International Journal of Psychology, 2014 DOI: 10.1002/ijop.12104

The associations between parenting styles and proactive and reactive aggression in Hong Kong children and adolescents Yu Gao1,2 , Wei Zhang2 , and Annis Lai Chu Fung3 1 Department 2 Department

of Psychology, Brooklyn College, Brooklyn, NY, USA of Psychology, Graduate Center of the City University of New York, New York,

NY, USA 3 Department

of Applied Social Sciences, City University of Hong Kong, Hong Kong, Hong

Kong

P

revious research suggests that reactive and proactive aggression may be differentially related to family contextual (e.g. parenting practices) factors. However, the existing research has focused largely on children and adolescents from Western countries, and no study has examined the parenting–aggression association using a parenting style measure sensitive to Asian culture. In this study parenting styles (i.e. warmth, control and guan/training) and proactive and reactive aggression were assessed in a large sample of school children in Hong Kong, China (N = 4,175, mean age = 11.75). We found that: (a) both low warmth (in boys only) and guan (i.e. high expectation and close supervision, in both boys and girls) were associated with elevated parent-reported proactive aggression, (b) high restrictive control (i.e. dominating and rejecting) was associated with high reactive aggression (in both boys and girls) based on parent- or child-report data, and with high proactive aggression (in boys only) based on parent-report data, and (c) guan was also positively associated with parent-reported reactive aggression. Findings provide more information about the Parenting Inventory using a large Asian sample, and extend existing research on familial correlates of different types of aggression. Keywords: Reactive aggression; Proactive aggression; Parenting; Guan; Asian; Youth.

The two types of aggression, reactive and proactive aggression, may be differentially related to many correlates, including family contextual (e.g. parenting practices) factors (Hubbard, McAuliffe, Morrow, & Romano, 2010). However, the existing research has focused largely on children and adolescents from Western countries and the few findings focusing on Asian samples did not use parenting measures sensitive to Asian population (Jia, Wang, & Shi, 2014; Xu, Farver, & Zhang, 2009; Xu & Zhang, 2008). The aim of this study was to extend existing research by investigating familial correlates of different types of aggression in a large sample of school children in Hong Kong, China. Proactive and reactive aggression Child and adolescent aggression has been conceptualised into two subtypes according to differences in function: reactive aggression refers to a response to provocation or a perceived threat, whereas proactive aggression is

described as purposeful behaviour aimed at gaining a reward or social dominance over others. Reactive aggression is characterised as involving high emotional arousal, impulsive, hostile and an inability to control affect (Crick & Dodge, 1996), whereas proactive aggression is theorised as a more regulated, predatory, instrumental form of aggression, with more positive expectances about the outcome of aggression (Crick & Dodge, 1996). Behavioural genetic studies have shown that genetic influences contribute more to proactive than reactive aggression (Baker, Raine, Liu, & Jacobson, 2008). Furthermore, one twin study has shown that only 12% of the environmental factors were the same for the two subtypes of aggression (Brendgen, Vitaro, Boivin, Dioone, & Perusse, 2006), suggesting that reactive and proactive aggression may result from largely separate environmental influences. Reactive and proactive aggression may be associated with different familial precursors. One study showed that parental substance abuse at age 7 predicted proactive but not reactive aggression at age 16 (Raine et al., 2006).

Correspondence should be addressed to Yu Gao, Department of Psychology, Brooklyn College, 5401 James Hall, 2900 Bedford Avenue, Brooklyn, NY 11210, USA. (E-mail: [email protected]).

© 2014 International Union of Psychological Science

2

GAO ET AL.

Dodge, Lochman, Harnish, Bates, and Pettit (1997) found that parental physical abuse was associated with reactive aggression but not proactive aggression. One study among a group of American children found that perceived positive parental affect was negatively related to proactive aggression but not to reactive aggression (Yeh, Chen, Raine, Baker, & Jacobson, 2011). In contrast, several studies have found that some familial factors, such as harsh and coercive parenting, are associated with both proactive and reactive aggression (e.g. Jia et al., 2014; Xu et al., 2009). To our knowledge, only two prior studies have examined the familial factors in relation to these two subtypes of aggression in Chinese children. Xu et al. (2009) found that harsh parenting was positively related to both proactive and reactive aggression in 401 children in Mainland, China. In a group of kindergarten children from China, Jia et al. (2014) found that hostile/coercive parenting was positively related to teacher-reported reactive and proactive aggression. However, the few findings focusing on Asian samples did not use parenting measures more sensitive to Asian population. Parenting styles in Asian population Numerous studies have been conducted to examine the effect of parenting on various outcomes in children, and many of these studies have adopted Baumrind’s typologies of parenting. Baumrind categorised parents into authoritatian, authoritative and permissive types. According to Baumrind, authoritative parenting style, characterised by high parental warmth and moderate control, is optimal for Caucasian children (Baumrind, 1972). Studies have generally supported this notion and shown the relationships between authoritative parenting and many positive outcomes including strong academic achievement, psychological adjustment and competence (Steinberg, Dornbusch, & Brown, 1992). In contrast, most Asian and Asian American families show authoritatian parenting style, which is characterised by high parental control and relatively lower levels of parental warmth (Lin & Fu, 1990; Steinberg et al., 1992). Although Asian American children often show superior academic achievement, fewer symptoms of poor adjustment and report less stress than do young Caucasian people (Crystal, Chen, Fuligni, & Stevenson, 1994), these authoritarian parenting styles have often been shown to be related to negative outcomes in Caucasian teenagers (Steinberg et al., 1992). These contradictory findings suggest that the traditional measures of parenting styles including strict control and warmth do not adequately capture important aspects of parenting in the families of Asian children and adolescents. To explain this paradox, researchers proposed the construct of “training” or guan (Chao, 1994) as a unique parenting style in Asian, and this construct was further

explored in a series of cross-cultural studies with samples of United States, Hong Kong and Pakistan (Stewart, Bond, Kennard, Ho, & Zaman, 2002; Stewart et al., 1998; Stewart et al., 1999). Stewart et al. suggested the widely recognised dimensions of warmth and control as the two basic dimensions, and introduced guan as the third dimension. In particular, guan refers to caring and monitoring and the teaching and training of children to engage in appropriate behaviours, with an emphasis on self-discipline and hard working. It was distinguished from dominating control, which involves not only a high level of demand on children, but also hostile, rejecting and somewhat uninvolved parental behaviours. Finally, the construct of warmth was defined to reflect parental understanding and loving of children. Researchers believe that these three style constructs can be considered as three major culturally sensitive constructs to be included in investigations on Chinese parenting (Stewart et al., 1998, 1999, 2002). The Parenting Inventory, one of the assessments to measure child-rearing attitudes in parents, was thus developed to assess these three dimensions of parenting styles perceived by Chinese participants (Stewart et al., 1998). Originally 22 items were included: 6 items were designed to assess parental warmth/acceptance, and another 8 were chosen to assess “restriction” or high demand without democratic exchange or negotiation. For these items, participants were asked to indicate how they perceived their parents at the present point in time. Finally, another eight items were developed from Chao’s study (1994), and included the behavioural measures of Chinese parenting that reflect guan or “training” (positive involvement and close supervision). For these eight items, participants were asked to describe parents’ attitudes and behaviours when they were very young (between ages 4 and 8 years). Each of the 22 items consisted of a statement to which participants were asked to indicate their degree of agreement on a scale ranging from 1 (strongly disagree) to 6 (strongly agree). However, one of the major issues in this line of research concerns the factor structure of the Parenting Inventory. Using small to moderate sample sizes, prior factor analyses suggest that there is a meaningful difference between parental control and warmth. However, it was unclear how many and which items should be included in each factor. For example, Stewart and Bond (2002) conducted a factor analysis in a group of 171 Hong Kong adolescents and retained five and three items for warmth and control subscale, respectively. In a group of 212, 14 to 15-year-old boys and girls from Bangladesh, Stewart, Bond, Abdullah, and Ma (2000) retained four and three items for warmth and control factor, respectively. More recently, Chan, Chan, and Chan (2013) selected 14 items from Stewart et al. (1998, 1999, 2002) and through a pilot testing with a small group of Chinese adolescents in Hong Kong retained 4 items for each of the © 2014 International Union of Psychological Science

PARENTING AND AGGRESSION

three factors. However, it was unclear which items were retained in that study. Therefore, one goal of this study was to test the factor structure of the Parenting Inventory in a large sample of Hong Kong children and adolescents.

research suggests that the aetiology of aggression may differ for males and females (Moffitt, Caspi, Rutter, & Silva, 2001), and that parenting practices may be different for the two sex groups (Stewart et al., 2000), we examined the above associations in boys and girls separately.

Parenting styles, reactive and proactive aggression As noted earlier, studies have shown that aversive parenting is associated with aggressive child behaviour in Chinese. It is consistent with the assumption that hostile/coercive parenting likely evokes feelings of hostility and opposition towards parents and therefore interferes with children’s proper internalisation of control (Hoffman, 1960), which then leads to engagement of aggressive behaviour. Social learning theory presumes that children learn and adopt aggressive strategies through watching and interacting with aggressive models (e.g. Crick & Dodge, 1996). In this context, parents serve as models from which children expect to learn less effective strategies, including using aggression, for social interaction. Consistently, studies have found that controlling parenting, psychologically or behaviourally, is associated with externalising and aggressive behaviour in Chinese culture (e.g. Nelson, Hart, Yang, Olsen, & Jin, 2006). It is therefore predicted that high parental control would be associated with elevated levels of both reactive and proactive aggression. In contrast, the concepts of guan have been argued to be more closely aligned with conceptions of authoritative parenting (Nelson et al., 2006) that combines high demands, firm control and warmth, and is perceived by children in the context of a caring, concerned relationship (Chao, 1994). Given that perceived positive parental affect has been found to be negatively related to proactive aggression but not reactive aggression (Yeh et al., 2011), and guan and warmth are both perceived as loving and positive (Chao, 1994), we predicted that guan and warmth would be negatively associated with proactive aggression in particular. The current study The aims of the current study were twofold: (a) to provide more information about the Parenting Inventory among Hong Kong parents using a large sample and (b) to examine the relative contribution of different parenting styles to children’s reactive and proactive aggression. Specifically, we hypothesised that parental control, reflecting restrictive, harsh and domineering type of parenting, would be positively associated with reactive and proactive aggression, and that guan and warmth, reflecting loving/acceptance, and high involvement and expectation, would be negatively associated with proactive aggression but not reactive aggression. Finally, as © 2014 International Union of Psychological Science

3

METHOD Participants Participants include 4,179 primary and secondary school children (54% males, mean age = 11.75, SD = 3.01) and their parents from Hong Kong, China; 68 secondary and 88 primary schools initially expressed an interest in participating, and 11 secondary and 12 primary schools were selected for participation based on their representativeness to the general population. Ethical approval was provided by the Ethical Committee Board of the City University of Hong Kong. Informed written consent was obtained from parents. Approval was also obtained from all principals, vice-principals and school administrators of all participating schools. Measures Reactive–Proactive Aggression Questionnaire (RPQ) This instrument has been used with groups ranging from 9 to 16 years and yields measures of reactive aggression, proactive aggression and total aggression (Raine et al., 2006). This 23-item scale contains 11 reactive items (e.g. “reacted angrily when provoked by others”) and 12 proactive items (e.g. “hurts others to win a game”). Scores were summated to assess reactive and proactive aggression together with an overall score of total aggression (the simple summation of reactive and proactive aggression). Internal consistency (Cronbach’s α) in adolescents has previously been reported as .86 for proactive aggression, .84 for reactive aggression and .90 for total aggression (Raine et al., 2006). Evidence for the reliability and validity of the self-report version in a Hong Kong population of schoolchildren can be found elsewhere (Fung, Raine, & Gao, 2009). In this study, RPQ was completed by both the caregiver (1,129 fathers and 3,050 mothers) and the child. Cronbach’s α was .86 (total aggression), .83 (proactive aggression) and .80 (reactive aggression) for self-report, and .88 (total aggression), .84 (proactive aggression) and .84 (reactive aggression) for parent-report. Means, standard deviation and scale ranges for each measure are listed in Table 1. Parenting Inventory As mentioned earlier, this questionnaire assesses three dimensions of parenting style (warmth, control and

4

GAO ET AL. TABLE 1 Means (standard deviations) and ranges of main study variables for overall sample, and for boys and girls, separately

Warmth Control Guan Parent-reported reactive aggression Parent-reported proactive aggression Child-reported reactive aggression Child-reported proactive aggression

Overall sample (N = 4,179)

Boys (N = 2,249)

Girls (N = 1,930)

4.56 (2.95) [4, 24] 3.17 (4.53) [6, 36] 4.77 (3.85) [6, 36] 5.79 (3.53) [0, 22] 0.94 (2.06) [0, 24] 4.87 (3.28) [0, 22] 0.94 (2.01) [0, 24]

18.22 (2.92) [4, 24] 19.17 (4.49) [6, 26] 28.51 (3.79) [6, 36] 6.16 (3.70) [0, 22] 1.13 (2.26) [0, 19] 5.15 (3.39) [0, 22] 1.22 (2.24) [0, 19]

18.29 (2.98) [4, 24] 18.87 (4.58) [6, 36] 28.69 (3.91) [6, 36] 5.36 (3.26) [0, 22] 0.71 (1.77) [0, 24] 4.54 (3.11) [0, 21] 0.61 (1.65) [0, 24]

t statistics −0.86 2.14* −1.52 7.42*** 6.74*** 6.09*** 10.00***

*p < .05, ***p < .001.

guan/training), and in this study parents were asked to indicate their degree of agreement on a 6-point Likert scale (Stewart et al., 1998). Because the original items were written for child or adolescent respondents to study perceived parenting styles, the wordings were changed to suit parent respondents by changing statements starting with “My parent … ” to first-person statements, for example “I am restrictive/controlling of my child” (Chan et al., 2013). For warmth and control, parents were asked to indicate how they perceived their parenting at present. For guan, parents were asked to describe their parenting when the child was young (around 4 to 8 years old). Statistical analyses RPQ—confirmative factor analysis (CFA) All model fitting and analyses were conducted using EQS 6 (Bentler, 2006) and SPSS 21.0. Following the prior research (Fung et al., 2009), the 23 items (12 proactive and 11 reactive) were subjected to CFA with the covariance matrix. The heterogeneous kurtosis estimation method was used to estimate the distribution of covariances in the evaluation of all models (Bentler, 2006) due to significant kurtosis for many of the items (Mardia’s normalised multivariate kurtosis was greater than 3). Three models were fitted. The first model was a null model, in which each of the items is assumed to represent completely independent and uncorrelated dimensions of aggression. Then two models were compared to the null model and then between themselves based on prior research: a one-factor model (general aggression) and a two-factor model (proactive and reactive aggression). In the one-factor model, all 23 items were loaded onto one latent factor (general aggression), and the factor variance was constrained at 1.00. In the

two-factor model, we loaded the 12 proactive aggression items onto one latent factor (proactive), and the 11 reactive aggression items onto another latent factor (reactive). The correlation between the two factors was set to be free. We set free fixed paths from each latent factor to their corresponding variables to be estimated, and error terms to be not intercorrelated in all the models. Five commonly used goodness-of-fit indexes to assess the fit of the models were reported: χ2 , the normed fit index (NFI), the nonnormed fit index (NNFI), the comparative fit index (CFI) and the root mean squared error of approximation (RMSEA) index. Certain values (e.g. CFI > .95 and RMSEA < .06) are indicative of a good-fitting model (Bentler, 2006). We made direct comparisons between the one- and two-factor models that are in a hierarchical relationship (i.e. nested) using the difference χ2 (Δχ2 ) test, with the one-factor model nested within the two-factor model. Separate model fittings were done for parent- and self-report RPQ across the two groups of genders. Parenting—exploratory factor analysis (EFA) The 14 items reflecting warmth and control were subjected to a maximum likelihood factor analysis with varimax rotation and a two-factor solution. The eight items reflecting guan were subjected to the similar analysis and a one-factor solution. Criteria for inclusion of items were factor loadings of .35 or higher on the rotated factor matrix (Spector PE). Separate models were fitted for the overall sample, and for boys and girls, respectively. Finally, the associations between parenting styles and aggression measures were examined using partial correlation. Each construct of parenting style was associated with the residualised aggression measure (Raine et al., 2006) after controlling for the other two parenting style constructs. © 2014 International Union of Psychological Science

PARENTING AND AGGRESSION

5

TABLE 2 Model-fitting results comparing the null model with the one-factor (general aggression) model and the two-factor (proactive–reactive aggression) model Parent-report RPQ Model Overall sample Null model One-factor model Two-factor model Boys Null model One-factor model Two-factor model Girls Null model One-factor model Two-factor model

Self-report RPQ

χ2

df

NFI

NNFI

CFI

RMSEA

42338.349 8388.801 2152.124

253 229 227

0.000 0.718 0.949

0.000 0.695 0.949

0.000 0.724 0.954

0.167 0.092 0.045

27085.234 4490.465 1307.278

253 229 227

0.000 0.730 0.952

0.000 0.713 0.955

0.000 0.740 0.960

14835.125 4658.685 931.768

253 229 227

0.000 0.653 0.937

0.000 0.628 0.946

0.000 0.663 0.952

df

NFI

NNFI

CFI

RMSEA

Null model 29008.298 One-factor model 4516.697 Two-factor model 1640.925

253 229 227

0.000 0.806 0.943

0.000 0.794 0.945

0.000 0.813 0.951

0.147 0.067 0.039

0.170 0.091 0.046

Null model 17740.226 One-factor model 2354.995 Two-factor model 1029.699

253 229 227

0.000 0.810 0.942

0.000 0.806 0.949

0.000 0.825 0.954

0.146 0.064 0.040

0.164 0.100 0.040

Null model 11310.871 One-factor model 3067.236 Two-factor model 713.529

253 229 227

0.000 0.735 0.937

0.000 0.723 0.951

0.000 0.749 0.956

0.152 0.080 0.033

RESULTS Reactive–Proactive Aggression Questionnaire (RPQ) Goodness-of-fit indexes for the three models are presented in Table 2. Based on RMSEA, the two-factor and one-factor models both fit better than the null models. In addition, the χ2 difference test indicated a highly significant improvement in fit for the two-factor (proactive–reactive) over the one-factor model for parent-report version (overall sample: Δχ2 (2) = 6236.677, p < .001; boys: Δχ2 (2) = 3183.187, p < .001; girls: Δχ2 (2) = 3726.917, p < .001) and self-report version (overall sample: Δχ2 (2) = 2875.772, p < .001; boys: Δχ2 (2) = 1325.296, p < .001; girls: Δχ2 (2) = 2353.707, p < .001). Correlations between the two latent factors were moderate to high for both versions: for self-report version, r = .597 in overall sample, r = .652 in boys and r = .550 in girls; for parent-report version, r = .609 in overall sample, r = .638 in boys and r = .573 in girls. Dimensions of parenting EFA revealed that the items that emerged in each factor were same for boys and girls. The items in the scales, and their factor loadings with the factor, are presented in Table 3. One item from the Control subscale (“I preach to my child”) and one from the Warmth subscale (“My child finds it hard to please me”) were eliminated due to cross loading. These scales were further purified by eliminating all items with a factor loading less than .35 (i.e. “I do not seem to notice when my child is unhappy,” “I like my child the way he/she is and do not try to change him/ her into someone else”). This procedure resulted © 2014 International Union of Psychological Science

Model

χ2

in a 4-item Warmth subscale (α = .67–.72) and a 6-item Control subscale (α = .70–.76). Guan items were assessed for coherence by computing α coefficients for the eight items. Two items with an item-scale correlation less than .35 were eliminated (i.e. “I used physical punishment when my child misbehaved,” “My child was allowed to sleep in my bed”). This procedure resulted in a final version of 6-item (α = .79–.80). In the following analyses using the guan construct, the mean score of the six items was used. Compared to girls, boys had higher scores on all aggression measures and parental control (see Table 1). Warmth was positively correlated with guan (r = .46–.53), and negatively correlated with control (r = −.07 to −.08). No significant correlation was observed between guan and control (r < −.03, ns). Parenting styles and aggression Parent-reported aggression After controlling for warmth and guan, control was positively related to reactive aggression in both sex groups (r > .05, p < .05) and proactive aggression in boys only (r = .05, p < .05) (see Table 4). After controlling for guan and control, warmth was negatively associated with proactive aggression, in particular in boys (r = −.10, p < .001). Finally, after controlling for warmth and control, guan was positively related to reactive aggression (r > .05, p < .05) but negatively related to proactive aggression (r < −.09, p < .001) in both sex groups. Self-reported aggression Parental control was positively related to reactive aggression in both sex groups (r > .05, p < .05). There was

6

GAO ET AL. TABLE 3 Factor loadings of items in control (factor 1) and warmth (factor 2) factors, and item-scale correlations for guan items Overall sample

Warmth items 1. I say nice things about my child to my friends 2. I enjoy spending time with my child 3. I let my child know through words or actions that I love him/her 4. I understand my child 5. My child finds it hard to please mea 6. I do not seem to notice when my child is unhappya Control items 1. My child is afraid of me 2. I am strict 3. My child worries that I will stop loving him/her if he/she does not live up to my expectations 4. I become angry at my child 5. I am critical of my child 6. I am restrictive/controlling of my child 7. I like my child the way he/she is and do not try to change him/ her into someone elsea 8. I preach to my childa Eigenvalues % of variance Cumulative variance Guan items 1. I emphasised self-discipline 2. I helped my child with his/her studies as much as my education allowed 3. I emphasised neatness and organisation 4. My main concern was my child’s needs 5. I emphasised the importance of hard work 6. I pointed out good behaviours in others as a model for my child 7. I used physical punishment when my child misbehaveda 8. My child was allowed to sleep in my beda a Items

Boys

Girls

Factor 1 (Control)

Factor 2 (Warmth)

Factor 1 (Control)

Factor 2 (Warmth)

Factor 1 (Control)

Factor 2 (Warmth)

0.029 −0.054 −0.033 −0.179 0.569 0.286

0.378 0.791 0.863 0.388 −0.297 −0.188

0.028 −0.040 −0.048 −0.182 0.577 0.272

0.396 0.798 0.864 0.354 −0.292 −0.132

0.033 −0.071 −0.015 −0.173 0.561 0.301

0.358 0.783 0.864 0.427 −0.302 −0.255

0.641 0.646 0.566

−0.125 −0.027 −0.194

0.653 0.647 0.562

−0.127 −0.043 −0.159

0.628 0.648 0.569

−0.123 −0.009 −0.232

0.557 0.474 0.448 −0.120

0.029 0.074 0.134 0.245

0.554 0.472 0.435 −0.129

0.062 0.079 0.146 0.265

0.557 0.476 0.460 −0.105

−0.011 0.065 0.119 0.222

0.150 2.366 16.897 16.897

0.590 2.274 16.244 33.142

0.166 2.368 16.911 16.911

0.562 2.226 15.904 32.814

0.143 2.359 16.852 16.852

0.623 2.353 16.808 33.660

.417 .481

.405 .486

.430 .476

.618 .529 .582 .441 .020 .303

.600 .517 .564 .442 .008 .302

.638 .542 .602 .439 .037 .305

were not included in the final version. TABLE 4 Partial correlation coefficients between parenting styles and aggression measures Overall sample Warmth

Parent-reported reactive aggression Parent-reported proactive aggression Child-reported reactive aggression Child-reported proactive aggression a p < .10,

−.02 −.08*** −.01 −.03

Control .19*** .03a .05** .01

Boys Guan

Warmth

.08*** −.10*** −.04** −.01

−.01 −.10*** .02 .04a

Control .21*** .05* .05* .02

Girls Guan .10*** −.09*** −.04a −.02

Warmth −.02 −.04a −.03 −.01

Control

Guan

.16*** −.01 .05* −.01

.05* −.11*** −.04a .01

*p < .05, **p < .01, ***p < .001.

a trend that warmth was positively correlated with proactive aggression in boys only (r = .04, p = .057), and guan was negatively associated with reactive aggression in both groups (r = .04, p < .07). DISCUSSION In this study, we aimed: (a) to provide more information regarding the constructs of Parenting Inventory in a large sample of Hong Kong parents, and (b) to examine the

relative contribution of different parenting styles, in particular guan, the more sensitive measure of Chinese parenting practice, to children’s reactive and proactive aggression. The main findings are: (a) both warmth (in boys only) and guan (in both sexes) were negatively associated with parent-reported proactive aggression, (b) high parental control was associated with high reactive aggression (in both sexes) based on parent- or child-report data, and with high proactive aggression (in boys only) based on parent-report data and (c) guan was also positively © 2014 International Union of Psychological Science

PARENTING AND AGGRESSION

associated with parent-reported reactive aggression in both sexes. Findings provide more information about the Parenting Inventory using a large Asian sample, and extend existing research on familial correlates of different types of aggression. One aim was to provide more empirical information about the constructs of the Parenting Inventory: Guan or training, a construct that has been proposed as being a significant influence on Chinese parenting behaviours and more relevant than the construct of authoritarian parenting in affecting the adaptation of Chinese young people, and warmth and control, two established parenting dimensions. Using a large sample size, we were able to retain in the Parenting Inventory 4, 6 and 6 items for the warmth, control and guan factor, respectively. Consistent with prior studies (e.g. Stewart et al., 1998), the guan items relating to physical punishment and sleeping in the parents’ bed were deleted due to their low item-scale correlation, and the six items that are derived from the construct of guan show coherence as reflected in acceptable α coefficient levels (.79–.80). These remaining six guan items seem quite consistent with the authoritative construct, known to be beneficial to Western young people (Baumrind, 1972). According to Stewart et al. (1998), control items were chosen to assess “restriction” or high demand without democratic exchange or negotiation, whereas guan items focus on high involvement and close supervision. In order to examine Chao’s hypothesis that the typical scales designed to measure the authoritarian construct contain both dominating control and guan items (1994), those control items that might overlap with guan component (involving positive involvement and close supervision) were specifically avoided (Stewart et al., 1998). Therefore, it is not surprising that in this study guan correlates highly with warmth but not with control (see Table 3). Partly consistent with our hypotheses, both warmth and guan were negatively associated with proactive aggression based on parent-report data. More importantly, we found that guan adds unique variance to predicting aggression, that is, after controlling for warmth and control, guan was still significantly correlated with parent-reported aggression measures (see Table 4). In addition, guan but not warmth was found to relate to reactive aggression in both sexes. Taken together, these findings suggest that although guan practices and parental warmth partly overlap, they may reflect different components of positive parenting. As a result, (aggression) research using parenting measures that do not include the dimension of guan may underestimate the influences of positive parenting on behavioural problems in Chinese families. However, cautions should be taken when interpreting the guan—reactive aggression association, because although the scores of guan were positively associated with parent-reported reactive aggression, this association © 2014 International Union of Psychological Science

7

was not as strong based on child-reported aggression data. Similar patterns were found for the warmth-proactive aggression association in males. That is, parental warmth was more strongly associated with parent-reported than with child-reported proactive aggression. In fact, there is a trend that these associations were in the opposite direction (see Table 4). Taken together, these discrepancies call for more comprehensive research adopting multiple informants’ approaches on the concept of guan and its relationship with the two subtypes of aggression. As expected, boys had higher scores on all aggression measures. Boys and girls’ parents reported very similar levels of warmth and guan, although boys’ parents rated themselves as more dominating in their control. This is consistent with the context of social norms regarding the need for the restrictive control of boys, with much higher expectation and more demand of boys than girls. Furthermore, high level of control and lack of warmth were both associated with high parent-reported proactive aggression in boys but not in girls, providing further support for the theory that the aetiology of aggression may differ for males and females (Moffitt et al., 2001; Nelson et al., 2006). There are several limitations of the study. First, this was a correlational study, and specifying the temporal ordering of variables is not possible. It is possible that overly restrictive and dominating type of parenting may lead the child to be more aggressive. It is also possible that children’s aggressive behaviour leads their parents to use more restrictive approaches to discipline. Second, aggression was assessed using both parent and self-reports to avoid relying on any single informant. However, such ratings are still susceptible to biases in reporting and would have been enhanced with the inclusion of observational measures. Third, in this study information about parenting was reported by the parent only, and results of parenting–aggression associations might be different for children as informants, because family members may experience their interactions differently and therefore have dissimilar views on parenting and parent–child relations (e.g. Lanz, Scabini, Vermulst, & Gerris, 2001). Finally, it is worth noting that the guan measure was retrospective, which may be different from current practices. Therefore, it was assumed that the parental style influences related to guan might be more evident in early years (Stewart et al., 1998). Nevertheless, the guan scale still contributes over and above parental warmth and control to the aggression measures in this study, indicating that guan is an independent dimension with predictive power beyond that provided by the “traditional” dimensions of warmth and control, at least in relation to aggressive behaviour. Future studies with multiple informants on parenting are needed to fully understand the relationships between parenting and aggression.

8

GAO ET AL.

In this study, the associations between guan and aggression were examined in a large sample from Hong Kong, China. It has yet to be determined if any of the associations can be generalised to Western countries and if the concept of guan can benefit the parenting and education practices in Westerners. Stewart et al. (2002) examined the concept of guan in participants from Hong Kong, Pakistan and United States, and found that guan items had adequate internal consistency in all samples. Guan was associated with positive outcomes of perceived health, relationship harmony and life satisfaction in Asian participants. More importantly, although these associations were not significant in US sample, they were not significantly different from those found in Asian samples. The authors concluded that the items of guan may “ … partly overlap with relevant and functional parental practices, but do not adequately capture the essential set of behaviours in the West” (p. 81). Future studies examining these associations in a large sample of Westerners should be conducted to answer these questions. In conclusion, using a large sample of Hong Kong children and adolescents, we were able to determine the appropriate items for the warmth, control and guan factor of the Parenting Inventory. More interestingly, we found that these constructs were differently associated with two subtypes of aggression, and these associations varied by informant and gender. Specifically, we found that in both males and females, reactive aggression was positively associated with parent- and child-reported control and parent-reported guan, and proactive aggression was positively associated with lack of guan (in both sexes), low warmth (in males only) and high level of control (in males only) based on parents’ report. Findings suggest that guan scale is an independent dimension of positive parenting when predicting aggressive behaviour in Chinese population, and that the two subtypes of aggression have different familial correlates in males and females.

ACKNOWLEDGEMENT This study was supported by Quality Education Fund of the HKSAR Education Bureau to Annis Lai Chu Fung. Manuscript received May 2014 Revised manuscript accepted September 2014

REFERENCES Baker, L. A., Raine, A., Liu, J., & Jacobson, K. C. (2008). Differential genetic and environmental influences on reactive and proactive aggression in children. Journal of Abnormal Child Psychology, 36(8), 1265–1278. doi:10.1007/s10802-008-9249-1. Baumrind, D. (1972). Socialization and instrumental competence in young children. In W. W. Hartup (Ed.), The young

child: Reviews of research (Vol. 2, pp. 202–224). Washington, DC: National Association for the Education of Young Children. Bentler, P. M. (2006). EQS 6 structural equations program manual. Encino, CA: Multivariate Software. Brendgen, M., Vitaro, F., Boivin, M., Dioone, G., & Perusse, D. (2006). Examining genetic and environmental effects on reactive versus proactive aggression. Developmental Psychology, 42(6), 1299–1312. doi:10.1037/0012-1649. 42.6.1299. Chan, D. W., Chan, L., & Chan, A. C. (2013). Parenting gifted children among Hong Kong Chinese parents: What differences does westernization make? Roeper Review, 35(3), 177–186. doi:10.1080/02783193.2013.794891. Chao, R. K. (1994). Beyond parental control and authoritarian parenting style: Understanding Chinese parenting through the cultural notion of training. Child Development, 65(4), 1111–1119. doi:10.1111/j.1467-8624.1994.tb00806.x. Crick, N. R., & Dodge, K. A. (1996). Social informationprocessing mechanisms in reactive and proactive aggression. Child Development, 67(3), 993–1002. doi:10.1111/ j.1467-8624.1996.tb01778.x. Crystal, D. S., Chen, C., Fuligni, A. J., & Stevenson, H. W. (1994). Psychological maladjustment and academic achievement: A cross-cultural study of Japanese, Chinese and American high school students. Child Development, 65(3), 738–753. doi:10.1111/j.1467-8624.1994.tb00780.x. Dodge, K. A., Lochman, J. E., Harnish, J. D., Bates, J. E., & Pettit, G. S. (1997). Reactive and proactive aggression in school children and psychiatrically impaired chronically assaultive youth. Journal of Abnormal Psychology, 106(1), 37–51. doi:10.1037/0021-843X.106.1.37. Fung, L. A., Raine, A., & Gao, Y. (2009). Cross-cultural generalizability of the Reactive-Proactive Aggression Questionnaire (RPQ). Journal of Personality Assessment, 91(5), 473–479. doi:10.1080/00223890903088420. Hoffman, M. L. (1960). Power assertion by the parent and its impact on the child. Child Development, 31(1), 129–143 Retrieved from http://www.jstor.org/stable/1126389. Hubbard, J. A., McAuliffe, M. D., Morrow, M. T., & Romano, L. J. (2010). Reactive and proactive aggression in children and adolescence: Precursors, outcomes, processes, experiences, and measurement. Journal of Personality, 78(1), 95–118. doi:10.1111/j.1467-6494.2009.00610.x. Jia, S., Wang, L., & Shi, Y. (2014). Relationship between parenting and proactive versus reactive aggression among Chinese preschool children. Archives of Psychiatric Nursing, 28(2), 152–157. doi:10.1016/j.apnu.2013.12.001. Lanz, M., Scabini, E., Vermulst, A. A., & Gerris, J. R. M. (2001). Congruence on child rearing in families with early adolescent and middle adolescent children. International Journal of Behavioral Development, 25(2), 133–139. doi:10.1080/01650250042000104. Lin, C. Y. C., & Fu, V. R. (1990). A comparison of child rearing practices among Chinese, immigrant Chinese, and Caucasian-American parents. Child Development, 61(2), 429–433. doi:10.1111/j.1467-8624.1990.tb02789.x. Moffitt, T. E., Caspi, A., Rutter, M., & Silva, P. (2001). Sex differences in antisocial behaviour: Conduct disorder, delinquency and violence in the Dunedin Longitudinal Study. Cambridge, U.K.: Cambridge University Press. © 2014 International Union of Psychological Science

PARENTING AND AGGRESSION

Nelson, D. A., Hart, C. H., Yang, C., Olsen, J. A., & Jin, S. (2006). Aversive parenting in China: Associations with child physical and relational aggression. Child Development, 77(3), 554–572. doi:10.1111/j.1467-8624.2006.00890.x. Raine, A., Dodge, K., Loeber, R., Gatzke-Kopp, L., Lynam, D. R., Reynolds, C., ... Liu, J. (2006). The Reactive-Proactive Aggression Questionnaire: Differential correlates of reactive and proactive aggression in adolescent boys. Aggressive Behavior, 32(2), 159–171. doi:10.1002/ab.20115. Steinberg, L., Dornbusch, S. M., & Brown, B. B. (1992). Ethnic differences in adolescent achievement: An ecological perspective. American Psychologist, 47(6), 723–729. doi:10.1037/0003-066X.47.6.723. Stewart, S. M., & Bond, M. H. (2002). A critical look at parenting research from the mainstream: Problems uncovered while adapting Western research to non-Western cultures. British Journal of Developmental Psychology, 20(3), 379–392. doi:10.1348/026151002320620389. Stewart, S. M., Bond, M. H., Abdullah, A. S. M., & Ma, S. S. L. (2000). Gender, parenting, and adolescent functioning in Bangladesh. Merrill-Palmer Quarterly, 46(3), 540–564 Retrieved from http://www.jstor.org/stable/23093744. Stewart, S. M., Bond, M. H., Kennard, B. D., Ho, L. M., & Zaman, R. M. (2002). Does the Chinese construct of guan export to the West? International Journal of Psychology, 37(2), 74–82. doi:10.1080/00207590143000162.

© 2014 International Union of Psychological Science

9

Stewart, S. M., Bond, M. H., Zaman, R. M., McBride-Chang, C., Rao, N., Ho, L. M., & Fielding, R. (1999). Functional parenting in Pakistan. International Journal of Behavioral Development, 23(3), 747–770. doi:10.1080/0165025 99383784. Stewart, S. M., Rao, N., Bond, M. H., McBride-Chang, C., Fielding, R., & Kennard, B. D. (1998). Chinese dimensions of parenting: Broadening Western predictors and outcomes. International Journal of Psychology, 33(5), 345–358. doi:10.1080/002075998400231. Xu, Y., Farver, J. A. M., & Zhang, Z. (2009). Temperament, harsh and indulgent parenting, and Chinese children’s proactive and reactive aggression. Child Development, 80(1), 244–258. doi:10.1111/j.1467-8624.2008.01257.x. Xu, Y., & Zhang, Z. (2008). Distinguishing proactive and reactive aggression in Chinese children. Journal of Abnormal Child Psychology, 36(4), 539–552. doi:10.1007/s10802007-9198-0. Yeh, M. T., Chen, P., Raine, A., Baker, L. A., & Jacobson, K. C. (2011). Child psychopathic traits moderate relationships between parental affect and child aggression. Journal of American Academy of Child & Adolescent Psychiatry, 50(10), 1054–1064. doi:10.1016/j.jaac.2011.06.013.

The associations between parenting styles and proactive and reactive aggression in Hong Kong children and adolescents.

Previous research suggests that reactive and proactive aggression may be differentially related to family contextual (e.g. parenting practices) factor...
496KB Sizes 43 Downloads 14 Views