This article was downloaded by: [Manuela Veríssimo]On: 05 April 2013, At: 03:00Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,37-41 Mortimer Street, London W1T 3JH, UK
Family SciencePublication details, including instructions for authors and subscription information:http://www.tandfonline.com/loi/rfsc20
Father involvement and peer play competence inpreschoolers: The moderating effect of the child'sdifficult temperamentNuno Torres a , Manuela Veríssimo a , Ligia Monteiro a , António J. Santos a & Inês Pessoa eCosta aa UIPCDE, ISPA-IU, Lisbon, PortugalVersion of record first published: 05 Apr 2013.
To cite this article: Nuno Torres , Manuela Veríssimo , Ligia Monteiro , António J. Santos & Inês Pessoa e Costa (2013): Fatherinvolvement and peer play competence in preschoolers: The moderating effect of the child's difficult temperament, FamilyScience, DOI:10.1080/19424620.2012.783426
To link to this article: http://dx.doi.org/10.1080/19424620.2012.783426
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any substantial or systematicreproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form toanyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contentswill be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses shouldbe independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims,proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly inconnection with or arising out of the use of this material.
Family Science, 2013http://dx.doi.org/10.1080/19424620.2012.783426
Father involvement and peer play competence in preschoolers: The moderatingeffect of the child’s difficult temperament
Nuno Torres, Manuela Veríssimo*, Ligia Monteiro, António J. Santos and Inês Pessoa e Costa
UIPCDE, ISPA-IU, Lisbon, Portugal
(Received 28 January 2013; final version received 16 February 2013)
This article focuses on the association between the relative involvement of the father with preschool children in two-parentfamilies, individual characteristics of children (age, gender, and difficult temperament) and the children’s peer play compe-tence. Relative involvement of the father was assessed using The Parental Involvement: Care and Socialization Questionnairecomposed of five subscales: Direct Care, Indirect Care, Teaching/Discipline, Play and Leisure Outdoors. Interactive peerplay competence was assessed using the Penn Interactive Peer Play Scale (PIPPS), with three subscales: Play Interaction,Play Disruption and Play Disconnection. Parental assessment of the child’s temperament was made using the preschool ver-sion of the Infant Characteristics Questionnaire’s (ICQ) subscale of Difficult Temperament. Multiple regression ordinaryleast squares (OLS) models showed that boys and children with difficult temperament tend to show more play disruptionand less play interaction, and also that difficult temperament was a moderator of the association between the relative fatherinvolvement in Direct Care, Play and Teaching/Discipline activities and individual differences in peer play Disruption andInteraction. Higher levels of relative father involvement in Direct Care and Play with temperamentally difficult children wereassociated with more disruptive play with peers, and higher levels of relative father involvement in Teaching/Discipline withdifficult children was associated with more play Interaction. The discussion elaborates on the differential styles of Interactionby fathers and mothers with temperamentally difficult children.
Keywords: father involvement; peer play competence; child temperament
Although families are still viewed as the major agent for thedevelopment of children’s social behavior, the theoreticalconceptualization of the family has undergone substan-tial reconsideration in the last decades (Tamis-LeMonda &Cabrera, 1999). The view of the mother-child dyad as thecentral unit has been replaced by a dynamic social sys-tems approach in which fathers, the marital relationshipand the wider social and cultural sphere are all consideredimportant in children’s socialization (Parke et al., 2008).
Evidence that fathers and mothers have both distinctand combined influences on the development of socialbehavior, such as attachment, peer friendships and socialskills (Parke et al., 2008), transformed parenting research.Fathers are no longer the forgotten links in developmen-tal psychology and are now viewed as central actors indevelopmental research (Lewis & Lamb, 2003).
It has been proposed that father involvement shouldbe a determinant of a range of psychobiological develop-mental outcomes, including social competence (Geary &Flinn, 2001; Hewlett, Lamb, Leyendecker, & Scholmerich,2000; Pleck, 2010), with paternal involvement becom-ing more salient at the preschool age (Lamb, 2010).Evolutionary theoretical models have proposed that fatherinvestment has rewards derived from their children’s social
*Corresponding author. Email: [email protected]
competence and social value, as the ability to capitalize onopportunities in the social environment (Geary & Flinn,2001; MacDonald, 1993). According to these models,father involvement would predict greater social and cogni-tive competence in children, such as academic achievementand capacity to both cooperate and compete with peers.
The literature points out that, in contrast to moth-ers, fathers engage in proportionately more play thancare-giving routines (Lamb, 2010). Furthermore, fathers’play with infants, toddlers and preschoolers also tendsto be more physically stimulating than mothers’ play(Bretherton, Lambert, & Golby, 2005). Fathers tend toengage in more unpredictable play with infants and tod-dlers than mothers do (Lamb & Lewis, 2010). Sometimesthe play climate of fathers can become disruptive, through,for example, ‘teasing episodes’, which were observedby Labrell (1994) as the father deliberately elicited anemotional reaction by destabilizing the infant. Fathersare also more likely than mothers to engage in ‘rough-and-tumble play’, and they are more likely to manifest‘directiveness’ such as issuing commands to the child(MacDonald & Parke, 1984; Tamis-LeMonda, Shannon,Cabrera, & Lamb, 2004). In the preschool age, fatherstend to engage significantly in less object-mediated play
© 2013 Taylor & Francis
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
2 N. Torres et al.
than mothers (MacDonald, 1987; MacDonald & Parke,1984).
Researchers have proposed that children may profitfrom interacting with people who have different behavioralstyles: fathers’ biological and socially reinforced masculinequalities predispose them to treat their children differentlythan mothers do. Fathers are also more likely to encour-age their children to be competitive and to take risks(Cabrera, Tamis-LeMonda, Bradley, Hofferth, & Lamb,2000; Geary, 2000; Geary & Flinn, 2001; Labrell, 1994;Tamis-LeMonda et al., 2004). There is evidence that fathersand mothers have both distinct and combined influenceon the development of peer interactions and social skills(Lewis & Lamb, 2003; MacDonald & Parke, 1984, Parkeet al., 2008). Research suggests that fathers may havegreater impact that mothers in domains of the socializationthat involve relations with the world outside the family andrelationships with peers (Lamb & Lewis, 2010).
The ability of the child to maintain positive play inter-actions with peers is considered a highly salient indicatorof social competence during early childhood (Fantuzzo,Mendez, & Tighe, 1998; Waters & Sroufe, 1983). Severalmeasures of social peer play have been associated withchild attributes, such as age, gender and temperament(Coplan, Rubin, & Findlay, 2006). Additionally, socioe-conomic status has been associated with increased socialparticipation in peer play (Rubin, Maioni, & Hornung,1976; Smilansky, 1968). Finally, in the last four decades,father involvement in child care has been receiving increas-ing attention in the literature as an important determinant ofsocial development in infancy and early childhood (Lamb,2010; Monteiro et al., 2010; Rutherford & Mussen, 1968;Tamis-LeMonda & Cabrera, 1999).
The initial evidence that fathers matter for peer rela-tionships came from studies of children whose fathers wereabsent; these children had poor-quality peer relationships,were less popular, and were reluctant to play rough games(Parke et al., 2008, Rubin, Bukowski, & Parker, 1998).In father-present families, there is some evidence that theamount and quality of father involvement has an impor-tant impact in long-term social development, namely inempathic capacities of children when they become adults(Koestner, Franz, & Weinberger, 1990).
The impact of the quality of the father-child relation-ship in two-parent families has mainly been examinedfrom two perspectives: (1) attachment research, which hasaddressed infant-father attachment and social adaptationin the peer group (e.g. Lamb, Frodi, Hwang, & Frodi,1983; Lewis & Lamb, 2003; Parke et al., 2008; Veríssimoet al., 2011); and (2) father-child interaction research, espe-cially that concerned with interactions during play, whichhas investigated relations with children’s peer relation-ships. This second tradition of research includes resultsdemonstrating that differences in the style of interactionand relationships of mothers and fathers with the child
are related to popularity, measured by sociometric ratingsand peer interaction patterns of boys and girls, and thatthe quality of the interactional styles with the father hasa differential impact from those of the mother on severalindicators of social competence and pro-social behaviors(e.g. MacDonald & Parke, 1984; Parke et al., 2008).
Research has shown that the father’s warm interac-tional styles are related to positive social outcomes (Parkeet al., 2008). Facilitative and affectionate fathering is alsoassociated with pro-social (e.g. sharing, helping, cooper-ating) sibling interaction (Volling & Belsky, 1992). Giventhe importance of play in father involvement with chil-dren, a large number of studies have researched the linksbetween the quality of father-child play interactions andsocial competence outcomes. Specifically, physically play-ful, affectionate and socially engaging father-child interac-tion predicts later child popularity, just as mothers’ verbalstimulation predicts popularity (Lamb & Lewis, 2010).Conversely, controlling and harsh father interactional stylesare related to negative social outcomes for children (Foster,Reese-Weber, & Kahn, 2007; Lamb & Lewis, 2010; Parkeet al., 2008).
A study by Carson and Parke (1996) showed thatexchanges of reciprocal negative affect (responding withnegative affect to the negative affect of an interactionpartner) in physical play specifically between father andchild (but not between mother and child) were associ-ated with low peer competency. Fathers who were morelikely to respond to their children’s negative affect dis-plays with negative affect of their own had children whoshared less, were more verbally aggressive, avoided othersand tended to be more physically aggressive. Other stud-ies have also reported stronger links between physical playand children’s peer competency for father-child rather thanmother-child dyads (Carson & Parke, 1996; MacDonald,1987, 1993).
There are few studies investigating the role of day-to-day routines in two-parent families regarding the relativeinvolvement of the father (i.e. the amount of involvementof father proportionate to the amount of involvement ofthe mother) in the development of interactive peer playcompetence. Although theoretical propositions have beenoffered to explain how different children might be affectedby similar experiences (Belsky, 2005; Klein Velderman,Bakermans-Kranenburg, Juffer, & van IJzendoorn, 2006),researchers have yet to investigate the role of chil-dren’s individual characteristics, including temperament,in moderating associations between paternal behaviors andchild outcomes (Carson & Parke, 1996; Ramchandani,van IJzendoorn, & Bakermans-Kranenburg, 2010; Tamis-LeMonda & Cabrera, 1999).
The goals of this study were to (1) assess fathers’routine relative involvement, compared to the mothers’,in child-related activities (child care and socialization) intwo-parent dual-earner families; (2) analyze the association
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 3
between the type and amount of fathers’ relative involve-ment and the development of preschooler’s social compe-tence during play with peers and (3) assess the role of theindividual characteristics of the child (age, gender and tem-perament) in the preschoolers’ social competence duringplay with peers.
The study was carried out in highly urbanized regionsof Portugal, where the vast majority of family couples aredual earners. According to the latest official national statis-tics in 2001 about family composition, 88.5% of familieswith children are two-parent families. Also, according tothe latest official national statistics, 85% of the Portuguesepopulation are Catholics, 3% have other religious affili-ations and 4% have no religious affiliation (8% did notprovide information).
Method
Participants
One hundred thirty-nine mothers and 118 fathers from two-parent families participated, out of a total of 151 families.All couples were either officially married or cohabiting.The range of mothers’ ages was 23–48 years (M = 34.4;SD = 4.5), and fathers’ ages ranged from 23 to 62(M = 36.6; SD = 6.1). Mothers’ education level variedbetween 2 and 22 years (M = 12.8; SD = 3.9) and fathers’between 4 and 19 (M = 11.7; SD = 3.8). In occupa-tional terms, 93% of the mothers worked between 1 and12 hours a day (M = 7.89; SD = 1.5), while 96.9% of thefathers were workers from 1 to 14 hours a day (M = 8.7;SD = 1.7). All families were ‘middle class’ in terms ofincome level, according to the standards of the local com-munity, and most had jobs which were characteristic of the‘middle class’ status; they were all of European descent.
Children were between 32 and 78 months old(M = 56.6; SD = 9.8): 49% were boys. All the childrenattended private daycare programs in the suburbs of Lisbon,Portugal, where families were recruited. The median ageof children attending childcare was 24 months (M = 22.8;SD = 14.3), and the majority of the children (57%) startedattending childcare between 12 and 36 months. The crite-ria for child selection were that he/she was attending thepreschool daycare centers where we collected the data andwas a biological child of the couple. Participants were toldthat when there were two children in the preschool daycarecenters, their replies should concern the firstborn child.
Both parents were asked to fill in a questionnaire sur-vey separately, which included items on routine fatherrelative involvement with one child from each family, tem-perament of the child and peer play competence of thereference child, plus demographic information (e.g. ageof the parents, parents’ educational level, gender and ageof the child). Of the total 151 families enrolled in thestudy, 136 mothers returned complete information on allvariables (90% complete response rate), contrasting with
58 fathers who returned complete information on all vari-ables (38% complete response rate), and 49 couples of bothfather and mother returned full information on all variables(32% couple complete full response rate). While the fatherinvolvement questionnaire had a higher response rate fromfathers (n = 118; 78% partial response rate for the fatherinvolvement questionnaire), the low complete response rateof fathers was due to the fact that only 61 fathers (40%response rate) returned the child’s temperament and peerplay competence questionnaires.
One-way analysis of variance (ANOVAs) showed thatthere were no differences between those who completedquestionnaires and those who did not on any of thedemographic variables except in one instance: specifically,fathers who replied to the peer play questionnaire had onaverage 2 years less education (M = 10.6; SD = 3.5) thanfathers who did not reply (M = 12.8, SD = 4.0), F(1,140) = 12.54, p = .001.
There were no significant differences between demo-graphic variables of the couples who had completeresponse rate of both father and mother in all question-naires and the couples who gave partial information eitherfrom father or mother, or a partial return of the question-naires from either parent.
Instruments
A summary of the theoretical constructs and their respec-tive measurement instruments, the instruments’ subscalesand examples of their items, and their reliability coeffi-cients [Cronbach’s alphas and intraclass correlation (ICC)coefficients] is shown in Table 1.
Relative involvement of the father was assessedusing The Parental Involvement: Care and SocializationActivities (Monteiro, Veríssimo, & Pessoa e Costa, 2008).It contains 26 items referring to the organization and imple-mentation of activities involving parent and child that occurin daily family life. The questionnaire assesses the rel-ative participation of one parental figure in relation tothe other. Participants are asked to answer on a 5-pointscale: (1) always the mother, (2) nearly always the mother,(3) both the mother and the father, (4) nearly always thefather and (5) always the father. Hence, this is not a sepa-rate measure of the father’s and the mother’s involvementper se; instead, the involvement of one parent is consideredthe proportion of involvement that is not attributed to theother one. Higher scores represent a greater involvement ofthe father. The questionnaire has five dimensions derivedfrom a previous factor analysis (Monteiro, Verissimo, &Pessoa e Costa, 2008): (1) Direct Care (5 items), relatedwith caretaking tasks, which imply direct contact andinteraction with the child (e.g. ‘Who feeds the child’);(2) Indirect Care (7 items) activities, which are related witharranging the resources to be available to the child, donot necessarily imply interaction (e.g. ‘Who usually buys
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
4 N. Torres et al.
Table 1. Summary of the theoretical constructs, measurement instruments used, subscales/dimensions used, sample items and reliabilitycoefficients.
Cronbach’salpha
Theoreticalconstruct
Measurementinstrument
Instrumentsubscales used Example items Mother Father
ICC withincouple
Relative father Parental Direct care ‘Who feeds the child’? .77 .77 .94involvement involvement: care
and socializationIndirect care ‘Who usually buys your Child’s
clothes’?.73 .69 .91
activities Teaching/Discipline ‘Who teaches the child newskills’?
.80 .70 .73
Play ‘Who plays physical games withthe child’?
72 .69 .93
Outdoor leisure ‘Who takes the child to the park’? 71 .76 .80
Child peer playcompetence
Penn InteractivePeer Play Scale(PIPPS)
Play disruption ‘Demands to be in charge’‘Is physically aggressive’
.77 .76 .73
Play interaction ‘Helps other children’ ‘Disagreeswithout fighting’
.75 .73 .73
Play disconnection ‘Wanders aimlessly’ ‘Refuses toplay when invited’
.68 .65 .39
Childtemperament
Infant CharacteristicsQuestionnaire(ICQ)
Difficulttemperament
‘How easy or difficult is it to calmdown or soothe your child whenhe/she is upset’?
.80 .79 .84
your child clothes’); (3) Teaching/Discipline (5 items),related with teaching skills and rules for the child (e.g.‘Who teaches the child new skills’) ; (4) Play (5 items),related with play activities between the child and theparent (e.g. ‘Who plays physical games with the child:football or rough and tumble’) and (5) Leisure Outdoors(4 items) activities done with the child outside the home(e.g. ‘Who takes the child to the park’). The Cronbach’salpha reached acceptable values both for mothers’ reports(between .71 and .80) and for fathers’ reports (between.69 and .77). Agreement between reports of parents ofthe same child, both of whom replied to the ParentalInvolvement questionnaire (n = 107), measured by ICCcoefficient was in the range between .73 and .94.
Peer play competence was assessed using the PennInteractive Peer Play Scale (PIPPS). The measure is a 32-item rating scale of preschool children’s competence asregards interactive peer play. Previous research (Fantuzzoet al., 1998) revealed three reliable dimensions, in which29 items loaded appreciably on only one dimension: PlayInteraction (10 items), Play Disruption (11 items) andPlay Disconnection (8 items). Items were scored on a4-point Likert scale (1–‘never’, 2–‘seldom’, 3–‘often’, 4–‘always’). Play Interaction includes items such as com-forting and helping other children, showing creativity inplay and encouraging others to join play (e.g. ‘Helps otherchildren’ ‘Disagrees without fighting’). Play Disruptionconsists of items relating to aggressive and dominant playbehaviors (e.g. ‘Demands to be in charge’ ‘Is physicallyaggressive’), while Play Disconnection consists of itemsdescribing withdrawn behavior and nonparticipation inpeer play (e.g. ‘Wanders aimlessly’ ‘Refuses to play when
invited’). Cronbach’s alpha of the three scales ranged from.68 to .77 for mothers and from .65 to .76 for fathers.Agreement measured by ICC coefficient between reportsof parents of the same couple, both of whom replied to thePIPPS questionnaire (n = 54), was between .39 and .73.
Parental assessment of the child’s temperament wasmade using the preschool version of the ICQ-InfantCharacteristics Questionnaire (Bates, Freeland, &Lounsbury, 1979). It assesses the temperament ofchildren according to parents’ representations about thechild’s typical behaviors. It is composed of 32 items, ona 7-point response scale. For the present study, only theDifficult Temperament scale was used (7 items), whichincludes statements about frequent and intense negativeaffect, and the degree of difficulty the infant presents tocaregivers. Cronbach’s alpha for mothers’ reports aboutdifficult temperament was .80, whereas that for fathers’reports was .79 Agreement measured by ICC coefficientbetween reports of parents of the same couple, both ofwhom replied (n = 101) to the Temperament questionnaire,was .84.
Data analysis plan Based on the response rates andCronbach’s alpha of mother and father reports, we decidedto use both parents’ responses on the father involvementquestionnaire, and to use only the mothers’ reports on tem-perament and peer play competence. Additionally, in orderto have an estimate of mother and father agreement, wecomputed ICC coefficients for all the variables, includ-ing in these calculations only the couples in which bothfather and mother replied. Missing data in items withinquestionnaires were negligible; it ranged from 0 to 6.5%.We handled missing data by imputing the mean value of the
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 5
scales on the missing item, which was reported as a suitablemethod for small amounts of missing data in self-reportscales (Shrive, Stuart, Quan, & Ghali, 2006).
Analyses were conducted in two steps. First, bivari-ate correlations were conducted to determine the extentto which relative father involvement, child attributes (age,gender, and difficult temperament) and family educationallevel were associated to peer play competence in the threedimensions of Peer Play (Disruption, Disconnection andInteraction).
A second step was to conduct a series of ordinary leastsquares (OLS) multiple regression analyses. Specifically,we regressed the three peer play scales on the fatherinvolvement, child attributes (age, gender, and difficulttemperament) and family educational level variables, andon the interaction terms of the five father relative involve-ment scales with child’s attributes (gender, age and difficulttemperament). This second step of the analysis had twoobjectives: (1) to statistically control the effects of allthe variables in the model, and (2) to estimate the possi-ble effects of the interaction between father involvementand child’s attributes on the scores of peer play compe-tence. For each dependent variable (the three peer playscales: Disruption, Interaction and Disconnection), we per-formed five OLS hierarchical multiple regression analysesusing as predictors each of the five father involvementscales and including in the regression models the otherfamily and child attribute variables, as well as interactionterms between child attributes and each of the five fatherinvolvement scales. The option to perform one separatemodel for each of the dimensions of father involvementwas selected on the basis of the high number of signifi-cant strong inter-correlations among the father involvementscales, which could represent a source of multi-collinearity,and also on the basis of having the statistical power todetect significant effects. The variables were entered in theOLS regression models in two hierarchical blocks – block1 was composed of linear relationships: family-level vari-ables (father involvement and parental education) and alsochild attributes (age, gender, and difficult temperament);block 2 was composed of the interaction terms of eachfather involvement dimension with the child attributes (ageX father involvement, gender X father involvement anddifficult temperament X father involvement).
Results
In Table 2, where the values for fathers’ reports onfather involvement estimates are in parentheses, bivariatecorrelation analyses show that the inter-correlationsof mothers’ reports of the child’s peer play scales ofDisruption, Disconnection and Interaction have small tomedium significant associations among them: Disruptionand Disconnection have positive correlations among them,and both have negative associations with Interaction. Infamilies with a higher level of education, the maternal
reports of children’s peer play Interaction scale tends tobe higher (r = .22; p < .01). In terms of the correlationsbetween child attributes and child peer play, boys tendto show more Disruption (r = .30, p <.01) and lessInteraction (r = −.22 p < .01), and children with higherscores on the maternal reports of difficult temperamentalso tend to have higher scores on Play Disruption(r = .45; p < .01) and lower scores on Play Interaction(r = –22; p < .01).
The descriptive statistics (means and standard devia-tions) at the bottom of the table show that according to bothfathers’ and mothers’ reports, the amount of father involve-ment is shared fairly equally with the mother in the dimen-sions of Teaching/Discipline, Play and Leisure Outdoors.The average values are between 2.65 and 3.14, which cor-respond to the response category ‘Both the mother and thefather’. On the other hand, fathers clearly have a lowerrelative involvement in the dimensions of Direct Careand Indirect Care (average values 2.13 and 2.44, whichcorrespond to the category ‘Nearly always the mother’).One-sample t-tests were conducted with a 99% confidenceinterval to determine if the sample means differed signifi-cantly from the value ‘3’,which represents perfect equalityin sharing tasks (the response category ‘both the motherand the father’). According to both parents’ reports, themean values of relative father involvement were signifi-cantly lower than 3 (meaning a lower relative involvementof the father than of the mother) in all scales with the excep-tion of Play and Teaching/Discipline; Play did not differsignificantly from 3 (t[138] = -1.60, p = .113) in mothers’reports, which means that this task is seen by the mothersto be shared equally with the father. Conversely, accord-ing to fathers’ reports, the mean for play had a significantlyhigher mean (t[117] = 3.21, p = .00) value than 3, thatis on average fathers consider themselves to be involvedin this task more than mothers do; finally, according tofathers’ reports, Teaching/Discipline did not differ signif-icantly from 3 (t[117] = .61, p = .54), which means thatthis task is seen by the fathers as being shared equally withthe mother.
In order to estimate the within-couple differencesof fathers’ and mothers’ reports on relative fatherinvolvement, we additionally performed repeated mea-sures ANOVA for the couples in which both the fatherand mother returned their versions of the questionnaire(n = 107). There were significantly lower values for moth-ers’ reports than for fathers’ reports in all dimensions, ascan be seen in Table 3. This means that fathers system-atically report higher values of father involvement thanmothers’ do, although the substantial differences are rathersmall. In summary, within couples, fathers’ and moth-ers’ reports agree that fathers engage less in all dimen-sions except play and teaching discipline. On the otherhand, fathers tend to report slightly higher scores in alldimensions of involvement with children than mothersreport they do.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
6 N. Torres et al.
Tabl
e2.
Cor
rela
tion
san
dde
scri
ptiv
est
atis
tics
ofpe
erpl
ayco
mpe
tenc
e,ch
ildr
en’s
attr
ibut
es,p
aren
ts’
educ
atio
nall
evel
and
rela
tive
fath
erin
volv
emen
t.
Chi
ldva
riab
les
Fam
ilyva
riab
les
Chi
ldpe
erpl
ayC
hild
atri
bute
sE
duca
tion
Rel
ativ
efa
ther
invo
lvem
ent
Var
iabl
es1
23
45
67
89
1011
12
1.P
lay
disr
upti
on−
−.15
∗.1
6∗.3
0∗∗−.
06.4
5∗∗−.
06.1
6∗.0
9.2
0∗.0
7.1
5∗(.
10)
(.09
)(.
19)∗
(.14
)(.
22)∗
2.P
lay
inte
ract
ion
−−.
30∗∗
−.31
∗∗.1
4−.
22∗∗
.22∗∗
.03
.04
−.01
.00
−.05
(.00
)(.
13)
(−.1
5)(.
03)
(−.0
1)3.
Pla
ydi
scon
nect
ion
−.0
1−.
15∗
.09
.02
.07
−.08
−.10
−04
−.07
(.09
)(−
.17)
∗(.
09)
(.09
)(−
.12)
4.S
ex(m
ale)
−−.
02.1
1−.
06.1
3−.
01.1
8∗.0
7.0
4(.
10)
(−.0
9)(.
04)
(.14
)(.
05)
5.A
ge−
−.19
∗.0
4−.
06−.
05−.
14.1
1−.
02(.
07)
(.06
)(−
.08)
(.29
)∗∗
(.09
)6.
Dif
ficu
ltte
mpe
ram
ent
−.0
0.1
3.0
5.2
0∗−.
06.0
5(.
03)
(.07
)(–
17)∗
(.09
)(.
12)
7.Fa
mily
educ
atio
n−
.21∗
.17∗
.17∗
.20∗
.14
(.18
)∗(.
16)∗
(.18
)∗.0
4.0
68.
Dir
ectc
are
−.4
9∗∗.5
9∗∗.5
9∗∗.5
2∗∗(.
27)∗∗
(.30
)∗∗(.
56)∗∗
(.31
)∗∗9.
Indi
rect
care
−.4
2∗∗.4
2∗∗.6
7∗∗(.
28)∗∗
(.11
)(.
50)∗∗
10.T
each
ing/
Dis
cipl
ine
−.6
3∗∗.5
5∗∗(.
26)∗∗
(.27
)∗∗11
.Pla
y−
.59∗∗
(.36
)∗∗12
.Lei
sure
outd
oors
−M
ean
Mot
hers
’re
port
s1.
602.
881.
38.4
956
.64
3.03
12.8
2.17
∗∗2.
13∗∗
2.65
∗∗2.
922.
70∗∗
Fath
ers’
repo
rts
(2.4
4)∗∗
(2.3
1)∗∗
(3.0
2)(3
.14)
∗∗(2
.85)
∗∗S
DM
othe
rs’
repo
rts
.33
.43
.33
.50
10.2
4.9
63.
43.6
1.5
0.6
2.5
7.6
2Fa
ther
s’re
port
s(.
65)
(.51
)(.
42)
(.48
)(.
51)
Not
es:V
alue
sin
pare
nthe
ses
are
for
fath
ers’
repo
rts
onfa
ther
invo
lvem
ent.
Dum
my
codi
ngfo
rge
nder
:boy
sco
ded
as‘1
’,gi
rls
code
das
‘0’.
∗ p<
.05;
∗∗p
<.0
1,on
e-ta
iled
test
sfo
rco
rrel
atio
nco
effi
cien
ts,t
wo-
tail
edte
sts
for
t-te
stdi
ffer
ence
sof
mea
nsfr
omth
eva
lue
3in
fath
erin
volv
emen
tsca
les.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 7
Table 3. Repeated measures ANOVA of mothers and fathersfor the couples in which both members returned their versionsof the questionnaire.
Mothers’reports
Fathers’reports F Significance
Direct care 2.22 2.41 31.64 .001(.58) (.57)
Indirect care 2.17 2.27 9.20 .001(.47) (.47)
Teaching/Discipline
2.70 2.99 45.77 .001(.56) (.36)
Play 3.00 3.12 9.89 .001(.48) (.45)
Leisureoutdoors
2.70 2.81 9.11 .001(.57) (.48)
Note: Standard deviations in parentheses.
As shown in Table 2, the inter-correlations of the fatherrelative involvement scales were all positive; in mothers’reports they were medium to strong in magnitude (between.42 and .67, with a mean of .55), while in fathers’ reportsthey ranged from small to strong (.11 to .56, with a mean of.32). Although some of the correlations between the scaleswere strong, they seem to be tapping different dimensionsof the father involvement construct, since the common vari-ance between the strongest correlations ranges from 30%to 45%.
A higher educational level of the family was signifi-cantly correlated with both parents’ reports on fathers’ rel-ative involvement in Direct Care (rmother = .21 and rfather =.18), Indirect Care (rmother = .17 and rfather = .16),Teaching/Discipline (rmother = .17 and rfather = .18) andwith Play according to mothers’ reports (rmother = .20) butnot according to fathers’ reports.
Relative involvement of fathers in Teaching/Disciplinewas significantly correlated with male gender, accordingto mothers’ reports (rmother = .18), but not according tofathers’ reports. Table 2 also shows significant positive cor-relations between Play Disruption and both parents’ reportsof father involvement in Teaching/Discipline (rmother =.20 and rfather = .19) and Leisure Outdoors (rmother =.15 and rfather = .22), and also between Play Disruption andfather involvement in Direct Care according to mothers’reports ( rmother = .18) but not according to fathers’ reports(rfather = .10).
Multiple regression models
All the OLS multiple regression models for Play Disruptionas well for Play Interaction attained a significant amountof explained variance; these models are summarized inTables 4 and 5. In contrast, none of the regression mod-els for Play Disconnection reached a significant amount ofexplained variance, as illustrated in Table 6.
As shown in Table 4, all the five models for PlayDisruption had a significant adjusted R2 for block 1, which
Tabl
e4.
Bet
aes
tim
ates
and
R2
ofre
gres
sion
mod
els
ofP
lay
Dis
rupt
ion,
wit
hm
othe
rs’
and
fath
ers’
repo
rts
onfa
ther
invo
lvem
ent.
Blo
ck1
Blo
ck2
Fam
ilyva
riab
les
Chi
ldat
trib
utes
Chi
ldat
trib
utes
XFa
ther
invo
lvem
ent
Fath
erin
volv
emen
tE
duca
tion
leve
lG
ende
r(m
ale)
Age
Dif
ficu
ltte
mpe
ram
ent
R2
Gen
der
XFa
ther
invo
lvem
ent
Age
XFa
ther
invo
lvem
ent
Dif
ficu
ltte
mpe
ram
ent
XFa
ther
invo
lvem
ent
R2
Cha
nge
Tota
lmod
elad
just
edR
2
Mod
el1
Dir
ectc
are
.02
−.02
.24∗
∗.0
3.3
9∗∗
.27∗
.07
.09
.22∗
∗.0
6∗.2
9∗∗
(.03
)(−
.12)
(.30
)∗∗(.
05)
(.46
)∗∗(.
35)∗∗
(−.0
5)(.
03)
(.23
)∗∗(.
05)
(.34
)∗∗M
odel
2In
dire
ctca
re.0
8−.
05.2
5∗∗.0
1.4
2∗∗.2
7∗−.
03.1
1.0
7.0
1.2
4∗∗(.
09)
(−.0
3)(.
30)∗∗
(.07
)(.
49)∗∗
(.35
)∗∗(.
01)
(.06
)(.
06)
(.01
)(.
31)∗∗
Mod
el3
Teac
hing
/D
isci
plin
e.0
2−.
04.2
5∗∗.0
6.4
1∗∗.2
7∗.1
0−1
1.1
2.0
3.2
6∗∗(.
05)
(−.0
4)(.
30)∗∗
(.07
)(.
47)∗∗
(.36
)∗∗(.
07)
(.01
)(.
03)
(.00
)(.
31)∗∗
Mod
el4
Pla
y.0
5−.
03.2
4∗∗.0
1.4
4∗∗.2
7∗−.
03.1
4.1
8∗(.
04)
27∗∗
(.09
)(−
.02)
(.28
)∗∗(.
04)
(.50
)∗∗(.
36)∗∗
(−.0
4)(.
07)
(.14
)(.
01)
(.33
)∗∗M
odel
5L
eisu
reou
tdoo
rs.0
5−.
05.2
6∗∗.0
3.4
2∗∗.2
8.1
5.1
6.0
8.0
3.2
7∗∗(.
02)
(−.0
2)(.
30)∗∗
(.07
)(.
48)∗∗
(.36
)∗∗(.
13)
(−.0
6)(.
12)
(.02
)(.
33)∗∗
Not
es:I
npa
rent
hesi
s,th
ebe
taes
tim
ates
offa
ther
invo
lvem
entu
sing
fath
ers’
repo
rts.
Nm
othe
rs=
136;
Nfa
ther
s=
101.
Dum
my
codi
ngfo
rge
nder
:boy
sco
ded
as‘1
’,gi
rls
code
das
‘0’.
∗ p<
.05;
∗∗p
<.0
1.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
8 N. Torres et al.
Tabl
e5.
Bet
aes
tim
ates
and
R2
ofre
gres
sion
mod
els
ofP
lay
Inte
ract
ion,
wit
hm
othe
rs’
and
fath
ers’
repo
rts
onfa
ther
invo
lvem
ent.
Blo
ck1
Blo
ck2
Fam
ilyva
riab
les
Chi
ldat
trib
utes
Inte
ract
ion
betw
een
chil
dat
trib
utes
and
fath
erin
volv
emen
t
Fath
erin
volv
emen
tE
duca
tion
leve
lG
ende
r(m
ale)
Age
Dif
ficu
ltte
mpe
ram
ent
R2
Gen
der
XFa
ther
invo
lvem
ent
Age
XFa
ther
invo
lvem
ent
Dif
ficu
ltte
mpe
ram
ent
XFa
ther
invo
lvem
ent
R2
Cha
nge
Tota
lmod
elad
just
edR
2
Mod
el1
Dir
ectc
are
.17
.18∗
−.29
∗∗.0
9−.
18∗
.18∗∗
−.16
.02
.04
.01
.14∗∗
(.16
)(.
15)
(−.2
2)∗∗
(.18
)(−
.24)
∗∗(.
21)∗∗
(−.2
0)(−
.01)
(.11
)(.
03)
(.18
)∗∗M
odel
2In
dire
ctca
re−.
03.2
1∗−.
28∗∗
.09
−.15
.18∗∗
.03
−.02
.13
.02
.15∗∗
(−.1
2)(.
15)
(−.1
9)∗
(.19
)∗(−
.23)
∗(.
21)∗∗
(.09
)(−
.10)
(.10
)(.
05)
(.21
)∗∗M
odel
3Te
achi
ng/
.04
.20∗
−.26
∗∗.0
9−.
20∗
.18
−.08
−.11
.10
.03
.16∗∗
Dis
cipl
ine
(−.0
8)(.
21)∗
(−.1
6)(.
17)
(−.3
2)∗∗
(.23
)∗∗(.
03)
(−.1
2)(.
23)∗
(.06
)∗(.
22)∗∗
Mod
el4
Pla
y.0
5.2
0∗−.
28∗∗
.10
−.18
∗.1
8∗∗−1
1.0
5−.
04(.
01)
.14∗∗
(−.0
3)(.
16)
(−.2
3)∗∗
(.21
)∗(−
.26)
∗∗(.
21)∗∗
(–.1
0)(.
08)
(−.1
0)(.
02)
(.18
)∗∗M
odel
5L
eisu
reou
tdoo
rs−.
09.2
1∗∗−2
7∗∗.1
0−.
17∗
.18∗∗
−.05
−.02
.04
.00
.13∗∗
(.00
)(.
17)
(−.2
2)∗
(.17
)(−
.24)
∗(.
21)∗∗
(.02
)(-
09)
(.03
)(.
01)
(.15
)∗∗
Not
es:I
npa
rent
hese
sth
ebe
taes
tim
ates
offa
ther
invo
lvem
entu
sing
fath
ers’
repo
rts.
Nm
othe
rs=
136;
Nfa
ther
s=
101.
Dum
my
codi
ngfo
rge
nder
:boy
sco
ded
as‘1
’,gi
rls
code
das
‘0’.
∗ p<
.05;
∗∗p
<.0
1.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 9
explained between 24% and 34% of the variance. This wasmostly due to the effects of the child’s gender and diffi-cult temperament, which had significant beta estimates inall models and similar magnitude for fathers’ and mothers’reports.
As to block 2, in regression model 1, for father involve-ment in Direct Care it attained a significant R2 change of.06 (F 3.71; p = .01) for mothers’ reports and a marginallysignificant R2 change of .05 (F 2.40; p = .07) for fathers’reports, and this was due to the significant interaction termof difficult temperament of the child with father involve-ment in Direct Care in both parents’ reports (betamother =.22; p = .005 and betafather = .23; p = .009). This inter-action term is illustrated in Figure 1. In order to illustratethe interaction terms, we transformed the continuous scoresof difficult temperament into a categorical variable withthree levels of difficult temperament: (1) scores below onestandard deviation of the mean (low difficult temperamentlevel, n = 23); (2) scores between one standard deviationbelow and one standard deviation above the mean (mediumdifficult temperament level, n = 95); and (3) scores aboveone standard deviation of the mean (high difficult tempera-ment level, n = 24).
To further explore the substantive meaning of this inter-action term, we tested if the simple slopes of the regression
of Play Disruption on Direct Care at the three values ofdifficult temperament were different from zero. For lev-els 1 and 2 the simple slopes were r = -.29 and r =.11, respectively, and not significantly different from zero,and for level 3 the simple slope was strong and signifi-cant (beta = .56; p = .005). The differences between thebetas of level 3 and levels 2 and 1 were significant (z = –2.16; p = .015 and z = 2.95; p = .016, respectively), andthe difference of the betas between levels 2 and 1 wasmarginally significant (z = 1.62; p = .053).
The Play Disruption regression model 4 (for fatherinvolvement in Play) according to mothers’ reports hada total adjusted R2 of 27% explained variance; block1 attained an R2 of .27 (F = 9.78; p = .00), and block2 attained an R2 change of .03 that was marginally signifi-cant (F = 2.17; p = .09), and this was mostly due to the sta-tistically significant interaction term of difficult tempera-ment with father involvement in Play (beta = .18; p = .02).These results are illustrated in Figure 2.
To further explore the substantive meaning of this inter-action term, we again tested the significance of the simpleslopes: for level 1 (low difficult temperament) the simpleslope was moderate and marginally significant (beta = –.35; p = .058), and for levels 2 and 3 (medium andhigh difficult temperament) the betas were .13 and .02,
2.75
2.50
2.25
2.00
1.75
Pla
y d
isru
pti
on
wit
h p
eers
1.50
1.25
1.00
.75
1.00 1.50 2.00
Father involvement in direct care
2.50 3.00 3.50
Low difficulttemperamentMedium difficulttemperamentHigh difficulttemperament
4.00
Figure 1. Interaction of difficult temperament with father involvement in direct care (mothers’ reports) on play disruption with peers.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
10 N. Torres et al.
Pla
y d
isru
pti
on
wit
h p
eers
2.50
2.25
2.00
1.75
1.50
1.25
Father involvement in play
1.00 1.50 2.00 2.50 3.00 3.50 4.00
Low difficulttemperament
temperament
temperament
Medium difficult
High difficult
Figure 2. Interaction of difficult temperament level with father involvement in play (mothers’ reports) on play disruption with peers.
respectively, and not significant (see Figure 2). However,the difference between the betas of levels 1 and 2 wassignificant (z = –1.96, p = .025).
Play interaction regression models
As shown in Table 5, the adjusted R2 for block 1 wassignificant in all the five regression models for play interac-tion, and explained between 13% and 22% of the variance,mostly due to the effects of the parents’ higher educationallevel, gender and temperament.
Using fathers’ reports in the model on Teaching/
Discipline, block 2 attained a significant R2 change of .06(F = 2.75; p = .04), and this was mostly due to the interac-tion term of difficult temperament of the child with fatherinvolvement in Teaching/Discipline (beta = .23; p = .02).This interaction term is illustrated in Figure 3. Again, tofurther explore the substantive meaning of this interactionterm, we tested the significance of the simple slopes of theregression for each level of difficult temperament. The betacoefficients were -.28 for level 1, -.13 for level 2 and .28 forlevel 3; none of them were significantly different from zero.However, the difference between the betas for levels 1 and3 was significant (z = –1.65, p = .05).
Play disconnection regression models
None of the regression models for play disconnectionaccounted for a significant amount of explained variance,in either blocks 1 and 2 or in total model adjusted R2. Therewere also no significant beta parameter estimates, as shownin Table 6.
Discussion
Findings from this study concerning the mean levels offather involvement illustrated in Table 2 show that, onthe whole, the families could be described as followingso-called ‘traditional’ gender roles; specifically, fatherstend to be more present and involved in play and inTeaching/Discipline than in other tasks, while mothers pro-vide both Direct and Indirect Care for children (Hwang &Lamb, 1997). Also, according to mothers’ reports, fatherstend to display higher involvement in Teaching/Disciplinewith boys than girls, which suggests a gender bias thathas been found in prior studies (e.g. McHale, Crouter, &Tucker, 1999).
The results of the correlation analyses support thenotion that the quality of interactive peer play is associatedwith child age, gender, and temperament (Coplan et al.,
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 11
4.00
3.50
3.00
2.50
Pla
y in
tera
ctio
n w
ith
pee
rs
2.00
1.50
Father involvement in teaching/discipline
2.00 2.50 3.00 3.50 4.00
Low difficulttemperamentMedium difficult
High difficulttemperament
temperament
Figure 3. Interaction of difficult temperament level with father involvement in teaching/discipline (fathers’ reports) on play interactionwith peers.
2006). In the present study, older children were reported bytheir mothers to engage in less isolated play, a result thatsupports the well-established notion that peer play socialparticipation increases with the children’s age (Howes,1983, 1988). As regards gender, the present findings arecongruent with previous literature showing that boys tendto display a more disruptive and aggressive play style thangirls (Johnson, Christie, & Yawkey, 1999). Finally, difficulttemperament was associated with more Play Disruptionand less Play Interaction, which is also congruent with pre-vious research on parental reports of difficult temperamentas associated with lower social competence measured bypeer acceptance in preschoolers (Szewczyk-Sokolowski,Bost, & Wainwright, 2005; Walker, Berthelsen, & Irving,2001). The mean levels of both Play Disruption and PlayDisconnection were rather low, which is to be expectedin a normative sample of preschoolers. That is, aggres-sive behavior is quite rare compared with positive playbehavior at this age and dominance hierarchies are alreadywell established which prevent frequent overt conflicts(Vaughn & Santos, 2008), while isolated play dimin-ishes with age and socio-emotional developmental progress(Rubin et al., 1976).
Results from the correlation analysis do not support theidea that higher father involvement has a linear and directassociation with positive social development of children(see Amato & Rivera, 1999, for a detailed discussion onthis topic). These results bring to mind Amato and Rivera’s(1999) caveat that only about 50% of studies show a sig-nificant effect of father involvement when the amount ofmother involvement is controlled for.
Due to the cross-sectional design of this study, it isnot possible to determine directional causality, and henceseveral alternative explanations are plausible for the associ-ation between more peer Play Disruption and greater fatherinvolvement. For instance, more disruptive children couldtend to call for greater father involvement in TeachingDiscipline, Leisure Outdoors, and Direct Care, due to themother’s greater need of support in dealing with this typeof more demanding children. The significant correlationsof father involvement in Teaching/Discipline with diffi-cult temperament (rmother = .20 and rfather = .19), whichare of the same magnitudes of the correlations of fatherinvolvement in Teaching/Discipline with Play Disruption(rmother = .20 and rfather = .17), partially support thehypothesis that fathers tend to be more involved in teaching
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
12 N. Torres et al.Ta
ble
6.B
eta
esti
mat
esan
dR
2of
regr
essi
onm
odel
sof
play
disc
onne
ctio
n,w
ith
mot
hers
’an
dfa
ther
s’re
port
son
fath
erin
volv
emen
t.
Blo
ck1
Blo
ck2
Fam
ilyva
riab
les
Chi
ldat
trib
utes
Inte
ract
ion
betw
een
chil
dat
trib
utes
and
fath
erin
volv
emen
t
Fath
erin
volv
emen
tE
duca
tion
leve
lG
ende
r(m
ale)
Age
Dif
ficu
ltte
mpe
ram
ent
R2
Gen
der
XFa
ther
invo
lvem
ent
Age
XFa
ther
invo
lvem
ent
Dif
ficu
ltte
mpe
ram
entX
Fath
erin
volv
emen
tR
2
Cha
nge
Tota
lm
odel
adju
sted
R2
Mod
el1
Dir
ectc
are
.12
.02
−.01
−.13
.05
.03
−.12
.03
.12
.02
.05
(.13
)(.
06)
(.06
)(−
.08)
(.12
)(.
05)
(−.1
1)(.
00)
(.06
)(.
00)
(.05
)M
odel
2In
dire
ctca
re.0
2.0
5−.
02−.
14.0
9.0
4−.
17−.
09.1
3.0
4.0
8(−
.06)
(.12
)(.
03)
(−.0
8)(.
13)
(.07
)(−
.16)
(.02
)(−
.06)
(.02
)(.
09)
Mod
el3
Teac
hing
/D
isci
plin
e.0
4.0
2−.
00−.
12.0
6.0
3.0
3.0
7.0
1.0
0.0
3(−
.12)
(.03
)(.
03)
(−.0
9)(.
17)
(.04
)(.
20)
(.05
)(−
.06)
(.02
)(.
06)
Mod
el4
Pla
y.0
8.0
2−.
00−.
14.0
7.0
3−.
06.0
4.0
3.0
0.0
3(.
09)
(.08
)(.
04)
(−.1
3)(.
14)
(.05
)(.
08)
(−.0
3)(−
.05)
(.00
)(.
05)
Mod
el5
Lei
sure
outd
oors
−.02
.04
−.01
−.13
.08
.03
−.06
−.09
−.08
.02
.05
(−.1
2)(.
09)
(.06
)(−
.05)
(.17
)(.
06)
(−.0
4)(−
.10)
(.09
)(.
02)
(.08
)
Not
es:I
npa
rent
hese
s,th
ebe
taes
tim
ates
offa
ther
invo
lvem
entu
sing
fath
ers’
repo
rts.
Nm
othe
rs=
136;
Nfa
ther
s=
101.
Dum
my
codi
ngfo
rge
nder
:boy
sco
ded
as‘1
’,gi
rls
code
das
‘0’.
∗ p<
.05;
∗∗p
<.0
1
and disciplinary matters with more temperamentally diffi-cult and with more play-disruptive children. However, thisholds for the dimension of Teaching/Discipline only, asfathers do not tend to be more involved with temperamen-tally difficult children in Leisure Outdoors or Direct Careaccording to both parents’ reports. Since fathers do nottend to engage more in Direct Care and Leisure Outdooractivities with temperamentally difficult children, the asso-ciation of Play Disruption with greater father involvementin Leisure Outdoor activities and in Direct Care (accord-ing to mothers’ reports) should be sought somewhere elsethan in the intervening variable of children’s difficult tem-perament. Perhaps greater involvement of the father inLeisure Activities and Direct Care could be causally linkedto higher Play Disruption with peers. The correlations,however statistically significant, are of a small magni-tude, which confirms previous research showing that thedifferential effects of the contributions of the father andmother to socialization in two-parent families are quite sub-tle (Lewis & Lamb, 2003), specially in the preschool years(Lamb, 2010).
In the second step of the OLS Regression Models,where we controlled for family education and childattribute variables and including interaction terms in block2, the complex interplay of father involvement and childvariables proposed by previous authors (Pleck, 2010;Tamis-LeMonda & Cabrera, 1999) is to some extentspelled out by the moderation analyses of the models’block 2. These show that the amount of relative fatherinvolvement in Direct Care interacts with child tempera-ment to explain a substantial amount of variance in PlayDisruption with peers. Findings also show that fatherinvolvement in Play interacts with temperament as regardsthe level of peer Play Disruption, as illustrated in Figure 3.
The significant interactions between father involve-ment in Direct Care and temperament, and in Play andtemperament suggest that children’s responses to parentalbehavior is co-determined, or moderated, by the child’sbehavioral tendencies (i.e. difficult temperament). Becausechildren perceived by fathers as temperamentally diffi-cult tend to induce negative and controlling responsesfrom their fathers, there is likelihood that their interac-tions will be more negative and contribute to a conflict-ual type of relationship between the father and the childthat can be reenacted in the child’s relationships withpeers. Previous research has shown that fathers are lessaffectionate and less responsive with children who areperceived as more difficult (Manlove & Vernon-Feagans,2002; Volling & Belsky, 1992). Consequently, in familieswhere the father is more involved, the plausible nega-tive interactional patterns with temperamentally difficultchildren can possibly reinforce disruptive behaviors withthe child’s peers. In a previous study by Foster et al.(2007), fathers’ negative expressiveness as a responseto children’s behavior was predictive of their children
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 13
being rated as more aggressive and disruptive by theirteachers.
Another possible feature of father involvement withmore challenging children is increased directiveness.In previous studies, paternal directiveness was associ-ated with negative social outcomes, such as seldom beingsought out by other children (MacDonald & Parke, 1984).The results of the present study showing that greater fatherinvolvement in Direct Care and Play with temperamen-tally difficult children was associated with more DisruptivePlay is congruent with findings by MacDonald and Parke(1984).
The relative father/mother involvement measure usedin the present study suggests that father involvement inDirect Care and in Play with children perceived as temper-amentally difficult is associated with more disruptive playwith peers than mother involvement, while, conversely,father involvement in Teaching/Discipline is associatedwith more interactive play with peers in children perceivedas temperamentally difficult. This differential effect may bethe result of the different styles of interaction of fathersand mothers. For instance, Bretherton et al. (2005) foundthat highly involved fathers of preschoolers tended to bestricter than mothers, and Carson and Parke (1996) showedthat reciprocal negative affect sequences in physical playspecifically between father and child (but not mother andchild) were associated with low peer interaction compe-tence (children who shared less, were more aggressive andavoided others). On the other hand, the higher strictness anddirectiveness of fathers in disciplinary matters with diffi-cult children could possibly result in an increase of socialcompetence as measured by peer play interaction scores.
Explanations derived from evolutionary theories ofresource control (Hawley, 2002) and the adaptive functionof aggressiveness (Vaughn & Santos, 2007, 2008) can alsobe used to understand our results. According to these theo-retical models, preschool children’s coercive strategies ofresource control (e.g. children’s dominance in the play-ground and assertive control of interesting toys) are partof social competence when combined with pro-social abili-ties from the child. Furthermore, social value is linked withabilities to not only cooperate but also compete with peersto gain access of social assets (Geary & Flinn, 2001). Theresults from the present study can also be understood inlight of the fact that men are more likely to encourage theirchildren to be competitive and to take risks. In the case oftemperamentally difficult children, this might result in theincrease of disruptive and aggressive strategies of compe-tition when interacting with other children. However, ourresults show that it is also possible that the positive effect ofhigher paternal investment in teaching and discipline mayincrease temperamentally difficult children’s acquisition ofsocial rules and more contained forms of solving conflictswith other children.
The generalization of the present findings is limited,given that the sample was mainly composed of dual-earnerfamilies (both father and mother working) and childrenattending a single daycare program. Another limitation per-tains to the use of a relative involvement rather than ameasure of overall time spent with the child. Accordingto Lamb, Pleck, Charnov, and Levine (1985), while themost reliable methods for measuring the amount of fatherengagement are time-use diaries reporting absolute timeof father involvement in several activities, other measuressuch as relative responsibility for childcare roles can bereliable and valid indirect proxies. This is especially rel-evant since time-use diaries are difficult to implement andcan lead to different problems of specification of activities(Pleck, 2010). Furthermore, the overall amount of paternalinvolvement time is correlated with a series of confound-ing variables including the amount of mother involvement(Cabrera et al., 2000; Lamb, 2010). Even using estimatesof absolute father involvement and statistically controllingfor mother involvement presents a problem in linear sta-tistical models since the respondents (i.e. the couple) arenot independent and hence the error terms are correlated.One of the ways of avoiding this can be the use of mea-sures asking for the relative father/mother involvement inco-resident parents (see Hwang & Lamb, 1997; Lamb et al.,1983; Pleck, 2010). These were the methodological reasonsfor our choice of a measure of relative father involvementin two-parent families, which was also the choice of previ-ous authors researching in the field of father involvement(e.g. Easterbrooks & Goldberg, 1984; Hwang & Lamb,1997).
ReferencesAmato, P. R., & Rivera, F. (1999). Paternal involvement and chil-
dren’s behavior problems. Journal of Marriage and Family,61(2), 375–384. doi:10.2307/353755
Bates, J. E., Freeland, C. B., & Lounsbury, M. L. (1979).Measurement of infant difficultness. Child Development, 50,794–803.
Belsky, J. (2005). Differential susceptibility to rearing influence:An evolutionary hypothesis and some evidence. In B. Ellis &D. Bjorklund (Eds.), Origins of the social mind: Evolutionarypsychology and child development (pp. 139–163). New York,NY: Guilford.
Bretherton, I., Lambert, J. D., & Golby, B. (2005). Involvedfathers of preschool children as seen by themselves and theirwives: Accounts of attachment, socialization, and compan-ionship. Attachment & Human Development, 7(3), 229–251.doi:10.1080/14616730500138341
Cabrera, N., Tamis-LeMonda, C. S., Bradley, R. H., Hofferth,S., & Lamb, M. E. (2000). Fatherhood in the twenty-first century. Child Development, 71(1), 127–136.doi:10.1111/1467–8624.00126
Carson, J. L., & Parke, R. D. (1996). Reciprocal negative affectin parent-child interactions and children’s peer competency.Child Development, 67(5), 2217–2226.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
14 N. Torres et al.
Coplan, R. J., Rubin, K. H., & Findlay, L. C. (2006). Social andnon social play. In D. P. Fromberg & D. Bergen (Eds.), Playfrom birth to twelve (2nd ed., pp. 75–86). New York, NY:Garland.
Easterbrooks, M. A., & Goldberg, W. A. (1984). Toddlerdevelopment in the family: Impact of father involvementand parenting characteristics. Child Development, 55(3),740–752.
Fantuzzo, J., Mendez, J., & Tighe, E. (1998). Parental assessmentof peer play: Development and validation of the parent ver-sion of the Penn Interactive Peer Play Scale. Early ChildhoodResearch Quarterly, 13, 655–672.
Foster, P. A., Reese-Weber, M., & Kahn, J. H. (2007). Fathers’parenting hassles and coping: Associations with emo-tional expressiveness and their sons’ socioemotional com-petence. Infant and Child Development, 16(3), 277–293.doi:10.1002/icd.507
Geary, D. C. (2000). Evolution and proximate expression ofhuman paternal investment. Psychological Bulletin, 126(1),55–77.
Geary, D. C., & Flinn, M. V. (2001). Evolution of human parentalbehavior and the human family. Parenting: Science andPractice, 1(1), 5–61.
Hawley, P. H. (2002). Social dominance and prosocial andcoercive strategies of resource control in preschoolers.International Journal of Behavioral Development, 26(2),167–176.
Hewlett, B. S., Lamb, M. E., Leyendecker, B., & Scholmerich, A.(2000). Parental investment strategies among Aka foragers,Ngandu farmers and Euro-American urban industrialists.In L. Cronk, N. Chagnon, & W. Irons (Eds.), Adaptationand human behavior: An anthropological perspective (pp.149–172). Hawthorne, NY: Aldine de Gruyter.
Howes, C. (1983). Patterns of friendship. Child Development,54(4), 1041–1053.
Howes, C. (1988). Peer Interaction of young children.Monographs of the Society for Research in ChildDevelopment, 53, 1–78.
Hwang, C. P., & Lamb, M. E. (1997). Father involvement inSweden: A longitudinal study of its stability and correlates.International Journal of Behavioral Development, 21(3),621–632. doi:10.1080/016502597384811
Johnson, J. E., Christie, J. F., & Yawkey, T. D. (1999). Playand early childhood development (2nd ed.). Longman, NY:Addison-Wesley.
Klein Velderman, M., Bakermans-Kranenburg, M. J., Juffer, F.,& van IJzendoorn, M. H. (2006). Effects of attachment-basedinterventions on maternal sensitivity and infant attachment:Differential susceptibility of highly reactive infants. Journalof Family Psychology, 20, 266–274.
Koestner, R., Franz, C., & Weinberger, J. (1990). The fam-ily origins of empathic concern: A 26-year longitudinalstudy. Journal of Personality and Social Psychology, 58(4),709–717. doi:10.1037/0022–3514.58.4.709
Labrell, F. (1994). A typical interaction behaviour between fathersand toddlers: Teasing. Early Development and Parenting,3(2), 125–130. doi:10.1002/edp.2430030209
Lamb, M., & Lewis, C. (2010). The development and significanceof father-child relationships in two-parent families. In M. E.Lamb (Ed.), The role of the father in child development (5thed.). Hoboken, NJ: Wiley.
Lamb, M. E. (Ed.) (2010). The role of the father in child develop-ment (5th ed.). Hoboken, NJ: Wiley.
Lamb, M. E., Frodi, M., Hwang, C.-P., & Frodi, A. M. (1983).Effects of paternal involvement on infant preferences formothers and fathers. Child Development, 54(2), 450–458.doi:10.2307/1129706
Lamb, M. E., Pleck, J. H., Charnov, E. L., & Levine, J. A.(1985). Paternal behavior in humans. American Zoologist,25(3), 883–894.
Lewis, C., & Lamb, M. E. (2003). Fathers’ influences on chil-dren’s development: The evidence from two-parent fami-lies. European Journal of Psychology of Education, 18(2),211–228.
MacDonald, K. (1987). Parent-child physical play with rejected,neglected, and popular boys. Developmental Psychology,23(5), 705–711. doi:10.1037/0012–1649.23.5.705
MacDonald, K. (1993). Parent-child play: Descriptions andimplications. Albany: State University of New York Press.
MacDonald, K., & Parke, R. D. (1984). Bridging the gap: Parent-child play interaction and peer interactive competence. ChildDevelopment, 55(4), 1265–1277. doi:10.2307/1129996
Manlove, E. E., & Vernon-Feagans, L. (2002). Caring forinfant daughters and sons in dual-earner households:Maternal reports of father involvement in weekday timeand tasks. Infant and Child Development, 11(4), 305–320.doi:10.1002/icd.260
McHale, S. M., Crouter, A. C., & Tucker, C. J. (1999). Familycontext and gender role socialization in middle childhood:Comparing girls to boys and sisters to brothers. Child devel-opment, 70(4), 990–1004.
Monteiro, L., Veríssimo, M., & Pessoa e Costa, I. (2008).Escala Envolvimento Parental: Actividades de Cuidados ede Socialização [Parental Involvement Questionnaire: Childcare and sociaization related tasks]. (Unpublished manual).ISPA, Lisbon.
Monteiro, L., Verissimo, M., Vaughn, B., Santos, A., Torres,N., & Fernandes, M. (2010). The organization of chil-dren’s secure base behaviour in two-parent Portuguesefamilies and father’s participation in child-related activi-ties. European Journal of Developmental Psychology, 7(5),545–560. doi:10.1080/17405620902823855
Parke, R. D., McDowell, D. J., Kim, M., Killian, C., Dennis, J.,Flyr, M. L., & Wild, M. N. (2008). Fathers’ contributionsto children’s peer relationships. In C. S. Tamis-LeMonda& N. Cabrera (Eds.), Handbook of father involvement:Multidisciplinary perspectives (pp. 141–167). Mahwah, NJ:Routledge.
Pleck, J. (2010). Paternal involvement: Revised conceptualizationand theoretical linkages with child outcomes. In M. E. Lamb(Ed.), The role of the father in child development (5th ed., pp.58–93). Hoboken, NJ: Wiley.
Ramchandani, P. G., van IJzendoorn, M., & Bakermans-Kranenburg, M. J. (2010). Differential susceptibility tofathers’ care and involvement: The moderating effect ofinfant reactivity. Family Science, 1(2), 93–101.
Rubin, K., Bukowski, W., & Parker, J. (1998). Peer interac-tion, relationships and groups. In W. Damon & N. Eisenberg(Eds.), Handbook of child psychology: Vol. 3. Social, emo-tional, and personality development (pp. 619–700). NewYork, NY: Wiley.
Rubin, K. H., Maioni, T. L., & Hornung, M. (1976). Free playbehaviors in middle- and lower-class preschoolers: Parten andPiaget revisited. Child Development, 47(2), 414–419.
Rutherford, E., & Mussen, P. (1968). Generosity in nurs-ery school boys. Child Development, 39(3), 755–765.doi:10.2307/1126981
Shrive, F. M., Stuart, H., Quan, H., & Ghali, W. A. (2006). Dealingwith missing data in a multi-question depression scale: Acomparison of imputation methods. BMC Medical ResearchMethodology, 13(6), 57–67.
Smilansky, S. (1968). The effects of sociodramatic play on dis-advantaged children: Preschool children. New York, NY:Wiley.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Family Science 15
Szewczyk-Sokolowski, M., Bost, K. K., & Wainwright, A.B. (2005). Attachment, temperament, and preschoolchildren’s peer acceptance. Social Development, 14,379–397.
Tamis-LeMonda, C. S., & Cabrera, N. (1999). Perspectiveson father involvement: Research and policy. Society forResearch in Child Development Social Policy Report, 13(2),1–32.
Tamis-LeMonda, C. S., Shannon, J. D., Cabrera, N. J., &Lamb, M. E. (2004). Fathers and mothers at play with their2- and 3-year-olds: Contributions to language and cogni-tive development. Child Development, 75(6), 1806–1820.doi:10.1111/j.1467-8624.2004.00818.x
Vaughn, B., & Santos, A. J. (2007). Why they don’t all get along:An evolutionary/ecological account of aggressive behaviorand trait aggression in human children and adolescents. InP. Hawley (Ed.), Aggression and adaptation: The brightside to bad behavior (pp. 31–63). Mahwah, NJ: LawrenceErlbaum.
Vaughn, B., & Santos, A. J. (2008). Behavioral structures gov-erning social transactions among young children: Affiliationand dominance in groups of preschool age children. In K. H.Rubin, W. Bukowski, & B. Laursen (Eds.), Handbook of peerinteractions, relationships, and groups (pp. 195–214). NewYork, NY: Guilford.
Veríssimo, M., Santos, A. J., Vaughn, B. E., Torres, N., Monteiro,L., & Santos, O. (2011). Quality of attachment to fatherand mother and number of reciprocal friends. Early ChildDevelopment and Care, 181(1), 27–38.
Volling, B. L., & Belsky, J. (1992). The contribution of mother-child and father-child relationships to the quality of siblinginteraction: A longitudinal study. Child Development, 63(5),1209–1222. doi:10.2307/1131528
Walker, S., Berthelsen, D., & Irving, K. (2001). Temperamentand peer acceptance in early childhood: Sex and social statusdifferences. Child Study Journal, 31(3), 177–191.
Waters, E., & Sroufe, L. A. (1983). Social competence as adevelopmental construct. Developmental Review, 3, 79–97.
Dow
nloa
ded
by [
Man
uela
Ver
íssi
mo]
at 0
3:00
05
Apr
il 20
13
Top Related