Conversations around homework: Links to parental mental health, family characteristics and child psychological functioning

Copyright © The British Psychological SocietyReproduction in any form (including the internet) is prohibited without prior permission from the Society

Conversations around homework: Links toparental mental health, family characteristicsand child psychological functioning

Lynne Murray1*, Matt Woolgar1, Carla Martins2, Anna Christaki1,Alison Hipwell3 and Peter Cooper1

1School of Psychology, The University of Reading, UK2Dep Psicologia IEP, Universidade Do Minho, Portugal3University of Pittsburgh, USA

Parents are increasingly expected to supplement their children’s school-based learningby providing support for children’s homework. However, parents’ capacities to providesuch support may vary and may be limited by the experience of depression. This mayhave implications for child development. In the course of a prospective, longitudinalstudy of children of postnatally depressed and healthy mothers, we observed mothers(N ¼ 88) and fathers (N ¼ 78) at home during maths homework interactions with their8-year-old children. The quality of parental communication was rated and analysed inrelation to child functioning. The quality of communication of each of the parents wasrelated to their mental state, social class and IQ. While postnatal depression was notdirectly related to child development, there was some evidence of the influence ofmaternal depression occurring in the child’s school years. Different aspects of parentalcommunication with the child showed specific associations with different childoutcomes, over and above the influence of family characteristics. In particular, childschool attainment and IQ were associated with parental strategies to encouragerepresentational thinking and mastery motivation, whereas child behaviouraladjustment at school and self-esteem were linked to the degree of parental emotionalsupport and low levels of coercion. Notably, the influence of maternal homeworksupport was more strongly related to child outcome than was paternal support, apattern reflected in mothers’ greater involvement in children’s schools and school-related activities. Some parents may need guidance in how to support their children’shomework if it is to be of benefit to child functioning.

Effects of depression on parenting and child developmentThe experience of clinical depression typically makes it hard for parents to engage with

their children in an appropriately responsive fashion and, in turn, evidence has

* Correspondence should be addressed to Professor Lynne Murray, School of Psychology, The University of Reading, 3 EarleyGate, Whiteknights, Reading, Berkshire, RG6 6AL, UK (e-mail: [email protected]).

TheBritishPsychologicalSociety

125

British Journal of Developmental Psychology (2006), 24, 125–149

q 2006 The British Psychological Society

www.bpsjournals.co.uk

DOI:10.1348/026151005X83568


accumulated to show that parental depression, including maternal depression occurring

after childbirth, is associated with a number of problems in early child development (see

Murray & Cooper, 2003). Because parent-child interactions in the context of depression

may show a wide range of difficulties that are not always evident in the volunteer

samples often studied in developmental psychology, examination of the links between

specific aspects of parenting style in these populations and different kinds of childoutcome may be particularly illuminating when seeking to understand the role of

parental communication in child functioning. For example, in our own prospective,

longitudinal study of the development of children of postnatally depressed and healthy

mothers (Murray, 1992), we found particular patterns of communication between

depressed mothers and their infants to be associated with specific child developmental

difficulties. Thus, whereas depressed mothers’ lack of active engagement with the infant

and difficulty in focusing on infant experience was associated with poorer infant

cognitive functioning (Murray, Fiori-Cowley, Hooper, & Cooper, 1996; Murray,Kempton, Woolgar, & Hooper, 1993), raised levels of hostility (also more prevalent

among postnatally depressed mothers than healthy mothers) were predictive of the

child’s feelings of low self-worth (Murray, Woolgar, Hipwell, & Cooper, 2001) and the

development of conduct problems (Morrell & Murray, 2003).

For some postnatally depressed mothers the depressive episode following delivery

quickly remits, whereas for others episodes are chronic and recurrent and, as the child

develops, mothers experiencing prolonged depression may find it particularly hard to

manage the changing demands on their capacity to support their child. We previouslyassessed the children in our longitudinal study as they adjusted to the challenges of

starting school at age five: we identified a number of child social difficulties associated

with maternal postnatal depression (Murray et al., 1999) as well as child adjustment

problems arising in the context of more recent, as well as postnatal, maternal episodes

(Sinclair & Murray, 1998). Our next assessment of this sample took place when the

children were 8 years old. At this point in their development, children are faced with

considerable demands to meet the academic standards of the school curriculum and, in

seeking to understand how maternal depression and any associated patterns ofparenting might affect the child’s development at this stage, we therefore considered it

would be particularly fruitful to focus on the way in which parents communicated with

their children during school-relevant homework tasks.

Homework as a context for studying parent-child interactionsChildren have been required to supplement their learning in school by doing homeworkever since the mid-nineteenth century (Gordon, 1980), the amount and nature varying

according to educational and political fashion. Research on the value of homework in

promoting school achievement suggests positive benefits at secondary school but the

effects are far less clear for younger children (Cooper, Lindsay, Nye, & Greathouse,

1998). In both the USA and the UK, parents have been increasingly encouraged to

involve themselves in their children’s education in recent years (Cowan & Traill, 1998;

UK Department of Education and Skills, 2004; Grolnick & Slowiaczek, 1994; Shumow,

1998; US Department of Education, 1994). Homework is considered a particularlyimportant bridge between home and school and one that may play a potentially

important role in enhancing the child’s performance (Grolnick & Slowiaczek, 1994;

Senechal & Le Fevre, 2002). On the parents’ side, homework is something that they

expect their children to do; parents also expect and aim to help their children with

Lynne Murray et al.126


homework (Epstein, 1986; Keys, Harris, & Fernandes, 1995; MacBeath & Turner, 1990;

Cowan & Hallam, 1999). Nevertheless, parents may not always feel well-equipped to

help their children and interactions with the child during attempts to help with

homework can be challenging and, at times, conflictual (Cowan & Traill, 1998; Levin

et al., 1997). This may be particularly likely in relation to mathematics, where the

strategies taught in school may differ from those used by the parents (Shumow, 1998).The degree to which depression may influence the parents’ ability to provide scholastic

support has been little researched to date. However, a study by Goldsmith and Rogoff

(1995) comparing the behaviour of dysphoric and non-dysphoric women in interactions

with 5- to 6-year-old children showed that those with dysphoric symptoms had

particular difficulties with structured teaching tasks, being less sensitive to both the

child’s emotional cues and the child’s understanding of the problem.

Dimensions of parental scholastic support and their role in child developmentThere is consistent evidence that the nature of parent-child problem-solving interactions

and support for children’s school-based activities is associated with longer term childoutcomes, including both academic achievement (Estrada, Arsenio, Hess, & Holloway,

1987) and behavioural adjustment and peer relationships (Johnson & Jason, 1994).

A number of dimensions of parental communication during problem-solving have been

considered important in promoting child progress and adjustment. One is that of the

motivational framework within which any tutoring takes place; in particular, the degree

to which the parent stimulates the child’s enthusiasm and sense of mastery (Diaz, Neal,

& Vachio, 1991; Nolen-Hoeksema, Wolfson, Mumme, & Guskin, 1995; Ratner & Stettnet,

1991; Vygotsky, 1962, 1978). A second dimension concerns the parents’ teaching role

and their ability to help the child understand the problem and find appropriate

strategies for its solution. In this regard, the degree to which the parent encourages

independent representational understanding, rather than simply identifying the

relevant elements of the task for them, has been emphasized (Diaz et al., 1991; Hubbs-

Tait, McDonald-Culp, Culp, & Miller, 2002; Landry, Smith, Swank, & Miller-Loncar, 2000;

Sigel, 1982). The issue of high rates of parental coercive control or intrusiveness during

problem-solving tasks has also been highlighted (Nolen-Hoeksema et al., 1995), with

evidence that such behaviour is associated not only with poor child cognitive outcomes(Culp, Hubbs-Tait, Culp, & Starost, 2000; Egeland, Pianta, & O’Brien, 1993) but also with

general child behaviour problems (Dodge, Pettit, & Bates, 1997; Nix et al., 1999). Finally,

a number of studies have shown the importance of parental emotional support during

problem-solving (Barocas et al., 1991; Diaz et al., 1991). Notably, however, it seems to be

low-level emotional support that is of particular benefit to the child, whereas clear

parental statements of praise may actually be unhelpful (Fagot & Gauvain, 1997).

The role of fathersIn contrast to the volume of research on the role that mothers play in their children’s

development, the contribution of fathers has seldom been studied (Lamb, 1997). With

regard to child scholastic achievements, this may, in part, stem from the fact thatmothers play a more active role than fathers in liaising between school and home

(David, 2002), and spend more time with children on homework-related activities

(Laasko, 1995; MacLeod, 1996). Aside from the issue of the relatively greater maternal

involvement in children’s schooling, there is also evidence that, despite broad

Conversations around homework 127

https://www.researchgate.net/publication/227717666_The_Relation_between_Mothers'_Hostile_Attribution_Tendencies_and_Children's_Externalizing_Behavior_Problems_The_Mediating_Role_of_Mothers'_Harsh_Discipline_Practices?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/232561601_Social_and_Interpersonal_Determinants_of_Developmental_Risk?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/233190680_Does_British_research_support_claims_about_the_benefits_of_parents_hearing_their_children_read_regularly_at_home_A_closer_look_at_the_evidence_from_three_key_studies?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/14068478_Mother-child_problem_solving_Continuity_through_the_childhood_years?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/233896797_Parents'_Reactions_to_Teacher_Practices_of_Parent_Involvement?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/232532129_Maternal_teaching_in_the_zone_of_proximal_development_A_comparison_of_low-_and_high-risk_dyads?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/232532129_Maternal_teaching_in_the_zone_of_proximal_development_A_comparison_of_low-_and_high-risk_dyads?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/223571059_Promoting_parental_attunement_to_children's_mathematical_reasoning_through_parent_education?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/31670039_The_Role_of_the_Father_in_Child_Development_M_R_Lamb?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==


communalities in the communicative strategies of mothers and fathers, there are key

differences (Jain, Belsky, & Crnic, 1996; Laasko, 1995). In the context of maternal

depression, the role of fathers may become particularly important. For example, studies

of infants of depressed mothers have shown that infants who become distressed and

withdrawn in the context of interactions with their mothers can be alert and well-

regulated when interacting with their healthy fathers (Hossain, Field, Gonzalez,Malphurs, & Del Valle, 1994). In the current study, therefore, we included fathers as well

as mothers and examined the nature of their homework interactions with the child, as

well as the extent of their overall involvement with the child’s schoolwork.

Design of current studyWe conducted naturalistic observations of the 8-year-old children of postnatallydepressed and control mothers as they completed mathematics problems in their home

in the presence of each of their parents. We focused on four key aspects of parental

discourse with the child during the homework tasks: (i) the promotion of positive

mastery motivation, (ii) the promotion of independent representational understanding,

(iii) the provision of general emotional support, and (iv) the presence of coercive

control. We examined these communication styles in relation to parental mental state

and family characteristics. While we anticipated that mothers who had experienced

postnatal depression might have residual difficulties in their interactions with theirchild, we predicted that mothers who had experienced depression during the child’s

school years would experience particular problems in negotiating the demands of

supporting their child during homework. Both previous results for this sample (Murray

et al., 1996, 1993; Murray, Hipwell, Hooper, Stein, & Cooper, 1996) and wider research

on the effects of depression on parent-child interactions (see review by Goodman &

Gotlib [1999]) have shown increased rates of hostility and coercion, and poorer capacity

to focus on the child’s agenda and promote the child’s cognitive development in

depressed mothers. We anticipated, therefore, that depressed maternal mood wouldinterfere with all four dimensions of parental homework support. In addition to

examining the role of parental mental state, other key family variables that could be

expected to have an impact on parenting and child outcomes were investigated, namely,

social class, parental conflict and IQ. We investigated the associations between parents’

behaviour with their child during homework and the child’s development, and, in

particular, their functioning in school. This included their performance on school

achievement tests, IQ, general adjustment to school and their level of self-esteem. Since

parental behaviour with the child is likely to be influenced by child, as well as family,characteristics, we also took into account the child’s functioning at the time of

the previous assessment at 5 years in each of the domains of child outcome assessed in

the current phase of the study. Given the central importance that parental support of the

child’s school functioning may have in child adjustment and performance, we predicted

that the nature of parental behaviour during homework would show associations with

child outcome, even when parental and family background factors and child

characteristics were taken into account. Moreover, in view of the accumulating

evidence for the specificity of relationships between parental and child behaviours, wepredicted that particular aspects of parenting during the homework task would be

differentially associated with distinct kinds of child outcome. Specifically, we expected

that the promotion of both positive mastery motivation and independent

representational thinking would be associated with positive child performance in



school achievement tests and high IQ. We also expected that the provision of general

emotional support and the absence of coercive control would be associated with high

self-esteem and good behavioural adjustment. Finally, in line with previous research, we

predicted that the contribution of maternal behaviour to child outcome would be

greater than that of paternal behaviour and that this would be reflected in differences in

the extent of mothers’ and fathers’ general involvement in their children’s schoolworkand activities.

Method

ParticipantsParticipants were originally recruited from a representative community sample of 702

primiparous women and their healthy, full-term infants. Mothers were screened at six

weeks with the Edinburgh postnatal depression scale (EPDS; Cox, Holden, & Sagovsky,

1987); those with scores indicating probable depression were interviewed. Sixty-one

women had experienced an episode of depression meeting research diagnostic criteria

(Spitzer, Endicott, & Robins, 1978) since delivery. After every alternate case was

identified (for the last six months of recruitment, after every case was identified),potentially psychiatrically well controls were randomly selected from those with low

EPDS scores (i.e. 9 or less) and these women were interviewed to confirm that no

depression had occurred either prior to or since delivery. Forty-two women were

selected as controls. All but three of the 103 women identified for recruitment

(each from the depressed group) agreed to participate in the study. Mothers and their

children were assessed at 18 months and again at 5 years. The current assessment took

place when the children were aged 8 years. At this point, 55 of the mothers and children

in the index group and 41 in the control group (94.8% and 97.6%, respectively, of thoseoriginally recruited) were seen again. We also attempted to assess as many of the

children’s fathers as possible: this was achieved for 46 (83.6%) of the index group and

35 (85.4%) of the controls. Sample characteristics are shown in Table 1.

ProcedureThe study children and their mothers initially visited the university research unit where

a range of assessments were conducted, including measurement of the child’s IQ, self-

esteem and their performance on a set of mathematics problems. The mother’s mental

state and IQ were also assessed and she was asked about her relationship with herpartner and her involvement with the child’s schoolwork and activities. Her permission

was sought to obtain teacher reports of the child’s behaviour and their performance on

school achievement tests. Two visits were subsequently made to the family home; on

one occasion the child was observed with their mother and on the other with their

father, each time completing a different set of mathematics problems. The order of both

the observation of mother and father and the allocation of the two sets of mathematics

problems was randomized. The parents were each asked to give the kind of back-up

support to the child that they might ordinarily do at home. The researcher set a video-camera to record a view of both child and parent and left the room for a 20 minute

period while the homework task was completed. In the case of fathers, who had not

previously attended the research unit assessment, the researcher then conducted

mental state and IQ assessments and interviewed the father to ascertain his involvement

in the child’s schoolwork and activities.



Measures

Parent measuresMental state. Maternal mental state when child was aged 8 years was assessed using the

affective disorders section of the structured clinical interview for DSM diagnoses (SCID;

First, Spitzer, & Williams, 1996) to determine whether or not the mother was currently

depressed. Mothers were also questioned using the same schedule about their moodsince the previous assessment at 5 years and the timing of any episodes of depression

was recorded. Mothers had similarly been interviewed at 5 years and 18 months;

together, the information from these three interviews, along with that identifying the

presence or absence of depression in the first three months post-partum, was used to

derive the following mental state variables: (i) post-partum depression, (ii) current or

recent depression (defined as occurring within the previous 12 months) (iii) depression

occurring during the child’s school years (i.e. between 5 and 8 years) and the overall

duration of the child’s exposure to depression (calculated in months). Fathers were alsointerviewed, using the full SCID. This was used, as for the mothers, to ascertain their

current and recent mental state as well as whether they had experienced an episode of

disorder in the child’s school years. However, since the fathers had not previously been

interviewed, it was not considered possible to obtain valid information concerning the

occurrence of any episodes prior to this time. Since the number of fathers who

experienced any one diagnosis in the time period concerned was small (depression:

N ¼ 18; alcohol abuse: N ¼ 3; an anxiety disorder: N ¼ 10), fathers were categorized

for each time period as having either experienced or not experienced any diagnosis.Maternal and Paternal IQ. The National Adult Reading Test (NART; Nelson &

Willison, 1991) was administered to provide a measure of overall IQ. In normal

populations, scores on the NART correlate highly with those on standard IQ tests. The

measure is widely used in research and clinical practice to estimate IQ in populations

with neurological or psychiatric disorders, as it overcomes the difficulty that symptoms

Table 1. Characteristics of the sample at eight and a half years

Control Index Statistics

No. assessed at 2 months/ 812 years 42/41 58/55

Age of child (in months) at 812 year

assessment (M [SD ])103.20 (3.30) 103.4 (2.34) tð93Þ ¼ 0:35

Proportion of boys 47% 54% x2ð1Þ ¼ 0:39Social class I, II and II non-manual 67% 64% x2ð1Þ ¼ 0:76Maternal mental state

Postnatal depression 0% 100%Current depression (81

2 years) 2.4% 14.5% x2ð1Þ ¼ 4:05*5–81

2 years depression 19.5% 49.1% x2ð1Þ ¼ 8:87**No. months depressed during

child’s lifetime (M [SD ])2.32 (4.05) 16.82 (13.61) tð94Þ ¼ 26:60***

Paternal mental stateCurrent disorder (81

2 years) 5.7% 17.4% x2ð1Þ ¼ 2:505–81

2 years disorder 17.1% 30.4% x2ð1Þ ¼ 1:89Parental conflict

Criticism score (M [SD ]) 3.66 (2.13) 5.53 (2.00) tð82Þ ¼ 24:14***

*p , :05; **p , :01; ***p , :001:



(e.g. retardation) may interfere with performance in standard IQ assessments and make

the estimation of IQ invalid.

Parental conflict. In order to assess the quality of parental relationship, the Dyadic

adjustment scale (DAS; Spanier, 1976) questionnaire was completed, with additional

items to assess perceived and felt criticism (Hooley & Teasdale, 1989).

Parental involvement. Parents were asked a series of structured questions about

their participation in the child’s school (e.g. running clubs, serving as parent governor,

assisting in class) and their involvement in the child’s school-related work and activities

(e.g. accompanying child to museums/sites relevant to school projects, helping child

find source materials from libraries/Internet sites). Parents were also asked about their

involvement in the child’s other non-school based activities (e.g. sports interests,

hobbies). Involvement in each area was rated on a 5-point scale.

Homework materialsThree sets of mathematics problems were compiled from the Primary Maths Bank

(Pollitt & Munro, 1989). This is a Rasch-calibrated set of items of varying difficulty

covering the range of mathematics taught to children in school between the ages of 7

and 13 years, from which questions may be selected and mixed to construct extended

tests. Each of the three sets in the current study comprised 25 items, including

computation, concepts and problem-solving, money, measures, time, fractions and

shape. The three sets were comparable in terms of their level of difficulty: Set A:

M ¼ 48:52 (SD ¼ 6:4), range 35.3–59.6; Set B: M ¼ 48:19 (SD ¼ 6:78), range 35.9–60.4;Set C: M ¼ 48:90 (SD ¼ 6:46), range 37.9–61.9.

Set A was used when the children visited the research unit to establish each child’s

level of ability; Sets B and C were used at home, one with each parent. To ensure that the

children were all posed with a similar, moderate degree of challenge, they were asked to

start the sets at home at the point just before they started failing items on Set A.

Parental homework support dimensions. Videotapes of parent-child discourse

during the homework sessions were transcribed1. Separate utterances were identified

on the basis of timing and intonational cues, as well as their content. Utterances were

examined in conjunction with the video and, for each one, the presence or absence of

each of the four dimensions of parental behaviour was recorded by a researcher who

had no other information concerning the children and their families. A random sample

of homework interactions for 10 of the children was scored by a second researcher.

(1) Promotion of mastery motivation: the parent shows positive enthusiasm for thetask, their willingness to help and encourages the child. For example: ‘Oh, let’s

have a go at that one, you can do it with the scales – OK?’; ‘Shall I just show you

that one to get you started?’

(2) Promotion of representational understanding: the parent may give information

about the principles involved in the problem. For example: ‘It’s like saying, for

example, that if there were two bowls and you have one of them, you have the

same proportion of bowls as if there were four and you had two; you have half in

each case.’ Alternatively, the parent may offer suggestions involving the principlesof the problem. For example: ‘There might be a number you can multiply the top

1Poor audio quality prevented scoring in a few cases.


https://www.researchgate.net/publication/240111878_Measuring_Dyadic_Adjustment_New_Scales_for_Assessing_the_Quality_of_Marriage_and_Similar_Dyads?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/20386901_Predictors_of_Relapse_in_Unipolar_Depressives_Expressed_Emotion_Marital_Distress_and_Perceived_Criticism?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==


one by. Then multiply the bottom one by the same number and it will equal one of

these’; ‘Perhaps you could use what you did for the question about the cost of

pencils’.

(3) Provision of general low-level emotional support: the parent demonstrates their

psychological availability and supportive awareness of the child by means of back-

up, affirmative comments (it is important to note that such comments were notconfined to times when the child was completing the problem successfully, but

could just as well be made as the child struggled with the task). For example: ‘yes,

that’s it’; ‘OK, you’re starting there, then’; ‘mm, mm, right’; ‘I see what you’re

doing’.

(4) Coercive control: The parent speaks forcibly and gives a strong directive for the

child to do something. For example: ‘Write that down there: : :put the 1 in the

right hand column’; ‘Now add 7 plus 8’.

Child measures: 8 year outcomesSchool academic achievement. Results of the children’s performance on the NationalCurriculum Tests at Key Stage 1 (SATS) in English, Mathematics and Science were

obtained. Scores across the three subjects were highly correlated (at least .57 between

any two tests) and the mean score for the three tests was used to characterize the level of

the children’s school achievement.

IQ. Four subscales from the WISC-III UK (Wechsler, 1992) were used – similarities

and vocabulary (both verbal), and block design and object assembly (both

performance); the prorated, scaled verbal and performance scores were summed to

give the Full scale IQ.Self-esteem. The Perceived competence scale for children (Harter, 1982) was

administered. The total score was used as an overall measure of child self-esteem.

Adjustment at school. Teachers completed the teacher form of the Rutter scale

(Rutter, 1967). This questionnaire has been widely used in research and clinical practice

to identify children with clinically significant behaviour problems in the context of

school. A total score of nine or above was taken to reflect the presence of at least a

moderate level of disturbance.

Child measures: 5 year outcomesIQ. The McCarthy scales of children’s abilities (McCarthy, 1972) were administered

(Murray et al., 1996); the total score on the General Cognitive Index (GCI) was used as a

measure of IQ at 5 years.

Self-esteem. The children’s behaviour during a competitive card game with a friend

was recorded at 5 years (Murray et al., 2001) and the child’s expression of low self-worth

during the procedure was used as an index of low self-esteem.

Adjustment at school. The Preschool Behaviour Checklist (PBCL; McGuire &

Richman, 1988) was completed by the children’s teachers (Sinclair & Murray, 1998).Total scores were used as a measure of the level of child behavioural disturbance.

Data analytic strategyFirst we conducted descriptive analyses of the parental homework communication

variables and compared the characteristics of mothers’ and fathers’ behaviour. We then

conducted univariate correlational analyses; these were used, first, to examine how


https://www.researchgate.net/publication/51327171_Effects_of_depression_on_children's_adjustment_to_school_Teacher's_reports?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/233896943_The_Perceived_Competence_Scale_of_Children?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/17153537_A_Children's_Behaviour_Questionnaire_for_Completion_by_Teachers?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/14187687_The_Impact_of_Postnatal_Depression_and_Associated_Adversity_on_Early_Mother_Infant_Interaction_and_Later_Infant_Outcome?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==

https://www.researchgate.net/publication/11659905_Cognitive_Vulnerability_to_Depression_in_5-year-old_Children_of_Depressed_Mothers?el=1_x_8&enrichId=rgreq-5d80a576-b18e-4a5a-b9f1-3949a0310865&enrichSource=Y292ZXJQYWdlOzIyNzgwNTg2NTtBUzoxMDEyODk5NzUzNTMzNTBAMTQwMTE2MDgyMDc3Mw==


parental mental state and wider family circumstances were related to parental behaviour

with the child and the child’s development and, second, to examine the relations

between each parent’s behaviour with the child during homework and the child’s

functioning. Finally, a series of regression analyses with blocked entry procedures was

performed to examine the influence of parental homework support on child outcome,

having taken relevant parental and family characteristics, as well as the child’s previousperformance in each domain of functioning, into account.

Results

Characteristics of homework communication in mothers and fathersThe total scores for the dimensions of homework support were expressed as

percentages of the total number of utterances. Inspection of the distributions of scores

on each of the parental homework support dimensions showed only those for coercive

control to be normally distributed.

Normality of distribution for the other variables was achieved by means of log

transformations. All dimensions showed good inter-rater reliability; kappas were as

follows: promotion of mastery motivation: .73; promotion of representationalunderstanding: .78; provision of emotional support: .94; coercive control: .88.

The characteristics of maternal and paternal homework communication (untrans-

formed scores) are shown in Fig. 1. There was considerable variability among the

parents; some interactions showed high levels of coercive control and low levels of

emotional support or promotion of the child’s motivation and understanding about the

task, whereas other parents were highly supportive and facilitated their children’s

endeavours. Overall, as can be seen from Fig. 1, mothers and fathers differed little from

each other on most dimensions of homework support (coercive control: z ¼ 0:67;promotion of independent representational understanding: z ¼ 1:76; emotional

support: z ¼ 1:54). If anything, fathers tended to show a more optimal pattern and

this was significant in the promotion of mastery motivation dimension (z ¼ 2:07,p , :05). While mothers’ and fathers’ styles of homework communication of were

broadly similar, the degree of general involvement in both the child’s school and their

school-related activities showed striking differences between the two parents, with

mothers being more involved than fathers (z ¼ 4.23, p , :001 and z ¼ 5:36, p , :001,for school and school-related activities, respectively). By contrast, the two parents’involvement in other aspects of the children’s lives did not differ (z ¼ 0:88; see Fig. 2).

Parental mental state, family characteristics, child gender and homework supportand their relation to child functioningTable 2 shows the univariate correlations (Pearson, point-biserial etc., as appropriate)

between parental and family variables, child gender and parental homework support

behaviours. It also shows the associations between these variables and the measures of

child functioning. Postnatal depression showed no association with any dimension of

maternal communication with the child, nor was it associated with child outcome. Thesame was true for the overall duration of the child’s exposure to maternal depression.

By contrast, current and more recent maternal depression was associated with poorer

support for the child in each dimension of homework discourse (i.e. more coercive

control, less emotional support and less promotion of mastery motivation and

representational understanding) and, with the exception of child IQ, generally poorer



child functioning. Paternal mental disorder showed only one significant association

with the nature of homework communication, with more coercive control occurring

in the context of current disorder. Current paternal disorder was also linked to a

greater likelihood of child behaviour problems in school. In general, middle-class

parents and those with higher IQs showed better patterns of homework support than

other parents; in turn, their children had better outcomes. These patterns ofassociation were similar for mothers and fathers, although links were generally

stronger for maternal variables.

The quality of parental homework communication (particularly mothers’ homework

communication) was associated with a number of child outcomes. With regard to child

IQ, mothers who promoted mastery motivation and representational understanding,

provided more emotional support and expressed less coercive control during the

interaction had children with higher scores, as did fathers who gave more emotional

support. Associations with child academic achievement, as assessed by the SATS,

broadly paralleled those for IQ; in the main, these were somewhat weaker than those

concerning child IQ, although some additional associations emerged. Thus, in addition

to the positive influence of maternal emotional support and low rates of coercive

Emotional support Coercive control

Representational understanding Mastery motivation

Mother Father

50

40

30

20

10

0

50

40

30

20

10

0

80

60

40

20

0

40

10

0

Figure 1. Homework support by mothers (N ¼ 78) and fathers (N ¼ 78).



control, paternal promotion of mastery motivation and low levels of coercive control

were associated with better results on the SATS.

Children’s self-esteem was related to most dimensions of maternal homework

communication, but not with paternal interactions. Finally, children whose mothers

exhibited more emotional support and whose fathers promoted more masterymotivation and expressed less coercive control were amongst the better adjusted at

school.

Multivariate analysesGiven the associations between different aspects of homework support, and between

mental state and other family characteristics, and the parents’ behaviour with the child,

the question arises of whether particular aspects of homework communication were

important and whether these influenced child functioning over and above the effects ofparental and family variables. This central question was addressed by means of

multivariate regressions. In addition, since child characteristics could themselves have

influenced parental behaviour, regression analyses also took into account the child’s

previous functioning in each domain of development. In the initial analysis (shown as

model 1 in Tables 3 to 6) we determined the relative contribution of the different

homework communication variables; in order to avoid prematurely deleting variables,

we retained those where p , :15 for subsequent analysis. In the subsequent, main

model (shown as model 2 in Tables 3 to 6), variables were entered in two blocks. Thefirst block consists of two types of variables: the developmentally equivalent variable to

that predicted at 8 years, as previously assessed at age 5 (i.e. for child IQ and academic

achievement, we controlled for previous child IQ [GCI]; for child self-esteem, we

controlled for earlier expressions of low self-worth; and for child behavioural

Maternal (N= 78) and paternal (N=78) involvement in school and home

0

0.5

1

1.5

2

2.5

3

3.5

Involvementat home

Involvementat school

Involvementin school work

MotherFather

Figure 2. Maternal (N ¼ 78) and paternal (N ¼ 78) involvement in school and home.



Table

2.

Ass

oci

atio

ns

among

par

enta

lhom

ework

dim

ensi

ons,

child

gender

,w

ider

fam

ilych

arac

teri

stic

san

dch

ildoutc

om

e

Mat

ernal

hom

ework

(N¼

88)

Pat

ernal

hom

ework

(N¼

78)

Child

outc

om

e

Mas

tery

motiva

tion

Rep

res.

under

stan

din

g

Em

otional

support

Coer

cive

contr

ol

Mas

tery

motiva

tion

Rep

res.

under

stan

din

g

Em

otional

support

Coer

cive

contr

ol

IQ

(N¼

88)

Aca

dem

ic

achie

vem

ent

(N¼

70)

Self-

este

em

(N¼

87)

School

mal

adju

stm

ent

(N¼

80)

Child

gender

(N¼

88)

.092

2.0

99

2.1

43

.020

2.1

27

2.1

45

.180

.079

.146

2.0

17

.011

.368**

Soci

alcl

ass

(N¼

88)

2.0

85

2.3

00**

2.3

47**

.368**

*2

.173

2.2

12þ

2.3

68**

.307**

2.3

79**

*2

.182

.093

.201þ

Mat

ernal

vari

able

s(N

¼88)

PN

D2

.026

2.1

46

.023

.019

––

––

.132

2.0

96

.019

.123

Curr

ent

dep

ress

ion

2.2

35*

2.2

20*

2.2

15*

.253*

––

––

2.1

60

2.3

15**

2.1

17

.162

5–8

year

sdep

ress

ion

2.1

57

2.0

47

2.2

40*

.169

––

––

2.0

76

2.2

90*

2.2

54*

.227*

No.m

onth

sdep

ress

ed2

.030

2.1

07

2.1

49

.046

––

––

.014

2.1

40

.011

.100

IQ.2

39*

.242*

.343**

2.3

61**

––

––

.364**

*.2

95*

.075

2.2

14þ

Pat

ernal

vari

able

s(N

¼78)

Curr

ent

dis

ord

er–

––

–2

.207þ

2.0

68

2.0

90

.278*

.064

2.0

40

2.2

18þ

.334**

5–8

year

sdis

ord

er–

––

–2

.010

2.1

20

.033

.058

.192þ

2.0

25

2.1

41

.328**

IQ–

––

–.2

10þ

.066

.285*

2.3

63**

.343**

.402**

2.0

57

2.1

68

Par

enta

lco

nflic

t(N

¼84)

2.0

49

2.0

27

2.0

96

.058

2.1

54

2.1

08

2.2

38*

.254*

2.0

26

2.1

10

2.1

89þ

2.0

35

Mat

ernal

hom

ework

(N¼

88)

Mas

tery

motiva

tion

–.4

27**

*.0

53

2.4

23**

*–

––

–.2

27*

.244*

.244*

2.0

83

Rep

res.

under

stan

din

g–

–.1

38

2.5

25**

*–

––

–.2

89**

.077

.117

2.1

35

Em

otional

support

––

–2

.526**

*–

––

–.3

63**

.183

.318**

2.3

68**

Coer

cive

contr

ol

––

––

––

––

2.3

05**

2.3

01*

2.3

89**

*.2

20þ

Pat

ernal

hom

ework

(N¼

78)

Mas

tery

motiva

tion

.129

––

––

.174

2.1

95þ

2.4

21**

*.0

50

.331**

.018

2.3

99**

Rep

res.

under

stan

din

g–

.234*

––

––

2.0

23

2.4

26**

*2

.083

.104

2.1

06

2.1

29

Em

otional

support

––

.353**

––

––

2.4

17**

*.2

80*

.173

.207þ

2.0

76

Coer

cive

contr

ol

––

–.4

56**

*–

––

–2

.221þ

2.4

06**

2.1

62

.266*

þp,

:10;*p

,:0

5;**p,

:01;**

*p,

:001.

Rep

res.

under

stan

din

g¼

repre

senta

tional

under

stan

din

g.



Table 3. Regression models for child IQ based on maternal, paternal and parental predictors

Maternal predictors

(N ¼ 88)

Paternal predictors

(N ¼ 78)

Parental predictors

(N ¼ 78)

b t b t b t

Model 1

M promotion of

mastery motivation

0.14 1.23 0.12 0.96

M promotion of repres.

understanding

0.21 1.70þ 0.30 2.10*

M emotional support 0.35 2.95** 0.20 1.47þþ

M coercive control 0.05 0.33 0.06 0.35

P promotion of mastery

motivation

0.04 0.28 20.03 20.22

P promotion of repres.

understanding

20.17 21.31 20.18 21.44

P emotional support 0.20 1.45 0.13 0.91

P coercive control 20.19 21.15 20.08 20.51

Model F Fð4; 83Þ ¼ 5:33,

p , :001

Fð4; 73Þ ¼ 2:37,

p , :10

Fð8; 69Þ ¼ 2:73,

p , :05

Model 2

Step 1

GCI (5 years) 0.49 5.33*** 0.43 4.18*** 0.43 4.20***

Child gender 0.15 1.78þ 0.17 1.78þ 0.18 1.86þ

Social class 20.19 22.04* 20.09 20.73 20.11 20.88

Maternal depression

(5–8 years)

0.17 1.89þ 0.15 1.50

Maternal IQ 0.21 2.26* 0.17 1.64

Paternal disorder

(5–8 years)

0.11 1.09 0.10 1.06

Paternal IQ 0.17 1.44 0.12 1.01

Model F Fð5; 80Þ ¼ 12:03,

p , :001, R 2 ¼ :43

Fð5; 70Þ ¼ 7:88,

p , :001, R 2 ¼ :36

Fð7; 68Þ ¼ 6:52,

p , :001, R 2 ¼ :40

Step 2

GCI (5 years) 0.46 4.98*** – – 0.40 3.89***

Child gender 0.19 2.23* – – 0.20 2.20*

Social class 20.13 21.31 – – 20.06 20.50

Maternal depression

(5–8 years)

0.17 1.93þ 0.17 1.73þ

Maternal IQ 0.15 1.62 0.09 0.86

Paternal disorder

(5–8 years)

– – 0.09 0.92

Paternal IQ – – 0.11 0.95

M promotion of mastery

motivation

– – – –

M promotion of repres.

understanding

0.19 2.13* 0.23 2.31*

M emotional support 0.12 1.26 0.12 1.12



adjustment in school at age 8, we controlled for the child’s score at 5 years on the PBCL),

and predictors comprising parental and family characteristics as well as child gender.

With regard to parental mental health, only the occurrence of disorder in the period 5 to

8 years was considered. This was, first, because neither postnatal depression nor

chronicity of child exposure showed any relation in univariate analyses with eithermaternal homework dimensions or children’s outcome and, second, only small numbers

of parents experienced current disorder. In the second block, as well as previous child

functioning and parental and family characteristics, we included those homework

communication variables retained in the initial step. For each child outcome, analyses

were first carried out for mothers2 and fathers separately and then for combined mother

and father variables.

Child IQ (see Table 3)In the analysis including maternal variables, high IQ scores at 8 years were, as expected,

predicted by better performance on the GCI at age 5; in addition, higher maternal IQ and

social class were associated with better child outcome. The occurrence of maternaldepression in the school years was, contrary to expectations, linked to higher child IQ

but this effect was not significant. Notably, maternal promotion of representational

understanding during the homework task contributed to the prediction of child IQ over

and above the influence of maternal IQ and social class; when this maternal homework

dimension was included, the effects of these other factors were no longer significant

Table 3. (Continued)

Maternal predictors

(N ¼ 88)

Paternal predictors

(N ¼ 78)

Parental predictors

(N ¼ 78)

b t b t b t

M coercive control – – – –

P promotion of mastery

motivation

– – – –

P promotion of repres.

understanding

– – – –

P emotional support – – – –

P coercive control – – – –

Model F Fð7; 78Þ ¼ 9:92,

p , :001, DR 2 ¼ :04,

Fð2; 78Þ ¼ 3:07,

p , :05

Fð9; 66Þ ¼ 6:14,

p , :001, DR 2 ¼ :05,

Fð2; 66Þ ¼ 3:29,

p , :05

þþp , :15;þp , :10; *p , :05; **p , :01; ***p , :001:Repres. understanding ¼ representational understanding.

2The analysis for combined mother and father variables includes only those mothers for whom father data were available (i.e.78 rather than the 88 mothers who were observed with their children and whose results are shown in the left hand columns ofTables 3 to 6). Each analysis in these tables was repeated for the 78 mothers for whom father data were available and, asthese were found to give identical results to those for the whole sample, they are not shown.



Table 4. Regression models for child academic achievement based on maternal, paternal and parental

predictors

Maternal predictors(N ¼ 88)

Paternal predictors(N ¼ 78)

Parental predictors(N ¼ 78)

b t b t b t

Model 1M promotion of mastery

motivation0.22 1.65þþ 0.27 1.95þ

M promotion of repres.understanding

20.11 20.84 20.05 20.35

M emotional support 0.10 0.72 20.04 20.26M coercive control 20.23 21.45 20.07 20.40P promotion of mastery

motivation0.19 1.22 0.17 1.07

P promotion of repres.understanding

20.07 20.54 20.13 21.00

P emotional support 0.05 0.34 0.04 0.25P coercive control 20.32 21.74þ 20.30 21.64þþ

Model F Fð4; 65Þ ¼ 2:50,p , :06

Fð4; 58Þ ¼ 3:45,p , :05

Fð8; 54Þ ¼ 2:45,p , :05

Model 2Step 1

GCI (5 years) 0.45 4.17*** 0.32 2.58* 0.32 2.62*Child gender 0.003 0.03 20.05 20.43 20.04 20.33Social class 0.08 0.66 0.07 0.45 0.14 0.95Maternal depression

(5–8 years)20.22 21.99þ 20.20 21.65

Maternal IQ 0.20 1.87þ 0.03 0.24Paternal disorder

(5–8 years)20.09 20.74 20.06 20.52

Paternal IQ 0.37 2.66** 0.37 2.61*

Model F Fð5; 63Þ ¼ 6:35,p , :001,R 2 ¼ :34

Fð5; 56Þ ¼ 3:86,p , :01,R 2 ¼ :26

Fð7; 54Þ ¼ 3:20,p , :01,R 2 ¼ :29

Step 2GCI (5 years) 0.46 4.28*** 0.29 2.36* 0.30 2.54*Child gender .001 0.003 20.02 20.13 20.02 20.18Social class 0.06 0.51 0.09 0.65 0.15 1.06Maternal depression

(5–8 years)20.19 21.81þ 20.16 21.40

Maternal IQ 0.18 1.69þ 20.02 20.18Paternal disorder

(5–8 years)20.05 20.43 20.02 20.20

Paternal IQ 0.29 2.06* 0.34 2.48*M promotion of mastery

motivation0.20 1.96þ 0.29 2.57*


– – – –

M emotional support – – – –M coercive control – – – –



(Model 2, Step 2). By contrast, when father variables were analysed, only the effect of

child IQ at 5 years predicted 8 year functioning. When maternal and paternal variables

were considered together, it was still the case that the mother’s promotion of

representational understanding was influential, while no other variables, apart from the

child’s gender and earlier performance, were significantly associated with child IQscores.

Academic achievement (see Table 4)In the analysis that included maternal variables, the only significant predictor of

child performance on the SATS was their GCI score at 5 years, with children with

higher scores at 5 years again achieving better results at 8 years. In addition, there

was a tendency for children whose mothers had been depressed at some point

between the child’s fifth and eighth year, and those with mothers with lower IQs,

to have poorer SATS results. When consideration of the mother’s interactions withthe child during homework was included, children whose mothers promoted

mastery motivation at 8 years tended to achieve better academic results,

independent of the other predictors. The analysis including father variables showed

a positive effect for higher paternal IQ and previous child performance, and a

tendency for children whose fathers showed low rates of coercive control to

perform well. When mother and father behaviours were considered together, along

with the influence of previous child behaviour and paternal IQ, only the mother’s

homework communication – the promotion of mastery motivation – wassignificantly associated with child SAT scores.

Self-esteem (see Table 5)In the analysis including maternal measures, poor child self-esteem at 8 years was

unrelated to the previous expression of low self-worth. However, low social class,


Maternal predictors(N ¼ 88)

Paternal predictors(N ¼ 78)

Parental predictors(N ¼ 78)

b t b t b t

P promotion of masterymotivation

– – – –


– – – –

P emotional support – – – –P coercive control 20.25 21.99þ 20.16 21.31

Model F Fð6; 62Þ ¼ 6:18,p , :001,DR 2 ¼ :04,

Fð1; 62Þ ¼ 3:85,p , :06

Fð6; 55Þ ¼ 4:05,p , :01,

DR 2 ¼ :05,Fð1; 55Þ ¼ 3:99,

p , :06

Fð9; 52Þ ¼ 4:00,p , :001,DR 2 ¼ :12,

Fð2; 52Þ ¼ 5:12,p , :01




Table 5. Regression models for child self-esteem based on maternal, paternal and parental predictors

Maternalpredictors(N ¼ 88)

Paternalpredictors(N ¼ 78)

Parentalpredictors(N ¼ 78)

b t b t b t


motivation0.18 1.56þþ 0.19þþ 1.46


20.14 21.18 20.13 20.92

M emotional support 0.18 1.50þþ 0.17 1.20M coercive control 20.29 22.01* 20.32 21.84þ

P promotion of masterymotivation

20.01 20.08 20.09 20.69


20.19 21.42 20.22 21.74þ

P emotional support 0.12 0.80 20.01 20.09P coercive control 20.20 21.14 20.10 20.62

Model F Fð4; 82Þ ¼ 5:11, p ,

:001Fð4; 72Þ ¼ 1:48,

nsFð8; 68Þ ¼ 2:68,

p , :05

Model 2Step 1

Child self-esteem(5 years)

20.11 21.04 20.15 21.24 20.07 20.62

Child gender 20.02 20.14 0.02 0.13 0.04 0.34Social class 0.21 1.80þ 0.06 0.44 0.22 1.53Maternal depression

(5–8 years)20.27 22.36* 20.32 22.68**

Maternal IQ 0.08 0.69 0.20 1.65Paternal disorder

(5–8 years)20.09 20.76 20.05 20.46

Paternal IQ 20.03 20.23 20.06 20.44

Model F Fð5; 81Þ ¼ 2:05,p , :10, R 2 ¼ :11

Fð5; 71Þ ¼ 0:63,ns, R 2 ¼ :04

Fð7; 69Þ ¼ 1:94,p , :10, R 2 ¼ :17

Step 2Child self-esteem

(5 years)0.04 0.36 – – 0.02 0.15

Child gender 0.05 0.49 – – 0.04 0.35Social class 0.34 3.19** – – 0.26 1.85þ

Maternal depression(5–8 years)

20.26 22.52* 20.30 22.66**

Maternal IQ 20.08 20.79 0.05 0.45Paternal disorder

(5–8 years)– – 20.06 20.58

Paternal IQ – – 20.09 20.68M promotion of mastery

motivation0.10 0.93 0.04 0.29


– – – –



maternal depression occurring in the child’s school years and the mother’s use of

coercive control during the homework task were associated with low self-esteem. In

the analysis including paternal variables, none were found to predict this childoutcome and when both maternal and paternal behaviours were considered

together, the effects of maternal depression and the mother’s use of coercive

control remained significant.

School maladjustment (see Table 6)The analysis including maternal measures showed child behaviour difficulties at school

at 8 years to be significantly predicted by problems at 5 years, as assessed by the PBCL,

and, within the maternal and family characteristics, both child gender and maternal IQ

were found to be related to child difficulties; thus, boys and children whose mothers had

lower IQ were more likely to exhibit school maladjustment than other children,independent of earlier child adjustment. Notably, the effect of maternal depression,

evident in univariate analyses, was not significant once these other factors were

considered. When the quality of maternal homework communication was taken into

account (Model 2, Step 2), it was found to make an independent contribution to child

behaviour in school, over and above the effects of child sex and previous adjustment,

and the effect of maternal IQ was no longer significant: mothers who provided low

levels of emotional support had children who showed difficulties. With regard to

paternal variables, the occurrence of some form of psychiatric disorder in the child’sschool years was associated with child difficulties in school, over and above the

influence of previous child difficulties. However, the father’s interactions with the child

during the homework task were unrelated to child outcome on this measure, once

family and paternal characteristics were taken into account. Due to the small number of





b t b t b t

M emotional support 0.23 1.91þ – –M coercive control 20.35 22.62* 20.45 23.34***P promotion of mastery

motivation– – – –


– – 20.18 21.57

P emotional support – – – –P coercive control – – – –

Model F Fð8; 78Þ ¼ 4:58,p , :001,DR 2 ¼ :21,

Fð3; 78Þ ¼ 7:93,p , :001

Fð10; 66Þ ¼ 3:11,p , :01,

DR 2 ¼ :16,Fð3; 66Þ ¼ 5:04,

p , :01




Table 6. Regression models for child school maladjustment based on maternal, paternal and parental

predictors




b Wald b Wald b Wald


motivation21.37 .11 25.04 .60


20.82 .05 0.86 .02

M emotional support 212.89 6.73** 219.12 5.28*M coercive control 20.004 .02 20.06 1.57P promotion of mastery

motivation222.06 7.32** 225.19 6.22*

P promotion of repres.Understanding

22.44 .46 26.39 1.48

P emotional support 27.49 2.21þþ 210.17 1.68P coercive control 20.02 .19 20.04 .68

Model F x2 ¼ 13:38,df ¼ 4, p , :01

x2 ¼ 16:53,df ¼ 4, p , :01

x2 ¼ 26:31,df ¼ 8, p , :001

Model 2Step 1

PBCL (years) 0.19 4.10* 0.19 2.35 0.18 1.59Child gender 3.12 5.36* 10.83 .03 11.59 .04Social class 1.10 1.49 0.99 .48 0.80 .22Maternal depression

(5–8 years)1.47 2.62 1.98 2.25

Maternal IQ 20.10 3.93* 20.07 .79Paternal disorder

(5–8 years)2.96 5.28* 3.01 4.24*

Paternal IQ 20.11 1.05 20.06 .22

Model F x2 ¼ 28:30,df ¼ 5, p , :001

x2 ¼ 31:78,df ¼ 5, p , :001

x2 ¼ 35:45,df ¼ 7, p , :001

Step 2PBCL (5 years) .22 4.51* 0.02 .01 – –Child gender 3.36 4.80* 16.57 .04 – –Social class .98 .93 1.45 .70 – –Maternal depression

(5–8 years)1.52 2.23 – –

Maternal IQ 2 .08 2.01 – –Paternal disorder

(5–8 years)7.45 2.39 – –

Paternal IQ 20.06 .06 – –M promotion of mastery

motivation– – – –


– – – –

M emotional support 216.48 4.04* – –



children exhibiting significant behavioural difficulties in school (N ¼ 9), analyses couldnot be conducted examining the joint contribution of maternal and paternal behaviours.

Discussion

In this study of 8-year-old children of postnatally depressed and well mothers, we used

an ecologically valid, homework-based procedure to conduct direct observations ofnaturalistic interactions of direct relevance to the demands currently placed on children

and their parents. A number of features of the study are notable and extend current

findings in the literature. First, whereas studies of parental support for child problem-

solving have generally focused on single aspects of parental interaction style, the current

study examined the relative contribution of different aspects of parental engagement

with the child and addressed the question of the specificity of each sort of input’s

effects. Second, since child characteristics may themselves have an impact on parental

behaviour, we capitalized on the longitudinal investigation of this sample to take childcharacteristics into account when evaluating the contribution of parental communi-

cation to child outcome. Finally, we directly observed the contribution of fathers as well

as mothers.

A number of aspects of our findings require further comment. First, in contrast to

findings for this sample earlier on in the children’s development, we found no

association between the original postnatal depressive episode and either current

maternal interactions with the child or child functioning. Maternal depression in the

child’s school years, by contrast, was associated with all aspects of maternal discourseexamined, as well as three of the four child outcomes considered. The link was

particularly evident in relation to low child self-esteem, where exposure to maternal

depression in this latter period was influential, even when taking other family

characteristics into account. For each of the other areas of functioning, however, the





b Wald b Wald b Wald

M coercive control – – – –P promotion of mastery

motivation244.08 2.50 – –


– – – –

P emotional support 24.51 .07 – –P coercive control – – – –

Model F x2 ¼ 33:91, df ¼ 6,p , :001,

Dx2 ¼ 5:62, df ¼ 1,p , :05

x2 ¼ 41:92, df ¼ 7,p , :001,

Dx2 ¼ 9:84, df ¼ 2,p , :01




strongest predictor of development at age 8 was the child’s previous performance at age

5 and when this was taken into account, along with other family characteristics, an

independent effect of maternal depression in the child’s school years was not identified.

The fact that postnatal depression was not associated with the current quality of mother-

child interactions is consistent with previous findings for this sample at 5 years (Murray

et al., 1999), together with other studies showing that remitted mothers generally show

improvements in the way they interact with their children (Campbell, Cohn, & Meyers,

1995; Mills, Puckering, Pound, & Cox, 1985). Nevertheless, the significance of postnatal

depression should not be overlooked: those mothers who experienced depression

during the child’s school years were generally mothers who had also experienced

postnatal depression. Insofar as its occurrence raises the risk of later episodes that are

themselves associated with both poor maternal interactions and child outcome, it is

therefore important that those experiencing postnatal depression are monitored

beyond the postnatal episode so that, should depression re-occur, clinical interventions

can be quickly mounted. It is also worth noting that the strong continuities in child

behaviour in school and cognitive functioning from 5 to 8 years may, in part, also reflect

an impact of postnatal depression; thus, the occurrence of behaviour problems at 5

years were predicted by postnatal episodes (Sinclair & Murray, 1998). Similarly,

particularly poor maternal interactions with the child arising in the context of the

postnatal episode were predictive of poor cognitive functioning in late infancy and

persisted to 5 years (Murray et al., 1996); such difficulties, in turn, predicted worse

outcomes on the measures of IQ and school achievements at 8 years. The fact that

patterns of achievement and behavioural adjustment evident within the first term of

school have such a strong bearing on future child performance emphasises the

importance of preschool and the earliest school experiences in launching children along

particular developmental pathways (Sylva, 1994).

With regard to the effects of parent-child discourse, we found evidence that the

current quality of parental conversations with the child during homework sessions

significantly contributed to each of the child outcomes investigated, over and above the

well-established effects of family characteristics such as family social class and parental

IQ. Indeed, in a number of instances, when the quality of parental communication during

homework was taken into account, the effects of family background variables ceased to

be significant. Anumber of the findings concerning the role of parental discoursewith the

child are particularly notable. First, for all but one child outcome (school maladjustment)

the univariate associations obtained between aspects of fathers’ engagement and child

outcome were also obtained for maternal discourse; for school maladjustment, this was

also true for one of the two dimensions of paternal discourse, namely, the presence of

coercive control. On balance, therefore, paternal and maternal discourse styles operated

in very similar ways in relation to child outcome. Nevertheless, once the joint

contributions ofmaternal andpaternal discoursewere considered, the nature ofmothers’

communication appeared to be of greater significance for child functioning than that of

fathers. This finding, based on direct observations of parent-child interactions, appears to

be part of a more general process reflected in the parents’ own reports of the extent and

nature of their involvement with their children. Thus, whereas fathers and mothers

reported similar levels of involvement in their children’s out-of-school activities, mothers

were significantly more likely than fathers to report playing an active role in their

children’s school and to spend more time assisting their children in school-related

activities. This pattern of difference in the roles of the two parents is consistent with



sociological evidence showing that mothers are far more active than fathers in

determining basic aspects of children’s educational experience (David, 2002).

A related question arises, however, of whether the different impact of mothers’ and

fathers’ conversations reflects specifically school-relevant aspects of the parents’

involvement with the child, as evidenced during the homework task, or else simply

reflects the generally greater amount of time mothers spent in interactions with their

children. In our previous investigations of the role of maternal interactions with the

child where significant associations with child cognitive development and behaviour

problems have been found, observations have been conducted during free play sessions

(Murray et al., 1996, 1993) or during an informal snack (Murray et al., 1999). This

suggests that the wider range of parent-child interactions where opportunities for

promoting the child’s cognitive development and regulating behaviour still occur may

be just as important as the specific opportunities afforded by homework tasks; this

hypothesis, however, requires a formal comparison of the effects of discourse in both

types of situation.

A third key aspect of the findings regarding the role of parental homework

communication was evidence for specificity of effects. Thus, broadly in line with our

initial predictions, we found child academic achievements and higher IQ to be promoted

by parental discourse that stimulated the child’s thinking and fostered enthusiasm for

tackling academic problems. By contrast, good emotional and behavioural outcomes

were best promoted by the parent being attentive and supportive of the child’s

endeavours and a parenting style that did not coerce or undermine the child’s efforts.

These findings are broadly consistent with those in the literature on the effects of

particular forms of parental interaction on different child outcomes, but extend previous

findings insofar as multiple measures were used simultaneously to identify the relative

importance of the different forms of input.

The findings of the present study concerning the role of parental psychiatric

problems and communication with the child about homework raise questions about

both clinical practice, particularly in relation to the long-term monitoring of those

mothers who experience depression in the postnatal months, and public policy on

education. Our observations showed striking variability in parents’ capacities to support

their children, including a significant number of parents who clearly felt helpless and

slipped into critical and coercive styles of relating that were associated with poor child

outcome. This suggests that it may be inappropriate for schools to assume parental

competence in this area and, indeed, that such an assumption may have the effect of

widening the divide between children from more privileged and less privileged

backgrounds. This suggestion is in line with previous research showing that the gap in

attainment between more and less disadvantaged children widens as a function of the

child’s reliance on home-based compared with school-based input (Heyns, 1978).

Together the findings indicate that parents themselves may require support as to

how to assist their children. Moreover, the study findings suggest areas of parental

behaviour where supportive interventions specific to the child outcomes of concern

might be targeted. If, for whatever reason, such interventions cannot be provided, then

additional school-based practices may be required to address inequalities in education

and personal opportunity.



Acknowledgements

The research was supported by the Tedworth Charitable Trust and the Medical Research Council.

We thank Wendy Gibbons, Paul Patterson and Steven Briers for assistance with coding and data

collection, and Alistair Pollitt for help with preparing homework materials.

References

Barocas, R., Seifer, R., Sameroff, A. J., Andrews, T. A., Croft, R. T., & Ostrow, E. (1991). Social and

interpersonal determinants of developmental risk. Developmental Psychology, 27, 479–488.

Campbell, S. B., Cohn, J. F., & Meyers, T. (1995). Depression in first-time mothers: Mother-infant

interaction and depression chronicity. Developmental Psychology, 31(3), 349–357.

Cooper, H., Lindsay, J., Nye, B., & Greathouse, S. (1998). Relationships among attitudes about

homework, amount of homework assigned and completed and student achievement. Journal

of Educational Psychology, 99(1), 70–83.

Cox, J. L., Holden, J. M., & Sagovsky, R. (1987). Detection of postnatal depression: Development of

the Edinburgh postnatal depression scale. British Journal of Psychiatry, 150, 782–786.

Cowan, R., & Hallam, S. (1999). What do we know about homework? Viewpoint No. 9. London:

Institute of Education.

Cowan, R., & Traill, D. (1998). Homework for primary schoolchildren: Ideals and reality.

Psychology of Education Review, 22(2), 20–27.

Culp, A., Hubbs-Tait, L., Culp, R. E., & Starost, R. (2000). Maternal parenting characteristics and

school involvement. Predictors of kindergarten cognitive competence among head start

children. Journal of Research in Childhood Education, 15, 5–17.

David, M. (2002). Equality in action. In J. Wilson (Ed.), Seminar 3: Addressing inequalities in

education. London: The Smith Institute.

Department of Education and Skills (DfES). (2004). Helping Parents Help their Children.

(Document 0137–2004). Nottingham, UK: Author.

Diaz, R. M., Neal, C. J., & Vachio, A. (1991). Maternal teaching in the zone of proximal development:

A comparison of low- and high-risk dyads. Merrill-Palmer Quarterly, 37(1), 83–108.

Dodge, K. A., Pettit, G. S., & Bates, J. E. (1997). How the experience of early physical abuse leads

children to become chronically aggressive. In D. Cicchetti & S. Toth (Eds.), Rochester

symposium on developmental psychopathology. Vol. 8: The effects of traumas on the

developmental process. Rochester, NY: University of Rochester Press.

Egeland, B., Pianta, R., & O’Brien, M. A. (1993). Maternal intrusiveness in infancy and child

maladaptation in early school years. Development and Psychopathology, 5, 359–370.

Epstein, J. (1986). Parents’ reactions to teacher practices of parent involvement. Elementary

School Journal, 86, 277–294.

Estrada, P., Arsenio, W. F., Hess, R. D., & Holloway, S. D. (1987). Affective quality of the mother-

child relationship: Longitudinal consequences for children’s school relevant cognitive

functioning. Developmental Psychology, 23(2), 210–215.

Fagot, B. I., & Gauvain, M. (1997). Mother-child problem solving: Continuity through the early

childhood years. Developmental Psychology, 33, 480–488.

First, M. B., Spitzer, R. L., &Williams, J. B. W. (1996). Structured clinical interview for DSM-IVaxis

I. New York: New York State Psychiatric Institute, Biometrics Research Department.

Goldsmith, D. F., & Rogoff, B. (1995). Sensitivity and teaching by dysphoric and nondysphoric

women in structured versus unstructured situations. Developmental Psychology, 31(3),

388–394.

Goodman, S. H., & Gotlib, I. H. (1999). Risk for psychopathology in the children of depressed

mothers: A developmental model for understanding mechanisms of transmission.

Psychological Review, 106(3), 458–490.

Gordon, P. (1980). Homework: Origins and justifications. Westminster Studies in Education, 3,

27–46.



Grolnick, W. S., & Slowiaczek, M. L. (1994). Parents’ involvement in children’s schooling. A

multidimensional conceptualization and motivational model. Child Development, 65(1),

237–252.

Harter, S. (1982). The perceived competence scale for children. Child Development, 53, 87–97.

Heyns, B. (1978). Summer learning and the effects of schooling. New York: Academic Press.

Hooley, J. M., & Teasdale, J. D. (1989). Predictors of relapse in unipolar depressives: Expressed

emotion, marital distress, and perceived criticism. Journal of Abnormal Psychology, 98,

229–235.

Hossain, Z., Field, T., Gonzalez, J., Malphurs, J., & Del Valle, C. (1994). Infants of depressed mothers

interact better with their non-depressed fathers. Infant Mental Health, 15(4), 348–357.

Hubbs-Tait, L., McDonald-Culp, A., Culp, R. E., & Miller, C. E. (2002). Relation of maternal

cognitive stimulation, emotional support, and intrusive behaviour during head start to

children’s kindergarten cognitive abilities. Child Development, 73(1), 110–131.

Jain, A., Belsky, J., & Crnic, K. (1996). Beyond fathering behaviours: Types of dads. Journal of

Family Psychology, 10, 431–442.

Johnson, J. H., & Jason, L. A. (1994). The development of a parent-tutor assessment scale. Urban

Education, 29(1), 22–33.

Keys, W., Harris, S., & Fernandes, C. (1995). Attitudes to school of top primary and first-year

secondary pupils. Slough: National Foundation for Educational Research.

Laasko, M.-L. (1995). Mothers’ and fathers’ communication clarity and teaching strategies with

their school-aged children. Journal of Applied Developmental Psychology, 16, 445–461.

Lamb, M. E. (1997). The role of the father in child development (3rd ed.). New York: Wiley.

Landry, S. H., Smith, K. E., Swank, P. R., & Miller-Loncar, C. L. (2000). Early maternal and child

influences on children’s later independent cognitive and social functioning. Child

Development, 71, 358–375.

Levin, I., Levy-Schiff, R., Appelbaum-Peled, T., Katz, I., Komar, M., & Meiran, N. (1997).

Antecedents and consequences of maternal involvement in children’s homework: A

longitudinal analysis. Journal of Applied Developmental Psychology, 18, 207–227.

MacBeath, J., & Turner, M. (1990). Learning out of school: Homework, policy and practice. A

research study commissioned by the Scottish Education Department. Glasgow: Jordan Hill

College.

Macleod, F. J. (1996). Does British research support claims about the benefits of parents hearing

their children read regularly at home? A closer look at the evidence from three key studies.

Research Papers in Education, 11(2), 173–199.

McCarthy, D. (1972).McCarthy scales of children’s abilities. NewYork: Psychological Corporation.

McGuire, J., & Richman, N. (1988). Pre-School Behaviour Checklist (PBCL) Handbook. Windsor:

nferNelson.

Mills, M., Pickering, C., Pound, A., & Cox, A. (1985). What is it about depressed mothers that

influences their children’s functioning? In J. E. Stevenson (Ed.), Recent research in

developmental psychopathology. Oxford: Pergamon Press.

Morrell, J., & Murray, L. (2003). Postnatal depression and the development of conduct disorder and

hyperactive symptoms in childhood: A prospective longitudinal study from 2 months to 8

years. Journal of Child Psychology and Psychiatry, 44(4), 489–508.

Murray, L. (1992). The impact of postnatal depression on infant development. Journal of Child

Psychology and Psychiatry, 33, 543–561.

Murray, L., & Cooper, P. (2003). Intergenerational transmission of affective and cognitive processes

associated with depression: Infancy and the preschool years. In I. M. Goodyer (Ed.), Unipolar

depression: A lifespan perspective. Oxford: Oxford University Press.

Murray, L., Fiori-Cowley, A., Hooper, R., & Cooper, P. J. (1996). The impact of postnatal depression

and associated adversity on early mother-infant interactions and later infant outcome. Child

Development, 67, 2512–2526.



Murray, L., Hipwell, A., Hooper, R., Stein, A., & Cooper, P. J. (1996). The cognitive development of

five year old children of postnatally depressed mothers. Journal of Child Psychology and

Psychiatry, 37, 927–935.

Murray, L., Kempton, C., Woolgar, M., & Hooper, R. (1993). Depressed mothers’ speech to their

infants and its relation to infant gender and cognitive development. Journal of Child

Psychology and Psychiatry, 34, 1083–1101.

Murray, L., Sinclair, D., Cooper, P., Ducournau, P., Turner, P., & Stein, A. (1999). The socio-

emotional development of five year old children of postnatally depressed mothers. Journal of

Child Psychology and Psychiatry, 40(8), 1259–1272.

Murray, L., Woolgar, M., Cooper, P. J., & Hipwell, A. (2001). Cognitive vulnerability in five year old

children of depressed mothers. Journal of Child Psychology and Psychiatry, 42(7), 891–899.

Nelson, H. E. & Willison, J. (1991). National Adult Reading Test (NART). Test Manual (2nd ed.).

Windsor: nferNelson.

Nix, R. L., Pinderhughes, E. E., Dodge, K. A., Bates, J. E., Pettit, G. S., & McFadyen-Ketchum, S. A.

(1999). The relation between mothers’ hostile attribution tendencies and children’s

externalizing behavior problems: The mediating role of mothers’ harsh discipline practices.

Child Development, 70, 896–909.

Nolen-Hoeksema, S., Wolfson, A., Mumme, D., & Guskin, K. (1995). Helplessness in children of

depressed and nondepressed mothers. Developmental Psychology, 31(3), 377–387.

Pollitt, A., & Munro, L. (1989). The Primary Maths Bank. Basingstoke: Macmillan Education.

Ratner, H. H., & Stettnet, L. J. (1991). Thinking and feeling: Putting humpty dumpty together again.

Merrill-Palmer Quarterly, 37, 1–26.

Rutter, M. (1967). A children’s behaviour questionnaire for completion by teachers: Preliminary

findings. Journal of Child Psychology and Psychiatry and Allied Disciplines, 8, 1–11.

Senechal, M., & LeFevre, J.-A. (2002). Parental involvement in the development of children’s

reading skill: A five-year longitudinal study. Child Development, 73(2), 331–666.

Shumow, L. (1998). Promoting parental attunement to children’s mathematical reasoning through

parent education. Journal of Applied Developmental Psychology, 19(1), 109–127.

Sigel, I. E. (1982). The relationship between parental distancing strategies and the child’s cognitive

behaviour. In L. M. Laosa & I. E. Sigel (Eds.), Families as learning environments for children

(pp. 47–86). New York: Plenum.

Sinclair, D., & Murray, L. (1998). The effects of postnatal depression on children’s adjustment to

school: Teacher reports. British Journal of Psychiatry, 172, 58–63.

Spanier, G. B. (1976). Measuring dyadic adjustment: New scales for assessing the quality of

marriage and similar dyads. Journal of Marriage and the Family, 38, 15–28.

Spitzer, R. C., Endicott, J., & Robins, E. (1978). Research diagnostic criteria: Rationale and

reliability. Archives of General Psychiatry, 35, 733–782.

Sylva, K. (1994). School influences on children’s development. Journal of Psychology and

Psychiatry and Allied Disciplines, 35(1), 135–170.

US Department of Education. (1994). Strong families, strong schools: Building community

partnerships for learning. Washington DC: Author.

Vygotsky, L. S. (1962). Thought and language. Cambridge, MA: Massachusetts Institute of

Technology Press.

Vygotsky, L. S. (1978). Mind in society. Cambridge, MA: Harvard University Press.

Wechsler, D. (1992). WISC-III Wechsler intelligence scale for children (3rd ed.). San Antonia, TX:

Psychological Corporation.

Received 23 December 2004; revised version received 26 October 2005


Conversations around homework: Links to parental mental health, family characteristics and child psychological functioning

Documents