Children's help seeking: The role of parenting

Learning and Instruction 18 (2008) 160e171www.elsevier.com/locate/learninstruc

Children’s help seeking: The role of parenting

Minna Puustinen a,*, Anna-Liisa Lyyra b, Riitta-Leena Metsapelto b, Lea Pulkkinen b

a University of Poitiers & CNRS (UMR 6215 LMDC), and IUFM Nord-Pas de Calais, Franceb University of Jyvaskyla, Finland

Received 20 July 2006; revised 29 January 2007; accepted 6 February 2007

Abstract

Ninety-nine families with a school-age child participated in this study, aimed at analysing the relationship between parentingand children’s help seeking. The parenting data included self-reports on parents’ child-rearing principles and behavioural observa-tions during parentechild interactions. To test help seeking, the children were placed in a problem-solving situation and had theopportunity to seek help from the experimenter. For girls, higher levels of parental nurturance were linked to longer thinking timespreceding help seeking and to lessened capacity to reuse previously received help. For boys, higher levels of fathers’ emotionalwarmth were related to higher rates of irrelevant help seeking.� 2007 Elsevier Ltd. All rights reserved.

Keywords: Help seeking; Parenting; Self-regulated learning

1. Introduction

Two decades ago, Nelson-Le Gall (1981, 1985) and Ames (1983) triggered revolution in the domain of help seeking.Instead of considering help seeking as a manifestation of dependence as usually done at that time, they focused on itsadaptive role in learning. The first studies inspired by this new idea did not take long to appear (e.g., Karabenick &Knapp, 1988; van der Meij, 1988), and help seeking was soon regarded as an instrument in educational settings. Quitenaturally, then, help-seeking research crossed paths with a simultaneously developing research area related to self-regulated learning (SRL), which describes the ways in which individuals regulate their own cognitive processes inan educational setting (e.g., Boekaerts, Pintrich, & Zeidner, 2000; Puustinen & Pulkkinen, 2001). At present, bothhelp-seeking researchers (e.g., Karabenick & Newman, 2006; Puustinen, Kokkonen, Tolvanen, & Pulkkinen, 2004)and SRL researchers (e.g., Pintrich, 2000) agree on the beneficial effects of selective help seeking on learning andunderstanding.

Given the abundance of help-seeking literature today, it is surprising how little attention has been paid to the issueof the development of the capacity to use help seeking as a strategy for SRL. Some researchers (e.g., Nelson-Le Gall,1985; Puustinen, 1998) have suggested that the help-seeking capacity develops through early parentechild

* Corresponding author. Universite de Poitiers & CNRS, Laboratoire Langage, Memoire, et Developpement Cognitif (UMR 6215 LMDC),

MSHS, 99 avenue du Recteur Pineau, 86000 Poitiers, France. Tel.: þ33 (0)549454890.

E-mail address: [email protected] (M. Puustinen).

0959-4752/$ - see front matter � 2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.learninstruc.2007.02.001

161M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

interactions, as suggested by Vygotsky (1981) in his theory of internalisation. According to this theory, all humancapacities first appear during interpersonal interactions and are progressively appropriated at the intra-psychologicallevel. The role of adult assistance and guidance in the internalisation process is essential; the zone of proximal devel-opment refers expressly to what the child is able to do in collaboration with an adult, but not yet alone.

Even less attention has been paid to what particular parenting properties or characteristics of the home environmentare related to children’s help-seeking behaviour. Newman’s (2000) thorough theoretical review on the role of social-isation in help seeking is a notable exception. He analysed the help-seeking literature in the light of the self-systemtheory (e.g., Connell, 1990), according to which three kinds of parental behaviour, parental involvement, support forautonomy, and support for the development of competence, influence children’s SRL. Parental involvement refers notonly to affective involvement but also to the parents’ interactions with the school (e.g., participation in parenteteacherconferences) and shared activities with their children (e.g., help with homework). Support for autonomy corresponds,for example, to joint attention and engagement, as well as conversational parentechild interactions in general, whichhelp the child develop a personal sense of agency (e.g., the feeling of being in charge of one’s own actions) and ef-fective questioning skills (e.g., formulating questions in accordance with social norms). Support for competence refersto the type of parental discourse; scaffolding-like discourse emphasising the importance of self-questioning and ques-tioning of others is considered particularly beneficial. According to Newman’s (2000) analysis, children potentiallylearn, via these three channels, that difficulties and failures may require assistance, for which they can rely on adults.In other words, the parents’ responsiveness to their child’s needs may help the young one develop a sense of controlvis-a-vis help seeking.

Stright, Neitzel, Sears, and Hoke-Sinex’s (2001) work on self-regulated learning has also shed new light on thisquestion. These authors coded parental instruction provided to children during a problem-solving task. The followingschool year (when the children entered third grade) they assessed the children’s SRL (e.g., frequency of help seeking)in the classroom. The results revealed that the quality of parental instruction predicted the frequency of child help-seeking behaviour in the classroom. For example, children whose parents presented their instructions in an under-standable manner (i.e., in small steps at an appropriate pace) were more likely than others to seek help from theirteacher or their peers when faced with a problem. In another study conducted with younger children, the same authors(Neitzel & Stright, 2003) showed further that when mothers provided their children with high-quality metacognitiveinformation (i.e., information that brought out the thinking behind the problem-solving process), guided their child-ren’s work in an appropriate and understandable way, and provided them with solid emotional support, the childrentended to ask their teacher for help in kindergarten.

To fully understand the implications of the preceding results, it is important to clearly define the underlying notions(i.e., help seeking, parenting) and to reflect upon the ensuing methodological issues. Parenting can be defined as the‘‘way in which parents construct, interpret, and form relationships, child-rearing practices and child-rearing environ-ments in interactions with children and adolescents’’ (Gerris, 2001, p. 9). In short, parenting is an inherently complexand multi-factorial construct, comprising both the social cognitions parents have formed about their actions, behav-iours, and practices as parents, and their actual behaviour when interacting with their child. However, studies on par-enting have often settled for only one of these aspects, either by reporting on parents’ general child-rearing principlesand practices, typically measured through self-report questionnaires gathered from the parents themselves(e.g., parental use of discipline and control; e.g., Kochanska, 1997), or by observing parents’ actual parenting behaviour(e.g., parental instruction during a joint problem-solving activity; e.g., Stright et al., 2001). This is all the more regret-table because previous findings have shown that self-report and observational measures are, at best, only moderatelyrelated (Kochanska, Kuczynski, & Radke-Yarrow, 1989; Metsapelto & Pulkkinen, 2005) and tend to provide differentkinds of information about child rearing. In addition, both of these measurement approaches have strengths and weak-nesses. For example, self-reports provide us with a general overview of parenting from a parent’s own viewpoint, whichis difficult to capture using another method since it is based on a wide range of experiences. At the same time, parentalself-reports may be biased by the parents’ wish to present their child-rearing practices in a favourable light (Schumm,1990). Behavioural observations, on the other hand, are known to be affected by the characteristics of the observer, thesubjects to be rated, and the setting (Cairns & Green, 1979). However, if parentechild interactions are examined bytrained observers over prolonged periods of time and across a broad array of situations, any regularity in the observedbehaviour patterns is usually regarded as a reliable and valid source of information on parenting.

Yet it seems important, when analysing help seeking, to consider it in all of its complexity, that is, as an SRL strategy.Self-regulated learners e and help-seekers e do not ask for help needlessly, that is to say, when they are capable

162 M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

of solving the problem by themselves. Furthermore, they confine their questions to just those hints and explanationsneeded to allow them to finish performing the task on their own. In other words, they avoid asking questions not aimedat understanding the underlying problem principles (e.g., ready-made answers; Puustinen et al., 2004). This suggeststhat the mere frequency of help seeking e regarded by Stright et al. (2001), for example, as an indicator of goodself-regulation in the classroom e might not be sufficient to reveal the real nature of the help-seeking strategy adoptedby a particular child. In fact, the above definition according to which the more a child asks for help the better his/her self-regulation is in the classroom (Stright et al., 2001), is contrary to the assumption that good or self-regulated help-seekersask only necessary questions.

Hence in the present study, our first aim was to obtain a more detailed picture of the relationship between parentingand children’s help seeking. We evaluated parenting from both self-reports and observation data, and we measuredhelp seeking using variables that fit with the definition of SRL. On the basis of previous results (Newman, 2000;Stright et al., 2001), we expected parents’ sensitivity to their children’s needs to be related to ‘‘better’’ (more self-regulated) help seeking (Hypothesis 1).

Recent reviews (e.g., Parke, 2002) have noted parenting differences between mothers and fathers. In general,mothers seem to take more responsibility for childcare, be more supportive, and show more affection than fathers,while fathers appear to be more play-oriented, directive, and authoritarian. Some evidence also suggests that mothersand fathers treat daughters and sons differently, with fathers differentiating more between sons and daughters thanmothers (e.g., Lytton & Romney, 1991). Such differential treatment has been reported to especially concern disciplinestrategies (i.e., physical punishment), with more punishment being directed at boys than at girls. Therefore, our secondaim was to analyse the role of parents’ and children’s gender on help seeking, by investigating the associationsbetween parenting and help seeking separately for each parentechild dyad (i.e., motheredaughter, mothereson,fatheredaughter, fathereson). On the basis of the above results, mothers’ parenting was expected to be related to‘‘better’’ help seeking in both girls and boys, and fathers’ parenting was expected to be associated with ‘‘better’’help seeking essentially in girls (Hypothesis 2).

Given that age differences have been reported in studies on children’s help seeking (e.g., Puustinen, 1998), we con-trolled for age effects on the results. Finally, given that the help-seeking tasks required mathematical reasoning, wecontrolled for the effect of the children’s intellectual ability on the results.

2. Method

2.1. Participants

Our study was part of the ongoing Jyvaskyla Longitudinal Study of Personality and Social Development (JYLS), inwhich the same individuals have been followed up since 1968 (e.g., Pulkkinen, 1982, 1998). The original sample con-sisted of 369 children aged 8 years (173 girls, 196 boys born in 1959) from 12 grade 2 classes randomly selected fromurban and suburban areas of a medium-size town in central Finland. In 2001, when 81% of the males and 85% of thefemales were still participating in the study, the participants were representative, for example, of their age cohort withrespect to employment rate and socioeconomic status (Pulkkinen, 2006). The present study was mostly based on the1997e1999 round of data collection, although we also used the results of a child-rearing questionnaire filled in by theparticipants in 1995. In 1997, all participants with elementary school children were contacted. Seventy-five percent ofthe families agreed to participate (a typical reason for refusal consisted of long distances between their home and ourlaboratory). There were 99 ethnically Finnish families in the final sample, which included the participants of the lon-gitudinal study ages 37e39 years (n¼ 99, 47 women and 52 men), their spouses whenever possible (n¼ 70, 40women and 30 men), and one of their 8- to 12-year-old children (n¼ 99, 42 girls and 57 boys). For mean ages andother descriptive statistics, see Table 1.

2.2. Procedure

The study was carried out in the Emotional and Behavioural Regulation research laboratory at the University ofJyvaskyla. The test sessions (two 3-h sessions) were divided into two 80-min test blocks separated by a 20-min break.The whole family participated in the first session. First, the parents were interviewed and filled in questionnaires whilethe children were administered tests designed to tap their cognitive, executive, behavioural, and emotional functioning.

Table 1

Help-seeking and parenting variables: counts, proportions, and reliability indexes for observed data, and descriptive statistics for imputed data

Variable Observed (%) Cronbach’s a M (SD) t-test

Girls (n¼ 42) Boys (n¼ 57) Girls (n¼ 42) Boys (n¼ 57) t Values df p

Age (years) 42 (100) 57 (100) e 10.1 (1.1) 10.4 (1.1) �1.23 97 0.223

Intellectual ability 42 (100) 56 (98.2) e 101 (11) 100 (10) 0.71 97 0.482

Help seeking

Mean thinking time (s) 39 (92.9) 49 (86.0) e 104 (101) 101 (79) 0.20 97 0.843

Reuse of received help 39 (92.9) 45 (78.9) e 2.21 (0.78) 2.30 (0.68) �0.57 97 0.570

Total number of answers, etca 42 (100) 57 (100) e 2.02 (2.16) 2.53 (2.67) �1.00 97 0.318

Parenting (self-report data)

Mothers’ nurturance 38 (90.5) 49 (86.0) 0.81 3.38 (0.37) 3.44 (0.37) �0.70 97 0.486

Fathers’ nurturance 34 (81.0) 47 (82.5) 0.82 3.36 (0.37) 3.34 (0.39) 0.23 97 0.822

Mothers’ restrictiveness 38 (90.5) 49 (86.0) 0.73 2.20 (0.36) 2.36 (0.40) �2.10 97 0.039

Fathers’ restrictiveness 34 (81.0) 47 (82.5) 0.74 2.28 (0.38) 2.43 (0.41) �1.89 97 0.061

Parenting (observation data)

Mothers’ emotional warmth 22 (52.4) 25 (43.9) 0.89 3.80 (0.42) 3.98 (0.57) �1.78 97 0.078

Fathers’ emotional warmth 20 (47.6) 32 (56.1) 0.91 3.57 (0.57) 3.42 (0.90) 0.93 97 0.357

Mothers’ parental guidance 22 (52.4) 25 (43.9) 0.82 3.22 (0.61) 3.37 (0.68) �1.17 97 0.246

Fathers’ parental guidance 20 (47.6) 32 (56.1) 0.80 3.03 (0.65) 2.86 (0.70) 1.22 97 0.227

a Total number of answers, confirmations, and questions not related to the problem.

163M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

Next, the parent who was included in the longitudinal study participated in two dyadic problem-solving situations withthe child. At the end of the first session, the whole family participated in a family discussion task. The second test ses-sion was solely for the children. It consisted of the help-seeking test and several self-regulation measures. The exper-imenters were psychologists and psychology students. All had received several years of training, including a period ofsupervised field practice.

2.3. Measures and variables

2.3.1. Intellectual abilityA short-form (Sattler, 1992) WISC-R (Wechsler, 1974) comprising the Vocabulary and Block Design tests was

administered to assess the children’s intellectual ability (see Table 1).

2.3.2. Help seekingThe children were placed in a problem-solving situation and had the opportunity for help seeking from the experi-

menter if needed. The material consisted of 12 number-series completion problems that formed four triplets of analo-gous problems. Examples of these problems are: 384 e 192 e 96 e 48 e __ e __, 416 e 208 e 104 e 52 e __ e __, and448 e 224 e 112e 56 e __ e __; for more details on the help-seeking materials and methodology see Puustinen (1998)and Puustinen et al. (2004). In each session, the children were first familiarised with the types of tasks and then began theactual test. The children were encouraged to do their best. They were told they could ask for help from the experimenterif needed. The experimenters, who were instructed to intervene only when the children requested, answered all ques-tions asked. Two standardised aids were available for each task, an explanation and a correct answer. The sessions wererecorded on audio- and videotape.

Five request categories were identified in the video protocols: (1) explanations, or general questions about the solv-ing process (e.g., ‘‘How should I do it?’’), (2) clarifications, or more detailed questions about the solving process (e.g.,‘‘Should I multiply by 6?’’), (3) confirmations, with which the children checked their answers (e.g., ‘‘It will be 18,won’t it?’’, ‘‘Is this right?’’), (4) answers, or direct requests for ‘‘ready-made’’ answers (e.g., ‘‘Could you tell methe answer, please?’’), and (5) other questions, which were not related to the experimental situation (e.g., ‘‘Whattime is it, please?’’). Interrater reliability was estimated by having two coders independently assess the same randomlyselected cases (approximately 10% of the data). Eighty-nine percent of their coding was identical.

The help-seeking variables reflected the following features of what we consider good or self-regulated help seek-ing. First of all, self-regulated help-seekers do not ask for help when they are capable of solving the problem on their

164 M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

own. In other words, they do not rush into questioning: self-regulated help seeking is associated with a long rather thanshort thinking time before actual question asking (see Puustinen et al., 2004). Thus, one of the help-seeking variableswas mean thinking time (in seconds) before asking the first task-relevant question (i.e., explanation, clarification,confirmation, or answer). Eleven participants did not ask any task-relevant questions and therefore did not have anobserved mean thinking time.

Another feature of self-regulated help-seekers is that they ask necessary questions only. So those children (n¼ 37)who asked questions aimed at understanding the solving process (i.e., explanations and clarifications) only during thefirst problem out of three in a triplet, and who were thus able to reuse received help on analogous problems instead ofasking for help again, were given a score of 3; those (n¼ 32) who asked for explanations and clarifications on the first,but also on either the second or the third problem in a triplet, were given a score of 2; and those (n¼ 15) who asked forexplanations and clarifications on the first, second, and third problems were given a score of 1. Children (n¼ 15) whodid not ask for explanations and clarifications on the first problem did not have a reused-help score.

Self-regulated help-seekers also avoid asking questions not aimed at understanding the solving process. Therefore, thetotal number of answers, confirmations, and questions unrelated to the current problem was used as a variable. We optedfor an overall sum variable instead of considering each of the three request categories separately because the categorieswere not normally distributed (the median was 0). The sum-variable distribution was somewhat less skewed (see Table 1).

2.3.3. ParentingBoth self-report and observation data on parenting were gathered. The self-report data were obtained using a

28-item child-rearing practices questionnaire (Metsapelto & Pulkkinen, 2003) that concerned parents’ behaviours, at-titudes, goals, and child-rearing values. In short, it measured the parents’ general parenting principles rather than theparenting of one particular child. The items, drawn from the Child Rearing Practices Report (Roberts, Block, & Block,1984) and Gerris et al.’s (1993) inventory, were worded in the first person. Parents used a four-point Likert scale rang-ing from 1 (not at all) to 4 (very much) to rate the degree to which the items described them as parents. An exploratoryfactor analysis conducted on the 1995 data yielded four factors labelled nurturance, restrictiveness, parenting stress,and parental knowledge (Metsapelto & Pulkkinen, 2003). In the present study, the composite nurturance and restric-tiveness scores were used because they reflect broad parenting dimensions known to be robust correlates of a child’sdevelopmental outcomes (e.g., Steinberg, 2001).

The nurturance score assesses the extent to which parents report being loving, responsive, and involved; the restric-tiveness score corresponds to the degree to which they report exerting psychological control over their children. More-over, previous studies on self-reported parenting have yielded satisfactory testeretest reliability levels for these twovariables (e.g., 0.89 for nurturance for both mothers and fathers, 0.82 for mothers’ restrictiveness, and 0.85 for fathers’restrictiveness over 2 years; Aunola & Nurmi, 2005). The remaining two variables were considered nonessential giventhe object of our study (parenting stress), or were relatively specific parenting dimensions theoretically unrelated tohelp seeking (parental knowledge). The nurturance measure was equal to the mean rating on the 10 items reflectingparental affection for, and support given to, the child (e.g., parent expresses appreciation, parent encourages indepen-dence, parent jokes with the child, parent shows affection, parent respects the child’s opinion). The restrictiveness mea-sure was equal to the mean rating on the 10 items reflecting demands for obedience and the use of punishment (e.g.,parent emphasises the proper behaviour the child should have with parents, parent considers child obedience to be im-portant, parent teaches the child that one has to pay for one’s misdeeds, parent emphasises family rules and customs,parent controls the child by warning about bad things). For Cronbach’s alphas, see Table 1.

In addition, parentechild interactions were observed during two dyadic parentechild problem-solving tasks,a crossword task (duration: M¼ 10 min 51 s, SD¼ 3 min 2 s) and a compound-word task (M¼ 8 min 50 s,SD¼ 1 min 52 s). In the crossword task (Poikkeus, 1996), the parentechild dyad was given a sheet of paper witha crossword puzzle and eight folded cards each containing three clues for doing the puzzle. The parent and childwere instructed to do the puzzle together using as few clues as possible. The dyad was allowed 15 min to performthe task, but the participants were not informed of this time limit. In the compound-word task (Metsapelto, Pulkkinen,& Poikkeus, 2001), which was performed immediately after the crossword task, the participants were given 30 single-word cards each containing a Finnish noun. The child was told to combine the nouns into 15 grammatically correctcompound words, which are common in Finnish. This time, the instructions were given solely to the child, and theparent was asked to provide help only when necessary. The participants were informed that they had 10 min to com-plete the task.

165M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

In both tasks, the entire observation period was coded based on videotapes of the interaction recorded from behindone-way mirrors. The coding scheme consisted of five items assessed on a five-point Likert scale (ranging from low tohigh). The items measuring emotional warmth were enjoyment of interaction (from lack of expressed enjoyment toclear and spontaneous expressions of joy), positive affect (from negative, nonchalant feelings to positive and warmfeelings expressed toward the child), and interest in the child’s activity (from the parent being uninvolved and showinglittle interest in the child to the highest involvement level in which the parent closely monitored the child’s activities).The items measuring guidance were democratic participation (from a high level of parental domination and control tojoint problem solving, during which the initiatives of both partners were taken into account) and assistance (from theparent giving no assistance to the parent providing the child with hints and advice facilitating problem solving; Met-sapelto et al., 2001). A mean-based composite score was calculated for both emotional warmth and parental guidance.

The videotapes were viewed several times in order to assign the parents a specific score on each behaviouraldimension of the two tasks. Parental behaviour was coded by a different observer for each task. Each score in theLikert-type coding scheme was described, and examples of parental behaviours were provided. The observers weretrained extensively prior to the coding process using videotaped interactions. Inter-observer reliability was estimatedby two coders who independently assessed the same randomly selected cases (20% of the data). Pearson’s correlationcoefficients between their coding of the crossword task and the compound-word task varied between 0.76 and 0.90.

2.4. Data editing and analysis

As shown in Table 1, complete data for the 99 children were not obtained for all variables. In fact, 17% of the fa-thers and 12% of the mothers were absent. In most cases, missing values did not exceed 20% of the data. Emotionalwarmth and parental guidance were exceptions, though, because they corresponded to behavioural observations inwhich only one of the parents (i.e., the participant of the longitudinal study) was present; these variables wereobserved for 47 mothers (47.5%) and 52 fathers (52.5%). Therefore, it was necessary to impute the missing values(Graham, Cumsille, & Elek-Fisk, 2002). The information lacking in the variables was missing completely at random,according to Little’s MCAR test, c2(93, N¼ 99)¼ 94.6, p¼ 0.433 (Roderick & Little, 1988). The imputation methodused here required normally distributed data. Skewed variables were thus normalised separately for girls and boysprior to imputation, using the PRELIS program (Joreskog & Sorbom, 1996). A logarithmic transformation was ap-plied for mean thinking time. The data were imputed by an iterative method called the expectation maximisation(EM) algorithm, using SPSS software based on the abundant longitudinal information available on parental monitor-ing and restrictiveness (10 and 3 items, respectively), the children’s opinions of their parents’ parenting styles (8items), the parents’ and children’s views of their home atmosphere (85 and 8 items), and the parents’ self-ratingsof their socio-emotional behaviour (8 items). Multi-group structural equation models were then estimated on the basisof the imputed data (i.e., complete data for all 99 children).

The imputed data were first analysed separately for girls and boys. Polychoric correlation matrices were calculatedby PRELIS so that we could examine the interrelationships between the parenting and help-seeking variables (Jore-skog & Sorbom, 1996). Gender differences in the correlations were tested using Fisher’s transformation of the cor-relation coefficient. The hypothesised relationships between parenting and help seeking were further studied usingstructural equation modelling (SEM) with LISREL 8.5 software (Joreskog, Sorbom, du Toit, & du Toit, 1999).SEM parameters were estimated using the maximum likelihood procedure, with the above-mentioned polychoric cor-relation matrices as input.

Gender-related similarities and differences in the study variable associations were then examined using the multi-group method. For the first stage, we investigated the similarity of the paths’ magnitudes. Parameters were estimatedto be equal when they were significant and had the same sign in the two gender groups. Those parameters whichproved significant in one gender group only were estimated freely without equality constraints. Then, dependingon the modification indexes, the equal-estimation constraints were deleted one at a time. The final models thus in-cluded statistically significant paths only (jtj � 2 for the estimates), estimated equally when possible and separatelyif necessary. The goodness of fit (to the data) of the models was evaluated using three indicators: the c2 test, theroot mean square error of approximation (RMSEA), which measures the approximation error of the model, andthe comparative fit index (CFI), which assesses the adequacy of a specified model in relation to the baseline model.The fit was considered good when the p-value associated with the c2 test was nonsignificant ( p> 0.05). RMSEAvalues under 0.08 were taken to suggest an acceptable fit, and values under 0.05, an excellent fit (Browne & Cudeck,

166 M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

1993). CFI values above 0.90 were taken to reflect an acceptable fit, and values above 0.95 an excellent fit (Hu &Bentler, 1995).

3. Results

3.1. Descriptive statistics

Any gender differences in the means presented in Table 1 were examined using the t-test. Boys’ mothers reportedbeing more restrictive than did girls’ mothers, t(97)¼�2.10, p< 0.05.

3.2. Correlations

Polychoric correlations between parenting and help seeking for girls and boys are given in Table 2. They showedthat the capacity to reuse received help in analogous tasks correlated negatively with ‘‘bad’’ help seeking (i.e., totalnumber of answers, confirmations, and questions unrelated to the task) for both girls and boys. There was a highlysignificant, positive correlation between reuse and mean thinking time for girls only.

The parenting variables did not correlate with the help-seeking variables for boys. For girls, maternal and paternalnurturance correlated negatively with the daughter’s capacity to reuse received help.

Reuse correlated positively with age, in both girls and boys, thus indicating older children’s greater ability to applyreceived help to analogous tasks. On the other hand, the negative correlation between ‘‘bad’’ help seeking and age, inboth girls and boys, indicated the younger children’s greater amount of irrelevant help seeking. Mean thinking timecorrelated positively with age in girls.

There was a highly significant positive correlation between the mothers’ and the fathers’ nurturance, and betweenthe mothers’ and the fathers’ restrictiveness, for both girls and boys; the former was significantly higher for girls thanboys. A highly positive correlation was also observed between maternal emotional warmth and parental guidance, andbetween paternal emotional warmth and parental guidance, for both girls and boys.

3.3. Structural equation models

In the SEMs, our aim was to try to explain children’s help seeking by parenting. One of the strengths of the SEMmethod is that, unlike correlations, it allows one to simultaneously investigate several independent and dependent var-iables in the same model. Models were computed separately for mothers’ and fathers’ parenting variables.

Fig. 1 shows that higher levels of both maternal and paternal nurturance were directly linked to a longer thinkingtime preceding help seeking, as well as to a lesser capacity to reuse received help in girls. Furthermore, Model A inFig. 1 shows that, for girls, higher levels of paternal emotional warmth were directly linked to a lower rate of ‘‘bad’’help seeking (i.e., total number of answers, confirmations, and other questions asked by the children).

In boys, the only significant connection between parenting and help seeking was between the fathers’ emotionalwarmth and their sons’ rate of ‘‘bad’’ help seeking. Higher levels of paternal emotional warmth were directly linkedto a higher rate of ‘‘bad’’ help seeking.

For girls and boys alike, being older was directly linked to an advanced capacity to reuse received help in analogoustasks, which, in turn, was directly related to a lower rate of ‘‘bad’’ help seeking (i.e., questions not aimed at under-standing the solving process), and for girls alone, it was associated to a longer thinking time preceding help seeking.In addition, higher age in both genders was directly related to a lower rate of ‘‘bad’’ help seeking and to greater in-tellectual abilities.

The goodness-of-fit values of the final models were excellent: for the fathers’ model c2(33, N¼ 99)¼ 33.2,p¼ 0.456, RMSEA¼ 0.01, and CFI¼ 0.95, and for the mothers’ model c2(23, N¼ 99)¼ 22.1, p¼ 0.512,RMSEA¼ 0.00, and CFI¼ 0.96. The variation in the girls’ received-help reuse explained by the paternal model(Model A in Fig. 1) was 38% (R2¼ 0.38), and that of ‘‘bad’’ help seeking was 36% (R2¼ 0.36). For boys, the corre-sponding figures were 24% and 33% (R2¼ 0.24 and R2¼ 0.33), respectively. On the other hand, the variation in thegirls’ received-help reuse explained by the maternal model (Model B in Fig. 1) was 42% (R2¼ 0.42), and that of‘‘bad’’ help seeking was 26% (R2¼ 0.26). For boys, the corresponding figures were 23% and 23% (R2¼ 0.23 andR2¼ 0.23), respectively.

Table 2

Polychoric correlations between the study variables

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

Help seeking

1. Mean thinking time e 0.52*** �0.26 0.08 0.21 �0.11 0.09 �0.18 �0.06 �0.20 �0.15 0.33* 0.11

2. Reuse of help 0.19 e �0.48** �0.39** �0.32* 0.00 0.10 �0.15 �0.16 �0.06 �0.23 0.51*** 0.14

3. Total number of answers etcb �0.19 �0.42** e 0.21 0.24 0.13 �0.02 �0.03 �0.20 �0.08 �0.02 �0.51*** �0.29

Parenting (self-report data)

4. Mothers’ nurturance �0.14 �0.01 0.00 e 0.74***a �0.07 0.13 �0.07 �0.07 �0.10 �0.10 0.04 0.22

5. Fathers’ nurturance �0.11 �0.15 0.16 0.45*** e �0.10 �0.02 �0.15 0.11 �0.20 0.11 0.07 0.28

6. Mothers’ restrictiveness �0.04 �0.11 0.06 0.22 0.10 e 0.57*** �0.06 0.00 �0.17 0.16 �0.22 �0.21

7. Fathers’ restrictiveness 0.05 �0.05 0.06 �0.13 0.07 0.66*** e �0.10 0.01 �0.14 0.04 �0.04 �0.24

Parenting (observation data)

8. Mothers’ emotional warmth 0.08 �0.06 �0.05 0.15 0.06 �0.02 �0.03 e �0.18 0.78*** �0.09 �0.03 0.19

9. Fathers’ emotional warmth �0.03 0.14 0.17 �0.07 0.09 �0.15 �0.18 �0.14 e �0.10 0.72*** �0.01 0.03

10. Mothers’ parental guidance 0.19 �0.02 �0.15 0.23 0.11 �0.06 �0.05 0.85*** �0.14 e �0.05 0.05 0.23

11. Fathers’ parental guidance �0.03 0.06 0.17 0.05 0.16 �0.07 �0.19 �0.08 0.77*** �0.08 e �0.08 0.11

Age (months) �0.06 0.46*** �0.28* 0.01 �0.14 �0.12 �0.12 �0.21 0.22 �0.14 0.09 e 0.60***

Intellectual ability 0.19 0.21 0.04 0.19 0.12 �0.03 �0.01 �0.03 0.32* 0.03 0.30* 0.56*** e

Note. N¼ 99 (girls: n¼ 42, boys: n¼ 57). *p< 0.05, **p< 0.01, ***p< 0.001. Girls’ correlations are above the diagonal and boys’ correlations, below the diagonal.a Significant gender difference in the correlations (p< 0.05).b Total number of answers, confirmations, and questions unrelated to the problem.

16

7M

.P

uustinenet

al./

Learning

andInstruction

18(2008)

160e171

Fathers’emotionalwarmth

Fathers’nurturance

Reuse ofreceived

help

”Bad”help

seeking

Meanthinking

time

Intellectualability

Age

.58GB***

Mothers’nurturance

Reuse ofreceived

help

”Bad”help

seeking

Meanthinking

time

Intellectualability

Age

Model A Model B

-.24GB*

.50GB***

-.26G*

.28B*

-.37GB***

-.36G** .66G

***

.43G***

.58GB***

-.22GB*

.49GB***

-.34GB***

-.41G***

.34G*

.65G***

Fig. 1. Multi-group SEM (n¼ 99; 42 girls and 57 boys) for fathers’ (Model A) and mothers’ (Model B) self-reported and observed parenting, and

girls’ (G) and boys’ (B) help seeking. ‘‘Bad’’ help seeking corresponds to the total number of answers, confirmations, and other questions asked by

the children.

168 M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

4. Discussion

The SEMs revealed that parenting was related to children’s help seeking, as reflected by their mean thinking timepreceding help seeking, their capacity to reuse received explanations in analogous tasks, and their questions not re-lated to understanding the solving process. At the same time, however, it is important to add that the girls’ and boys’results were quite different and only partially confirmed our hypotheses.

For girls, the parents’ increased sensitivity to their daughters’ needs, as assessed by maternal and paternal nurtur-ance (self-report data) and paternal emotional warmth (observation data), was related to a longer mean thinking timepreceding help seeking for the former and to fewer questions unrelated to understanding the solving process for thelatter. It appears as if a warm, child-centred parenting style contributes to the development of emotionally stable andself-confident girls who dare to face difficult problem-solving tasks calmly: they do not hesitate to take time to thinkabout the problems by themselves before deciding to ask for help, and when seeking assistance, they do not ask ques-tions that avoid the problem-solving situation (i.e., ready-made answers, irrelevant questions) or that reflect a lack ofself-confidence (i.e., confirmations).

At the same time, however, increased maternal and paternal nurturance was linked in girls to a lesser capacity toreuse received help on analogous tasks. This seemingly contradictory finding might reflect the fact that even thoughparental sensitivity to the child’s needs seems to have beneficial effects on the daughters’ help seeking from the emo-tional point of view, it could be an impediment to the development of the more ‘‘cognitive’’ ability to autonomouslyapply received hints and explanations to similar tasks. However, more research is obviously needed to fully understandthe relationship between parenting (e.g., how parents support the development of their child’s autonomy) and girls’help seeking.

Help seeking by boys, on the other hand, seems to depend on factors other than their mothers’ or fathers’ generalchild-rearing practices (i.e., self-report data). An analysis of the observation data further confirmed that the mothers’parenting did not explain their sons’ help seeking. It showed, however, that the fathers’ increased emotional warmth

169M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

was linked in boys to a higher rate of ‘‘bad’’ help seeking (i.e., answers, confirmations, and other questions), suggest-ing that warm and caring fathers may jeopardise the development of help seeking as an effective learning strategy intheir sons. The paternal parenting observed here led to different help-seeking behaviours in girls and boys: for girls,the fathers’ increased emotional warmth was related to a more positive type of help seeking than it was for boys. Didthe fathers display their emotional warmth differently with their son and daughter, or did the girls and boys experiencetheir fathers’ expressed emotional warmth differently? Further analyses are necessary to answer this question.

It is interesting to note that nurturance and emotional warmth, which correspond to two different approaches to thestudy of parenting, yielded rather consistent results in girls. Both variables measure a warm and caring type of par-enting, but while nurturance was reflected here in the parents’ self-reported child-rearing principles and practices ingeneral, emotional warmth was based on behavioural observations of parenting during actual parentechild interac-tions. Previous studies have shown that self-report and observational measures are at best moderately related. Metsa-pelto and Pulkkinen (2005), for example, reported a lack of association between parents’ self-reported nurturance andtheir observed child-centredness during parentechild interactions. In the present study, the relationships between thefathers’ emotional warmth and their daughters’ help seeking were consistent with those obtained using the corre-sponding self-report variable (i.e., nurturance): in both cases, paternal parenting was related to more serenity andself-confidence in the daughters’ help seeking.

However, one detail calls for attention: nurturance was related solely to the mean thinking time and the reuse ofreceived help, whereas emotional warmth was related only to the total number of ‘‘irrelevant’’ questions. More re-search is clearly needed to fully understand the subtle dynamics that relate parenting to help seeking. As far as par-enting is concerned, it would be profitable to strive for a multi-faceted methodological approach, in an attempt tocapture this construct in all of its complexity. Thus, in addition to parental self-report questionnaires and behaviouralobservations of parentechild interactions, as used in the present study, it might be interesting in future studies to in-clude children’s self-reports about their experiences of parenting. In fact, recent evidence suggests that children arecapable of reporting on their social relationships (Galinsky, 2000).

Some other limitations of the study are related to the longitudinal context of our study. For example, the presentstudy involved a secondary analysis of data drawn from a more extensive ongoing longitudinal study. Consequently,the only parenting measures that were available assessed rather general parental orientations. As pointed out by McCalland Appelbaum (1991), reliance on collected data often limits the focus of a study, but the findings can neverthelessexpand our understanding of the phenomena. In the future, it would be useful to collect more specific parenting data tomeasure, for example, the types of help children seek from their parents and the types of help the parents provide duringinteractions. In addition, our study raises the question of the stability of the parents’ self-reports over time. The lon-gitudinal study participants filled in the child-rearing questionnaire in 1995, but the parentechild observations andthe help-seeking test took place in 1997e1999. Even though previous studies on self-reported parenting have sug-gested reasonable levels of over-time stability, it might be wise to opt for single-wave data collection in future studies.

Finally, more attention should also be paid in the future to the choice of a help-seeking methodology (see Newman,2002). Choosing an experimental design necessarily implies lowering ecological validity. On the other hand, inobservational studies like those conducted by Stright and her colleagues (e.g., Stright et al., 2001), it is difficult togo beyond simple frequency counts, which do not reflect all aspects of the help-seeking process such as the learners’capacity to reuse received help in other tasks or their thinking time preceding the decision to seek help. As severalauthors (e.g., Nelson-Le Gall, 1985; Puustinen et al., 2004) have highlighted, help seeking is ‘‘a process-like activity,in which the actual question-asking behaviour is only one element among several others’’ (Puustinen et al., 2004, p. 241).

Acknowledgements

This research was a part of the ‘‘Human Development and its Risk Factors’’ project (http://www.jyu.fi/humander/)funded by the Academy of Finland (Finnish Centre of Excellence Program No. 40166 for 1997e1999 and No. 44858for 2000e2005).

References

Ames, R. (1983). Help-seeking and achievement orientation: perspectives from attribution theory. In B. DePaulo, A. Nadler, & J. D. Fisher (Eds.),

New directions in helping, Vol. 2 (pp. 165e186). New York: Academic.

170 M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

Aunola, K., & Nurmi, J.-E. (2005). The role of parenting styles in children’s problem behavior. Child Development, 76, 1144e1159.

Boekaerts, M., Pintrich, P. R., & Zeidner, M. (Eds.). (2000). Handbook of self-regulation. San Diego, CA: Academic.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A. Bollen, & J. S. Long (Eds.), Testing structural equationmodels (pp. 136e162). Newbury Park, CA: Sage.

Cairns, R. B., & Green, J. A. (1979). How to assess personality and social patterns: observations or ratings? In R. B. Cairns (Ed.), The analysis of

social interactions: Methods, issues, and illustrations (pp. 209e226). Hillsdale, NJ: Erlbaum.

Connell, J. P. (1990). Context, self, and action: a motivational analysis of self-system processes across the life span. In D. Cicchetti, & M. Beeghly

(Eds.), The self in transition: Infancy to childhood (pp. 61e97). Chicago, IL: University of Chicago Press.

Galinsky, E. (2000). Ask the children. New York: Quill.

Gerris, J. R. M. (2001). Introduction and overview. In J. R. M. Gerris (Ed.), Dynamics of parenting (pp. 9e13). Leuven, Belgium: Garant-

Uitgivers.

Gerris, J. R. M., Vermulst, A. A., van Boxtel, D. A. A. M., Janssens, J. M. A. M., van Zutphen, R. A. H., & Felling, A. J. A. (1993). Parenting in

Dutch families: A representative description of Dutch family life in terms of validated concepts representing characteristics of parents, children,

the family as a system and parental socio-cultural value orientations. Nijmegen, The Netherlands: Institute of Family Studies, University of

Nijmegen.

Graham, J. W., Cumsille, P. E., & Elek-Fisk, E. (2002). Methods for handling missing data. In I. B. Weiner (Series Ed.), J. A. Schinka & W. F.

Velicer (Vol. Eds.), Handbook of psychology. Research methods in psychology, Vol. 2 (pp. 87e114). New York: Wiley.

Hu, L., & Bentler, P. M. (1995). Evaluating model fit. In R. H. Hoyle (Ed.), Structural equation modeling: Concepts, issues and applications(pp. 76e99). Thousand Oaks, CA: Sage.

Joreskog, K. G., & Sorbom, D. (1996). PRELIS 2. User’s reference guide. Chicago, IL: Scientific Software International.

Joreskog, K. G., Sorbom, D., du Toit, S., & du Toit, M. (1999). Lisrel 8: New statistical features. Chicago, IL: Scientific Software International.

Karabenick, S. A., & Knapp, J. R. (1988). Help seeking and the need for academic assistance. Journal of Educational Psychology, 80, 406e408.

Karabenick, S. A., & Newman, R. S. (Eds.). (2006). Help seeking in academic settings. Goals, groups, and contexts. Mahwah, NJ: Erlbaum.

Kochanska, G. (1997). Mutually responsive orientation between mothers and their young children: implications for early socialization. Child

Development, 68, 94e112.

Kochanska, G., Kuczynski, L., & Radke-Yarrow, M. (1989). Correspondence between mothers’ self-reported and observed child-rearing practices.

Child Development, 60, 56e63.

Lytton, H., & Romney, D. M. (1991). Parents’ differential socialization of boys and girls: a meta-analysis. Psychological Bulletin, 109, 267e296.

McCall, R. B., & Appelbaum, M. I. (1991). Some issues of conducting secondary analyses. Developmental Psychology, 27, 911e917.

van der Meij, H. (1988). Constraints on question asking in classrooms. Journal of Educational Psychology, 80, 401e405.

Metsapelto, R.-L., & Pulkkinen, L. (2003). Personality traits and parenting: neuroticism, extraversion, and openness to experience as discrimina-

tive factors. European Journal of Personality, 17, 59e78.

Metsapelto, R.-L., & Pulkkinen, L. (2005). The moderating effect of extraversion on the relation between self-reported and observed parenting.

Journal of Applied Developmental Psychology, 26, 371e384.

Metsapelto, R.-L., Pulkkinen, L., & Poikkeus, A.-M. (2001). A search for parenting style: a cross-situational analysis of parental behavior.

Genetic, Social, and General Psychology Monographs, 127, 169e192.

Neitzel, C., & Stright, A. D. (2003). Mothers’ scaffolding of children’s problem solving: establishing a foundation of academic self-regulatory

competence. Journal of Family Psychology, 17, 147e159.

Nelson-Le Gall, S. (1981). Help-seeking: an understudied problem-solving skill in children. Developmental Review, 1, 224e246.

Nelson-Le Gall, S. (1985). Help-seeking behavior in learning. Review of Research in Education, 12, 55e90.

Newman, R. S. (2000). Social influences on the development of children’s adaptive help seeking: the role of parents, teachers, and peers. Devel-

opmental Review, 20, 350e404.

Newman, R. S. (2002). What do I need to do to succeed. when I don’t understand what I’m doing!? Developmental influences on students’

adaptive help seeking. In A. Wigfield, & J. Eccles (Eds.), Development of achievement motivation (pp. 285e306). San Diego, CA: Academic.

Parke, R. D. (2002). Fathers and families. In M. H. Bornstein (Ed.), Handbook of parenting (2nd ed.), Being and becoming a parent, Vol. 3

(pp. 27e73). Mahwah, NJ: Erlbaum.

Pintrich, P. R. (2000). The role of goal orientation in self-regulated learning. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook ofself-regulation (pp. 451e502). San Diego, CA: Academic.

Poikkeus, A. -M. (1996). Process vs. product orientation in joint problem solving of mothers and their sons with learning disability. Unpublished

manuscript.

Pulkkinen, L. (1982). Self-control and continuity from childhood to late adolescence. In P. B. Baltes, & O. G. Brim Jr. (Eds.), Life-span devel-opment and behavior, Vol. 4 (pp. 63e105). New York: Academic.

Pulkkinen, L. (1998). Levels of longitudinal data differing in complexity and the study of continuity in personality characteristics. In R. B. Cairns,

L. R. Bergman, & J. Kagan (Eds.), Methods and models for studying the individual (pp. 161e183). Beverly Hills, CA: Sage.

Pulkkinen, L. (2006). The Jyvaskyla Longitudinal Study of Personality and Social Development. In L. Pulkkinen, J. Kaprio, & R. J. Rose (Eds.),

Socioemotional development and health from adolescence to adulthood (pp. 29e55). New York: Cambridge University Press.

Puustinen, M. (1998). Help-seeking behavior in a problem-solving situation: development of self-regulation. European Journal of Psychology of

Education, 13, 271e282.

Puustinen, M., Kokkonen, M., Tolvanen, A., & Pulkkinen, L. (2004). Children’s help seeking and impulsivity. Learning and Individual Differ-

ences, 14, 231e246.

Puustinen, M., & Pulkkinen, L. (2001). Models of self-regulated learning: a review. Scandinavian Journal of Educational Research, 45, 269e286.

Roberts, G. C., Block, J. H., & Block, J. (1984). Continuity and change in parents’ child-rearing practices. Child Development, 55, 586e597.

171M. Puustinen et al. / Learning and Instruction 18 (2008) 160e171

Roderick, J. A., & Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the Amer-

ican Statistical Association, 83, 1198e1202.

Sattler, J. M. (1992). Assessment of children (3rd ed.). San Diego, CA: Author.

Schumm, W. R. (1990). Evolution of the family field: measurement principles and techniques. In J. Touliatos, B. F. Perlmutter, & M. A. Straus

(Eds.), Handbook of family measurement techniques (pp. 23e36). Newbury Park, CA: Sage.

Steinberg, L. (2001). We know some things: adolescenteparent relationships in retrospect and prospect. Journal of Research on Adolescence, 11,

1e20.

Stright, A. D., Neitzel, C., Sears, K. G., & Hoke-Sinex, L. (2001). Instruction begins in the home: relations between parental instruction and child-

ren’s self-regulation in the classroom. Journal of Educational Psychology, 93, 456e466.

Vygotsky, L. S. (1981). The genesis of higher mental functions. In J. V. Wertsch (Ed.), The concept of activity in Soviet psychology (pp. 144e188).

Armonk, NY: Sharpe.

Wechsler, D. (1974). Wechsler Intelligence Scale for Children-Revised. New York: The Psychological Corporation.