Research paper-How parenting style affect child development

Applied Developmental Psychology 26 (2005) 85–106

Authoritative parenting, parental scaffolding of long-division

mathematics, and children’s academic competence in fourth grade

Jonathan F. Mattanah a,c,*, Michael W. Pratt

b , Philip A. Cowan

c , Carolyn P. Cowan

c

a Department of Psychology, Towson University, 8000 York Rd., Towson, MD, 21252, United States

b Wilfrid Laurier University, Canada

c University of California at Berkeley, United States

Abstract

The current study examined the relationships among authoritative parenting, parental scaffolding of long-division

math problems, and children’s academic competence. In a sample of 70 two-parent middle class families

participating in a longitudinal study on the transition to school, authoritative parenting was assessed globally at the

beginning of fourth grade, when the children were 9–10 years old, from a broad sample of laboratory-based parent–

child interactions. Microanalytic assessments of parent scaffolding were conducted during a long-division task.

Teacher reports and child self-reports of the children’s competence were obtained at the end of fourth grade.

Extending past studies, results showed that mothers’scaffolding behavior early in the fourth grade predicted teachers’

reports and children’s own reports of the children’s academic competence at the end of the year, after controlling for

the effects of authoritative parenting and prior math achievement on those outcome variables. Discussion focuses on

the role of scaffolding as a tutoring strategy that may promote academic competence in school-aged children.

D 2004 Elsevier Inc. All rights reserved.

Keywords: Parent–child interaction; Authoritative parenting; Scaffolding behavior; Children’s academic competence;

Longitudinal study

0193-3973/$ -

doi:10.1016/j.a

* Correspo

21252, United

E-mail add

see front matter D 2004 Elsevier Inc. All rights reserved.

ppdev.2004.10.007

nding author. Present address: Department of Psychology, Towson University, 8000 York Rd., Towson, MD,

States. Tel.: +1 410 704 3208; fax: +1 410 704 3800.

ress: [email protected] (J.F. Mattanah).

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10686

1. Introduction

Research on the links between parent–child interactions and children’s academic achievement in school

has been conducted at two very different levels of analysis. At the macro-level, researchers have examined

the effects of different styles of parenting on children’s achievement. These macro-level parenting styles

are presumed to have indirect effects on children’s achievement, providing a context for specific micro-

level parenting practices and guidance strategies that directly affect children’s academic outcomes

(Darling & Steinberg, 1993). The current study examined the interrelationship between authoritative

parenting, a macro-level parenting style variable, and parental scaffolding behavior, a micro-level

behavior, and the effects of each of these variables on children’s academic competence in the fourth grade.

1.1. Authoritative parenting and children’s academic competence

Authoritative parenting has been studied extensively as a general parenting style that is linked with

greater academic and social competence in children from 3.5 to 18 years old (Baumrind, 1971, 1991;

Steinberg, Elmen, & Mounts, 1989; Steinberg, Lamborn, Darling, Mounts, & Dornbusch, 1994; Weiss &

Schwarz, 1996). Baumrind (1971) originally identified authoritative parenting as a parenting style in

which parents set firm limits and demand high expectations from their children while maintaining a

relationship of warmth, mutual support, and appropriate autonomy-granting. Baumrind and others found

that authoritative parents, in contrast to parents who were more authoritarian or permissive, had children

who appeared more academically motivated, socially competent, and well-liked by peers and teachers

(Baumrind, 1971, 1991; Steinberg et al., 1994; Weiss & Schwarz, 1996). This style has predicted

competence in preschool, grade school, and adolescent offspring, although the findings appear most

robust with Caucasian, middle class samples (Chao, 2001; Park, 2002). Although the authoritative

parenting style has been generally associated with positive academic outcomes in school age children,

some research has shown more mixed results of parental involvement in children’s academic activities at

school, suggesting that the specific manner and type of help offered by parents may be critical for

children’s success (Pomerantz & Eaton, 2001). The current study sought to specify more precisely the

aspects of authoritative parenting that might be most helpful to children’s success at school.

Recently, authoritative parenting has been conceptualized as composed of three primary dimensions,

termed demandingness (i.e., setting appropriate limits and having high expectations), responsiveness

(i.e., responding appropriately to children’s needs with warmth and support), and autonomy-support (i.e.,

permitting and actively encouraging children’s autonomy and independence in appropriate contexts)

(Gray & Steinberg, 1999; Maccoby & Martin, 1983; Steinberg et al., 1989). Steinberg and others have

argued that a focus on these three dimensions, rather than examining the construct as a whole, allows one

to determine the precise aspects of authoritative parenting that are most critical for children’s

development, and which aspects may be relatively less important at certain ages or with certain types of

children (Gray & Steinberg, 1999; Steinberg et al., 1989). Using this dimensional approach, a number of

studies have shown that all three aspects of authoritative parenting are related to academic competence in

adolescence (Barber & Olsen, 1997; Herman, Dornbusch, Herron, & Herting, 1997; Steinberg et al.,

1989). A fewer number of studies have shown links between these dimensions and children’s academic

competence during middle and late childhood (Grolnick & Ryan, 1989; Mattanah, 2001). In keeping

with Steinberg’s suggestion, the current study focused on the three dimensions of authoritative parenting

as those were related to academic competence in grade-school aged children.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 87

1.2. Scaffolding behavior and children’s competence

To further specify the types of parental involvement most helpful to children’s academic success, the

current study examined the relationship between authoritative parenting and a specific behavioral

dynamic observed when parents and other adult tutors work with children on academic tasks. This

behavioral dynamic has been termed scaffolding and was inspired by Vygotsky’s writings (e.g., 1978) on

the role of adult tutors in children’s cognitive development. Vygotsky theorized that all learning of value

takes place in the child’s bZone of Proximal DevelopmentQ (ZPD), a cognitive space where children are not yet able to function independently but are able to function with the support of more expert adults.

Scaffolding, as a metaphor, refers to a transactional process by which parents or other experienced tutors

discover the optimal level at which to instruct the less experienced learner, intervening in a such a way

that the learner can succeed at the task but also gradually learn the skills to complete the task

independently (Conner & Cross, 2003; Conner, Knight, & Cross, 1997; Neitzel & Stright, 2003; Plumert

& Nichols-Witehead, 1996; Pratt, Green, MacVicar, & Bountrogianni, 1992; Pratt, Kerig, Cowan, &

Cowan, 1988; Rogoff, 1990; Rogoff, Ellis, & Gardner, 1984; Wood, 1988; Wood & Middleton, 1975).

Previous research consistent with a scaffolding approach has shown that adult tutors, including parents

and others, titrate the amount and specificity of help they provide to children, according to the child’s age,

the nature of the task, and the adult’s developing understanding of the child’s ability at the task (Plumert &

Nichols-Witehead, 1996; Rogoff et al., 1984; Wertsch, McNamee, Mclane, & Budwig, 1980). Early, now

classic, studies of scaffolding showed mothers of 2.5-year-old children more actively regulated the child’s

completion of a difficult puzzle-task than mothers of 4.5-year-old children (Wertsch, McNamee, Mclane,

& Budwig 1980). Similarly, Rogoff, Ellis, and Gardner (1984) demonstrated that mothers provided greater

assistance to younger children completing a difficult school-like task involving sorting common objects

than they did to older children doing the same task or to older or younger children completing an easier task

that involved putting grocery items away in a kitchen. More recent work has shown that adults, in

particular, are better able to provide sensitive scaffolding to children during cognitive tasks than are similar

aged peers, in part because adults makes a greater effort to involve the child in joint decision-making

processes and guided completion of the task, rather than focusing exclusively on getting the task done

(Radziszewska & Rogoff, 1988, 1991; Rogoff, 1990).

One series of studies on the scaffolding concept has examined more specifically the links between

scaffolding behaviors by parents and children’s achievement on the task being taught. In an early study

with mothers of 3.5- to 5-year-old children, Wood and Middleton (1975) observed the dynamic teaching

interactions between mothers and children during a block-building task. As the task progressed, mothers

gradually developed a pattern of providing more specific instructions to the child when the child was

failing at the task and providing less specific instructions when the child was succeeding. The first part

of this pattern ensures some child success, whereas the second component helps the child to operate

more independently. They termed this overall process the contingent shift principle (the CS principle)

and assessed how often mothers engaged in such contingent shifting with their child. Wood and

Middleton found that mothers who used the CS principle more often had children who succeeded at the

task with their mothers present and on a similar task completed by themselves after the tutoring session

was over in comparison to children with mothers who used the CS principle less often. Two other studies

with young children have shown that both mothers’ and fathers’ scaffolding behaviors are correlated

with children’s success on a number of tasks, including block-building, puzzle construction, and story

telling (Conner, Knight, & Cross, 1997; Pratt, Kerig, Cowan, & Cowan, 1988).

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10688

More recently, Pratt and his colleagues have examined parents’ scaffolding behavior with 9- to 11-

year-old children, using a developmentally appropriate task, long-division math homework (Pratt,

Green, Vicar, & Bountrogianni 1992; Pratt & Savoy-Levine, 1998). Long division is a relatively new

skill for children of this age. Thus one would expect parents to be more likely to assist their children

in learning of this task. Inspired by Wood’s research with young children (Wood & Middleton, 1975),

Pratt et al. developed a coding system for this task that examined the dynamic interaction between

the parent’s and the child’s behavior during the completion of this task. Parents were rated for the

amount and specificity of help they provided for their child, and children were rated for whether they

could successfully follow the parent’s instructions in completing the task. In support of the construct

validity of this system, Pratt, Green et al. (1992) found that parents tended to provide more specific

help around difficult subtasks of a long-division math problem (such as estimating the correct divisor)

as compared with easier subtasks of the problem (such as subtraction), and that parents provided

more specific help for children who initially had weaker math skills (as measured by a pretest of

long-division math skills). In keeping with Wood and Middleton’s (1975) findings with younger

children, Pratt, Green et al. (1992) also found that mothers’ greater use of the CS principle was

positively correlated with the child’s gains in division skills from a pre- to a posttest of independent

long-division problem solving.

1.3. Goals of the present research

The current study furthers past research on authoritative parenting, parental scaffolding behavior,

and academic competence in 9- to 10-year-old children in three specific ways: (1) it explores

whether scaffolding behavior has implications for children’s motivation for learning and self-

reported confidence in their own academic abilities; (2) it examines links between mothers’ and

fathers’ authoritative parenting and scaffolding behavior and children’s academic competence; and

(3) it examines the incremental effects of scaffolding behavior in predicting academic competence,

over and above the effects of authoritative parenting. Each of these aims is described in detail

below.

First, past studies of scaffolding have generally only examined links between scaffolding behavior

and an immediate posttest of the task being taught. Although such a posttest is a good measure of the

immediate effects of scaffolding as a teaching strategy, it provides little information about the relation

between scaffolding behavior and children’s broader academic competence and enthusiasm for learning

in the school environment. One informal study has suggested that preschool children who were taught

using scaffolding teaching techniques were more enthusiastic about learning the material and reported

greater enjoyment of the task (Wood, 1988), but systematic data on this were not reported. The current

study tests these ideas more formally by exploring links between parental scaffolding behavior and

subsequent teacher reports of academic competence in the fourth grade, which take into account the

child’s motivation for challenging academic work.

To further examine the motivational implications of parental scaffolding behavior, the current study

also examined the relationship between parental scaffolding and children’s self-reports of their academic

competence. A number of studies have reported a strong relationship between children’s self-reported

academic competence and greater academic achievement, greater intrinsic interest in school activities,

preference for challenging work, and lower rates of school dropout (Goldberg & Cornell, 1998; Harter,

1992; Miserandino, 1996). We expected that scaffolding would increase children’s sense of competence,

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 89

because children taught using scaffolding techniques might experience a greater sense of subjective

mastery over challenging academic tasks.

Second, the current study examined separately the effects of mothers’ and fathers’ authoritative

parenting and scaffolding behavior on children’s academic competence. Much previous research on

authoritative parenting as a composite style combined mothers’ and fathers’ parenting behaviors in the

authoritative parenting construct, making it impossible to discern their separate effects (see, for example,

Weiss & Schwarz, 1996). A few studies have separately examined the relations between mothers’ and

fathers’ authoritative parenting, or dimensions of authoritative parenting, and academic-related

outcomes. Pratt, Kerig, Cowan, and Cowan (1992) found that mothers’ authoritarian parenting, but

not fathers’ parenting, predicted lower levels of linguistic skill (as measured by achievement test scores

in reading, spelling, and story recall) development over a 2-year period from 3 to 5 years old. Fathers’

authoritative parenting, but not mothers’ authoritative parenting, predicted better nonlinguistic skill

development (achievement test scores in math and general information) over that same time period.

Mothers’ demandingness and fathers’ demandingness, responsiveness, and autonomy-support have been

associated with academic competence as rated by teachers, in a study of 9- to 10-year-old children

(Mattanah, 2001).

Two studies have examined the scaffolding behavior of mothers and fathers with young children

(Conner, Knight, & Cross, 1997; Pratt, Kerig, Cowan, & Cowan, 1988). Conner et al. found that mothers

and fathers were equally effective in their use of scaffolding teaching techniques; there were no mean

differences in their use of the CS principle. Interestingly, however, mothers’ use of the CS principle was

more closely associated with children’s performance at the block-building task than was fathers’ use of

the principle. Pratt et al. found also that mothers and fathers were equally effective at using the CS

principle but that fathers were actually less likely to make bextreme shiftsQ, jumping too quickly from very specific to very general levels of help, than were mothers. Unlike in Conner et al., both mothers’

and fathers’ use of the CS principle correlated with the child’s success at the task. The current study

furthers past research by exploring links between mothers’ and father’s authoritative parenting, mothers’

and fathers’ scaffolding behaviors, and multiple measures of academic competence.

Finally, the current study examined the incremental effects of parental scaffolding behavior, over and

above the effects of authoritative parenting. Studies with young children have shown that more effective

scaffolding behavior by parents is associated with higher scores on the three dimensions of authoritative

parenting (Pratt, Green, et al., 1992; Pratt et al., 1988). Thus, it is important to know whether scaffolding

as a specific teaching strategy has unique effects on children’s academic competence, after accounting

for the known association between authoritative parenting and academic competence (Baumrind, 1971,

1991; Steinberg et al., 1989). To test this hypothesis, we examined whether parental scaffolding behavior

predicted children’s academic competence in fourth grade, after controlling for the effects of

authoritative parenting and children’s prior math achievement.

2. Method

The current study was conducted as part of the Schoolchildren and Their Families Project (the SAF

Project), a longitudinal investigation of family factors in children’s adaptation to elementary school

(Johnson, Cowan, & Cowan, 1999; Mattanah, 2001). Families entered the study prior to the oldest

child’s entry into kindergarten, and were followed up during their child’s kindergarten, first grade, and

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10690

fourth grade years. The current study focuses on the fourth grade visit, when data on parental scaffolding

were collected for the first time.

Families for this study were recruited from neighborhood preschools and day care centers and through

brief announcements in the media. Families were sent screening packets and were accepted into the

study provided that they met the basic recruitment requirements: (a) an oldest child entering

kindergarten; and (2) two parents living with the child. This was required because of the interest in

the effects of couple relationships on children’s adjustment.

The sample was a nonclinical sample of families who were not in treatment and not seeking treatment

when the study began. The SAF Project included a preventive intervention component to study the

effects of helping normal families cope with the stresses of the transition to school. The intervention

consisted of a 16-week couples group offered in random order to two-thirds of the participating families

in the year before the child’s transition to kindergarten. One-third of the families, serving as the control

group, were told they could meet with the mental health staff once a year about any concerns they were

having during their child’s transition to elementary school.

The current analyses were conducted on data collected 4 years after the intervention ended.

Participants in the intervention group showed no differences from the comparison group on parenting

predictor variables, including scaffolding behavior and authoritative parenting (all ts b 1.50, ps N 0.05),

or on child outcome variables, including math achievement and teacher- and child-reported academic

competence (all ts b 1.50, ps N 0.05). Given these findings, intervention status was not considered in

analyses presented below.

2.1. Participants

The current study included 70 families from the initial sample of 84 families who completed the

transition to school phase of the SAF Project and all aspects of the fourth grade visit. Of these 70

families, data from five mother–child and one father–child interaction were not used because they

contained fewer than 10 parent–child teaching sequences; an additional seven fathers did not participate

in the parent–child interaction. Thus, the final sample size included 65 mothers and 62 fathers from the

same families and their children.

Of these 65 families, the average age of the fathers was 42.3 years and of the mothers was 40.5 years.

There were 84.1% Caucasian and 15.9% non-Caucasian children (6.3% African-American, 8% Asian-

American, and 1.6% Hispanic). The sample was highly educated, with 86.4% of the women and 81% of the

men having attained at least a bachelor’s degree. The mean family income was $78,000 (range: $22,000–

$300,000), suggesting a relatively affluent sample. However, the median income for American dual-

worker families (most families in this study were dual-income families) in the years 1990–1993 was

$58,000 (State of California, Department of Finance, 2000). Also, despite the relatively high mean income

in this sample, 22% of the participating couples had incomes below the median income for American

families during the time period in which the study was conducted. The child sample included 35 boys and

30 girls, who were an average of 9.7 (SD = 0.32) years old at the time of the fourth grade follow-up visit.

2.2. Measures and procedures

Parents and children participated in the homework interactions in late fall or winter of fourth grade.

Teachers reported on children’s academic competence in the late spring of fourth grade, toward the end

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 91

of the school year. Children were administered an achievement test and asked about their perceptions of

their own academic competence during interviews conducted in their home, during the early part of the

summer, shortly after the end of fourth grade. Due to personnel issues, interviews with students could

only be completed during the summer time; however, we conducted these interviews during the early

part of the summer when students’ memories of their fourth grade year should still have been fairly fresh.

Relationships between parental scaffolding behavior and children’s subsequent academic outcomes thus

were examined over a 6- to 9-month period. In addition, children’s math achievement scores obtained at

the end of first grade were used as a control variable in some analyses.

2.2.1. Authoritative parenting

Authoritative parenting was assessed during a visit to the project laboratory, in which mother–child

and father–child dyads participated separately in a 40-min interaction. The long-division arithmetic task

was one of four activities completed during the interaction. Only data from the long-division task are

reported in the present study. The other three activities were: (1) a playful, hand-slapping game, (2) a

social-coaching task in which the dyad watched a short videotape of peer conflict and discussed how to

resolve the conflict, and (3) a puzzle-construction task. These activities required the parent and child to

complete constructive tasks (such as homework and a puzzle), have conversations (during the social-

coaching task), and engage in playful interactions (during the hand-slapping game). We hoped that

through these tasks, the parents would demonstrate a variety of behaviors typical of their parenting style,

including warmth and responsiveness (during play and conversation), teaching, making demands, and

setting limits (during the constructive tasks), and supporting the child’s autonomy (during constructive

tasks in the sense that they let the child lead and complete the task as independently as possible). In this

nonclinical, upper middle-class sample, children were rarely off-task and in need of limit-setting, which

may have limited our ability to assess some aspects of the demandingness dimension of authoritative

parenting. The order of mother–child and father–child interactions was counterbalanced across the years

of data collection. T-tests suggested no difference in mean levels of parenting behaviors between mothers

who went first versus mothers who went second or between fathers who went first versus fathers who

went second (all ts b 1.50, ps N 0.05).

Two advanced graduate students, trained and supervised by the senior investigators of the study (Drs.

Philip and Carolyn Cowan), administered and scored all data collected during the parent–child

interaction (training was facilitated by having the students watch videotapes of parent–child interactions

collected at earlier time points of the study). These research assistants watched the entire parent–child

interaction behind a one-way mirror and rated the parent’s behavior at the end of the visit, using a 17-

item global observational rating system developed by Cowan and Cowan (1992). Each scale in the

system is rated on a five-point scale, (1 = very low levels of the behavior and 5 = high or optimal levels).

Inter-rater reliability (average alpha reliability coefficients were 0.92 for mothers and 0.94 for fathers)

and validity indices of this observational system have been established in previous studies, in which

parenting behaviors coded from the system have predicted children’s academic and social outcomes in

kindergarten (Cowan, Cowan, Schulz, & Heming, 1994; Miller, Cowan, Cowan, Hetherington, &

Clingempeel, 1993) and first grade (Cowan et al., 1994).

To assess authoritative parenting, the current study focused on the ratings of parental responsiveness,

demandingness, and autonomy-support, the three primary dimensions of authoritative parenting

identified in past studies (Gray & Steinberg, 1999; Mattanah, 2001; Steinberg et al., 1989). Three

variables in the coding system best capture the responsiveness dimension: warmth, responsiveness, and

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10692

pleasure in parental role. The warmth scale captures parental expressions of praise, loving gestures, and

physical affection toward their child during the interaction. The responsiveness scale captures how

smoothly and sensitively the parents respond to children’s needs expressed during the interaction.

Finally, the pleasure scale captures parental demonstrations of pleasure and enjoyment of their child

during the course of the interaction.

In order to create a single rating of the responsiveness dimension, parent’s scores on warmth,

responsiveness, and pleasure were standardized and then summed. This responsiveness dimension was

found to evidence good internal consistency reliability for mothers (alpha = .86; range of correlations

among individual variables = .47 to .71) and fathers (alpha = .83; range of correlations = .59 to .80).

Four variables in the coding system best capture the demandingness dimension of parenting: limit-

setting, maturity demands, structure, and confidence in parental role. The limit-setting scale assesses

how well the parent establishes, maintains, and follows through on limits set during the interaction. The

maturity demands scale assesses parental expectations for mature and appropriate behavior from their

10-year-old child (e.g., expecting their child to sit in their chair during the course of the interaction). The

structure scale assesses the parent’s capacity to organize the activities and lead the child in such a way

that the child is able to accomplish tasks and feel positive about his or her experience. Finally, the

confidence in parental role scale assesses how confidently and effectively the parent works with the

child, including setting limits around child misbehavior. In order to create a single rating of

demandingness, parent’s scores on limit-setting, maturity demands, structure, and confidence were again

standardized and summed. This dimension was internally consistent for mothers (alpha = .75; range of

correlations = .29 to .62) and fathers (alpha = .66; range of correlations = .22 to .44).

Finally, one variable in the coding system assesses autonomy-support, parental respect for child

autonomy. This scale captures parents’ attempts to both allow and actively encourage the child to take a

leadership role during the interaction, by, for example, asking the child what he/she would like to do

first, and checking with the child before offering assistance on a task.

2.2.2. Assessment of parental scaffolding behavior

Microanalytic ratings of scaffolding behavior were made from one task during the whole parent–child

interaction, a 10-min long-division or word problem task in which the parent was instructed to help the

child as much or as little as needed. Because mothers and fathers followed each other in doing the

homework task, we used different types of division problems for each interaction. The first was a set of

six standard long-division problems (with one to two digit divisors and two to four digit dividends, e.g.,

760 divided by 12). The second was a set of five word problems involving division. An example of a

word problem was bAlbert has 216 pieces of candy and he wants to give an equal number of them to the 32 friends at his birthday party. How many pieces of candy will he have left over?Q Pratt and Savoy- Levine (1998) showed that parental scaffolding behavior predicted subsequent child success on both of

these kinds of math problems. Problem-type for each mother–child and father–child interaction was

counterbalanced throughout the data collection.

The videotaped interactions were transcribed by the coders and used along with the child’s long-division

worksheet. The interaction between the parent and child in solving the problems was fluid and dyadic, with

the parent sometimes initiating the interaction by a gesture or comment or with the child initiating the

interaction by simply starting the problem without the parents’ help. However, to develop a consistent

coding system, and following Pratt, Green et al. (1992), we divided each parent–child interaction sequence

into a discrete teaching episode, consisting of a parental behavior followed by a child response. Because we

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 93

were primarily interested in children’s responses to parental directives and behaviors, the parent’s behavior

was always coded first followed by the child’s response. If the child began solving the problems with no

direction from the parent, the parental (non) behavior would be rated as a d0T in the coding system described below followed by a child response (either success or failure depending on whether the child was

successfully solving the problem on their own). Because long-division mathematics consists of three core

subskills, i.e., estimation, multiplication, and subtraction, each episode was classified according to its

content as an estimation, multiplication, or subtraction step. Occasionally, the parent or child would make a

comment that was not related to completing the homework assignment, such as a comment about what they

were going to do after the visit was over, or that they were hungry. These comments were not coded because

they could not easily be classified as a helping or teaching comment. There was an average of 39 (SD = 12)

teaching episodes for mother–child interactions, and 46 (SD = 18) teaching episodes for father–child

interactions. These means differed significantly, t(60) = 3.04, p b 0.01, suggesting that father–child

interactions consisted of more teaching episodes than mother–child interactions. It is possible that children

were slightly less likely to discuss non-homework related topics with their fathers than with their mothers,

thereby yielding more teaching episodes for fathers.

Next, parental behaviors for each episode were scored using a nine level coding system developed by

Pratt, Green et al. (1992) (see Table 1). The primary goal of this coding system is to rate each parental

behavior for the amount and specificity of help offered by the parent in solving the problem. Parental

behaviors could take a variety of forms, ranging from simple hand gestures to initiate the problem to

complex explanations of how to do long division. The key issue for coding purposes was the type and

amount of information conveyed. At the low end of the coding scale (levels 0–2), parents provide

minimal help by either saying nothing, using simple nonverbal gestures to begin or continue the

problem, or providing very general statements about how to start or continue the problem step (see

examples in Table 1). Scores in the mid-range of the coding scale (levels 3–5) indicate that the parent

Table 1

Examples a of parental levels of intervention for task components of the long-division problems

Level Task component

Estimation Multiplication Subtraction

0 No directive No directive No directive

1 General verbal/nonverbal

(bTry this oneQ) General verbal/nonverbal (gestures

with hand to begin problem)

General verbal/nonverbal

(bDo it pleaseQ) 2 General hints (bHow many timesQ) General hints (What’s next?Q) General hints (bWhat do we doQ) 3 Label subcomponent

(bDivide themQ) Label subcomponent

(bMultiply themQ) Label subcomponent

(bSubtract themQ) 4 Specify step of component (bHow

many times does 2 go into 4Q) Specify step of component

(bMultiply 41 by 6Q) Specify step of component

(bSubtract 182 from 326Q) 5 Hint about step

(bThat looks like too manyQ) Hint about step (bThat is too bigQ) Hint about step

(bI don’t think you did that rightQ) 6 Give step answer or recording

(bIt’s 2 timesQ) Give step answer or recording

(bPut it hereQ) Give step answer or recording

(b8 minus 3 is 5Q) 7 Give step answer and recording

(bIt’s 5, put it hereQ) Give step answer and recording

(example similar)

Give step answer and recording

(example similar)

8 Parent demonstrates Parent demonstrates Parent demonstrates

a From Pratt, Green, Mac Vicar, and Bountrogianni (1992). Reprinted by permission.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10694

offers some help but is definitely not doing the problem step for the child. For example, the parent might

break the problem down into smaller steps (such as bHow many times does 2 go into 4Q) or provide a specific hint, such as telling the child they have done a step of the problem incorrectly. Finally, at the

high end of the coding system (levels 6–8) parents actually do part or all of the problem step for the

child, e.g., telling the child the answer, indicating where to record the answer, or doing the entire

problem step.

Finally, the child’s response to each parent’s behavior was coded as either successful or unsuccessful,

following guidelines in Pratt, Green et al. (1992) and Pratt, Kerig et al. (1992). A response was considered

successful if the child moved forward correctly in solving the problem, and unsuccessful if the child did

something incorrect or did nothing in response to the parent’s statement. To evaluate the reliability of the

coding system, two undergraduate coders rated all the data gathered from the study and percent agreements

were calculated between their exact scores. Results were as follows: for mother–child interactions, percent

exact agreement for segmentation of the transcript was 86.8% (j = 0.63), agreement for coding of parental statements was 74.3% (j = 0.70), and agreement for coding of child responses was 90.7% (j = 0.80). For father–child interactions, segmentation agreement was 86.3% (j = 0.65), coding of parental statements was 74.5% (j = 0.70), and coding of child responses was 87.5% (j = 0.75).

To effectively summarize the pattern of parent scaffolding, we focused our assessment on a general

index applicable to this type of problem-contingent shift ratio (CS ratio) introduced by Pratt, Green et al.

(1992). Based on the coding of parental statements into one of the nine levels and the coding of child

responses as successful or unsuccessful, the CS ratio was calculated by dividing the number of

contingent shifts made by the parent by the total number of shifts made throughout the transcript. A

parental shift was considered contingent when it followed the rule of providing more specific help

immediately after a child failure or less specific help immediately after a child success (Wood &

Middleton, 1975). Shifts of greater than three levels (e.g., from levels 1 to 5) were considered extreme

and hence noncontingent, even if they were in the right direction, following Pratt, Green et al. (1992).

Levels 0 and 1 (see Table 1) were not distinguished for purposes of calculating the CS ratio, and staying

at level 0 or l after a child success was considered contingent, as no movement to less specificity was

possible.

2.2.3. Measures of children’s academic competence

2.2.3.1. Child success at the long-division task. As an initial assessment of the child’s academic

competence, we calculated the percentage of successful responses made by the child across the long-

division task. Although this measure has been used in past studies to assess children’s academic

performance on the task itself (e.g., Pratt, Green, et al., 1992; Pratt, Kerig, et al., 1988), it is more

accurately a measure of the parent and child’s mutual success at completing the task.

2.2.3.2. Math achievement. To assess the child’s overall math skills, children completed the Peabody

Individual Achievement Test (PIAT-R; Markwardt, 1989) at the end of first and fourth grade, which

contains subscales on mathematics, reading, spelling, and general fund of knowledge. The mathematics

subtest uses a multiple-choice format to assess a wide range of math skills, from early matching and

numeral recognition abilities to advanced concepts of geometry and trigonometry. Children begin with

items that are easily mastered and continue until they fail consistently. The PIAT is a well-validated and

reliable instrument that has been shown to correlate positively with school performance and with other

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 95

measures of achievement and intelligence, such as the Wide-Range Achievement Test and the WISC-R

(Sattler, 1992).

2.2.3.3. Teacher-rated academic competence. To assess the child’s academic competence in school,

teachers completed the Child Adaptive Behavior Inventory (CABI) on all students in their

classroom in the spring of fourth grade (unaware of the specific target child in our study). This

inventory consists of 106 five-point Likert scale items adapted by Cowan and Cowan (1990) from

Schaefer and Hunter’s (1983) 60-item scale, with 17 additional items from the downward extension

of Quay and Peterson’s Behavior Problem Checklist (1983), and 23 items from Achenbach and

Edelbrock’s (1987) Child Behavior Checklist (CBCL). The Cowans created six additional items

specifically for the SAF Project. Teachers’ scores were converted to z-scores to represent the

behavior of the target child relative to other same-sex children in the classroom, thereby controlling

for teachers’ overall positive or negative bias and for different classroom compositions. The

subscales of the CABI have shown good internal consistency reliability (average Cronbach’s alpha = 0.81;

range = 0.66–0.90) and have been validated in a number of studies (Cowan et al., 1994; Katz & Gottman,

1993).

The Academic Competence factor, one of six factors that has emerged in repeated factor analyses of

the instrument, is composed of the following three 5-item scales: (a) Intelligence Scale—this scale

assesses the child’s academic potential as displayed in the classroom. Sample items are bis smart for his/ her ageQ; bunderstands difficult wordsQ; and blearns new skills to cope with new situations or problemsQ; (b) Creativity Scale—this scale assesses the child’s novel and creative approach to academic tasks.

Sample items include bis always asking questionsQ; bthinks up interesting things to doQ; and bhas good ideas about how to do things or make thingsQ; (c) Task Orientation Scale—this scale assesses how interested and motivated the child is to do well at academic tasks. Items include bdoesn’t seem interested in things (reverse scored)Q; bworks carefully and does his/her bestQ; and boften works quietly at an activity for a long timeQ. The current study focused on the factor as a whole, as the three scales that make it up are highly intercorrelated (overall alpha = 0.85).

2.2.3.4. Self-reported academic competence. The child’s self-perceived academic competence was

assessed the summer after fourth grade using the Academic Competence subscale of Harter’s (1985)

Self-Perception Profile for Children. The scale contains six items, rated on a 1–4 scale. A sample item is:

bSome kids feel they are very good at their school work, but other kids worry about whether they can do the school work assigned to themQ. Children’s answers to the six items were averaged to form one score for self-reported academic competence. Harter has reported good reliability for this subscale (0.80–0.85)

in elementary school children. The scale showed good internal consistency for our sample, as well,

Cronbach’s alpha = 0.84.

3. Results

3.1. Descriptive statistics

Means (and SDs) for all parent predictor and child outcome variables are presented in Table 2. To

compare mothers and fathers on their average levels of authoritative parenting and scaffolding behavior,

Table 2

Mean (and SD) scores for mothers’ and fathers’ behaviors and for child competence variables

Variables Mothers Fathers

M SD M SD

Parenting predictors

Responsiveness (range = 1–5) 3.28 0.54 3.15 0.64

Demandingness (range = 1–5) 3.37 0.41 3.44 0.39

Autonomy-support (range = 1–5) 3.43 0.60 3.32 0.64

Use of CS principle 50.9% 13.3% 48.3% 12.7%

Child outcome measures

Child success on long-division task 62.3% 14.8% 60.5% 13.9%

Math achievement (range = 82–145) 118.30 16.17

Teacher-rated academic competence a

0.04 2.63

Self-reported academic competence b

3.28 0.56

a Standard score.

b Range = 1–4.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10696

four 2-way ANOVAs were computed, using, in turn, responsiveness, demandingness, autonomy-

support, and percent usage of the CS principle as dependent variables, with parent sex as a within-

subjects variable, and child gender as a between-subjects variable. Results indicated no mean differences

between the parents on the three dimensions of authoritative parenting or on the measure of scaffolding

behavior. Consistent with past studies (Conner, Knight, & Cross, 1997; Pratt, Kerig, Cowan, & Cowan,

1988;), these data suggest that mothers and fathers were equally effective in their scaffolding of their

children’s learning of a cognitive task.

Table 3 shows the average percentages of success by the child following each level of maternal

intervention defined in the coding hierarchy, for each of the three separate components of the math

problem. Table 4 shows the same information following paternal interventions. According to the

principles of sensitive scaffolding, parents should carefully adjust the amount of help they provide

Table 3

Percentage of success by scaffolding level and task component for mothers

Level Task component

Estimation Multiplication Subtraction

0 58 (190) a

88 (105) 82 (124)

1 61 (307) 79 (66) 72 (50)

2 50 (362) 57 (42) 50 (34)

3 48 (141) 65 (17) 74 (66)

4 58 (395) 75 (104) 69 (67)

5 61 (234) 60 (43) 58 (31)

6 76 (120) 78 (50) – (5)

7 – (4) – (5) – (0)

8 b

– (10) – (13) – (8)

a Numbers of interventions at each level in parentheses; no success percentages calculated for levels with fewer than 10

interventions. b Children do not respond for level 8 demonstrations so there was no success score possible.

Table 4

Percentage of success by scaffolding level and task component for fathers

Level Task component

Estimation Multiplication Subtraction

0 59 (192) a

90 (70) 89 (93)

1 55 (357) 73 (44) 79 (67)

2 53 (376) 54 (46) 75 (88)

3 52 (162) 52 (23) 74 (36)

4 57 (425) 70 (94) 73 (80)

5 49 (278) 48 (33) 51 (35)

6 66 (142) 81 (72) – (3)

7 – (8) – (4) – (0)

8 b

– (56) – (27) – (11)

a Numbers of interventions at each level in parentheses; no success percentages calculated for levels with fewer than 10

interventions. b Children do not respond for level 8 demonstrations so there was no success score possible.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 97

according to the child’s ongoing need for help across the problem (Wood, 1988). As can be seen from

the tables, children were relatively successful across the different amounts of help the parent offered,

suggesting that the parents were providing the amount of help the child needed to succeed much of the

time, without providing so much help that the child could never learn from his or her mistakes and move

towards independent mastery over the problem. Across the different subcomponents and levels of the

hierarchy, children’s average success rates were 66% with their mothers and 65% with their fathers.

These percentages were not significantly different from each other (t = 0.238, p N 0.10), suggesting that

children were equally successful with their mothers and fathers.

3.2. Correlations between authoritative parenting, parental scaffolding behavior, and children’s

academic outcomes

Table 5 presents correlations between the three dimensions of authoritative parenting, scaffolding

behavior, and the four child outcome variables. The first hypothesis of this study was that the

three dimensions of authoritative parenting, rated globally across the entire parent–child interaction,

would be associated but not redundant with displays of more effective scaffolding behavior during

the homework task. As seen in Table 5, mothers’ use of the CS principle was positively correlated

with all three dimensions of authoritative parenting, especially with autonomy-support. Fathers’ use

of the CS principle was also significant and positive with autonomy-support, but not with the other

two authoritative parenting dimensions. These results suggest that mothers who are responsive,

demanding, and support autonomy during a parent–child interaction are also more likely to scaffold

their child’s learning of a cognitive task (Pratt, Green, et al., 1992; Pratt et al., 1988). For fathers,

it appears that autonomy-support is the dimension of authoritative parenting most closely linked to

scaffolding behavior during a cognitive task.

The second hypothesis of this study was that the three dimensions of authoritative parenting and

scaffolding behavior would each be associated with children’s academic competence. Surprisingly,

mothers’ and fathers’ responsiveness and demandingness were not significantly correlated with child

success on the task, math achievement, teacher reports of children’s academic competence, or

Table 5

Correlations between authoritative parenting, parental scaffolding behaviors, and children’s academic competence scores

Variable 1 2 3 4 5 6 7 8 9 10 11 12 13

(1) Mother’s

responsiveness

– 0.58***0.40***0.43***0.40***0.36** 0.25* 0.02 0.08 0.03 0.21 0.12 0.10

(2) Mother’s

demandingness

– 0.19 0.36***0.43***0.33** 0.30* 0.07 0.11 0.09 0.18 0.16 0.09

(3) Mother’s

autonomy-support

– 0.22 0.31* 0.26* 0.36** 0.29*0.13 0.21 0.50*** 0.18 0.06

(4) Father’s

responsiveness

– 0.61***0.47***0.19 0.16 0.05 0.15 0.22 0.15 0.06

(5) Father’s

demandingness

– 0.31* 0.33** 0.14 0.06 0.20 0.12 0.22 0.13

(6) Father’s

autonomy-support

– 0.44***0.28*0.18 0.11 0.32** 0.22 0.07

(7) Mother’s use of

CS principle

– 0.25 0.38**0.28* 0.27* 0.39** 0.43***

(8) Father’s use of

CS principle

– 0.20 0.47***0.29* 0.23 0.18

(9) Child success on

long-division task

with mother

– 0.29* 0.13 �0.07 0.11

(10) Child success

on long-division

task with father

– 0.30* 0.28* 0.34**

(11) Math

achievement on

the PIAT

– 0.43***0.26*

(12) Teacher-

reported academic

competence

– 0.26*

(13) Child-reported

academic

competence

–

*p b 0.05. **p b 0.01. ***p b 0.001; all two-tailed tests.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–10698

children’s own reports of their academic competence. Maternal and paternal autonomy-support were

linked with greater math achievement by children in fourth grade. By contrast, mothers’ use of the

CS principle was positively correlated with the children’s success on the task, children’s math

achievement on the PIAT, and with teacher and children’s reports of academic competence. Fathers’

use of the CS principle was correlated significantly with greater success on the task and with math

achievement on the PIAT.

Looking across parents, the association between parental use of the CS principle and child-reported

academic competence tended to be somewhat stronger for mothers than for fathers [r(63) = 0.43 for

mothers versus r(60) = 0.18 for fathers; z = 1.53, p b 0.07). Despite the fact that mothers and fathers

were equally effective in their scaffolding behavior, mothers’ scaffolding performance was somewhat

more closely associated with children’s expressed confidence in their own academic skills than was

fathers’ scaffolding performance.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 99

3.3. Regression analyses predicting fourth grade academic competence from authoritative parenting,

parental scaffolding behavior, and prior math achievement

The third hypothesis of this study was that parental scaffolding behavior would predict children’s

academic competence in fourth grade, over and above the effects of authoritative parenting and

children’s prior math achievement, as assessed in first grade. A series of hierarchical regression

equations were used to test this hypothesis. For each dependent variable (child success on long-division

task, math achievement in fourth grade, teacher-rated academic competence, and child-reported

academic competence), math achievement on the PIAT in first grade was entered on step one; parental

responsiveness, demandingness, and autonomy-support were entered as a block on step two; and

parental use of the CS principle were entered on step three.

Table 6 presents results of these regression equations for mothers’ parenting variables. For child

success at the task, only mothers’ use of the CS principle added unique variance on the third step of the

equation. The child’s math achievement in grade one strongly predicted grade four math achievement in

the first step of the second equation, suggesting that individual differences in children’s basic math

Table 6

Summary of regression analyses predicting academic competence from math achievement in first grade, maternal authoritative

parenting, and maternal scaffolding behavior

Dependent variables and step predictors entered Beta t Increment in R 2

Total R 2

Child success on long-division task

(1) Grade one math achievement 0.00 0.042 0.00 0.00

(2) Responsiveness 0.13 0.703 0.05 0.05

Demandingness 0.06 0.317

Autonomy-support 0.09 0.521

(3) Use of the CS principle 0.37* 2.587* 0.11* 0.16

Grade four math achievement

(1) Grade one math achievement 0.66*** 6.853*** 0.55*** 0.55***

(2) Responsiveness 0.02 0.143 0.04 0.59***

Demandingness 0.06 0.511

Autonomy-support 0.17 1.530

(3) Use of the CS principle 0.03 0.275 0.00 0.59***

Teacher-rated academic competence

(1) Grade one math achievement 0.37*** 2.868*** 0.18*** 0.18***

(2) Responsiveness �0.04 �0.241 0.02 0.20*** Demandingness 0.08 0.526

Autonomy-support �0.04 �0.294 (3) Use of the CS principle 0.30* 2.217* 0.07* 0.27***

Self-reported academic competence

(1) Grade one math achievement 0.39*** 3.212*** 0.17*** 0.17***

(2) Responsiveness 0.07 0.472 0.03 0.20***

Demandingness 0.05 0.318

Autonomy-support �0.30* �2.089* (3) Use of the CS principle 0.43*** 3.476*** 0.15*** 0.35***

*pb0.05. ***pb0.001.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106100

abilities are quite stable over time. Mothers’ authoritative parenting and use of the CS principle during

the homework task did not add additional variance to the prediction of grade four math achievement over

the effects of prior math achievement in grade one. For both teacher and self-reported academic

competence, math achievement in grade one accounted for significant variance on step one of those

equations, authoritative parenting added no additional explanatory variance on step two, but mothers’

use of the CS principle added significant variance on step three. These results suggest that mothers’

scaffolding performance was an independent predictor of the child’s academic competence on the long-

division task itself and on academic competence as rated by teachers and the children (though not math

achievement, per se), over and above the effects of earlier math achievement and authoritative parenting.

Table 7 presents the findings of parallel regression equations using fathers’ parenting variables as

predictors of child outcome variables. In the first equation, fathers’ scaffolding behavior added

significant unique variance to the prediction of the child’s success at the long-division task, over and

above the significant effects of grade one math achievement. Fathers’ authoritative parenting and

effective scaffolding (greater use of the CS principle) scores did not add to the ability to predict grade

four math achievement, teacher-reported academic competence, or child-reported academic competence

Table 7

Regression analyses predicting academic competence from math achievement in first grade, paternal authoritative parenting,

and paternal scaffolding behavior

Dependent variable and step predictors entered Beta t Increment in R 2

Total R 2

Child success on long-division task

(1) Grade one math achievement 0.33** 2.728** 0.18** 0.18**

(2) Responsiveness 0.05 0.307 0.02 0.20*

Demandingness 0.13 0.842

Autonomy-support �0.04 �0.273 (3) Use of the CS principle 0.41** 3.290** 0.17** 0.37**

Grade four math achievement

(1) Grade one math achievement 0.62*** 6.137*** 0.46*** 0.46***

(2) Responsiveness 0.14 1.049 0.07 0.53***

Demandingness �0.14 �1.145 Autonomy-support 0.20 1.654

(3) Use of the CS principle 0.09 0.838 0.01 0.54***

Teacher-rated academic competence

(1) Grade one math achievement 0.47** 3.477** 0.25*** 0.25***

(2) Responsiveness �0.04 �0.208 0.05 0.30** Demandingness 0.13 0.850

Autonomy-support 0.18 1.121

(3) Use of the CS principle �0.01 �0.091 0.00 0.30**

Self-reported academic competence

(1) Grade one math achievement 0.32* 2.376* 0.14** 0.14**

(2) Responsiveness �0.08 �0.424 0.03 0.17* Demandingness 0.13 0.824

Autonomy-support 0.12 0.743

(3) Use of the CS principle 0.12 0.880 0.01 0.18

*p b 0.05. **p b 0.01. ***p b 0.001.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 101

at the end of fourth grade, after accounting for the significant effects of first grade math achievement.

Thus, it appears that fathers’ scaffolding behavior is an important independent predictor of the child’s

immediate success at the task being taught but not of the child’s overall academic competence in fourth

grade, over and above the effects of the child’s earlier academic achievement.

4. Discussion

The current study examined the links between parental behaviors and 10-year-old children’s academic

competence at two very different levels of analysis. At the macro-level, we examined dimensions of

authoritative parenting as those related to teacher-rated and self-reported academic competence in fourth

grade. At the micro-level, we examined parents’ sensitive scaffolding of a math homework assignment

and links between that parental scaffolding behavior and children’s academic outcomes. Additionally, we

paid attention to the relationship between a general authoritative parenting style and specific parental

scaffolding behaviors.

Addressing this last issue first, we found that dimensions of authoritative parenting were moderately

correlated with parental scaffolding behaviors. Consistent with past research (Pratt, Green, et al., 1992;

Pratt et al., 1988), these findings support the idea that parents who are generally warm, supportive,

demanding, and able to encourage children’s growing autonomy will appear more sensitive to their

child’s specific needs for help during a cognitive task. The modest nature of these correlations suggests,

however, that authoritative parenting and scaffolding behavior are by no means redundant measures of

the same behavioral construct. Interestingly, in our analyses, it was the dimension of authoritative

parenting assessing parental support for autonomy that was most strongly associated with sensitive

scaffolding, supporting the idea that parents who value their child’s independent mastery over difficult

tasks are most likely to use a scaffolding approach to tutoring cognitive tasks (cf., Wood, 1988).

Although the three dimensions of authoritative parenting focused on in this study were associated

with scaffolding behavior, we found very limited relationships between these three dimensions and

children’s academic outcomes. Only maternal and paternal autonomy-support was associated with

greater academic achievement on the PIAT. This finding is consistent with past research suggesting

that a lack of intrusiveness by parents is linked with greater academic success by children (Pomerantz

& Eaton, 2001). The general lack of association between dimensions of authoritative parenting and

children’s academic competence was unexpected, and, given the well documented links between

authoritative parenting and children’s academic competence during this time period (Grolnick &

Ryan, 1989; Mattanah, 2001; Steinberg et al., 1989, 1994) calls into question the validity of our

procedures for assessing authoritative parenting. It is possible that our laboratory procedures did not

include tasks in which certain dimensions of authoritative parenting would become most salient, such

as setting limits or demanding compliance. It is also possible that dimensions such as warmth,

responsiveness, and setting limits are quite subtle in the parent–child interactions of late childhood-

aged children and their parents, and were not salient to observers during a laboratory interaction. In

support of this idea, many studies finding links between authoritative parenting and academic

competence during late childhood and adolescence assess the parenting dimensions using child self-

report rather than laboratory procedures, perhaps providing a more psychologically salient insider’s

perspective on the parent–child relationship (Barber & Olsen, 1997; Herman et al., 1997; Steinberg et

al., 1989, 1994).

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106102

Despite the limited relations found between authoritative parenting and children’s academic

competence, the current study did find significant associations between mothers’ and fathers’ scaffolding

behavior and children’s academic performance in the immediate term, as shown by the significant

correlations between parental scaffolding behavior and child success at the task. Additionally, mothers’

scaffolding behavior was significantly associated with academic competence in the longer term, as

shown by the links between scaffolding behavior and teacher- and child-report of academic abilities.

These findings suggest that the beneficial effects of scaffolding may not be limited to immediate gains in

the learning of a specific task (Pratt, Green, et al., 1992; Wood & Middleton, 1975), but may extend to

the achievement of broader academic goals desired by teachers and parents. It is possible that parents

who routinely teach their children tasks using a scaffolding method boost their children’s self-esteem

regarding academic tasks and, in turn, help their children remain motivated to succeed at school. In

support of the claim that scaffolding would be related to child motivation, we found that maternal

scaffolding behavior correlated equally strongly with the task-orientation subscale of teacher-reported

academic competence (r = 0.31) as with the intelligence subscale of this composite factor (r = 0.39).

These findings are consistent with the more informal observations of Wood (1988), who reported that

experiences of more effective scaffolding appeared to encourage young children’s mastery motivation. In

the present study, however, these results were demonstrated in a much larger sample of school-aged

children and in real classroom settings across a period of several months. The applied implications of

these results are substantial.

The current study also extends work that has examined differences between mothers’ and fathers’

scaffolding behavior (Conner, Knight, & Cross, 1997; Pratt, Kerig, Cowan, & Cowan, 1988;).

Consistent with past studies of younger children and parents, we found no mean differences between

mothers’ and fathers’ scaffolding patterns in teaching long-division math to their preadolescent children.

Fathers are as skilled as mothers in this domain of helping their children with math homework. Similar to

Conner et al., we found that mothers’ scaffolding behavior was somewhat more closely associated with

children’s self-reported academic competence than was fathers’ scaffolding behavior. Additionally,

unlike mothers’ scaffolding behavior, fathers’ scaffolding did not add unique variance to the prediction

of teacher-reported or self-reported academic competence, after controlling for first-grade math

achievement. Perhaps mothers in this sample are more involved in their children’s day-to-day homework

assignments than are fathers, and thus their scaffolding behavior may have greater effects on children’s

academic competence in the classroom than fathers’ scaffolding behavior.

Finally, in the current study, we found that effective scaffolding behavior was a unique predictor of

children’s academic competence, beyond the effects of authoritative parenting. This result suggests that

scaffolding, although associated with authoritative parenting, is a unique correlate of greater academic

competence in school-aged children. However, it is difficult to draw this conclusion definitively in the

current analyses, given that the dimensions of authoritative parenting focused on in this study did not

predict the academic outcome variables.

4.1. Limitations of the study

Although these findings suggest important links between scaffolding behavior and children’s

academic competence, a number of methodological limitations should be noted. First, given that this

study included highly educated parents of children with above average academic abilities, it is hard to

know whether a teaching style like scaffolding that places great emphasis on independent functioning is

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 103

applicable to children who have greater difficulties with academic tasks or to parents who are less well

educated themselves, and therefore may have difficulty scaffolding a task like long division. Not

surprisingly, raters and experimenters in this study anecdotally noted that parents who had a poorer grasp

of the concepts of long division had difficulty following the principles of scaffolding instruction for the

present task.

Second, in cases with a substantial number of 0 and 1 level interventions, in which the parent said

very little while the child completed the task largely on his or her own, the parent may have received a

high score on use of the CS principle just because the child was already fairly competent at the task. In

these cases, the children’s academic competence, and not the parent’s scaffolding behavior, could have at

least partly accounted for the correlation between scaffolding and competence at school. This potential

confound between parent and child effects calls into question the claim that scaffolding behavior

promotes academic competence in the classroom. Nevertheless, the results of the regression analyses,

specifically those for mothers, suggest that scaffolding behavior adds power to the prediction of

children’s academic competence beyond direct assessments of child performance on the PIAT

instrument.

Third, authoritative parenting and scaffolding behavior were rated from the same parent–child

interaction, increasing the likelihood that a correlation would be found between these two sets of

variables. However, because these variables were rated at two very different levels of analyses by

different coders, using microanalytic analysis in the case of scaffolding behavior and global ratings in the

case of authoritative parenting, it is significant that they were positively correlated, and that mother’s

scaffolding behavior accounted for unique variance in educational variables, over and above the effects

of authoritative parenting.

Finally, it should be noted that the parent–child interaction used in this study focused on a very

specific task, long-division mathematics. One should use caution in generalizing from this specific task

to statements about the general effects of scaffolding behavior on children’s academic competence.

Parents’ confidence level and abilities in math no doubt influence their ability to be effective in this task,

and other tasks may bring out more or less effective teaching strategies for some parents. Although this

study found some links between scaffolding during this task and children’s subsequent academic

competence, the modest level of these correlations may be related to the specific nature of the task.

Future studies should consider other types of tutoring interactions to determine whether there are specific

links between scaffolding around certain tasks and children’s competence at those tasks or whether a

more general parental teaching style exists across different tasks (cf. Pratt et al., 1988).

4.2. Conclusion

Results of the current study have important implications for our understanding of the role parents play

in promoting their children’s academic competence at school. It is clear that parents who set high

standards for their children’s achievement remain involved in their children’s school lives, and create a

home environment of warmth and structure which help their children succeed in the classroom

(Baumrind, 1991; Steinberg, 2001). However, as shown in this study, parents also may promote

children’s academic competence by the specific manner in which they help their children learn new

academic material. One of the key ways in which parents support children’s schooling during the middle

elementary years is through homework. The results of this study suggest that when parents get involved

with helping their children with homework, they must simultaneously provide their children the

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106104

guidance they require to solve the problem, vary their level of involvement, and leave room for their

children to develop mastery of the problem. Parents who use such a scaffolding approach to teaching

their children appear to have children who feel more confident about their own academic abilities and are

seen as more academically motivated and competent in the classroom by their teachers.

Acknowledgements

Work on this project was supported by the National Institute of Mental Health Grant MH 31109,

awarded to Philip A. Cowan and Carolyn Pape Cowan, and by a National Research Service Award F31-

MH11467, awarded to Jonathan Mattanah. The authors would like to thank Stephen Hinshaw and

Allyson Mattanah for their feedback on earlier versions of this article. We also acknowledge the

enthusiastic participation of the parents and children in the Schoolchildren and their Families Project,

without whom this research would not have been possible.

References

Achenbach, T. M., & Edelbrock, C. (1987). Manual for the youth self-report and profile. Burlington7 University of Vermont,

Department of Psychiatry.

Barber, B. K., & Olsen, J. A. (1997). Socialization in context: Connection, regulation, and autonomy in the family, school, and

neighborhood, and with peers. Journal of Adolescent Research, 12, 287–315.

Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology Monographs, 4, 1–103.

Baumrind, D. (1991). Effective parenting during the early adolescent transition. In P. A. Cowan, & E. M. Hetherington (Eds.),

Advances in family research (Vol. 2, pp. 111–163) Hillsdale, NJ7 Erlbaum.

Chao, R. K. (2001). Extending research on the consequences of parenting style for Chinese American and European Americans.

Child Development, 72, 1832–1843.

Conner, D. B., & Cross, D. R. (2003). Longitudinal analysis of the presence, efficacy and stability of maternal scaffolding

during informal problem-solving interactions. British Journal of Developmental Psychology, 21, 315–334.

Conner, D. B., Knight, D. K., & Cross, D. R. (1997). Mothers and fathers’ scaffolding of their 2-year-olds during problem-

solving and literacy interactions. British Journal of Developmental Psychology, 15, 323–338.

Cowan, P. A., & Cowan, C. P. (1990). Becoming a family: Research and intervention. In I. Sigel, & G. Brody (Eds.), Methods of

family research: Biographies of research projects (pp. 1–51). Hillsdale, NJ7 Erlbaum.

Cowan, P. A., & Cowan, C. P. (1992). Schoolchildren and their families project: Description of parenting style ratings.

Unpublished manuscript, University of California at Berkeley.

Cowan, P. A., Cowan, C. P., Schulz, M., & Heming, G. (1994). Prebirth to preschool family factors predicting children’s

adaptation to kindergarten. In R. Parke, & S. Kellam (Eds.), Exploring family relationships with other social contexts:

Advances in family research (Vol. 4, pp. 75–114) Hillsdale, NJ7 Erlbaum.

Darling, N., & Steinberg, L. (1993). Parenting style as context: An integrative model. Psychological Bulletin, 113, 487–496.

Goldberg, M. D., & Cornell, D. G. (1998). The influence of intrinsic motivation and self-concept on academic achievement in

second- and third-grade students. Journal for the Education of the Gifted, 21, 179–205.

Gray, M., & Steinberg, L. (1999). Unpacking authoritative parenting: Reassessing a multidimensional construct. Journal of

Marriage and the Family, 61, 574–587.

Grolnick, W. S., & Ryan, R. M. (1989). Parent styles associated with children’s self-regulation and competence in school.

Journal of Educational Psychology, 81, 143–154.

Harter, S. (1985). Manual for the self-perception profile for children. Denver, CO7 University of Denver.

Harter, S. (1992). The relationship between perceived competence, affect, and motivational orientation within the classroom:

Processes and patterns of change. In A. K. Boggiano, & T. S. Pittman (Eds.), Achievement and motivation: A social-

developmental perspective (pp. 77–114). New York7 Cambridge Univ. Press.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106 105

Herman, M. R., Dornbusch, S. M., Herron, M. C., & Herting, J. R. (1997). The influence of family regulation, connection, and

psychological autonomy on six measures of adolescent functioning. Journal of Adolescent Research, 12, 34–67.

Johnson, V., Cowan, P. A., & Cowan, C. P. (1999). Children’s classroom behavior: The unique contribution of family

organization. Journal of Family Psychology, 13, 355–371.

Katz, J. F., & Gottman, J. M. (1993). Patterns of marital conflict predict children’s internalizing and externalizing behaviors.

Developmental Psychology, 29, 940–950.

Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family: Parent–child interaction. In P.

Mussen, & E. M. Hetherington (Eds.), Handbook of child psychology: vol. 4. Socialization, personality, and social

development (4th ed.) (pp. 1–101). New York7 Wiley.

Markwardt, F. C. (1989). Peabody individual achievement test—revised. Circle Pines, MN7 American Guidance Service.

Mattanah, J. F. (2001). Parental psychological autonomy and children’s academic competence and behavioral adjustment in late

childhood: More than just limit-setting and warmth. Merrill-Palmer Quarterly, 47, 355–376.

Miller, N. B., Cowan, P. A., Cowan, C. P., Hetherington, E. M., & Clingempeel, W. G. (1993). Externalizing in preschoolers

and early adolescents: A cross-study replication of a family model. Developmental Psychology, 29, 3–18.

Miserandino, M. (1996). Children who do well in school: Individual differences in perceived competence and autonomy in

above-average children. Journal of Educational Psychology, 88, 203–214.

Neitzel, C., & Stright, A. (2003). Mother’s scaffolding of children’s problem-solving: Establishing a foundation of academic

self-regulatory competence. Journal of Family Psychology, 17, 147–159.

Park, H. -S. (2002). Parenting practices, ethnicity, socioeconomic status and academic achievement in adolescents. School

Psychology International, 23, 386–395.

Plumert, J. M., & Nichols-Witehead, P. (1996). Parental scaffolding of young children’s spatial communication. Developmental

Psychology, 32, 523–532.

Pomerantz, E. M., & Eaton, M. M. (2001). Maternal intrusive support in the academic context: Transactional socialization

processes. Developmental Psychology, 37, 174–186.

Pratt, M. W., Green, D., MacVicar, J., & Bountrogianni, M. (1992). The mathematical parent: Parental scaffolding, parenting

style, and learning outcomes in long-division mathematics homework. Journal of Applied Developmental Psychology, 13,

17–34.

Pratt, M. W., Kerig, P., Cowan, P. A., & Cowan, C. P. (1988). Mothers and fathers teaching 3-year-olds: Authoritative parenting

and adult scaffolding of young children’s learning. Developmental Psychology, 24, 832–839.

Pratt, M. W., Kerig, P. K., Cowan, P. A., & Cowan, C. P. (1992). Family worlds: Couple satisfaction, parenting style, and

mothers’ and fathers’ speech to young children. Merrill-Palmer Quarterly, 38, 245–262.

Pratt, M. W., & Savoy-Levine, K. M. (1998). Contingent tutoring of long-division skills in fourth and fifth

graders: Experimental tests of some hypotheses about scaffolding. Journal of Applied Developmental Psychology, 19,

287–304.

Quay, H. C., & Peterson, D. R. (1983). Interim manual for the revised behavior problem checklist. Coral Gables, FL7 University

of Miami, Applied Social Sciences.

Radziszewska, B., & Rogoff, B. (1988). Influence of adult and peer collaborators on the development of children’s planning

skills. Developmental Psychology, 24, 840–848.

Radziszewska, B., & Rogoff, B. (1991). Children’s guided participation in planning imaginary errands with skilled adults or

peer partners. Developmental Psychology, 27, 381–389.

Rogoff, B. (1990). Apprenticeship in thinking: Cognitive development in social context. New York7 Oxford Univ. Press.

Rogoff, B., Ellis, S., & Gardner, W. (1984). Adjustment of adult–child instruction according to child’s age and task.

Developmental Psychology, 20, 193–199.

Sattler, J. M. (1992). Assessment of children (3rd ed.). San Diego7 Jerome M. Sattler Publishers.

Schaefer, E. S., & Hunter, W. M. (1983). Classroom behavior inventory. Unpublished instrument. University of North Carolina

at Chapel Hill.

State of California, Department of Finance. (2000). City/County Population and Housing Estimates, 1991–2000, with 1990

Census Counts. Sacramento, CA, May.

Steinberg, L. (2001). We know some things: Parent–adolescent relationships in retrospect and prospect. Journal of Research on

Adolescence, 11, 1–19.

Steinberg, L., Elmen, J. D., & Mounts, N. S. (1989). Authoritative parenting, psychosocial maturity, and academic success

among adolescents. Child Development, 60, 1424–1436.

J.F. Mattanah et al. / Applied Developmental Psychology 26 (2005) 85–106106

Steinberg, L., Lamborn, S., Darling, N., Mounts, N., & Dornbusch, S. (1994). Over-time changes in adjustment and competence

among adolescents from authoritative, authoritarian, indulgent, and neglectful families. Child Development, 65, 754–770.

Vygotsky, L. S. (1978). Mind in society. Cambridge, MA7 Harvard Univ. Press.

Weiss, L. H., & Schwarz, J. C. (1996). The relationship between parenting types and older adolescents’ personality, academic

achievement, adjustment, and substance use. Child Development, 67, 2101–2114.

Wertsch, J., McNamee, G., Mclane, J., & Budwig, N. (1980). The adult–child dyad as a problem solving system. Child

Development, 51, 1215–1221.

Wood, D. (1988). How children think and learn. Oxford, UK7 Blackwell.

Wood, D., & Middleton, D. (1975). A study of assisted problem-solving. British Journal of Psychology, 66, 181–191.