The Relation of Parenting Style to Adolescent School Performance

Author(s): Sanford M. Dornbusch, Philip L. Ritter, P. Herbert Leiderman, Donald F.

Roberts and Michael J. Fraleigh
Source: Child Development, Vol. 58, No. 5, Special Issue on Schools and Development (Oct.,
1987), pp. 1244-1257
Published by: Wiley on behalf of the Society for Research in Child Development
Stable URL: https://www.jstor.org/stable/1130618
Accessed: 01-10-2018 19:22 UTC

REFERENCES
Linked references are available on JSTOR for this article:
https://www.jstor.org/stable/1130618?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked references.

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Society for Research in Child Development, Wiley are collaborating with JSTOR to
digitize, preserve and extend access to Child Development

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
The Relation of Parenting Style to
Adolescent School Performance

Sanford M. Dornbusch, Philip L. Ritter, P. Herbert

Leiderman, Donald F. Roberts, and Michael J. Fraleigh
Stanford Center for the Study of Youth Development

DORNBUSCH, SANFORD M.; RITTER, PHILIP L., LEIDERMAN, P. HERBERT; ROBERTS, DONALD F.; and
FRALEIGH, MICHAEL J. The Relation of Parenting Style to Adolescent School Performance. CHILD
DEVELOPMENT, 1987, 58, 1244-1257. This article develops and tests a reformation of Baumrind's
typology of authoritarian, permissive, and authoritative parenting styles in the context of adolescent
school performance. Using a large and diverse sample of San Francisco Bay Area high school
students (N = 7,836), we found that both authoritarian and permissive parenting styles were nega-
tively associated with grades, and authoritative parenting was positively associated with grades.
Parenting styles generally showed the expected relation to grades across gender, age, parental
education, ethnic, and family structure categories. Authoritarian parenting tended to have a stronger
association with grades than did the other 2 parenting styles, except among Hispanic males. The full
typology best predicted grades among white students. Pure authoritative families (high on authorita-
tive but not high on the other 2 indices) had the highest mean grades, while inconsistent families
that combine authoritarian parenting with other parenting styles had the lowest grades.

A recent review of research on the family The research of Baumrind is particularly

and school as educational institutions notes pertinent because she attempts to link compo-
an increasing emphasis on "process" studies
nents of family interaction to cognitive com-
petence. She postulates three family par-
that seek to identify those features of the fam-
ily environment through which socioeco- enting styles (authoritative, authoritarian, and
nomic and cultural background have an im- permissive) that have consequences for the
pact on mental development and school development of cognitive and social compe-
achievement. Hess and Holloway (1984) ana- tence. These three family types differ in the
lyzed results from studies of preschool, pri-
values, behaviors, and standards that children
mary, and middle-school children and are expected to adopt; in the ways these
identified five processes linking family andvalues, behaviors, and standards are transmit-
school achievement: (1) verbal interaction be-
ted; and in parental expectations about the
tween mother and children, (2) expectation behavior
of of children. In this study we extend
parents for achievement, (3) positive affectiveBaumrind's typology to a large and ethnically
relationships between parents and children, diverse sample of adolescents.
(4) parental beliefs and attributions about the
child, and (5) discipline and control strategies. Baumrind, in a series of studies of pre-
Among these various processes, discipline school children and their families (Baumrind
and control strategies appeared to have a ma- & Black, 1967), and later in studies of some-
jor influence on school achievement (Baum- what older children, delineated three modes
rind, 1973; Hess & McDevitt, 1984; Mar-of family interaction that we will reformulate
joriebanks, 1979). for use in this study of adolescents and their

This research was supported by the Hewlett Foundation, the Irvine Foundation, the Bank of
America Foundation, individual trustees of the California Family Foundation, the Stanford Center
for the Study of Youth Development, and Father Flanagan's Boys' Home. It was part of the joint
project known as the Study of Stanford and the Schools. The principals of the six cooperating schools
participated actively at every stage, from project design to analysis: Verdis Crockett, Samuel John-
son, Jr., Gary McHenry, Robert Palazzi, Charles Perotti, Gary Poulos, Joyce Rosenstiel, and Jesus
Sanchez. We are indebted to Lee J. Cronbach, Helena Kraemer, Steven H. Chaffee, Michael W.
Kirst, Michael Garet, W. Richard Scott, Robert C. Calfee, Shirley Feldman, Eleanor E. Maccoby,
Martin Ford, and Albert H. Hastorf for their suggestions and criticisms. Jean Kanerva, Barbara
Prescott, Lindsay White, Lisa Shaffer, Fox Vernon, Robert Macaulay, Ivan Fukumoto, Angela
Valenzuela, and Worku Negash assisted in data collection and analysis. Send requests for reprints to
the first author at The Stanford Center for the Study of Youth Development, Stanford University,
Stanford, CA 94305.

[Child Development, 1987, 58, 1244-1257. ? 1987 by the Society for Research in Child Development, Inc.
All rights reserved. 0009-3920/87/5805-0003$01.00]

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1245

parents. (We will not describe the harmoni-

sible and more independent than other chil-
ous and nonconforming patterns, which dren. we
Male children were as independent as
do not use.) the other children were, and they appeared to
be socially responsible. At ages 8 and 9, both
The authoritarian style of parenting had
the following characteristics: parents maleat-
and female offspring of authoritative
tempted to shape, control, and evaluate the were high in social and cognitive
parents
behavior and attitudes of their childrencompetence
in ac- (Baumrind, in preparation).
cordance with an absolute set of standards; The studies of Baumrind and others have
parents emphasize obedience, respect for au- on preschool children and children in
focused
thority, work, tradition, and the preservation
elementary school. Studies of family pro-
of order; verbal give-and-take between cesses
parent and school achievement beyond child-
and child is discouraged. Baumrind's study of rare. A recent study showed that the
hood are
preschool children found that such a mode
effectof of parental control processes persisted
family interaction was associated with low performance among children 12
in school
levels of independence and social respon-
years of age (Hess & McDevitt, 1984). In ad-
sibility. dition, there is suggestive evidence that high
Baumrind later described the authorita- achievement in the adolescent years is associ-
rian pattern, somewhat more formally, as be- ated with at least one family process, high
ing high in demandedness on the part of the identification with parents (Kandel & Lesser,
parents and low in parental responsiveness to 1969; Morrow & Wilson, 1961; Rickberg &
the child. She continued her studies of chil- Westby, 1967; Shaw & White, 1965; Swift,
dren, this time with children 8-9 years old 1967; Weinhert & Trieber, 1982).
(Baumrind, 1971, 1973). She found that the This article develops and tests Baum-
authoritarian pattern, high in demandedness rind's conceptualization of family processes in
and low in parental responsiveness, had dif- the context of adolescent school performance.
ferent consequences for girls and for boys. The study is unusual in that it extends Baum-
Girls, but not boys, who came from authorita- rind's typology of authoritative, authoritarian,
rian families were more socially assertive. For and permissive parenting to a very large and
both sexes, intrusive-directiveness was asso- diverse sample of adolescents, using high
ciated with lower cognitive competence school grades as the criterion variable. A
(Baumrind, in preparation). large-scale questionnaire study of adolescents
A second pattern is permissive parenting, in high schools was used to derive indirect
in which parents are tolerant and accepting measures of the style of parenting. In assign-
toward the child's impulses, use as little pun- ing scores on the three parenting styles, we
ishment as possible, make few demands for relied on the face validity of questions and
mature behavior, and allow considerable self- response categories. The reliability of two of
regulation by the child. In the study of pre- our three measures and the consistency of our
school children, Baumrind found the children findings increase our confidence in the utility
of permissive parents were immature, lacked of this approach.
impulse control and self-reliance, and
evidenced a lack of social responsibility and Sources of Data
independence. In the follow-up studies at 8-
9 years of age, these children were low in The major source of data for this study is
a questionnaire completed by 7,836 adoles-
both social and cognitive competence.
cents enrolled in six high schools in the San
Authoritative parenting is the third type Francisco Bay area, approximately 88% of the
described by Baumrind. This pattern contains total enrollment of those schools, in Spring
the following elements: an expectation of ma- 1985. The questionnaire contained numerous
ture behavior from the child and clear setting items. Those used in this article include stu-
of standards by the parents; firm enforcement dent background characteristics, self-reported
of rules and standards, using commands and grades, perceptions of parental attitudes and
sanctions when necessary; encouragement of behaviors, and family communication pat-
the child's independence and individuality; terns. From this questionnaire we used per-
open communication between parents and ceptions of family processes to construct indi-
children, with encouragement of verbal give- ces of parenting style, background variables
and-take; and recognition of the rights of both to serve as controls, and self-reported grades
parents and children. as the dependent variable.
Female children of authoritative parents Some questionnaire items were not an-
in the preschool sample were socially respon- swered by all students. Small variations in

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 1246 Child Development

sample size across tables reflect this fact. Wegree; (6) graduate or professional degree; and
chose to present all the available data rather (7) don't know. The question that was used in
than including only those cases where thethe parent survey had slightly finer grada-
data were complete. tions, and was recoded to match the break-
down shown here. Mother's education and
The data from that questionnaire are aug-
mented by information on parental education father's education were then averaged to
from two additional sources. First, a studentcreate a single parental education measure for
questionnaire had been administered in each family. Finally, for qualitative analyses,
mean parental education was trichotomized
Spring 1983 to a sample of students at five of
the six participating schools. The studentsso as to produce categories of clear social
who participated in both surveys gave us ameaning: up to 3.5 = low education, 4 to 4.5
substantial pool of students for whom we had = middle education, and 5 and above = high
education. Families whose mean education
parental education. Second, parental re-
sponses to a family questionnaire mailed todid not reach attendance at a college were in
the homes of all students in our sample pro-
the low-education group; the middle-
vided information on parental education for education group included college attendance
additional cases in our sample. For thosebut not receiving a 4-year degree; the high
analyses, such as multiple regressions, which group had at least a 4-year college degree.
required the inclusion of parental education Family structure.-Our measure of fam-
as a measure of social stratification, a substan-ily structure came from student reports of who
tial portion of our cases had to be excludedis present in the household. In the analyses
because of the absence of information on pa-presented here, family structure consists of
rental education. five categories: two natural parents, single
For one school in our sample we had cur- mother, mother and stepfather, single father,
rent grade point averages for every student. and father and stepmother. All other family
Those data enabled us to assess the validity of forms were too infrequent to provide a sample
the self-reported grades that we used as a de- large enough for analysis.
pendent variable.
Measures of Parenting Style
Measures
Three parenting style indices were de-
veloped to roughly conform with Baumrind's
Demographic Variables three styles of parenting (authoritarian, per-
Ethnicity.-Each high school student missive, and authoritative). Twenty-five items
was asked to select one of nine categoriesorfor
sets of items were identified in the student
questionnaire as closely reflecting one of the
ethnic identification: Asian, black, Filipino,
Pacific Islander, American Indian, Latino three
or styles, and each index was constructed
Hispanic, white, and other. Vietnamesebyre- taking the means of the appropriate items.
spondents were combined with the Asian No question was allowed to contribute to
subgroup. Sample sizes provide sufficient
more than one of the indices, so that the three
scores
cases for the analysis in this article of re- are not forced to be correlated with
sponses from four groups: Asian, black, eachHis-other.
panic, and (non-Hispanic) white. The authoritarian index was based on the
Parental education.-Our measure ofmean pa- response to the following eight ques-
rental status or social class was parental edu-
tions concerning the frequency of certain fam-
cation. There was no information on parental
ily behaviors: in their family communication,
education in the student questionnairethe used
parents tell the youth not to argue with
as the basis for most of this article. Two other
adults, that he or she will know better when
sources, responses to a questionnaire mailed
grown up, and that the parents are correct and
to parents and responses to a previous student
should not be questioned; as a response to
questionnaire, were used to determine paren-
poor grades, the parents get upset, reduce the
tal education for a subset of the total sample.
youth's allowance, or "ground" the youth; as a
Together, these two additional sources response
pro- to good grades, parents tell the
vided information on parental education youth
for to do even better, and note that other
4,053 cases, or 52% of the total sample. grades should be as good.
The parental education categories usedThe permissive index was the mean of
in the previous student questionnaire were:
eight responses: hard work in school is not
(1) not a high school graduate; (2) high school
important to the parents (the mean for four
graduate; (3) vocational, trade, or business
academic subjects), the parents don't care if
school; (4) some college; (5) 4-year collegethe
de-student gets bad grades, they don't care if

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1247

the student gets good grades, there are nodefined, with a family included in a
were
rules concerning watching television, and (us-
pure family style category if it scored in the
ing the highest involvement of the possible
top one-third on one parenting style index
parent figures) the parents are not involved in in the top one-third on either of the
and not
other two indices.
education, they do not attend school programs
for parents, they do not help with homework,
Half of the families (50%) could not be
and they do not check the child's homework.
characterized as having a pure parenting
The authoritative index was calculated
style, while 18% (1,321) were categorized as
from the mean frequency of nine responsespure permissive, 17% (1,218) were pure au-
concerning family behavior: in their family
thoritative and 15% (1,064) were pure au-
communication, parents tell the youth to look
thoritarian. Thus, pure parenting styles apply
at both sides of issues, they admit that to
theonly half of the families in the total sample.
youth sometimes knows more, they talk about
politics within the family, and they empha- In addition, we created a variable in
which every family was assigned to one of the
size that everyone should help with decisions
possible combinations of pure parenting
in the family; as a response to good grades,
styles. These combinations range from being
parents praise the student, and give more
freedom to make decisions; as a responsehigh
to on all three pure parenting styles to be-
ing high on none of them.
poor grades, they take away freedom, encour-
age the student to try harder, and offer to
Measures of Student Performance
help. Self-reported grades.-The measure of
These three indices of parental style student performance used throughout this ar-
were used as continuous variables throughout ticle is the response by the student to a ques-
the article as the main measures for the three tion that asks for the selection of a category
parenting styles. that represents the usual grade the student
receives. The categories were: mostly A's,
The reliability of these three quantitative about half A's and half B's, mostly B's, about
indices of parenting style was assessed using half B's and half C's, mostly C's, about half
Cronbach's alpha. The alpha coefficients C's and half D's, mostly D's, and mostly be-
were .70 for the eight items of the authorita- low D. A numerical scale of self-reported
rian index, .60 for the eight items of the per- grades was then related to these responses,
missive index, and .66 for the nine items of with 4.0 representing the top category.
the authoritative index. The alphas for the au-
thoritarian and authoritative indices were We have consulted with educators about
moderately high and satisfactory, and thethe use of grades as a measure of school per-
alpha for the permissive index was onlyformance. Their consensus was that grades,
slightly lower. unlike scores on intelligence tests and mea-
sures based on standardized achievement
The slightly lower reliability for the in- tests, provide the most appropriate measure of
dex of permissive parenting may be a product current school performance. Grades have
of the limited nature of the indicators of per-their difficulties as a measure of intellectual
missiveness within our questionnaire. Theperformance, for they often represent rela-
concept of permissiveness may be tapping tively arbitrary assessments by a teacher. But
two distinct and identifiable parental at- the typical grade, usual grade, or mean grade
titudes. Permissiveness may refer to a par-is the summation of many judgments about
enting attitude that is essentially neglectful
the extent to which a student is responding to
and uncaring, or it may refer to parenting thatthe school curriculum.
is caring and concerned but ideologically
genuinely permissive. It is impossible to dis- Grade-point averages.-We found that
entangle these differing orientations in our grade-point averages were available in most
permissiveness scale. Ideally, researchers of our schools only for seniors approaching
should construct scales and measure these graduation. One school had up-to-date grade-
two separable orientations. In the meantime, point averages for all its students. We there-
we urge caution in interpreting those portionsfore compared the questionnaire response,
of our results that feature indicators of permis- the self-reported grade, to the grade-point av-
siveness.
erage for each student in that school.
In addition to these quantitative The
mea-correlation between grade-point av-
sures, types of families were constructed erages and self-reported grades was .76 (N =
based on the scores on the three indices. In1,146). We were concerned that there might
particular, three "pure" styles of familiesbe a systematic inflation of self-reported

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
1248 Child Development

grades for students whose academic perfor- ences among ethnic groups are seen in that
mance was low. Accordingly, we examinedanalysis. Asian, black, and Hispanic families
the responses of students at each grade level.were higher on the authoritarian index for
There was only a slight tendency to overstate both sexes than were white families.
grades when one reached grades near the bot- Families of Asians, Hispanics, and black
tom of the distribution-mean grades of Cmales were lower on the authoritative index
and below. than were white families. For permis-
siveness, the ethnic differences were more
Accordingly, throughout this article we
complex. Compared to whites, blacks were
will use a single measure of school perfor-
mance that was available for almost all stu- lower on permissiveness, Hispanics were
higher, and Asians were slightly higher. Six-
dents in the sample. Self-reported grades give
teen of the 18 differences were statistically
a close approximation to the distribution of
significant.
grades on the transcript and will be used as
the measure of school performance in all anal- The means on each parenting index in
yses.
Table 1 also showed a clear relation to paren-
tal education. Comparing within each sex,
Results
families with higher parental education
Parenting Style tended to be somewhat lower in authoritarian
In Table 1 we report the mean on each and
ofpermissive parenting and higher on au-
the three parenting style indices by sex, thoritative
age, parenting. These differences in
ethnicity, parental education, and familyparenting styles among parental education
structure. The extent to which different groups are interesting in themselves, even
groups were reported by their children tothough
em- the association of parenting styles
ploy each style of parenting is interesting withingrades will be shown to apply across all
parental education groups.
itself, in addition to its relation to high school
grades. Since each index is based on a differ-
ent set of questions, scores on one indexWith respect to family structure, single
mothers showed a higher level of permissive
should not be compared with scores on an-
parenting than did two natural parents. For
other. Rather, comparisons should focus on
their sons only, single mothers showed lower
group differences in the means for a single
index. levels of authoritarian parenting when com-
pared to households containing both natural
There were small sex differences in the parents. Single fathers were also more per-
parenting styles reported by the students. Fe-
missive for both sexes, while they were less
authoritarian for females and less authorita-
males, compared to males, reported a slightly
lower level of authoritarian parenting, a dif-
tive for males than families containing both
ference that was statistically significant.
natural parents. Step-families, compared to
There was no gender difference in the reports
families with two natural parents, tended to
of permissive parenting. These small genderbe more authoritarian and more permissive,
differences in means will not be discussed and, for males only, less authoritative. Of the
further. Most of our analyses relating grades
24 comparisons between two-natural-parent
to parenting styles will not show gender dif-
families and other types of families, 12 were
ferences in the results. statistically significant.
Family parenting style does appear to beParenting Styles and Grades
related to the age of the adolescent. There For both sexes, the correlations between
was a decline in the mean score on the au-
grades and the three indices of parenting
thoritarian index with increased age; permis-
style strongly support earlier studies on the
siveness, on the other hand, was higher in cognitive
the impact of parenting styles. The
older age groups. The authoritative index did
negative correlation of authoritarian parenting
not show a clear relation to age. This suggests
to grades was -.18 for males and -.23 for
that, while there may be shifts in the levelfemales.
of For permissive parenting, the corre-
authoritarian or permissive parenting as mat-
lations were -.09 for males and -.17 for fe-
uration takes place, the authoritative style
males. Finally, authoritative parenting had
may represent an ideological commitment positive correlations with grades of .08 for
that does not readily change as children grow
males and .13 for females. All correlations
up.
were significant at the .001 level. The relation
In Table 1, the mean on each parenting of authoritarian parenting to grades was the
index for each ethnic-sex group is comparedstrongest of the three correlations for both
to the appropriate mean for whites. Differ- sexes.

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1249

One question that could be asked about There were, however, ethnic differences
these relations of parenting styles to grades is strength of the correlations between
in the
whether they apply equally well within parenting styles and grades. For Asians, the
groups that differ on age, ethnicity, family
correlations of grades with both the authorita-
structure, or education. Correlational analyses
tive and the permissive styles were near zero.
within categories provide a series of indepen-
For Hispanic males, authoritarian parenting
dent tests of the relation between parenting
showed almost no relation to grades (-.03),
styles and grades. We will later present multi-
even though the relation was strongly nega-
ple regressions for the total sample and
tive among Hispanic females (-.26). Among
within ethnic groups, but here we will assess
whites, our largest ethnic group, and blacks,
the consistency of these relations in specific
our smallest, all correlations were as ex-
categories of students. pected. Asians appear to be the ethnic group
for whom our typology applies least well. Al-
though our approach does not seem to be lim-
The bulk of the literature on parenting
ited in application to only a single ethnic
styles is based on studies of young children.
In that younger age group, the age ofgroup,
each data from Asians appear to offer sup-
child is more likely to be a central variable
port only for the relation of authoritarian par-
enting to grades.
than in our studies of an adolescent popula-
tion in high school. We did not expect that the
Although we have only a smaller sample
relations between parenting style and grades
of students for whom we know the education
"would be highly dependent on the age of the
of their parents, it seems appropriate to use
student, but we tested the possibility by look-
that information to see whether our parenting
ing at the relation between parenting style
style indices relate to grades across social
and grades for males and females in each of
classes. There are low, middle, and high pa-
the five largest age groups in our sample: 14,
rental education families within each sex,
15, 16, 17, and 18. There were no important
making six independent subgroups within
fluctuations among age groups in the associa-
which to examine the relation of the three
tion between parenting style and grades. All
parenting styles to grades. All 18 correlations
30 correlations (three scores by five ages and
were in the expected direction, with 11 statis-
two sexes) were in the expected direction,
tically significant. All correlations of au-
and 29 were statistically significant.
thoritarian parenting with grades were statis-
tically significant. The data support the view
The Baumrind typology was developed that the parenting style typology applies fairly
from the intensive analysis of parenting in across the social classes.
well
largely middle-class, white families. We can
take advantage of the size of our sample and We can simultaneously control for ethnic-
ity
its diversity to see whether, controlling for theand parental education and thereby pro-
duce numerous correlations of parenting style
sex of the child, the four main ethnic groups
in our study show similar relations between with grades, although many were based on a
each style of parenting and grades, and small
thus number of cases. There are four ethnic
examine the extent to which a conceptualiza-groups, three parental education groups, two
tion developed in one cultural arena appliessexes, and three styles of parenting. Exclud-
to groups with possibly divergent norms ing groups with fewer than 10 students, there
and
values. were 63 remaining correlation coefficients to
examine. Of the 63, 48 were in the expected
The data indicate that, across ethnic direction (positive for the authoritative index
groups, authoritarian and permissive stylesand negative for the authoritarian and permis-
were associated with lower grades, and an siveau-indices) and 15 in the opposite direction,
a ratio better than three to one. Looking only
thoritative style was associated with higher
grades. All eight correlation coefficientsatfor
correlations that were statistically sig-
the two sexes and four ethnic groups nificant,
were 21 were in the predicted direction
negative when the authoritarian parenting and only one (authoritative parenting for sons
style was related to grades, and the same "of low-education Asians) was in the opposite
was
direction.
true when the permissive parenting style was These correlations supported the
hypothesized relations between each par-
related to grades. For the authoritative style
enting style and grades.
the correlation to grades was positive in seven
out of eight ethnic-sex groups, with the only As American society has exhibited a de-
failure among Asian females. Thirteen of thein the proportion of children living with
cline
24 correlations were statistically significant
both atnatural parents, we wish to see if our
the .05 level. reformulation of Baumrind is applicable to

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 0) 8 m 9 w t~- m Nc. 0 In
m t N -C O I

C O
* *
4-J4

H C6 - s fs C6 cfs-cfs C5- C6 C6-C6-

CA

C) In 00 in 00 in C) In 0 in ie n C) in C) in 00 in 0 0
z

z
i i
?
0

o co ?s fsc~
c 0C 0 CI O--;
CIOcs
0 e --: C -.:
0 q 0fe n vC
in -;,C6
in v 0 6 -: C -, f
z

in~~~
CO40..
If3C).C4
. .-~ . 0. ?
- CO
- c~
v 00
v - 0 00 v O e

cr;
-4-"
z
w H

&*
z 0in t NM
eq eq coc I ~ c00
C6 ci o I C3 l 6 Oi C6

> ? ? ?

C? cs C? (i s ? C C ? C
Z
z*D-it)
*D *` n0 D L D~E C nO( ` `
in In In01-

CEf??)f(
09
U)
c , c _. f r
z3
0 c
z: ?r - " : " " ? r?4?
W -, , o - c V jiciv c

1250

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 O~ ~ U P--4 t. -4 ~
t~- CCO 04 '44 '4 cc 0 -4 C:C

m CO t- C) -4 "-4 05(D cd
*j-4 .? - C

* *1 o0 0m 0 00 q0t
v 1 li 11000 11 - .
C05 'i 6 C C6 6 fi C

* o

C6 C oi c C6 c~i - -
c.1 CM8
x O

W Wo o

%.

+ ~ c'-

be 6
W) 0

w $0

0 0$
* ~ ~ 0
0 vmv0 m
C)00C 0-4m 00 0

1251

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
1252 Child Development

children living in other types of families. Forfamily structures do not limit the scope of ap-
five types of family structure, the three par- plication of parenting styles.
enting styles were related to grades for both
sexes. Of those 30 correlations, only twoMultiple Regressions
failed to be in the hypothesized direction. Table 2 contains two multiple regres-
sions in which a series of structural variables
The two failures occurred in the least com-
mon family structure-the child living withare
a combined, in the first regression, with the
male single parent-where sampling fluctua-
indices of parenting style, or, in the second
tion is greatest. Within the more frequent
regression, with a set of measures of pure par-
family structures-two natural parents, a sin-enting style (in the top one-third on one par-
gle mother, or a mother and stepfather in theenting style index and not in the top one-third
household-15 of the 18 correlations were on either of the other two indices) in order to
predict
statistically significant at the .05 level using a grades. In a separate article we will
two-tail test. Since all 17 statistically show that family processes, of which par-
enting style is just one element, are more
significant correlations were in the predicted
powerful than structural variables in the ex-
direction, the evidence suggests that diverse

TABLE 2

DETERMINANTS OF GRADES, USING STRUCTURAL VARIABLES AND EITHER PARENTING STYL

PURE PARENTING STYLES

WITH PARENTING STYLE INDICES WITH PURE PARENTING STYLES

b Beta F b Beta F

Female .............. .084*** .054 13.2*** .093*** .060 15.8**

(.023) (.024)
Parental education .... 117*** .171 112.2*** .125*** .183 126.9***
(.011) (.011)
Black ............. -.188** - .045 8.9** -.205** -.049 10.3**
(.063) (.064)
Hispanic ............. - .117* - .040 6.4* - .122** - .042 6.7**
(.047) (.047)
Asian ................ .485*** .186 151.6*** .466*** .179 137.4***
(.039) (.040)
Age .................. - .023* - .037 5.9* - .023* - .036 5.6*
(.010) (.010)
Single parent ......... - .213*** -.107 48.6*** - .200*** -.100 42.3***
(.031) (.031)
Stepparent ........... - .163*** - .064 17.6*** -.175*** - .068 19.8***
(.039) (.039)
Authoritarian ......... - .303
(.020)
Permissive ........... -.127*** -.088 26.9***
(.025)
Authoritative ......... .053* .037 5.0*
(.024)
Pure Authoritarian .... .. ... ... -.295*** -.177 134.2***
(.025)
Pure Permissive . ......... - .143"** -.083 27.0***
(.027)
Pure Authoritative .... ... ... ... .045 .027 3.2
(.025)

Constant ............ 3.941 ... ... 2.825

R2 .................. .176 .156
N .................... 3,752 3,752

NOTE.-Numbers in parentheses are stand

+ p < .10, two-tailed.
* p < .05, two-tailed.
** p < .01, two-tailed.
*** p < .001, two-tailed.

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1253

Using measures of pure parenting style

planation of variability in grades. Indeed, the
proportion of variance explained in that produced
analy- similar results in the second equa-
tion.
sis was surprisingly high, .34 to .38, given Pure authoritarian parenting showed a
that
no measure of intellectual performance or relation to grades than either of the
stronger
previous school performance was used as atwo parenting types. In this equation,
other
predictor. pure authoritarian parenting was approxi-
mately equal to parental education in strength
In Table 2, the only family processes
as a predictor of grades. The relations be-
used as predictors are those related to our
tween the pure parenting variables and
measures of parenting styles, so that we are
grades were in the expected directions.
not expecting to explain a high proportion of
the variance in grades. The utility of the mul- We also examined ethnic differences in
tiple regression technique is that it permits the impact of parenting styles. Table 3 pres-
the simultaneous operation and statistical ents multiple regressions within each ethnic
control of all the structural variables we have group, using the same structural variables and
used in the preceding analyses, and that itparenting indices that were used in the first
enables us to assess the relative strength ofequation of Table 2. (We also did regressions
the relation between grades and each of thewithin each of the three most frequent family
three parenting styles when the structuralstructures. We do not present the tables of
variables are all taken into account. results, but the findings show the same pat-
tern in each family structure, with authorita-
The first finding we note in Table 2 is
rian and permissive parenting negatively as-
that the proportion of the variance in grades
sociated with grades, and with authoritative
explained by the predictors was slightly lower
parenting positively associated with grades.)
when the pure parenting styles were sub-
stituted for the scores on the parenting style Within the Asian group, authoritarian
indices. One explanation of this result will be
parenting was the strongest predictor of
discussed in our presentation of Table 4. grades, but the other parenting indices were
The standardized beta weights provide a not significantly related to grades. Within
means for assessing the relative contribution blacks, the group with the smallest number of
of each of the predictor variables. Focusing cases, no parenting index was significantly as-
sociated with grades.
first on the structural variables, we note that
the most powerful ethnic predictor in both Among Hispanics, an interesting result
equations was Asian. This gives further sup-emerged. Being female was significantly asso-
port to our conclusion that the parenting ciated with high grades. Yet the interaction of
styles we have studied do little to explain the
females and authoritarian parenting was asso-
high grades of the Asians in our sample. ciated with low grades, significant if a one-tail
Parental education was also a relatively test was used, and with the same magnitude
powerful predictor, with betas averaging ap- of beta weight as being female. This interac-
proximately .18. Other relatively powerful tion of gender and authoritarian parenting re-
structural predictors were our two measuresversed the relation of authoritarian parenting
of family structure. Being in a single-parent itself to grades within the Hispanic sample.
household or in a household containing a The failure of authoritarian parenting to affect
step-parent was negatively associated with Hispanic males was noted earlier. This may
partially explain why, among Hispanics, fe-
grades. (A separate article will examine the
males were only slightly higher than males in
processes within those family structures that
produced these results.) In addition, female
mean grades, while the difference was much
students tended to get higher grades than more substantial in the other ethnic groups.
males, and black, Hispanic, and older stu-
We can speculate on the reasons that au-
dents tended to get lower grades.
thoritarian parenting is gender-specific in its
Turning to the parenting style indices inimpact on Hispanics. Perhaps this reflects cul-
the first regression, we note once again thetural orientations that produce major gender
differences within the Hispanic population.
relatively stronger relation of the authorita-
rian index to grades, with a beta weightFor example, Hispanic informants suggest
higher than the betas for the permissive and that disobedience is expected among male
authoritative indices. Moreover, that the au-children in authoritarian households but not
thoritarian index is stronger than parental expected from females. Males see themselves
education as a predictor indicates that this as future heads of households; their subordi-
process variable was a better predictor than nation is only temporary. Others suggest the
the usual measure of social status. importance of considering the lifelong orien-

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
1254 Child Development
TABLE 3 TABLE 3 (Continued)
DETERMINANTS OF GRADES, USING STRUCTURAL b Beta F
VARIABLES AND PARENTING STYLE WITHIN
ETHNIC GROUPS
Whites:d
Female ......... .080** .053 8.5**
b Beta F
(.028)
Parental edu-
Asians:a
cation......... .140*** .188 103.0***
Female ........ .001 .001 .0
(.069) (.014)
Parental edu- Age ............ -.025* -.041 4.7*
cation......... .086** .139 7.7** (.011)
(.031) Single parent .... -.240*** -.123 44.0***
Age ............ - .070* - .129 6.3* (.036)
(.028) Stepparent ...... -.138** - .058 9.7**
(.044)
Single parent .... -.269* -.125 6.4* Authoritarian.... -.344*** -.262 200.7***
(.107)
Stepparent ...... -.654*** -.186 14.3*** (.024)
Permissive...... - .142*** - .096 22.2***
(.173)
(.030)
Authoritarian .... -.228*** -.190 14.6***
Authoritative .... .073* .050 6.4*
(.060)
Permissive...... - .073 - .057 1.0 (.029)
(.074)
" Intercept = 5.534; N = 370; R2 = .141.
Authoritative .... - .089 - .070 1.5
b Intercept = 1.777; N = 135; R2 = .121.
(.073) ' Intercept = 1.624; N = 285; R2 = .082.
Blacks:b d Intercept = 3.993; N = 2,592; R2 = .157.
Female ......... .354** .266 9.5** + p < .10, two-tailed.
(.115) * p < .05, two-tailed.
Parental edu- ** p < .01, two-tailed.
cation......... .076 .128 1.9 *** p < .001, two-tailed.
(.054)
Age ............ .017 .029 .1 tations of Hispanic females, emphasizing
(.053)
femininity and family. Whatever the explana-
Single parent.... -.019 -.013 .0
(.123)
tion, we have here clear additional evidence
Stepparent ...... .131 .065 .5 of difficulty in directly applying the parenting
(.180) typology across diverse cultures.
Authoritarian.... -.142 -.127 2.2
Finally, looking at the results for whites
(.095)
Permissive ...... .076 .064 .5 in Table 3, we note how well the pattern of
(.109) findings reflects the original formulation. Au-
Authoritative.... -.001 -.001 .0 thoritarian and permissive parenting are asso-
(.097) ciated with low grades, and authoritative par-
Hispanics: C enting is associated with high grades. The
Female ......... 1.070* .662 3.87* typology derived from a predominantly white
(.544) sample of children obviously continues to fit
Parental edu-
the white adolescent population fairly well.
cation......... .037 .059 .95
(.038) Pure and Inconsistent Parenting Styles
Age ............ -.007 -.010 .03 All families in our sample can be
(.038) categorized as either predominantly practic-
Single parent.... -.167 -.085 2.06 ing one form of parenting or practicing a com-
(.116)
bination of parenting styles. The data in
Stepparent ...... -.209 - .078 1.73
(.159)
Table 4 include the mean grades of students
Authoritarian .... .117 .088 .24 from the families that could be categorized as
(.240) pure authoritarian, pure permissive, and pure
Permissive...... - .048 - .039 .33 authoritative. We find that, for both sexes, the
(.085) mean grades of the children from pure au-
Authoritative .. .170+ .125 3.19 thoritative families were much higher than
(.095) the mean grades of children from pure au-
Authoritarian
thoritarian or pure permissive families (all dif-
x female ..... -.280 + - .666 3.28 + ferences significant at the .001 level).
(.155)
Looking at the same three pure family
parenting styles, we examined mean grades

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1255

TABLE 4

MEAN GRADE OF EACH COMBINATION OF HIGH ON PARENTING STYLE INDICES, BY SEX

MALE FEMALE TOTAL

INDICES ON WHICH Mean Mean Mean

HOUSEHOLD IS HIGH Grade N Grade N Grade N

All indices high ................. 2.34 92 2.42 54 2.37 146

(.73) (.79) (.75)
Authoritarian and permissive ..... 2.42 349 2.49 328 2.45 677
(.82) (.83) (.83)
Authoritarian and authoritative ... 2.54 444 2.65 303 2.58 747
(.77) (.77) (.77)
Permissive and authoritative ..... 2.84 129 2.94 133 2.89 262
(.76) (.69) (.72)
Authoritarian only (pure) ........ 2.62 555 2.68 509 2.65 1,064
(.79) (.79) (.79)
Permissive only (pure) .......... 2.61 673 2.70 648 2.66 1,321
(.90) (.85) (.87)
Authoritative only (pure) ........ 2.96 552 3.08 666 3.02 1,218
(.77) (.72) (.75)
No index high .................. 2.80 917 3.00 908 2.90 1,825
(.82) (.74) (.79)

within ethnic,whose family parenting style is

parental inconsistent,
educat
especially with combinations
structure categories. For that include
both au-
thoritarian
the four ethnic parenting. The combination
groups, three of au-
thoritarian,
tion groups, and the permissive, and authoritative
three mos par-
ily structures,enting
there(all indices high)
wereand the combination
no ex
of authoritarian
ordering of mean and permissive parenting
grades for t
Pure authoritative
were associated withfamilies
the lowest mean grades. a
highest mean Authoritarian
grades. combined with authoritative
parenting was also associated with low
Table 4 also includes, for th
grades. Only the combination of authoritative
and the total sample, mean g
and permissive parenting (not including au-
dents coming
from families ex
thoritarian parenting) had mean grades higher
possible combination of high
than pure authoritarian or pure permissive.
three parenting style indices.
of a family with Thus, inconsistency,
a pure when including
paren a
high index
that the family be on authoritarian
in the parenting, is asso-
top o
parenting ciated with
style the lowest grades.and
index We speculateno
one-third that inconsistency
the other in the home
on twoenvironmentind
every familycreates
is thus
anxiety assigned
among children, and that anxi-
ety reduces the high
following groups: relation betweenon the stu- all
high on two dent's
indices
effort in school and(authorit
the grade received.
We have examined some other data and
missive, authoritarian from our au
sampleauthoritative);
permissive and and note a slight tendency for incon-
sistent communications
single index (our pure from parents to be as-
authorit
sociated with aauthoritativ
missive, and pure lower correlation between
not high on anyhours of homework
index. and grades.

The mean grades Discussion in Table 4 e

tern that helps to explain the
predictive power This article
of has theprovided evidence tha
equati
parenting styles Baumrind'sin Table
typology of parenting2. Th
styles, ori
of students from inally formulated
pure to explain social and cogn
author
were clearly the tive development
highest. among young children, t
But ca
of students insuccessfully pure be applied to adolescents and r
authoritari
missive families were not the lowest. The lated to their academic performance in hig
lowest grades were found among students school. Students from a wide range of back

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
1256 Child Development
grounds tended to get lower grades when to grades, the Asians as a group were receiv-
their descriptions of family behavior indi-
ing high grades in school. In addition, while
cated more authoritarian parenting, more per- authoritarian parenting was significantly asso-
missive parenting, or less authoritative par- ciated with lower grades among Asians, there
enting. The association between grades and
was no significant relation between grades
the index of authoritarian parenting was stron-
and the other two parenting styles. This arti-
ger than the association between grades and cle concludes with more questions than an-
the indices of the other styles of parenting. swers in examining Asian parenting practices
and school performance.
The measurement of parenting styles
from data derived from the child's percep- The typology of parenting styles that we
tions creates a potential problem. For ex-have adapted was primarily devised for the
ample, if students who are more estranged study of middle-class white families and their
from their parents do less well in school andchildren. Indeed, the parenting typology did
also tend to assign negative authoritariantend to be more associated with grades among
characteristics to their parents, that might ex-
whites than among the other ethnic groups.
plain some of our findings. But some of ourYet, with the exception already noted for His-
results do not fit this explanation focused onpanic males, in all ethnic groups authoritarian
bias in reporting. First, some combinations ofparenting showed the expected relation to
parenting styles, such as a highly authorita-grades. Permissive and authoritative par-
rian style mixed with high levels of permis-enting were not as consistently related to
siveness, were associated with lower grades grades across ethnic lines.
than a pure authoritarian style. That students
It is impressive that the diverse measures
reporting such mixed or inconsistent par-
of parenting styles were associated with
enting styles did less well in school suggests
that the reports are more than a reflection ofgrades across a wide variety of social catego-
attitudes toward parents. Second, families ofries. The two sexes, the five age groups, the
five types of family structure, and the three
different ethnic background or different pa-
rental education markedly diverged in their parental education groups all exhibited the
use of parenting styles. Yet, without any al-
same predicted pattern. The families that
lowance for the values and norms of each were high in authoritarian or permissive par-
enting tended to have students who did less
group with respect to authoritarian parenting,
it continued to be negatively associated withwell in high school, and the families that were
grades across diverse groups. Such results high in authoritative parenting had children
suggest that we are dealing with more thanwho a got higher grades in school. There were
global positive or negative perception of par-major differences between the sexes, among
ents and their behavior. Nevertheless, a sur-the age groups, among the family structures,
vey such as this one cannot answer objections and among the parental education groups in
the extent to which the different styles of par-
to using such perceptual data. Observational
enting were employed. Yet, regardless of
data, preferably longitudinal, are needed to
check on these results. each group's mean scores on the parenting
styles, the relation of each style to school per-
Even as we stress the applicability of this
formance exhibited the predicted pattern
typology of parenting styles across a variety within
of each group.
social groups, there are numerous findings
There is a need for further investigations
that call for further investigation. For ex-
that will help increase our understanding of
ample, the mean level of authoritarianism was
these parenting styles and their conse-
about the same in families of Hispanic males
quences. Certainly, longitudinal studies that
and of Hispanic females, yet authoritarianism
can unscramble the causal pattern are crucial.
was much more associated with poor school
To some extent, parental behavior is a prod-
performance among the Hispanic females.
uct of school performance by children, and
Similarly, our data show clearly that the
that relation probably is inflating our correla-
success of Asian children in our public tional analysis. In addition, determining
schools cannot be adequately explained inwhich parent or step-parent is engaging in
terms of the parenting styles we have studied.
which type of parenting style may help us to
Compared to whites, Asian high school stu-
delineate the meaning of various parenting
dents of both sexes reported that their behaviors. Finally, careful studies of the
families were higher on the index of au-meanings of specific behaviors as interpreted
thoritarian parenting and lower on the index by members of various social groups, particu-
of authoritative parenting. Yet, counter to the
larly ethnic groups, could produce a major ad-
general association of such parenting patterns
vance in our knowledge. Both better data and

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms
 Dornbusch et al. 1257
better conceptualizations are needed to ad-
Kandel, D. B., & Lesser, G. S. (1969). Parental and
vance our knowledge of parent-adolescent re- peer influences on educational plans of adoles-
lationships. cents. American Sociological Review, 34, 213-
223.
Marjoriebanks, K. (1979). Family environments. In
References H. J. Walberg (Ed.). Educational environments
and effects (pp. 15-37). Berkeley: McCuthan.
Baumrind, D. (1971). Current patterns of parental
Morrow, W. R., & Wilson, R. C. (1961). Family rela-
authority. Developmental Psychology Mono- tions of bright high achieving and under-
graph, 4, 1-103. achieving high school boys. Child Develop-
Baumrind, D. (1973). The development of instru-ment, 32, 501-510.
mental competence through socialization. In
Rickberg, R. A., & Westby, D. L. (1967). Parental
A. D. Pick (Ed.), Minnesota symposium on encouragement, occupation, education and
child psychology (Vol. 7, pp. 3-46). Min- family size: Artifactual or independent deter-
neapolis: University of Minnesota Press. minants of adolescent educational expecta-
Baumrind, D., & Black, A. E. (1967). Socializationtions. Social Forces, 49, 362-374.
practices associated with dimensions of compe-
Shaw, M. E., & White, D. L. (1965). Relationship
tence in preschool boys and girls. Child Devel-between child parent identification and aca-
opment, 38, 291-327. demic underachievement. Journal of Clinical
Hess, R. D., & Holloway, S. D. (1984). Family andPsychology, 21, 10-13.
school as educational institutions. In R. D.
Swift, D. F. (1967). Family environment and 11 +
Parke (Ed.), Review of child development re- success: Some basic preconditions. British
search (Vol. 7, pp. 179-222). Chicago: Univer- Journal of Educational Psychology, 37, 10-12.
sity of Chicago Press. Weinhert, F. E., & Trieber, B. (1982). School
Hess, R. D., & McDevitt, T. M. (1984). Some cogni-socialization and cognitive development. In W.
tive consequences of maternal intervention Hartup (Ed.), Review of child development re-
techniques: A longitudinal study. Child Devel-search (Vol. 6, pp. 704-758). Chicago: Univer-
opment, 55, 2017-2030. sity of Chicago Press.

This content downloaded from 171.65.48.27 on Mon, 01 Oct 2018 19:22:28 UTC
All use subject to https://about.jstor.org/terms