The developmental dynamics of children’s academic performance and mothers’ homework-related affect and practices

The Developmental Dynamics of Children’s Academic Performance andMothers’ Homework-Related Affect and Practices

Gintautas Silinskas, Noona Kiuru, Kaisa Aunola, Marja-Kristiina Lerkkanen, and Jari-Erik NurmiUniversity of Jyväskylä

This study investigated the longitudinal associations between children’s academic performance and theirmothers’ affect, practices, and perceptions of their children in homework situations. The children’s (n �2,261) performance in reading and math was tested in Grade 1 and Grade 4, and the mothers (n � 1,476)filled out questionnaires on their affect, practices, and perceptions while their children were in Grades 2,3, and 4. The results showed, first, that the more help in homework the mothers reported, the slower wasthe development of their children’s academic performance from Grade 1 to Grade 4. This negativeassociation was true especially if mothers perceived their children not to be able to work autonomously.Second, children’s good academic performance in Grade 1 predicted mothers’ perception of child’sability to be autonomous and positive affect in homework situations later on, whereas poor performancepredicted mothers’ negative affect, help, and monitoring. Finally, mothers’ negative affect mediated theassociation between children’s poor performance, maternal practices, and perceptions of their children.

Keywords: homework, parental involvement, parental affect, academic performance, evocative effect

Homework introduces new challenges for children from thebeginning of their school career, because it demands time andeffort that was previously spent in play. Parental assistance withhomework becomes an important factor in their children’s school-ing, because parents provide a model of how to do homework (Xu& Corno, 1998). The ways in which parents help their childrenwith homework have been found to contribute to children’s aca-demic performance (Cooper, Lindsay, & Nye, 2000; Patall, Coo-per, & Robinson, 2008; Pomerantz & Eaton, 2001; Silinskas,Kiuru, et al., 2013). Conversely, children’s performance may alsoimpact the ways in which parents assist their children with theirhomework (Bell, 1968; Hartup & Laursen, 1991; Silinskas, Niemi,Lerkkanen, & Nurmi, 2013). Most previous research in the fieldhas focused on the associations between parental practices inhomework situations and children’s academic performance (Coo-per et al., 2000; Fan & Chen, 2001; Pomerantz, Moorman, &Litwack, 2007; Taylor, Clayton, & Rowley, 2004). Only a fewstudies have investigated the role of parental affect (e.g., Levin etal., 1997; Pomerantz, Wang, & Ng, 2005; Xu, 2012; Xu & Corno,1998).

The key limitation of most previous research is, however, thatthe studies conducted on children’s academic performance, and

parental affect and practices in homework situations have beencross-sectional (Nurmi & Silinskas, 2014). Thus, little is knownabout the developmental dynamics of children’s academic perfor-mance, and maternal practices and affect across time. Also, mostof the studies have examined adolescents (Dumont et al., 2012,2014; Hill & Tyson, 2009; Pomerantz & Eaton, 2001), while therelation of parental homework-related affect and practices withchildren’s achievement in early school years has gained less at-tention. The aim of the present study was therefore to investigatethe longitudinal associations between primary schoolchildren’sacademic performance, and their mothers’ affect, practices, andperceptions of the child in relation to homework. We tested amodel where prior academic performance predicts mothers’homework-related practices, perceptions of the child, and affect,which then contributes to children’s performance later on. Further-more, the reciprocal relations between maternal practices andaffect across Grades 2, 3, and 4 were examined.

Parental Help and Monitoring During Homeworkand Parental Perception of Child’s Ability

to Be Autonomous

Two distinct forms of parental assistance with their children’shomework have been documented in previous literature: monitor-ing and help (Pomerantz & Eaton, 2001; Silinskas, Niemi, et al.,2013). Help is typically defined as active teaching, instructing orguiding a child to do his or her homework, whereas monitoring isconceptualized as checking child’s homework to ensure it is com-pleted and correct (Pomerantz & Eaton, 2001). However, help andmonitoring can be administered in a child supporting or thwartingway. For instance, Grolnick and Pomerantz (2009) have suggestedthat homework assistance can take a form of control or structure.Controlling forms of homework assistance (e.g., parents simplytelling students what to do and exactly how) has been shown tohave detrimental effects for child development, whereas structure

Gintautas Silinskas, Noona Kiuru, and Kaisa Aunola, Department ofPsychology, University of Jyväskylä; Marja-Kristiina Lerkkanen, Depart-ment of Teacher Education, University of Jyväskylä; Jari-Erik Nurmi,Department of Psychology, University of Jyväskylä.

This study has been carried out in the Centre of Excellence in Learningand Motivation Research, financed by the Academy of Finland (Nr.213486 for 2006–2011) and another grant from same funding agency (Nr.125811 for 2008–2009).

Correspondence concerning this article should be addressed to GintautasSilinskas, University of Jyvaskyla, Department of Psychology, P.O. Box 35,FI-40014, Finland. E-mail: [email protected]

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

Developmental Psychology © 2015 American Psychological Association2015, Vol. 51, No. 4, 419–433 0012-1649/15/$12.00 http://dx.doi.org/10.1037/a0038908

419

(e.g., clear communication of expectations, appropriate scaffoldingduring tasks, rationales for why tasks might be important, andacknowledgment and sympathy when frustration is experienced)would bring favorable outcomes for the developing child.

According to the self-determination theory (SDT; Deci &Ryan, 2008; Deci, Vallerand, Pelletier, & Ryan, 1991; Ryan &Deci, 2000), the ways in which parents assist their children withhomework influence children’s autonomous self-regulation andperformance. The theory postulates that there are three basicpsychological needs—the feelings of autonomy, competence,and relatedness—that are important for the development ofintrinsically motivated behavior. Overall, parenting practices thatsupport autonomy have been suggested to promote children’smotivation and performance, while overly controlling parentingpractices have been suggested to have the opposite effect (Pomer-antz et al., 2007). In a similar vein, parental practices that supportchildren’s competence and relatedness have been suggested toincrease intrinsic motivation in academic contents and thereforelead to favorable child outcomes, such as enhanced academicperformance (Deci & Ryan, 2008; Deci, Vallerand, Pelletier, &Ryan, 1991; Ryan & Deci, 2000).

It can be assumed that the two different types of homework-related practices (help and monitoring) differently satisfy chil-dren’s basic psychological needs (autonomy, competence, andrelatedness). For example, parental help may have positive conse-quences by promoting children’s feelings of competence via bet-tering their skills (Deci & Ryan, 2008; Deci et al., 1991). Help viaparental involvement, whereby a parent shows interest in thechild’s schooling, may also increase a child’s feelings of related-ness. However, despite best intentions, direct help from parents inhomework situations can also have paradoxical consequences. Forinstance, if parents try to help too much or become overinvolved,help might also be detrimental to a child’s feelings of competenceand autonomy, because extensive direct instruction may signal toa child that she or he is not competent enough to cope with his orher homework independently. In turn, monitoring can be assumedto provide more autonomy than direct help, because monitoringincludes checking child’s homework to ensure it is completed andcorrect rather than including extensive direct instruction. Conse-quently, it can be assumed that monitoring brings more benefits fora developing child than help does. However, because parent—child interaction is not that intense, it is possible that child’s needfor relatedness is not satisfactorily addressed. In the present study,monitoring was operationalized as checking children’s homework.In turn, help was operationalized as teaching and guiding a child incompleting homework.

By contrast, parental perception of child’s ability to be autono-mous in homework situations and take responsibility of theirschoolwork is likely to be manifested by parental expressions oftrust that the child will be able to work independently, withoutadditional pressure being applied by a parent. This type of involve-ment may benefit children’s skills through supporting intrinsicmotivation for schoolwork (e.g., Aunola, Viljaranta, Lehtinen, &Nurmi, 2013). Parental perception of child’s ability to be autono-mous transmits the message that parents trust their children byallowing them to feel autonomous and competent enough to handletheir homework (Moorman & Pomerantz, 2008). Parental percep-tion of child’s ability to be autonomous was assumed to provide

the basis for parents’ autonomy granting to the child (although thiswas not directly measured).

In addition to examining the role of help, monitoring, andparental perception of child’s ability to be autonomous in chil-dren’s academic development, we also examined the extent towhich parental perception of child’s ability would moderate theimpacts of help and monitoring on children’s academic perfor-mance. Mothers may use help and monitoring in either a needsupporting or need suppressing way. Thus, it is possible that ifmothers do not trust their child’s ability to be autonomous, mater-nal monitoring and help would manifest in a more controllingways than in the cases where mothers believe in their child’sability to be autonomous (Chen & Stevenson, 1989; Cooper et al.,2000; Levin et al., 1997; Silinskas, Leppänen, Aunola, Parrila, &Nurmi, 2010). Therefore, it can be expected that monitoring andhelp without perception of child’s ability to be autonomous may bemore detrimental for children’s academic performance than in thecases where mothers believe in their children’s ability to be au-tonomous.

Parental Affect in Children’s Homework Situations

Homework activities are emotionally charged events. Bothchildren and parents experience a wide range of emotions inthese situations, including frustration, tension, gratitude andhappiness (Levin et al., 1997; Pomerantz et al., 2007; Pomer-antz et al., 2005; Silinskas, Kiuru, et al., 2013; Xu, 2011; Xu &Corno, 1998). Children’s uncertainty about how to do theirhomework (Pomerantz, Wang, & Ng, 2005) or lack of self-regulation in homework situations (Xu & Corno, 1998) cantrigger irritation or annoyance in parents. However, despitesuch feelings, parents often try to remain positive and supporttheir children (Pomerantz et al., 2005; Silinskas, Kiuru, et al.,2013; Xu & Corno, 1998). Previous research on the role ofmothers has shown that positive emotions strengthen the bondbetween mothers and children (Levin et al., 1997) and protectchildren against the detrimental influence of mothers’ negativeaffect (Pomerantz et al., 2005). Positive affect in homeworksituations can also be expected to promote children’s feelings ofrelatedness and thereby promote their learning motivation andperformance (Pomerantz et al., 2005; Xu & Corno, 1998).

Previous research has frequently made a clear distinctionbetween positive and negative affect in the homework context.Positive affect has been described as the enjoyable, loving, andsupportive interaction that parents experience and maintainwith their children in homework situations, whereas negativeaffect has been described as the irritation, annoyance, frustra-tion, criticism or hostility experienced and transmitted to theirchildren by parents (Pomerantz et al., 2007). Because it hasbeen shown that both positive affect and negative affect can bepresent in the same homework situation to some degree (Po-merantz et al., 2005; Xu, 2012), the absence of positive affectin homework situations does not necessarily imply the presenceof negative affect, and vice versa.

Parental affect is an important factor, because it may triggercertain homework-related practices and perceptions from theparents’ repertoire (Pomerantz et al., 2007). For example, neg-ative parental affect, such as irritation and frustration, feelingstressed, pressured or overly concerned with their children’s

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

420 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

performance may lead to the use of less sensitive and morecontrolling parenting practices in the homework context (Po-merantz & Eaton, 2001; Pomerantz et al., 2007). Parents’ neg-ative affect in this case may origin from children’s problems inan academic task, such as doing their homework. In similarmanner, positive parental affect and lack of negative affect canbe assumed to provide a basis for more sensitive and lesscontrolling parenting practices and more positive perceptions ofthe child (Pomerantz et al., 2007). For example, children aca-demic competence may evoke positive affect such as pride,optimism and joy among parents that then increase parentalperception of child’s ability to be autonomous.

Overall, only a few empirical studies have directly examinedthe dynamics of parental practices, perceptions, and affect inspecific homework situations. As an exception, Pomerantz andEaton (2001) found that the more mothers worried about theirchildren’s academic performance, the more they monitored andhelped their fourth- to sixth-graders. In another study, Levin etal. (1997) showed that parental cost and tension in homeworksituations was associated with increased levels of maternal helpamong first- and third-graders. However, these studies have notinvestigated the relations of parental practices and affect acrossmultiple time-points and have not taken into account differentkinds of practices and perceptions, and both negative andpositive affect, in the same study. Thus, in the present study, wetested a model that included longitudinal associations betweenparental practices and affect across multiple time-points.

Parental Affect and Practices in Homework Situationsand Children’s Academic Performance

The fact that the degree of direct parental involvement andcontrol exercised differ between help and monitoring may explainprevious research findings on the association in the homeworkcontext between parental practices and children’s academic per-formance. It has been found that parental help characterized bydirect involvement is related to low standardized test scores andgrades (Cooper et al., 2000; Hill & Tyson, 2009; Silinskas, Niemi,et al., 2013). Moreover, parental monitoring, which involves lessdirect parental influence, has been found to be weakly negativelyrelated to children’s academic performance (Patall et al., 2008;Silinskas, Niemi, et al., 2013). One possible reason for thesefindings that parental help and monitoring have a negative effecton children’s academic performance is that they decrease chil-dren’s intrinsic motivation (Deci & Ryan, 2008; Deci et al., 1991;Ryan & Deci, 2000). Low intrinsic motivation then leads to theslower development of academic performance. By contrast, previ-ous research has suggested that granting autonomy contributespositively to children’s academic performance. For instance, Coo-per et al. (2000) demonstrated that autonomy support was posi-tively related to standardized test scores and class grades. Simi-larly, Ng, Kenney-Benson, and Pomerantz (2004) showed thatencouraging autonomy enhanced the performance of children,especially those who had learning difficulties. Moreover, parents’trust on their children’s potential and ability to be self-regulatedhas been shown to bring positive outcomes for child’s academicdevelopment (for a review, see Pomerantz et al., 2007; see also,Aunola et al., 2013). One possible reason for these findings is thatautonomy granting increases children’s intrinsic motivation (Deci

& Ryan, 2008; Deci et al., 1991; Ryan & Deci, 2000), which thenleads to the faster development of academic performance.

Previous research has shown that mothers’ affect in homeworksituations is associated with children’s academic performance(Levin et al., 1997; Pomerantz & Eaton, 2001; Pomerantz et al.,2005; Pomerantz et al., 2007). For example, Pomerantz and Eaton(2001) found that mothers’ worrying about children’s academicperformance was negatively related to both previous and subse-quent levels of academic achievement. Similarly, maternal cost,which is similar to negative affect, was negatively related to theacademic performance of first- and third-graders (Levin et al.,1997).

In the present study, we assumed that doing homework andhomework assistance on behalf of the parents are associated withthe development of children’s academic performance, becausesuccessful completion of homework can strengthen children’s ac-ademic skills and provide positive learning experiences that facil-itate learning process also in classroom situations.

Children’s Evocative Impact on Their ParentalInvolvement and Affect

Although it has been widely assumed that parental involvementin homework contributes to children’s academic outcomes, it maybe that the interaction also works in the reverse direction, and thatchildren’s characteristics have an impact on their parents’ affectand practices (Bell, 1968; Hartup & Laursen, 1991; Pomerantz &Eaton, 2001; Rutter, 1997; Scarr & McCartney, 1983; Stattin &Kerr, 2000). Note that this “evocative effect,” as Scarr and Mc-Cartney (1983) termed it, refers to the extent to which the char-acteristics of the children (e.g., their academic performance andsocioemotional characteristics), not their active efforts (i.e.,agency), influence the behavior of their significant others. It mightbe assumed, for example, that children’s poor academic perfor-mance activates higher levels of control efforts among parents, asevidenced in their efforts to help and monitor their children,whereas well-performing children are perceived as being moreable to be autonomous. Children’s academic performance hasalready been previously shown to activate parental homework-related practices, for example, the amount of help (Levin et al.,1997; Pomerantz & Eaton, 2001; Silinskas, Leppänen, et al.,2010).

Similarly, children poor academic performance may activatenegative affect, such as concern among parents, while good per-formance is likely to evoke positive affect. Previous research hasshown that children’s poor academic performance predicts paren-tal negative affect and worrying in homework situations (Levin etal., 1997; Pomerantz & Eaton, 2001). It is even possible thatparental affect may mediate the impact of children’s academicperformance on their parents’ practices in the homework situa-tions, as affect often activates the likelihood of certain kinds ofbehavior (Pomerantz & Eaton, 2000). However, despite someresearch on the evocative effect of children’s performance onparental practices and affect, there has so far been no systematicexamination in the same study of the degree to which children’sacademic performance relates to both parental homework-relatedaffect and practices.

Previous research on the associations between parental affect,parenting practices, and parents’ perception of the child in home-

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

421MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

work situations, and children’s academic performance has someimportant limitations. First of all, the majority of studies have beencross-sectional (e.g., Cooper et al., 2000; Chen & Stevenson, 1989;Patall et al., 2008). Thus, they do not provide an opportunity toexamine the direction of impact between mothers’ affect, theirpractices, and perceptions, and children’s academic performance(Nurmi & Silinskas, 2014). Second, parental homework-relatedpractices, perception of the child, and affect have never beeninvestigated in the same study as possible antecedents and conse-quences of children’s academic performance. Finally, studies onparental homework involvement have mainly focused on olderchildren, early adolescents at best. This is particularly problematic,because the role of different parenting practices may differ de-pending on the age of the child, which may explain some incon-sistencies in previous conceptualizations and findings. Thus, ourlongitudinal investigation aimed at shedding light on the interre-lations between parental practices, parents’ perception of the child,and affect (Grades 2, 3, and 4), and children’s performance(Grades 1 and 4) at the primary school level.

The Aims of the Present Study

The following research questions were examined:(1) To what extent do maternal affect (positive and negative),

practices (help and monitoring) and perception of the child (per-ception of child’s ability to be autonomous) in homework activitiespredict the development of children’s academic performance? Weexpected first that help would have the most negative impact onchildren’s academic performance, because it can be seen as moredetrimental to a child’s feelings of autonomy and competence thanmonitoring (Hill & Tyson, 2009). We also expected that monitor-ing should have negative impact on children’s academic perfor-mance. Maternal perception of child’s ability to be autonomous, inturn, was expected to have a positive impact on child’s academicperformance. Moreover, we expected to find that maternal nega-tive affect would inhibit and positive affect would promote chil-dren’s academic performance (Ng et al., 2004; Pomerantz et al.,2005).

To what extent maternal perception of child’s ability to beautonomous moderates the associations between parental help andmonitoring, and children’s academic performance? Because helpand monitoring can be administered in either a need supporting orneed thwarting way, we expected that, if mothers do not trust theirchild’s ability to be autonomous, maternal monitoring and helpwould have more negative effect on children’s academic perfor-mance in Grade 4 than when mothers trust their child’s ability tobe autonomous (Chen & Stevenson, 1989; Cooper et al., 2000;Levin et al., 1997; Silinskas, Leppänen, et al., 2010).

(2) To what extent does children’s academic performance pre-dict mothers’ positive and negative affect, and their homework-related help and monitoring, and parental perception of child’sability to be autonomous? As it has been suggested that children’sacademic performance has an evocative effect on their mothers’practices in homework situations (Bell, 1968; Scarr & McCartney,1983; Silinskas, Niemi, et al., 2013), we expected that the lowerthe children’s academic performance, the more negative affectmothers would experience in homework situations, and the morethey would engage in monitoring and help (Chen & Stevenson,1989; Cooper et al., 2000; Levin et al., 1997; Silinskas, Leppänen,

et al., 2010). By contrast, we assumed that the higher the children’sacademic performance, the more their mothers would perceivetheir children to be able to be autonomous in homework situations(Cooper et al., 2000).

(3) To what extent does mothers’ positive and negative affectpredict their subsequent help, monitoring, and perception ofchild’s ability to be autonomous in respect to their children’shomework? Based on previous research, we expected that moth-ers’ negative affect in homework situations would be positivelyrelated to subsequent help (Levin et al., 1997) and monitoring(Pomerantz & Eaton, 2001), and negatively related to mothers’perception of child’s ability to be autonomous. By contrast, weexpected that mothers’ positive affect would be negatively relatedto help and monitoring but positively related to autonomy. We alsoexpected that negative affect in particular would mediate thepredictions from children’s academic performance to their moth-ers’ help, monitoring, and perception of their children’s ability tobe autonomous (Levin et al., 1997; Pomerantz & Eaton, 2001).

Method

Participants and Procedure

This study is a part of a large-scale longitudinal study (FirstSteps Study; Lerkkanen et al., 2006–2011) that followed about2,000 children from the beginning of their kindergarten year toGrade 4, with simultaneous data gathering from their mothers andteachers. The main purpose of this follow-up was to investigate thedevelopment of children’s learning and motivation in the familyand school contexts. The follow-up took place in four municipal-ities in different parts of Finland (one urban, one rural, and twocontaining both urban and semirural environments). In three ofthese, an invitation to participate was sent to the whole age cohort,and in the fourth (urban) municipality to approximately half of theage cohort. The sample was highly homogeneous in ethnic andcultural background (e.g., Finnish-speaking schools and students),which is characteristic of a school population outside the metro-politan region and Swedish-speaking areas in Finland. Detailedinformation on the demographics of the participants is presented inTable 1. Mothers were asked to give their written consent to theirchild’s and their own participation in the study. Only childrenwhose parents gave their written consent to participate (80%) weretested.

Children. For the purpose of this study, children’s readingand math performance was tested twice: at the end of Grade 1(when they were 7 to 8 years old; April 2008; n � 2,048) and atthe end of Grade 4 (when they were 10 to 11 years old; April 2011;n � 1,954). Trained research assistants administered group tests inreading and math in classroom settings. Children who were absentfrom school on the day of the tests—for instance, because ofillness—were tested as soon as they returned to school.

Finnish children enter elementary school in the year of theirseventh birthday. Before that, at the age of 6, Finnish children areentitled to attend kindergarten. Kindergarten education is volun-tary and free of charge, and almost all children participate in it(96%; National Board of Education, 2004). Finnish kindergartencurriculum emphasizes personal and social development, learningby play, and integrated activities. Although children are encour-aged to play with letters and numbers, it is only in Grade 1 that

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

422 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

Finnish children are exposed to a formal and systematic instructionin reading and math.

Because decoding in the highly transparent Finnish language ismastered at the very beginning of a child’s school career (Lep-pänen, Niemi, Aunola, & Nurmi, 2004; Lerkkanen, Rasku-Puttonen, Aunola, & Nurmi, 2004; Silinskas, Parrila, et al., 2010),other aspects of reading, such as fluency and comprehension, areemphasized in Grades 1 to 4. The aim of the math curriculumduring Grades 1 to 4 is to support the learning of basic mathemat-ical concepts and arithmetic skills as well as the development ofmathematical thinking (National Board of Education, 2004).

Mothers. The mothers of the children were asked to fill outquestionnaires during the spring term of Grade 2 (March, 2009),the spring term of Grade 3 (March, 2010), and the spring term ofGrade 4 (March, 2011). Mothers were asked to complete thequestionnaires at home independently, without consulting othermembers of the family. Questionnaires were received from 1,476mothers for Grade 2, from 1,452 mothers for Grade 3, and from1,286 mothers for Grade 4. Dropout analyses were conducted tocompare the performance of children whose mothers’ reports wereavailable with the performance of children whose mothers’ reportswere missing. The results showed, for example, that the children ofmothers whose reports were not available in Grade 2 performedmore poorly in math (p � .05) than the children of mothers whohad provided reports. Mothers with missing data did not differfrom mothers with complete data in their children’s reading inGrade 2. Overall, the pattern of missing data seems to imply arestriction of range, as mothers of poorly performing children weremore likely to drop out. This would attenuate the strength of theassociations between the variables of maternal involvement withthe variables of children’s performance.

Data on the demographic background characteristics of thesample were collected in the children’s kindergarten year (seeTable 1). Information on the demographics of the families whojoined the study later was collected at the time of their firstmeasurement point.

Measures

The measures used to assess mothers’ practices and affect wereidentical across the measurement points. For children’s readingand math tests, grade-adapted versions of the same tests were usedin Grades 1 and 4. The psychometric properties of the mothers’questionnaire and the children’s measures (i.e., valid number ofcases, means, standard deviations, reliabilities [Cronbach’s alpha],potential and actual ranges of the values, and skewness) are pre-sented in Table 2.

Mothers’ Questionnaire

Mothers’ homework practices and perceptions (Grades2–4). To measure the type of practices mothers employ concern-ing their children’s homework (help and monitoring) and mothers’perception of child’s ability to be autonomous, mothers were askedto answer 10 questions on a 5-point scale (1 � never at all, 2 �not much, 3 � sometimes, 4 � almost always, and 5 � always).The items are presented in Table 3. Mothers’ help (i.e., directinstruction) in their children’s homework was measured by fouritems; mothers’ monitoring of children’s homework was measuredby three items; and mothers’ perception of child’s ability to beautonomous was assessed by three items. Our questions werebased on the items used previously by Pomerantz and Eaton (2001)

Table 1Demographic Background Characteristics of the First Steps Study

n % M SD Actual range

Family characteristicsFamily structure 1,953

Both parents with biological children 75.2Family with children from previous marriages 9.1Single parents with children 13.9Other 1.8

Number of children in the family 1,561 2.25 1.04 1–8Maternal characteristics

Mother’s age 1,861 38.41 5.27 25–56Mother’s educationa 1,861 4.19 1.50 1–7Mother’s work status 1,480

Employed 74.1Unemployed 25.9

Child characteristicsChild’s age in months (Kindergarten; September 1, 2006) 1,859 73.89 3.52 61–89Child’s gender 2,261

Girl 47.6Boy 52.4

Birth order of a child 1,515First born child with siblings 37.9Only child 17.0Middle child 16.0Youngest child 29.1

a Mothers’ education was assessed on a 7-point scale: no vocational education; short vocational courses; avocational school qualification; a college qualification; a bachelor’s degree from university; a master’s degree;and a licentiate or doctoral degree.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

423MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

and Pomerantz and Ruble (1998). The items relating to the scalesof help and monitoring have already been published by Silinskas,Niemi, et al. (2013).

Mothers’ affect in homework situations (Grades 2–4).Mothers’ affect in their children’s homework situations was as-sessed by four questions on a 5-point scale (1 � not at all; 5 �very much). The items are listed in Table 3. Positive and negativeaffect were measured by two items each. The questions were basedon those used by Pomerantz et al. (2005) and Poulou and Norwich(2002) and have been published previously by Silinskas, Kiuru, etal. (2013).

Children’s Tests

Reading performance (Grades 1 and 4). A reading perfor-mance test battery (Lindeman, 1998) was used to measure chil-dren’s reading performance at the end of Grades 1 and 4. In theword-reading test, the child was asked to select the correct wordfrom four phonologically similar alternatives and link this to apicture by drawing a line between the two. The task comprised amaximum of 80 items to be attempted within a 2-min time limit,and the score was the number of correct responses marked withinthe time limit (maximum 80 points). The measure of readingperformance was the same at both test points in terms of thenumber of items and the time limit. The reading performance testbattery is a standardized instrument for Finnish first- to sixth-graders (Lindeman, 1998).

Math performance (Grades 1 and 4). Math performancewas tested in a group situation at the end of Grades 1 and 4. Mathperformance was assessed by an arithmetic test (Räsänen &

Aunola, 2007; see also Räsänen, Salminen, Wilson, Aunio, &Dehaene, 2009). In this test, a maximum of 28 items, consisting of14 addition and 14 subtraction tasks, is attempted within a 3-mintime limit. The task difficulty increases gradually across the test.Although the number of items and time limit is identical in Grades1 and 4, more advanced items are added to the task in Grade 4. Thescore was the total number of correct answers (maximum 28points).

To create a performance variable, the children’s reading andmath performance scores were first standardized (Z scores), andthen the mean score of the two standardized variables was com-puted.

Data Analysis

Structural equation modeling was utilized in seeking answers tothe research questions. The data analyses were carried out asfollows. First, descriptive analyses were conducted to investigatemean level differences in homework-related affect, homework-related perceptions and practices over time (Grades 2–4). Second,separate measurement models for each construct across Grades 2,3, and 4 were estimated. Third, all five measurement models werecombined into a final measurement model consisting of five latentconstructs across 3 time-points. Finally, a structural equationmodel was constructed by including the stabilities of the con-structs, cross-lagged associations between mothers’ affect andpractices, and concurrent associations. The final model also in-cluded children’s performance (see Figure 1).

All the analyses were conducted using the Mplus statisticalpackage (Version 6.12; Muthén & Muthén, 1998–2010). The

Table 2Psychometric Properties of the Major Study Variables

Reliability(Cronbach’s �)

Range

Variable n M SD Potential Actual Skewness

Maternal practices, perception, andaffect

Help (Grade 2) 1,456 2.96 .74 .88 1–5 1–5 .77Help (Grade 3) 1,358 2.82 .66 .86 1–5 1–5 .90Help (Grade 4) 1,285 2.67 .60 .83 1–5 1–5 .45Monitoring (Grade 2) 1,460 3.95 .83 .85 1–5 1–5 �.50Monitoring (Grade 3) 1,357 3.62 .82 .85 1–5 1–5 �.13Monitoring (Grade 4) 1,286 3.29 .81 .83 1–5 1–5 �.01Perception of child’s ability to be

autonomous (Grade 2) 1,459 3.82 .85 .82 1–5 1–5 �.63Perception of child’s ability to be

autonomous (Grade 3) 1,357 3.79 .85 .85 1–5 1–5 �.62Perception of child’s ability to be

autonomous (Grade 4) 1,285 3.93 .81 .85 1–5 1.33–5 �.62Negative affect (Grade 2) 1,474 1.64 .81 .84 1–5 1–5 1.31Negative affect (Grade 3) 1,451 1.72 .84 .85 1–5 1–5 1.10Negative affect (Grade 4) 1,277 1.91 .92 .89 1–5 1–5 .78Positive affect (Grade 2) 1,476 3.96 .80 .72 1–5 1–5 �.67Positive affect (Grade 3) 1,452 3.91 .79 .75 1–5 1–5 �.14Positive affect (Grade 4) 1,275 3.74 .83 .76 1–5 1–5 �.29

Child performanceReading (Grade 1) 2,048 18.28 8.96 .97a 0–80 0–58 .66Reading (Grade 4) 1,954 36.06 9.25 .95a 0–80 0–66 �.02Math (Grade 1) 2,049 10.50 4.13 .98a 0–28 0–28 .32Math (Grade 4) 1,953 17.03 4.09 .85a 0–28 0–27 �.63

a The Kuder–Richardson reliability, a measure of internal consistency for dichotomous variables.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

424 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

proportions of missing data for the main study variables rangedfrom 0% to 37% (M � 26.32%, SD � 13.24%). Our data was notmissing completely at random: Little’s (1988) missing-completely-at-random test: �2(792) � 968.85, p � .001. Thus, the standardmissing-at-random approach was applied. This missing datamethod uses all the data that are available in order to estimate themodel without imputing data. Because the distributions of thevariables were skewed, the model parameters were estimated usingthe MLR estimator. The MLR estimator produces standard errorsand chi-square test statistics for missing data with non-normaloutcomes by means of a sandwich estimator.

Model fit was examined using the Comparative Fit Index (CFI),Tucker–Lewis Index (TLI), root mean square error of approxima-tion (RMSEA), and standardized root mean square residual(SRMR). CFI and TLI values above .95, a RMSEA value below.06 and a SRMR value below .08 indicate good model fit (Hu &Bentler, 1999; Muthén, 1998–2004). Only CFI and TLI valuesbelow .90 and RMSEA and SRMR values above .10 are indica-tions of poor model fit (Kline, 2005).

Results

Descriptive Analyses

Two-way repeated ANOVAs were conducted to investigatemean differences across time (Grades 2, 3, and 4) in help, moni-toring, and mothers’ perception of child’s ability to be autono-mous. The analyses revealed a significant interaction of Time �Type of Practice, F(4, 1,079) � 147.58; p � .001; partial �2 � .35.The results showed both a significant decrease in mothers’ help,F(2, 1,082) � 149.93, p � .001, partial �2 � .22, and monitoring,F(2, 1,085) � 440.31, p � .001, partial �2 � .45, across Grades 2,3, and 4, and a significant increase in mothers’ perception of

child’s ability to be autonomous, F(2, 1,083) � 32.93, p � .001,partial �2 � .06, across the grades.

Two-way repeated ANOVAs were also conducted to examinemean differences in maternal affect in the homework context overtime (Grades 2, 3, and 4), and differences between the type ofaffect (positive vs. negative affect). The analyses showed a signif-icant interaction of Time � Type of Affect, F(2, 1,043) � 78.55;partial �2 � .13; p � .001. Maternal negative affect significantlyincreased across time, F(2, 1,046) � 43.73; partial �2 � .08; p �.001, whereas maternal positive affect significantly decreased. Ateach time-point, however, mothers reported more positive affectthan negative affect (p � .001).

Measurement Models

To answer our key research questions we first constructedmeasurement models for parental practices, perceptions and affectacross Grades 2 to 4. We started by estimating a separate mea-surement models for each construct. The resulting measurementmodels were then combined to form the final measurement model.In order to ensure time invariance, the factor loadings of the sameitems were set equal across the different measurement points(Grades 2, 3, and 4). Latent factors were allowed to be correlatedwith each other. Additionally, we also allowed autocorrelationsbetween the residuals of the same items. The measurement modelsfor each construct as well as the measurement model for all 5constructs in one model showed model fits within criteria range ofgood model fit (TLI .95; CFI .95; RMSEA � .08; SRMR �.06). Thus, no other modifications to the measurement modelswere implemented. The factor loadings of the measurement modelthat included all 5 constructs are presented in Table 3. The corre-lation matrix between latent constructs is presented in Table 4.

Table 3Factor Loadings (Standardized Estimates) of the Confirmatory Factor Analysis for Maternal Practices, Perception, and Affect

Items Grade 2 Grade 3 Grade 4

Maternal practices and perception concerning children’s homeworkHelp

1. Do you instruct your child in his/her homework? .84 .82 .772. Do you help or guide your child in his/her homework? .89 .88 .833. Do you help your child in his/her homework related to reading? .73 .69 .654. Do you help or guide your child in his/her homework related to mathematics? .80 .72 .67

Monitoring5. Do you make sure that your child has done his/her homework? .63 .59 .526. Do you check your child’s homework? .92 .94 .957. Do you check your child’s homework together with your child? .86 .87 .87

Perception of child’s ability to be autonomous8. Do you know that the child remembers to do his/her home assignments? .80 .86 .869. Do you trust that the child takes care of his/her home assignments by

himself/herself?.85 .87 .88

10. Do you have to force your child to do the home assignments? �.65 �.65 �.64

Maternal affect in homework situationsNegative affect

1. I feel hopeless .61 .62 .622. I feel stressed/frustrated .87 .92 .96

Positive affect1. I feel satisfaction .80 .80 .832. I feel joy .89 .90 .94

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

425MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

Structural Model

Next, structural equation models were constructed. First, stabil-ity paths across the 3 time-points within the same constructs werespecified. The cross-lagged paths from mothers’ practices andperceptions concerning children’s homework (help, monitoring,and mothers’ perception of child’s ability to be autonomous) atTime T to mothers’ affect (positive and negative affect) at TimeT 1, and vice versa were also estimated. Next, the associationsbetween the latent constructs in Grade 2 and associations betweenthe error terms of the latent constructs in Grades 3 and 4 wereestimated. Then, to examine the impact of children’s academicperformance on their mothers’ homework-related affect and prac-tices, we added children’s performance in Grade 1 as a predictor ofmothers’ practices, perceptions and affect in Grade 2. To examinethe impact of mothers’ practices, perceptions and affect, we alsoestimated paths from mothers’ practices, perceptions and affect inGrade 4 to their children’s performance in Grade 4, after control-ling for the children’s performance in Grade 1. The model showedgood fit (TLI .95; CFI .95; RMSEA � .08; SRMR � .06).The model is presented in Figure 1.

Maternal Homework-Related Affect, Practices andPerceptions, and Children’s Performance

Our first research question asked to what extent mothers’ affect,perceptions, and practices are related to children’s subsequentperformance. The results (see Figure 1) showed that when parentalhomework practices, perceptions, and both types of affect wereincluded in the model, only mothers’ help predicted children’slater academic performance: The more help mothers reported, thepoorer their children’s performance was in Grade 4, after control-ling for their previous level of performance in Grade 1 as well asother maternal practices and affect.

To make sure that our analyses were not merely examiningconcurrent associations between mothers’ practices and affect inGrade 4 and children’s performance in Grade 4, we modified ourprevious models in such a way that mothers’ practices, perceptionsand affect in Grade 3 (instead of mothers’ practices and affect inGrade 4) would directly predict children’s performance in Grade 4.The results of these analyses revealed almost identical results tothe previous model: The more help mothers reported giving inGrade 3, the poorer their children’s performance was in Grade 4(� � –.18, p � .001). Similarly, as in the previous model, maternalmonitoring, maternal perception of child’s ability to be autono-mous, and positive affect in Grade 3 were not significantly relatedto children’s performance in Grade 4. However, there was oneadditional result: Mothers’ negative affect negatively predictedchildren’s subsequent performance. That is, the more negativeaffect mothers reported in Grade 3, the poorer their children’sperformance was later on in Grade 4 (� � –.06, p � .01).

We also ran three separate models for each of the maternalpractices and perceptions. In these models, all the model specifi-cations were the same except for the fact that only one parentalpractice or perception was included at a time in each model. Theresults of these separate models showed that all constructs signif-icantly predicted children’s performance: The more help and mon-itoring mothers reported giving in Grade 4, the poorer their chil-dren’s performance was (�help � –.21, p � .001; �monitoring �–.11, p � .001). By contrast, the more the mothers perceived theirchildren to be able to be autonomous in homework situations, thebetter their children’s performance was (�autonomy � .11, p �.001). In all of these models, mothers’ negative affect also signif-icantly predicted children’s performance in Grade 4 (�negative affect

ranging from –.06 to –.09, p � .05).Because it is possible that help which is provided by the parent

in an autonomy supporting or controlling manner may make a

Table 4Correlations Between the Latent Constructs of Mothers’ Practices, Perception, and Affect

1 2 3 4 5 6 7 8 9 10 11 12 13 14

1. Help (Grade 2)2. Help (Grade 3) .70���

3. Help (Grade 4) .64��� .75���

4. Monitoring (Grade 2) .54��� .42��� .38���

5. Monitoring (Grade 3) .47��� .59��� .49��� .73���

6. Monitoring (Grade 4) .48��� .55��� .61��� .61��� .77���

7. Perception of child’sability to be autonomous(Grade 2) �.35��� �.31��� �.34��� �.18��� �.21��� �.22���

8. Perception of child’sability to be autonomous(Grade 3) �.31��� �.39��� �.41��� �.16��� �.24��� �.26��� .77���

9. Perception of child’sability to be autonomous(Grade 4) �.29��� �.35��� �.41��� �.13��� �.22��� �.28��� .66��� .79���

10. Negative affect (Grade 2) .28��� .24��� .27��� .03 .09�� .10�� �.40��� �.40��� �.34���

11. Negative affect (Grade 3) .24��� .25��� .24��� .02 .09�� .10��� �.37��� �.43��� �.35��� .72���

12. Negative affect (Grade 4) .19��� .23��� .26��� �.01 .06� .09��� �.30��� �.35��� �.39��� .67��� .72���

13. Positive affect (Grade 2) .10��� .04 .05 .16��� .14��� .13��� .21��� .18��� .18��� �.43��� �.36��� �.34���

14. Positive affect (Grade 3) .04 .04 .02 .13��� .13��� .12��� .21��� .26��� .22��� �.38��� �.44��� �.40��� .63���

15. Positive affect (Grade 4) .06� .04 .05 .12��� .13��� .13��� .16��� .22��� .22��� �.33��� �.34��� �.43��� .58��� .65���

� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

426 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

difference, we run additional statistical analyses to test the inter-action of perceived ability to be Autonomous � Help, and theinteraction of perceived ability to be Autonomous � Monitoring.The results showed that the interaction term of the latent variablesperceived ability to be Autonomous � Help was statisticallysignificantly in predicting child performance in Grade 4 (unstan-dardized estimate � .083, SE � .042, p � .043). An interpretationof this interaction is shown in Figure 2: Mothers’ help for theirchildren was less detrimental if mothers perceive their children tobe able to work on the school tasks autonomously than if mothersperceive their children not to be able to work autonomously.

Moreover, because it is possible that mothers’ practices andperceptions have different impacts on children’s academic perfor-mance depending on the child’s level of academic performance,we tested if children’s performance in Grade 1 acts as a moderatorfor the associations between mothers’ homework-related practicesand perceptions, and children’s performance. To investigate this,we applied a multisample procedure to our final model by con-trasting 50% of low performing children against 50% high per-formers (based on Grade 1 median split). Models were contrastedagainst each other, one at a time: freely estimated multisamplemodel versus a model that had one constrained path. The paths thatwere constrained, one at a time, were (a) mothers’ help in Grade 4and child performance in Grade 4, (b) mothers’ monitoring inGrade 4 and child performance in Grade 4, or (c) mothers’ per-ceptions about children’s ability to be autonomous in Grade 4 andchild performance in Grade 4. The difference in the models wasnonsignificant for the model with constrained path from mothers’help (��2 � .455, �df � 1, p � .500), but significant for the

models of constrained path from monitoring (��2 � 11.154,�df � 1, p � .001) and mothers’ perception of children’s abilityto be autonomous (��2 � 3.92, �df � 1, p � .047). A closerinvestigation of the beta coefficients for low and high achievinggroups revealed that mothers’ monitoring was negatively related tochildren’s performance among low performers (� � –.10, p �

Figure 1. Longitudinal associations between mothers’ practices (help, monitoring), perception of child’s abilityto be autonomous, affect (negative affect and positive affect), and children’s performance: standardized solutionof structural equation model. Concurrent associations and nonsignificant cross-lagged paths are not shown.� p � .05. �� p � .01. ��� p � .001.

Figure 2. The moderating role of mothers’ perception of child’s ability tobe autonomous on the association between mothers’ help in Grade 4 andchildren’s performance in Grade 4.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

427MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

.01), while it was positively associated with children’s perfor-mance among high performers (� � .09, p � .01). Moreover,mothers’ perception of their child’s ability to be autonomous wasnot related to children’s performance in Grade 4 among lowperformers (� � –.01, p .10), whereas it was positively asso-ciated with children’s performance in Grade 4 among high per-formers (� � .10, p � .01).

Children’s Performance, and Mothers’ Practices,Perceptions, and Affect

Our second research question concerned the extent to whichchildren’s performance predicts mothers’ practices, perceptions,and affect. The results (see Figure 1) for the associations betweenchildren’s performance and maternal homework affect and prac-tices showed that academic performance in Grade 1 predicted bothmaternal positive and negative affect: The weaker the children’sperformance was, the more negative and the less positive affecttheir mothers’ reported in homework situations. Academic perfor-mance in Grade 1 also predicted all types of practices and percep-tions: The poorer the children’s performance was, the more in-volved in helping and monitoring their mothers were and the lessthey perceived their children to be able to be autonomous inhomework situations.

Maternal Homework-Related Practices, Perceptions,and Affect

Our third research question concerned the cross-lagged associ-ation between maternal practices and affect. The results showed(see Figure 1), first, that mothers’ negative affect in homeworksituations positively predicted the amount of help they gave fromGrade 2 to Grade 3, and from Grade 3 to Grade 4: The morenegative affect mothers reported, the more they engaged in helpingtheir children with homework. The results showed, however, thatmothers’ help and positive affect were not related longitudinally.Second, mothers’ negative and positive affect in homework situ-ations in Grade 2 positively predicted subsequent monitoring: Themore both negative and positive affect mothers reported, the moreinvolved they became in monitoring their children’s homework.The same was true for mothers’ negative affect in Grade 3: Themore negative affect mothers reported, the more they monitoredtheir children’s homework. Third, mothers’ negative affect inGrade 2 negatively predicted maternal perception of child’s ability tobe autonomous in Grade 3: The higher the amount of negative affectwas, the less mothers trusted to their children’s ability to be autono-mous later on. Moreover, parental perception of child’s ability to beautonomous negatively predicted mothers’ negative affect acrossGrades 2 and 3: The more mothers trusted in their children’sability to be autonomous, the less negative affect they experiencedlater on. Also, parental perception of child’s ability to be autono-mous positively predicted positive affect across Grade 2 and Grade3, and across Grades 3 and 4: The more mothers trusted in theirchildren’s ability to be autonomous, the more positive affect inhomework situations they experienced later on.

We also investigated the extent to which mothers’ affect inhomework situations mediated the impact of children’s academicperformance on maternal practices. The results showed that theeffect from children’s performance in Grade 1 on mothers’ help

(standardized indirect estimate � –.021, p � .01) and monitoring(standardized indirect estimate � –.022, p � .001), and mothers’perception of child’s ability to be autonomous (standardized indi-rect estimate � .033, p � .001) in Grade 3 was partially mediatedvia mothers’ negative affect in Grade 2. Children’s poorer perfor-mance predicted higher maternal negative affect, which, in turn,predicted higher subsequent levels of maternal help and monitor-ing and a lower subsequent level of perception of child’s ability tobe autonomous. Positive affect did not mediate any of the rela-tionships.

Controlling for the Background Variables

We conducted a few sets of analyses to account for the role ofother child characteristics (e.g., gender, birth order) and mother orfamily characteristics (e.g., work status, education, family struc-ture) in the association between children’s performance and ma-ternal affect, practices, and perceptions (see Table 1 for the list ofdemographics). Entering these control variables in the model didnot, however, change any of the results reported above.

Discussion

The present study examined the developmental dynamics be-tween mothers’ affect, their practices relating to their children’shomework, their perceptions of their children’s ability to be au-tonomous, and the children’s academic performance when movingacross Grades 1 to 4. The results showed, first, that when differentparental practices were entered simultaneously in the same model,an increase in the amount of maternal help in homework situationswas associated with slower progress in children’s academic per-formance. This negative prediction was true especially when moth-ers simultaneously perceived their children as not being able towork autonomously. Second, mothers’ perceptions of their chil-dren’s ability to be autonomous predicted increased performanceparticularly among good performers. Moreover, mothers’ monitor-ing positively predicted children’s performance among high per-formers, but negatively among poor performers. Third, children’slow academic performance in Grade 1 predicted an increase inmothers’ help, monitoring and negative affect, and a decrease inmaternal perception of child’s ability to be autonomous and pos-itive affect. Moreover, negative affect was shown to mediate theassociation between children’s poor academic performance andparental practices and perceptions.

The Role of Maternal Homework Practices,Perceptions, and Affect in Children’s Performance

We first asked to what extent parental practices, perceptions andaffect would predict children’s academic performance. Our find-ings indicated that a high amount of maternal help in homeworksituations predicted slower progress in children’s academic per-formance. This association was particularly true for those motherswho simultaneously perceived their children as not being able towork autonomously. This result supported the assumption thatmothers’ engagement in help in controlling ways that involveslittle trust for children’s ability to be autonomous is a risk forchildren’s academic development. One explanation for our findingis that the kind of maternal homework assistance that includes a

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

428 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

high degree of direct involvement and low degree of trust tochildren’s competence inhibits children’s autonomous motivationand feeling of competence, which then leads to poor performance(Deci & Ryan, 2008; Deci et al., 1991; Ryan & Deci, 2000). Theresults of the present study are in line with some previous findingsamong primary schoolchildren (Cooper et al., 2000; Levin et al.,1997; Silinskas, Leppänen, et al., 2010, Silinskas et al., 2012) andearly adolescents (see Hill & Tyson, 2009). Previous research hasshown that intrusive homework assistance does not benefit chil-dren’s performance (Dumont et al., 2012, 2014; Pomerantz et al.,2007). For instance, controlling parenting leads a child to experi-ence a lack of autonomy and a low sense of agency in the learningprocess (Pomerantz et al., 2007), which may negatively influencechildren’s skill development. Moreover, if a child perceives thatthe mother is investing a lot of effort in assisting him or her withhomework without a trust for child’s ability to be autonomous, heor she may develop unfavorable cognitions, such as low self-concept of ability (Aunola et al., 2002; Frome & Eccles, 1998).This may then lead to lower engagement in learning tasks, which,in turn, can lead to poorer academic performance as compared toother children (Onatsu-Arvilommi & Nurmi, 2000).

Our results showed further that also mothers’ perceptions oftheir children’s ability to be autonomous and mothers’ monitoringcontributed to children’s academic development but that theseeffects were dependent on the child’s level of academic perfor-mance. For example, the results indicated that children who benefitmost of their mothers’ trust in their ability to be autonomous arethose who show good academic performance. One possible expla-nation for this result is that mothers’ perception of their children’sability to be autonomous leads to a higher level of autonomygranting. However, to benefit from this kind of autonomy mayrequire good academic skills. By contrast, children who show lowacademic performance may benefit from more structured parentingpractices. It is noteworthy that although we had a measure ofmothers’ perception of their child’s ability to be autonomous wedid not have a measure of mothers’ actual autonomy support inhomework situations. Consequently, compared to an investigationof parental beliefs an investigation of the autonomy supportingparenting practices might yield somewhat different findings (seealso Ng et al., 2004)

Also, mothers’ monitoring predicted children’s performancedifferently among high performers and low performers. Amonghigh performers mothers’ monitoring contributed positively tochildren’s academic development, whereas among poor perform-ers high level of monitoring decreased the developmental progressin academic skills. One possible explanation for this result is thatpart of the low performers are children with learning difficultiesand the kind of monitoring that mothers are able to provide themis not efficient enough to help them to do better in reading andmath. It is also possible that the mothers of low-achieving childrenhold negative beliefs about their children’s potential, and, there-fore, when they become involved in their children’s homework bymonitoring, mothers mainly focus on their children’s difficulties(Pomerantz et al., 2007). This kind of involvement may decreasechildren’s motivation leading to poor academic performance. Inturn, mothers of high-achieving children may hold more positivebeliefs about their children’s potential and abilities and that alsocommunicates their children that they are capable to do well. Such

positive message is then likely to foster skill development as well(Aunola et al., 2013; Aunola et al., 2002).

In addition to maternal homework-related practices, maternalaffect in homework situation also played an important role in thedevelopment of children’s academic performance. The results ofthe present study showed evidence that maternal negative affect inhomework situations inhibited the positive development of chil-dren’s academic performance. In a previous study by Pomerantz etal. (2005), maternal negative affect in homework situations wasfound to be negatively related to a child’s homework persistenceand mastery orientation, and to be positively related to a child’shelplessness. Such beliefs and behavior may later influence chil-dren’s performance. However, contrary to the results of someprevious studies (Pomerantz et al., 2005; Pekrun, 2006), we foundno evidence of mothers’ positive affect impacting children’s per-formance.

It should be kept in mind, however, that the role of maternalhelp, monitoring, and perception of child’s ability to be autono-mous may change during different developmental periods of child-hood and adolescence, underlining the importance of future re-search among children at different ages. On the mean level, forexample, our results showed a decrease in mothers’ help andmonitoring across Grades 2, 3, and 4, as well as an increase inmaternal perception of child’s ability to be autonomous.

The Role of Children’s Academic Performance inMaternal Practices, Perceptions, and Affect inHomework Situations

Overall, the findings of the present study that children’s aca-demic performance predicted their mothers’ practices, perceptions,and affect in homework situations is in line with both the notionand the growing body of empirical literature confirming thatchildren’s characteristics have an evocative effect on their parents’reactions toward them (Pomerantz & Eaton, 2001; Scarr & Mc-Cartney, 1983; Silinskas, Niemi, et al., 2013; Stattin & Kerr,2000). In other words, the development of children’s academicperformance should be understood as continuous interplay andtransaction between the parent being responsible for guiding andsupporting the child, and the child herself/himself (Sameroff,2010). In line with our expectations, the results of our studyshowed that the poorer their children’s academic performance wasin Grade 1, the more mothers monitored and tried to help theirchildren in Grade 2. These results are similar to those of someprevious studies (Chen & Stevenson, 1989; Cooper et al., 2000;Levin et al., 1997). For instance, Silinskas, Leppänen, et al. (2010)showed that, from Grade 1 onward, parents of poor performersengage in more teaching of reading and math than parents of highachieving pupils. Similarly, Levin et al. (1997) reported that moth-ers increased their help with homework if their children weredoing poorly at school. Cooper et al. (2000) interpreted theircross-sectional findings as indicating that parents of high achieverssupport autonomy or avoid interfering in their child’s independentcompletion of assignments.

Children’s performance also predicted their mothers’ affect inhomework situations: The lower the children’s academic perfor-mance, the more negative affect their mothers’ experienced. Thisresult is in line with our expectation that unfavorable results atschool (e.g., a low skill level) would trigger negative emotions in

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

429MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

mothers. This has also been found in previous research (Levin etal., 1997; Pomerantz & Eaton, 2001). Moreover, we found positiveassociations between children’s performance and mothers’ posi-tive affect (see also, Ng et al., 2004).

Maternal Practices, Perceptions, and Affect inHomework Situations

As expected, the results of the present study showed that themore negative affect mothers reported, the more they engagedin monitoring and help, and the less they perceived their chil-dren to be able to be autonomous in homework situations lateron. Similar associations have been reported by Levin et al.(1997), and Pomerantz and Eaton (2001). The present studyadds to the literature by showing that mothers’ negative affectin homework situations longitudinally predicts their highamounts of help and monitoring, and low level of perception ofchild’s ability to be autonomous after controlling for the pre-vious levels of help and monitoring.

The association between mothers’ negative affect and help isnot surprising given the fact that direct maternal help is oftenrelated to a tense atmosphere when doing homework (Levin etal., 1997). The association between negative affect and help isalso reflected in the fact that mothers’ direct involvement isespecially needed for children who are struggling academically(Levin et al., 1997; Pomerantz & Eaton, 2001; Silinskas, Lep-pänen, et al., 2010; Silinskas et al., 2012). This is because suchchildren are likely to be uncertain about how to do theirhomework (Pomerantz & Eaton, 2001) and often express neg-ative attitudes and feelings of helplessness (e.g., frustration andgiving up; Xu & Corno, 1998). As these children show poorattitudes and possible avoidance behavior, their mothers maystart to experience negative emotions, which then evoke aparticular repertoire of homework assistance (e.g., more helpand monitoring, less maternal perception of child’s ability to beautonomous).

The results showed further that when mothers perceived theirchildren to be able to be autonomous, mothers’ negative affectin homework situations decreased and their positive affectincreased. These results suggest that mothers find trust forchildren’s autonomy to be pleasant and gratifying, possiblybecause autonomy support is targeted, in particular, to childrenwho do well in the classroom. By trusting child’s ability to beautonomous and possibly granting autonomy in homework sit-uations, parents support their children’s internal motivation. Asa consequence, children may learn to take more responsibilityfor their schoolwork, thus removing the need for parents tocontinually issue reminders about it. Despite the freedom achild is exposed to concerning homework, parental perceptionof child’s ability to be autonomous does not imply permissive-ness, neglect, or promotion of detachment (Joussemet, Landry,& Koestner, 2008). On the contrary, parents who grant auton-omy need to remain interested in their children’s schooling andprovide help if a child needs or asks for it. In sum, as a proxyof the least involving form of homework practices, parentalperception of child’s ability to be autonomous relates to adecrease in negative affect and increase in positive affect in thehomework context. This type of homework assistance wasfound to increase gradually across Grades 2 to 4. This is not

surprising, given the fact that as children grow older andbecome experienced in completing homework assignments un-aided, granting children autonomy in doing their homeworkseems to become the most frequent parental practice.

Maternal Affect as a Mediator Between Children’sPerformance and Parenting Practices

We also investigated the possibility that maternal affect func-tions as a mediator between children’s performance and mothers’practices. The results showed evidence for the mediator hypothe-sis: A low level of academic performance on the part of the childfirst triggers a mother’s negative affect, which then increases herhelp and monitoring and thus decreases the freedom of action thatshe allows her children. Although intuitive, this is an interestingand important finding. As far as we know, this is the first study toexamine these kinds of mediator effects of parental involvementand their affect. This finding provides some explanation of thelimited benefits of help on children’s performance. That is, lowperformance may trigger negatively loaded maternal help whenassisting their children with homework. Thus, despite the bestintentions of satisfying children’s need for relatedness (beingpresent during homework, showing interest in children’s schoollife) and increasing feeling of competence (by teaching, in-structing, guiding), negatively loaded help may transmit theopposite idea. For instance, signs of mothers being irritated,annoyed, and frustrated may inhibit children’s feeling of relat-edness (Deci & Ryan, 2008; Deci et al., 1991; Ryan & Deci,2000). Overall, our results showed that the level of children’sacademic performance evokes certain kinds of parental involve-ment in homework situation, and that parental affect plays arole in this connection.

Limitations

This study has limitations that restrict the generalizability of ourresults. First of all, we used self-reports to investigate mothers’homework-related affect and practices. Although similar measureshave been used in previous research on maternal homework-related affect and practices (Pomerantz et al., 2005; Levin et al.,1997; Silinskas, Leppänen, et al., 2010), the results may sufferfrom the social desirability effect and reporter bias. Moreover,self-reports may not capture the richness of the situations in whichhomework takes place. Thus, future studies should consider usingquestionnaires that aim at tapping appropriate parental homeworkbehaviors and that also make an effort to differentiate between theactual and the intentional parental behavior. Some qualitative andobservational studies have explored mother—child interaction inhomework situations (e.g., Xu & Corno, 1998), but while captur-ing the richness of the homework situation, their sample sizes weresmall and data often cross-sectional. Consequently, future studieswould benefit from combining both quantitative and qualitativemeasures.

The second limitation is the attrition of mothers whose childrenwere already struggling. Although attrition is a common problemin longitudinal research, one needs to be cautious in making anygeneralizations on the strength of the associations found in thisstudy. Future research should pay a particular attention to main-taining the interest of mothers who have academically struggling

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

430 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

children. Obtaining data on how mothers with academically strug-gling children act and behave in homework situations would beparticularly valuable in shedding light on the maternal practicesthat are the most beneficial for these children’s academic perfor-mance.

Third, although SDT (Deci & Ryan, 2008; Deci, Vallerand,Pelletier, & Ryan, 1991; Ryan & Deci, 2000) was used to providea theoretical rationale for our hypotheses, our data did not providepossibilities for direct testing of the extent to which homeworkpractices would impact children’s psychological needs (e.g., aca-demic competence, the strength of the parent—child relationshipin relation to homework activities). Obtaining empirical evidencethat is tightly related to the SDT could be the target for the futurework.

The fourth limitation of our study is that it was conducted in onecountry, Finland, which may influence the associations foundbetween mothers’ affect, perceptions, practices, and children’sperformance. For instance, the educational system in Finland dif-fers from that of many other countries in several respects, one ofthem being the unusually late age of formal school entry (7 yearsof age). Possible differences of this kind in the educational back-ground, such as no systematic instruction in reading and mathbefore Grade 1 and, possibly, great variation in time parents areexpected to spend with their children, suggest that caution isnecessary in generalizing the findings.

Conclusions

The results of the present longitudinal study add to theexisting literature in several ways. First, poor academic perfor-mance on the part of the children was found to prompt increasedlevels of maternal help and monitoring along with negativeemotions; good performance, on the other hand, led to increasedindependence from maternal control and maternal positive af-fect in homework situations. Second, from among the parentalpractices, perceptions and affect studied, mothers’ help (i.e.,direct instruction), in particular, contributed to their children’sslower academic development, particularly when mothers didnot trust their children’s ability to be autonomous. In the case ofmonitoring and perception of children’s ability to be autono-mous, children’s academic performance seems to act as a mod-erator. Mothers’ perceptions of their children’s ability to beautonomous as well as monitoring predicted good performanceparticularly among high achievers. Third, negative affect wasshown to mediate the association between children’s poor ac-ademic performance and the repertoire of parental practicesemployed when assisting their children with homework.

The results of this study suggest that educators working withacademically struggling children should make sure that parents areaware of the dangers of the potentially vicious circle of increasednegative emotion and increased help, and their limited benefits forchildren’s performance. The detrimental effect of overinvolvementand negative affect could at least partly be overcome by teachingparents how to help their children with homework so that chil-dren’s needs for autonomy, relatedness, and competence are notviolated and a positive atmosphere is maintained in challenginghomework situations.

References

Aunola, K., Nurmi, J.-E., Niemi, P., Lerkkanen, M.-K., & Rasku-Puttonen,H. (2002). Developmental dynamics of achievement strategies, readingperformance, and parental beliefs. Reading Research Quarterly, 37,310–327. http://www.jstor.org/stable/748231

Aunola, K., Viljaranta, J., Lehtinen, E., & Nurmi, J.-E. (2013). The role ofmaternal support of competence, autonomy and relatedness in children’sinterest and mastery orientation. Learning and Individual Differences,25, 171–177. http://dx.doi.org/10.1016/j.lindif.2013.02.002

Bell, R. Q. (1968). A reinterpretation of the direction of effects in studiesof socialization. Psychological Review, 75, 81–95. http://dx.doi.org/10.1037/h0025583

Chen, C. S., & Stevenson, H. W. (1989). Homework: A cross-culturalexamination. Child Development, 60, 551–561. http://dx.doi.org/10.2307/1130721

Cooper, H., Lindsay, J. J., & Nye, B. (2000). Homework in the home: Howstudent, family, and parenting-style differences relate to the homeworkprocess. Contemporary Educational Psychology, 25, 464–487. http://dx.doi.org/10.1006/ceps.1999.1036

Deci, E. L., & Ryan, R. M. (2008). Facilitating optimal motivation andpsychological well-being across life’s domains. Canadian Psychology,49, 14–23. http://dx.doi.org/10.1037/0708-5591.49.1.14

Deci, E. L., Vallerand, R. J., Pelletier, L. G., & Ryan, R. M. (1991).Motivation and education: The self-determination perspective. Educa-tional Psychologist, 26, 325–346. http://dx.doi.org/10.1080/00461520.1991.9653137

Dumont, H., Trautwein, U., Lüdtke, O., Neumann, M., Niggli, A., &Schnyder, I. (2012). Does parental homework involvement mediate therelationship between family background and educational outcomes?Contemporary Educational Psychology, 37, 55–69. http://dx.doi.org/10.1016/j.cedpsych.2011.09.004

Dumont, H., Trautwein, U., Nagy, G., & Nagengast, B. (2014). Quality ofparental homework involvement: Predictors and reciprocal relationswith academic functioning in the reading domain. Journal of Educa-tional Psychology, 106, 144–161. http://dx.doi.org/10.1037/a0034100

Fan, X., & Chen, M. (2001). Parental involvement and students’ academicachievement: A meta-analysis. Educational Psychology Review, 13,1–22. http://dx.doi.org/10.1023/A:1009048817385

Frome, P. M., & Eccles, J. S. (1998). Parents’ influence on children’sachievement-related perceptions. Journal of Personality and Social Psy-chology, 74, 435–452. http://dx.doi.org/10.1037/0022-3514.74.2.435

Grolnick, W. S., & Pomerantz, E. M. (2009). Issues and challenges instudying parental control: Toward a new conceptualization. Child De-velopment Perspectives, 3, 165–170. http://dx.doi.org/10.1111/j.1750-8606.2009.00099.x

Grolnick, W. S., & Ryan, R. M. (1989). Parent styles associated withchildren’s self-regulation and competence at school. Journal of Educa-tional Psychology, 81, 143–154. http://dx.doi.org/10.1037/0022-0663.81.2.143

Hartup, W. W., & Laursen, B. (1991). Relationships as developmentalcontexts. In R. Cohen & A. W. Siegel (Eds.), Context and development(pp. 253–279). Hillsdale, NJ: Erlbaum.

Hill, N. E., & Tyson, D. F. (2009). Parental involvement in middle school:A meta-analytic assessment of the strategies that promote achievement.Developmental Psychology, 45, 740–763. http://dx.doi.org/10.1037/a0015362

Hu, L.-T., & Bentler, P. (1999). Cutoff criteria for fit indices in covariancestructure analysis: Conventional criteria versus new alternatives. Struc-tural Equation Modeling, 6, 1–55. http://dx.doi.org/10.1080/10705519909540118

Joussemet, M., Landry, R., & Koestner, R. (2008). A self-determinationtheory perspective on parenting. Canadian Psychology, 49, 194–200.http://dx.doi.org/10.1037/a0012754

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

431MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES

Kline, R. B. (2005). Principles and practice of structural equation mod-eling (2nd ed.). New York, NY: Guilford Press.

Leppänen, U., Niemi, P., Aunola, K., & Nurmi, J.-E. (2004). Developmentof reading skills among preschool and primary school pupils. ReadingResearch Quarterly, 39, 72–93. http://dx.doi.org/10.1598/RRQ.39.1.5

Lerkkanen, M.-K., Niemi, P., Poikkeus, A.-M., Poskiparta, M., Siekkinen,M., & Nurmi, J.-E. (2006–2011). The First Steps Study [Alkuportaat],ongoing. Jyväskylä, Finland: University of Jyväskylä.

Lerkkanen, M.-K., Rasku-Puttonen, H., Aunola, K., & Nurmi, J.-E. (2004).Predicting reading performance during the first and the second year ofprimary school. British Educational Research Journal, 30, 67–92. http://dx.doi.org/10.1080/01411920310001629974

Levin, I., Levy-Shiff, R., Applebaum-Peled, T., Katz, I., Komar, M., &Merian, N. (1997). Antecedents and consequences of maternal involve-ment in children’s homework: A longitudinal analysis. Journal ofApplied Developmental Psychology, 18, 207–227. http://dx.doi.org/10.1016/S0193-3973(97)90036-8

Lindeman, J. (1998). Allu–Ala-asteen lukutesti [Reading test for primaryschool]. Turku, Finland: University of Turku.

Little, R. J. A. (1988). A test of missing completely at random formultivariate data with missing values. Journal of the American Statis-tical Association, 83, 1198–1202. http://dx.doi.org/10.1080/01621459.1988.10478722

Moorman, E. A., & Pomerantz, E. M. (2008). The role of mothers’ controlin children’s mastery orientation: A time frame analysis. Journal ofFamily Psychology, 22, 734–741. http://dx.doi.org/10.1037/0893-3200.22.5.734

Muthén, B. O. (1998–2004). Mplus technical appendices. Los Angeles,CA: Muthén & Muthén.

Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus user’s guide (6thed.). Los Angeles, CA: Muthén & Muthén.

National Board of Education. (2004). National core curriculum for basiceducation 2004. Vammala, Finland: Vammalan Kirjapaino Oy. Re-trieved from http://www.oph.fi/english/sources_of_information/publications

Ng, F. F.-Y., Kenney-Benson, G. A., & Pomerantz, E. M. (2004). Chil-dren’s achievement moderates the effects of mothers’ use of control andautonomy support. Child Development, 75, 764–780. http://dx.doi.org/10.1111/j.1467-8624.2004.00705.x

Nurmi, J.-E., & Silinskas, G. (2014). Parents and their children’s schoollives—Commentary on the special issue, “Parents’ role in children’sschool lives.” British Journal of Educational Psychology, 84, 454–458.http://dx.doi.org/10.1111/bjep.12046

Onatsu-Arvilommi, T., & Nurmi, J.-E. (2000). The role of task-avoidantand task-focused behaviors in the development of reading and mathe-matical skills during the first school year: A cross-lagged longitudinalstudy. Journal of Educational Psychology, 92, 478–491. http://dx.doi.org/10.1037/0022-0663.92.3.478

Patall, E. A., Cooper, H., & Robinson, J. C. (2008). Parent involvement inhomework: A research synthesis. Review of Educational Research, 78,1039–1101. http://dx.doi.org/10.3102/0034654308325185

Pekrun, R. (2006). The control-value theory of achievement emotions:Assumptions, corollaries, and implications for educational research andpractice. Educational Psychology Review, 18, 315–341. http://dx.doi.org/10.1007/s10648-006-9029-9

Pomerantz, E. M., & Eaton, M. M. (2000). Developmental differences inchildren’s conceptions of parental control: “They love me, but they makeme feel incompetent.” Merrill-Palmer Quarterly, 46, 140–167. http://www.jstor.org/stable/23093346

Pomerantz, E. M., & Eaton, M. M. (2001). Maternal intrusive support inthe academic context: Transactional socialization processes. Develop-mental Psychology, 37, 174–186. http://dx.doi.org/10.1037/0012-1649.37.2.174

Pomerantz, E. M., Moorman, E. A., & Litwack, S. D. (2007). The how,whom, and why of parents’ involvement in children’s academic lives:More is not always better. Review of Educational Research, 77, 373–410. http://dx.doi.org/10.3102/003465430305567

Pomerantz, E. M., & Ruble, D. N. (1998). The role of maternal control inthe development of sex differences in child self-evaluative factors. ChildDevelopment, 69, 458–478. http://dx.doi.org/10.1111/j.1467-8624.1998.tb06202.x

Pomerantz, E. M., Wang, Q., & Ng, F. F. (2005). Mothers’ affect in thehomework context: The importance of staying positive. DevelopmentalPsychology, 41, 414–427. http://dx.doi.org/10.1037/0012-1649.41.2.414

Poulou, M., & Norwich, B. B. (2002). Cognitive, emotional and behav-ioural responses to students with emotional and behavioural difficulties:A model of decision-making. British Educational Research Journal, 28,111–138. http://dx.doi.org/10.1080/01411920120109784

Räsänen, P., & Aunola, K. (2007). Test of arithmetic skills. Unpublishedtest material. First Steps Study. Jyväskylä, Finland: University of Jyväs-kylä.

Räsänen, P., Salminen, J., Wilson, A. J., Aunio, P., & Dehaene, S. (2009).Computer-assisted intervention for children with low numeracy skills.Cognitive Development, 24, 450 – 472. http://dx.doi.org/10.1016/j.cogdev.2009.09.003

Rutter, M. (1997). Nature-nurture integration: The example of antisocialbehavior. American Psychologist, 52, 390 –398. http://dx.doi.org/10.1037/0003-066X.52.4.390

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and thefacilitation of intrinsic motivation, social development, and well-being.American Psychologist, 55, 68–78. http://dx.doi.org/10.1037/0003-066X.55.1.68

Sameroff, A. (2010). A unified theory of development: A dialectic inte-gration of nature and nurture. Child Development, 81, 6–22. http://dx.doi.org/10.1111/j.1467-8624.2009.01378.x

Scarr, S., & McCartney, K. (1983). How people make their own environ-ments: A theory of genotype greater than environment effects. ChildDevelopment, 54, 424–435. http://www.jstor.org/stable/1129703

Silinskas, G., Kiuru, N., Tolvanen, A., Niemi, P., Lerkkanen, M.-K., &Nurmi, J.-E. (2013). Maternal teaching of reading and children’s readingskills in Grade 1: Patterns and predictors of positive and negativeassociations. Learning and Individual Differences, 27, 54–66. http://dx.doi.org/10.1016/j.lindif.2013.06.011

Silinskas, G., Leppänen, U., Aunola, K., Parrila, R., & Nurmi, J.-E. (2010).Predictors of mothers’ and fathers’ teaching of reading and mathematicsin kindergarten and Grade 1. Learning and Instruction, 20, 61–71.http://dx.doi.org/10.1016/j.learninstruc.2009.01.002

Silinskas, G., Lerkkanen, M.-K., Tolvanen, A., Niemi, P., Poikkeus, A.-M.,& Nurmi, J.-E. (2012). The frequency of parents’ reading-related activ-ities at home and children’s reading skills during kindergarten and Grade1. Journal of Applied Developmental Psychology, 33, 302–310. http://dx.doi.org/10.1016/j.appdev.2012.07.004

Silinskas, G., Niemi, P., Lerkkanen, M.-K., & Nurmi, J.-E. (2013). Chil-dren’s poor academic performance evokes parental homework assis-tance: But does it help? International Journal of Behavioral Develop-ment, 37, 44–56. http://dx.doi.org/10.1177/0165025412456146

Silinskas, G., Parrila, R., Lerkkanen, M.-K., Poikkeus, A.-M., Niemi, P., &Nurmi, J.-E. (2010). Mothers’ reading-related activities at home andlearning to read during kindergarten. European Journal of Psychology ofEducation, 25, 243–264. http://dx.doi.org/10.1007/s10212-010-0014-9

Stattin, H., & Kerr, M. (2000). Parental monitoring: A reinterpretation.Child Development, 71, 1072–1085. http://dx.doi.org/10.1111/1467-8624.00210

Taylor, L. C., Clayton, J. D., & Rowley, S. J. (2004). Academic social-ization: Understanding parental influences on children’s school-related

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

432 SILINSKAS, KIURU, AUNOLA, LERKKANEN, AND NURMI

development in the early years. Review of General Psychology, 8,163–178. http://dx.doi.org/10.1037/1089-2680.8.3.163

Xu, J. (2011). Homework emotion management at the secondary schoollevel: Antecedents and homework completion. Teachers College Re-cord, 113, 529–560.

Xu, J. (2012). Predicting students’ homework environment management atthe secondary school level. Educational Psychology, 32, 183–200. http://dx.doi.org/10.1080/01443410.2011.635639

Xu, J., & Corno, L. (1998). Case studies of families doing third-gradehomework. Teachers College Record, 100, 402–436.

Received July 26, 2013Revision received December 1, 2014

Accepted January 19, 2015 �

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

433MOTHERS’ HOMEWORK-RELATED AFFECT AND PRACTICES