item, the factor loadings
measured from mother and youth reports) (lambdas) for the control constructs in the
because ofthe smaller sample size available mother-youth model were very similar, indi-
where both mother and youth data are pres- cating that mother and youth data contrib-
ent. Psychological control was measured by uted relatively equally to these constructs.
mother and youth responses to two ques- This implies that the relations obtained
tions from the CRPBI: were not an artifact of who was answering
the questions. It should also be noted that
1. I am a person who is always trying to this analysis allowed the error terms for the
change my child./My mother is a person who is pairs of mother and child indicators to co-
always trying to change me. vary, as suggested by Bank et al. (1990), to
control for informant variance.
2. I am a person who says, if my child really
cared for me, s/he would not do things that cause The nonsigni^cant association between
me to worry./My mother is a person who says, if internalized and externalized problems in
I really cared for her, I would not do things that both analyses is not consistent with prior ev-
cause her to worry.
idence of comorbidity between these types
of problems. Much of the evidence for co-
Behavioral control was measured by mother morbidity, however, has come from studies
and youth responses to two items from the that have utilized clinical samples that could
CSRFFI measuring disengagement: suffer from sampling errors associated with
selecting extreme cases (i.e., those seeking
1. Members of our family generally go their or referred to treatment). Garber, Quiggle,
own way. Panak, and Dodge (1991) used a community
2. Each family member does as he or she sample similar to ours and found substantial
wishes without concern about the others. comorbidity (.42) between depression and
aggression using the CBC. This is not sur-
In essence these items represent shared per- prising, though, given that they used total
ceptions of control of psychological auton- scaie scores from the CBC which are highly
omy and love withdrawal as psychological correlated and which may contain items that
Barber, Olsen, and Shagle 1131
Pressure to Change
(youth) .55
Loneiy
Pressure to Change
(mother)
Confused
Love Withdrawal
(youth)
Depressed

Love Withdrawal
(mother)

Family Members
Go Their Own Way
(youth)
Family Members Drugs
Go Their Own Way
(mother)
Truant
Family Members
Do As They Wish
(youth) Swear
Family Members
Do As They Wish
(mother) Adj QFi =.92 • p < .01
standardized soiution

FIG. 3.—Fainily control and adolescent problem behaviors using combined mother and youth
reports for family variables.

are not conceptually distinct. Our analyses chological and emotional hold over the
which used raw items from the CBC appear child.
to have identified sets of internalized and
externalized items that in these data are not Finally, we conducted three specific
predictably interrelated. analyses to assure that the model would re-
main viable when considering the potential
The negative relationship between role of parental supportive behavior. We
psychological and behavioral control in used the 10-item acceptance scale from the
these data indicate that parents that exert CRPBI as a measure of parental support.
high levels of psychological control have a This is a commonly used scale for parental
tendency to be uncontroUing behaviorally. nurturance and consists of items such as:
This finding would be counterintuitive if "makes me feel better afler talking over my
one maintains an undiSerentiated conceptu- worries with her," "smiles at me very often,"
alization of control and its effects; that is, if "enjoys doing things with me," "believes in
parents are controlling in one area they showing her love for me," etc. First, we con-
would be likely to be controlling in other ducted a factor analysis ofthe 10 acceptance
areas. Our attempt has been, however, to items, the 10 psychological control items,
demonstrate that control of psychological and the 5 monitoring items in order to deter-
and behavioral processes are different phe- mine the independence of the acceptance
nomena with different effects. The negative items from the two control scales. Oblimin
association between psychological control rotation revealed a four-factor solution: Fac-
and behavioral control actually reinforces tor 1 contained the 10 acceptance items. Fac-
the distinction by indicating that parents tor 2 contained 7 ofthe psychological control
who intrude on the psychological develop- items, and Factor 3 contained the 5 monitor-
ment of their children do not necessarily re- ing items. There were no meaningful sec-
strict their behavior. In fact, they appear to ondary loadings for any of the items from
be predictably less restrictive. After replica- these factors. The three love withdrawal
tion, thisfindingmerits further study. It may items from the psychological control scale
be that psychologically controlling parents loaded independently on Factor 4, indicat-
are less concerned about what their child is ing some empirical and conceptual distinc-
doing as long as they feel they have a psy- tion among the psychological control items.
1132 Child Development
However, there was no evidence that either The theorized effects were evident not
ofthe control scales overlapped with the ac- only according to the adolescents' construc-
ceptance scale. tion of their family interaction, but when
perceptions of both mothers and youth were
Second, we incorporated a latent vari- combined. It was also noteworthy that each
able for parental acceptance indicated by the control construct could be measured at both
10 acceptance items from the CRPBI into the dyadic and systems level. This was evi-
the LISREL analysis using the youth-only dent in the second-order factor analysis
model because ofthe larger sample size. Ac- where factors representing both levels of
ceptance was significantly related to both measurement clustered together as indica-
psychological (-.36, p < .001) and behav- tors of psychological and behavioral control.
ioral contiol (.41, p < .001), indicating that It was also refiected in the equivalent func-
parents who are more accepting tend to ex- tioning of the behavioral control construct
hibit less psychological control and more from the youth-only analysis, which used
awareness of their adolescents activities. Ac- a dyadic measure of monitoring, to the
ceptance had no significant direct associa- mother-youth analysis, which used a system-
tion with internalized problems, but was level measure of disengagement. These
mildly predictive of externalized problems findings have implications for the potential
( — .17, p < .05). Most importantly, the pres- utility of collaborative work between clini-
ence of the acceptance variable did not sub- cally oriented family researchers who focus
stantially alter the key theoretical paths of on the family system and developmental
the model. psychologists and social psychologists who
Further, in order to test for possible in- focus more on interacting dyads in the
teractions between acceptance and control, family.
we conducted a median split on the accep-
tance variable and tested the model sepa- The fact that these control processes can
rately on families with high versus low ac- be measured at multiple levels within the
ceptance. The chi-square test for invariance family strengthens the validity of these con-
indicated that the model functioned the structs as markers of salient family processes
same in both groups (x^ difference = 3.23, affecting individual development. The va-
df = 4). The results of these three analyses lidity is also reinforced because the findings
demonstrated that the theorized pattern of are consistent with other studies that employ
associations between control and adolescent very different methodologies. Most relevant
difficulty exists independent of the levels of in this regard is the recent work of Baumrind
supportive relationships between parents (1991a, 1991b). Her intensive observation of
and their adolescents. They also demon- parents and children and her focus on devel-
strated that the control variables served as oping parenting typologies contrast dramati-
key mediators of parental acceptance in pre- cally with our self-reported survey method-
dicting adolescent difficulty. ology and our focus on discrete parental
constructs. Nevertheless, both methodolo-
Discussion gies have furnished quite similar results in
demonstrating associations between psycho-
This study had two central purposes: logically intrusive family experience and
first, to distinguish parental psychological internalized adolescent difficulties, and
control and behavioral control, and, second, between behaviorally undercontrolled fam-
to validate this difference by demonstrating ily patterns and adolescent externalized
that the two types of control have contrasting problems.
associations with internalized and external-
ized problem behaviors in youth. It was the- These findings provide encouraging ev-
orized that pattems of family interaction that idence for the usefulness of greater specifi-
inhibit or intrude on the psychological de- cation in the study of family and adolescent
velopment of youth pose particular risk for development. In particular, it is evident that
internalized problems, whereas insufficient patterns of control need to be studied not
structure or regulation of youth behavior only in terms of absolute frequency or quan-
would be more strongly associated with ex- tity of control, but in terms of different types
ternalized problems. Results provided sup- of control. In so doing it is critical to con-
port for the theory, in that psychological con- sider carefully the specific domains of child
trol and behavioral control were empirically development likely to be affected by the dis-
independent dimensions of family inter- tinct varieties of control. It is apparent from
action that evidenced the predicted con- these findings that control-related disrup-
trasting effects on youth characteristics. tions in the development of psychological
Barber, Oben, and Shagle 1133
autonomy and the failure to provide ade- ent meanings and associations in these set-
quate behavioral regulation have quite dif- tings (Spencer & Dornbusch, 1990).
ferent effects on youth characteristics.
This study had a relatively narrow focus.
We recognize that this is only an initial Our measurement of family processes was
test of the theory, and that these findings intentionally restricted to two specific types
must be replicated on other samples. The of control. Clearly, adolescent experience in
response rate from mothers in our study, for families is more complex than this. Already,
example, is a limitation that needs to be cor- the model has been shown to remain valid
rected in future tests. But we are encouraged when tested in the context of other institu-
at the prospects of more precisely under- tions of socialization (peers, school, and reli-
standing the match between specific social- gion) (Barber & Shagle, 1992). Future work
ization environments and discrete develop- will progressively incorporate other aspects
mental outcomes. Future efforts should also of family and individual experience to more
test this model utilizing other types of out- comprehensively understand the develop-
comes and conceive of other ways in which ment of youth difficulty. We suspect, for ex-
socialization can be understood to have ef- ample, that non-normative stress, such as
fects on specific outcomes. Costanzo and negative life events, will exacerbate the neg-
Woody's (1985) work on parental expecta- ative effects of both psychological and be-
tions and values as they relate to specific do- havioral control (Barber, 1992). It will also
mains of behavior such as children's obesity be important to test the functioning of the
is an excellent example of the type of speci- model across key normative transitions
ficity that is warranted in socialization re- known to have an impact on family and ado-
search. Such efforts appear to have consider- lescent development, such as puberty,
able potential payoff for the refinement of school changes, and increased interaction
theory as well as for prevention and inter- with peers (Brown, 1990; Simmons, Burge-
vention. son, Carlton-Ford, & Blyth, 1987; Steinberg,
1988). In this regard, our preliminary finding
that the model worked as expected for ado-
We acknowledge other limitations to lescents (eighth and tenth graders), but not
this study. First, the direction of the theo- for preadolescents (fifth grade), suggests that
rized effects of family interaction on psycho- these family control patterns are particularly
logical and behavioral characteristics of salient for youth who are experiencing fun-
youth cannot be confirmed due to the cross- damental changes in personality and social
sectional nature of the data. It is possible development coincident with adolescence.
that the direction is actually reverse, that is, This is consistent with much of the theory
that parents adopt control styles as a function that generated the model regarding the inhi-
of their children's characteristics. Thus, par- bition of psychological autonomy which
ents who perceive their children to be so- emerges strongly through the adolescent in-
cially isolated or withdrawn may increase dividuation process, and the need for behav-
their attempts to shape the personality of ioral regulation and supervision as adoles-
their child through psychological controls. cents increase their behavioral freedom.
Also, some parents whose adolescents are Longitudinal work is needed to better spec-
acting out may determine that their children ify the age at which these effects begin to
are beyond their control and therefore be- take hold. Finally, the model needs to be
come quite lax in their supervision. Baum- tested using more severe forms of problem
rind's (1991a) finding that externalizing ado- behaviors. The theorized patterns were evi-
lescents came from historically unstructured dent with our relatively mild forms of inter-
and disorganized homes, however, provides nalized and externalized problems, but the
some evidence of causal precedence of fam- question remains open whether these forms
ily experience. It is most likely that recipro- of family control are as useful in understand-
cal and cyclical associations exist between ing youth with more chronic and severe
family interaction and youth behavior (e.g., problems, or whether these more extreme
Patterson, 1982). A more precise under- forms of difBcvilty have more complex or dif-
standing of these effects await longitudinal ferent etiological patterns.
data. The study is also characteristically
limited by a white, middle-class sample. We
caution against generalizing to other sam- References
ples with different ethnic, cultural, and so- Achenbach, T. M. (1985). Assessment and taxon-
cial compositions due to the possibility that omy of child and adolescent psychopathol-
both control and deviance may have differ- ogy. Beverly Hills, CA: Sage.
1134 Child Development
Achenbach, T. M., & Edelbrock, C. (1987). Man- At the threshold: The developing adolescent
ual for the youth self-report and profile. Burl- (pp. 171-198). Cambridge, MA: Harvard Uni-
ington: University of Vermont, Department of versity Press.
Psychiatry. Burger, G. R., & Armentrout, J. A. (1971). A factor
Baldwin, A. L. (1955). Behavior and development analysis of 5th and 6th graders' reports of pa-
in childhood. New York: Dreyden. rental child-rearing behavior. Developmental
Bandura, A. (1964). The stormy decade: Fact or Psychology, 4, 483.
fiction? Psychology in the Schools, 1, 224- Cicchetti, D., & Toth, S. L. (Eds.). (1991). Inter-
231. nalizing arid externalizing expressions of dys-
Bandura, A., & Walters, R. H. (1959). Adolescent function: Rochester Sym.posium on Develop-
aggression. New York: Ronald. mental Psychopathology (Vol. 2). Hillsdale,
Bank, L., Dishion, T., Skinner, M., & Patterson, NJ: Erlbaum.
C. R. (1990). Method variance in structural Cooper, C. R., Grotevant, H. D., & Condon, S. M.
equation modeling: Living with "glop." In (1982). Methodological challenges of selectiv-
C. R. Patterson (Ed.), Depression and aggres- ity in family interaction: Addressing temporal
sion in family interaction (pp. 247-279). patterns of individuation. JoumaZ of Marriage
Hillsdale, NJ: Erlbaum. and the Family, 44, 749-754.
Barber, B. K. (1992). Family, personality, and ado- Cooper, C. R., Grotevant, H. D., & Condon, S. M.
lescent problem behaviors, journal of Mar- (1983). Individuality and connectedness in
riage and the Family, 54, 69-79. the family as a context for adolescent identity
Barber, B. K., & Shagle, S. C. (1992). Adolescent formation and role-taking skill. In H. D. Gro-
problem behaviors: A social-ecological analy- tevant & C. R. Cooper (Eds.), New directions
sis. Family Perspective, 26, 493-515. for child development: Vol. 22. Adolescent
Baumrind, D. (1967). Child care practices ante- development in the family (pp. 43-59). San
ceding three patterns of preschool behavior. Francisco: Jossey-Bass.
Genetic Psychology Monographs, 75, 43-88. Costanzo, R. P., & Woody, E. Z. (1985). Domain-
Baumrind, D. (1971). Current patterns of parental specific parenting styles and their impact on
authority. Developmental Psychology Mono- the child's development of particular devi-
graphs, 4, 1-102. ance: The example of obesity proneness.
Baumrind, D. (1991a). The influence of parenting Journal of Social and Clinical Psychology, 3,
style on adolescent competence and sub- 425-445.
stance use. Journal of Early Adolescence, 11, Dishion, T. J., & Loeber, R. (1985). Adolescent
56-95. marijuana and alcohol use: The role of par-
Baumrind, D. (1991b). Effective parenting during ents and peers revisited. American Journal of
the early adolescent transition. In P. A. Drug and Alcohol Abuse, 11, 11-25.
Cowan & M. Hetherington (Eds.), Family Dombusch, S. M., Carlsmith, J., Bushwall, S., Rit-
transitions (pp. 111-163). Hillsdale, NJ: ter, P., Liederman, P., Hastorf, A., & Gross, R.
Erlbaum. (1985). Single parents, extended households,
Baumrind, D., & Black, A. E. (1967). Socialization and the control of adolescents. Child Devel-
practices associated with dimensions of com- opment, 56, 326-341.
petence in preschool boys and girls. Child Dombusch, S. M., Ritter, P., Liederman, P., Rob-
Development, 38, 291-327. erts, D., & Fraleigh, M. (1987). The relation
Beavers, W. R. (1982). Healthy, midrange, and se- of parenting style to adolescent school perfor-
verely dysfunctional families. In F. Walsh mance. Child Development, 58, 1244-1257.
(Ed.), Normal family processes. New York: Erikson, E. E. (1968). Identity: Youth and crisis.
Guilford. New York: Norton.
Becker, W. C. (1964). Consequences of different Freud, A. (1958). Adolescence. Psychoanalytic
kinds of parental discipline. In M. L. Hoffhian Study of the Child, 13, 255-278.
& W. W. Hoffman (Eds.), Review of child de- Garber, J., Quiggle, N. L., Panak, W., & Dodge,
velopment research (Vol. 1, pp. 169-208). K. A. (1991). Aggression and depression in
New York: Russell Sage Foundation. children: Comorbidity, specificity, and social
Bloom, B. L. (1985). A factor analysis of self-report cognitive processing. In D. Gicchetti & S. L.
measures of family functioning. Family Pro- Toth (Eds.), Internalizing and externalizing
cess, 24, 225-239. expressions of dysfunction: Rochester Sym-
Bios, P. (1979). The adolescent passage: Develop- posium on Developmental Psychopathology
mental issues. New York: International Uni- (Vol. 2, pp. 225-264). Hillsdale, NJ: Erlbaum.
versities Press. Hauser, S. (1991). Families and their adolescents.
Brown, B. B. (1990). Peer groups and peer cul- New York: Free Press.
tures. In S. S. Feidman & C. R. Elliot (Eds.), Hauser, S., Powers, S. L, Noam, G., Jacobson, A.,
Barher, Olsen, and Shagle 1135
Weiss, B., & Follansbee, D. (1984). Familial E. J. Mash, L. A. Hamerlynck, & L. C. Handy
contexts of adolescent ego development. (Eds.), Behavior modification and families: I.
Child Development, 55, 195-213. Theory and research (pp. 267-316). New
Joreskog, K. G., & Sorbom, D. (1989). LISREL York: Brunner-Mazel.
VII: A guide to the program and applications Patterson, G. R. (1982). Coercive family processes.
(2d ed.). Chicago: SPSS Inc. Eugene, OR: Castalia.
Lewin, K. (1935). A dynamic theory of personal- Patterson, G. R., Capaldi, D., & Bank, L. (1989).
ity. New York: McGraw-Hill. An early starter model for predicting delin-
Josselson, R. L. (1980). Ego development in ado- quency. In D. Pepler & K. H. Rubin (Eds.),
lescence. In J. Adelson (Ed.), Handbook of The development and treatment of childhood
adolescent psychology. New York: Wiley. aggression (pp. 139-168). Hillsdale, NJ:
Loeber, R., & Dishion, T. J. (1984). Boys who fight Erlbaum.
at home and school: Family conditions influ- Patterson, G. R., & Stouthamer-Loeber, M. (1984).
encing cross-setting consistency. Journal of The correlation of family management prac-
Consulting and Clinical Psychology, 52, tices and delinquency. Child Development,
759-768. 55, 1299-1307.
Lorenz, F. O., Conger, R. D., Simons, R. L., & Pulkinnen, L. (1982). Self-control and continuity
Whitbeck, L. (1991). Economic pressure and from childhood to adolescence. In P. B. Baltes
marital quality: An illustration of the method & O. G. Brim (Eds.), Life-span development
variance problem in the causal modeling of and behavior (Vol. 4, pp. 63-105). New York:
family processes, youma/ of Marriage and the Academic Press.
Family, 53, 375-388. Roe, A., & Siegelman, M. (1963). A parent-child
Maccoby, E. E., & Martin, J. A. (1983). Socializa- questionnaire. Child Development, 34, 355-
tion in the context of the family: Parent-child 369.
interaction. In E. M. Hetherington (Ed.), Rohner, R. P. (1986). The warmth dimension:
P. H. Mussen (Series Ed.), Handbook of child Foundations of parental acceptance-
psychology: Vol. 4. Socialization, personal- rejection theory. Beverly Hills, CA: Sage.
ity, and social development (pp. 1-101). New Rollins, B. C , & Thomas, D. L. (1979). Parental
York: Wiley. support, power, and control techniques in the
Marcia, J. E. (1980). Identity in adolescence. In J. socialization of children. In W. R. Burr, R.
Adelson (Ed.), Handbook of adolescent psy- Hill, F. I. Nye, & I. L. Reiss (Eds.), Contem-
chology. New York: Wiley. porary theories about the family: Vol. 1. Re-
McCord, J. (1979). Some child-rearing anteced- search based theories (pp. 317-364). New
ents of criminal behavior in adult men. Jour- York: Free Press.
nal of Personality and Social Psychology, 37, Sabatelli, R. M., & Mazor, A. (1985). Differentia-
1477-1486. tion, individuation, and identity formation:
McCord, J. (1990). Problem behaviors. In S. S. The integration of family system and individ-
Feidman & G. Elliot (Eds.), At the threshold: ual developmental perspectives. Adoles-
The developing adolescent (pp. 414-430). cence, 20, 619-633.
Cambridge, MA: Harvard University Press. Schaefer, E. (1965). Children's reports of parental
Miller, B. C , McCoy, J. K., Olson, T. D., & Wal- behavior: An inventory. Child Development,
lace, C. M. (1986). Parental control attempts 36, 413-424.
and adolescent sexual behavior. Journal of Seligman, M. E., & Peterson, C. (1986). A learned
Marriage and the Family, 48, 503-512. helplessness perspective on childhood de-
Minuchin, S. (1974). Families and family therapy. pression: Theory and research. In M. Rutter,
Cambridge, MA: Harvard University Press. C. E. Izard, & P. B. Read (Eds.), Depression
Olson, D. H., McCubbin, H. I., Barnes, H. L., in young people (pp. 223-249). New York:
Larsen, A. S., Muxen, M. J., & Wilson, M. A. Guilford.
(1983). Families: What makes them work. Siegelman, M. (1965). College student personality
Beverly Hills, CA: Sage. correlates of early parent-child relationships.
Olweus, D. (1980). Familial and temperamental Journal of Consulting Psychology, 29, 558-
detenninants of aggression behavior in ado- 564.
lescents—a causal analysis. Developmental Simmons, R. G., Burgeson, R., Carlton-Ford, S., &
Psychology, 16, 644-660. Blyth, D. A. (1987). The impact of cumulative
Parker, G., Tupling, H., & Brown, L. B. (1979). A change in early adolescence. Child Develop-
parental bonding instrument. British Journal ment, 58, 1220-.-1234.
of Medical Psychology, 52, 1-10. Spencer, M. B., & Dombusch, S. M. (1990). Chal-
Patterson, G. R. (1976). The aggressive child: Vic- lenges in studying minority youth. In S. S.
tim and architect of a coercive system. In Feidman & G. R. Elliot (Eds.), At the thresh-
1136 Child Development
old: The developing adolescent (pp. 123- Stierlin, H. (1974). Separating parents and ado-
146). Cambridge, MA: Harvard University lescents. New York: Quadrangle.
Press. Thomas, D. L., Gecas, V., Weigert, A., & Rooney,
Steinberg, L. (1988). Reciprocal relations between E. (1974). Family socialization and the ado-
parent-child distance and pubertal matura- lescent. Lexington, MA: Lexington Books.
tion. Developmental Psychology, 24, 122- Volk, R. J., Edwards, D. W., Lewis, R. A., & Spren-
128. kle, D. H. (1989). The assessment of adapt-
Steinberg, L. (1990). Autonomy, conflict, and har- ability and cohesion in families of adolescent
mony in the family relationship. In S. S. Feid- drug abusers and non-abusers. Family Rela-
man & G. R. Elliot (Eds.), At the threshold: tions, 38, 266-272.
The developing adolescent (pp. 255-276). White, J. L. (1989). The troubled adolescent. New
Cambridge, MA: Harvard University Press. York: Pergamon.

Вам также может понравиться