Children's self‐reports and parents' reports of internalising and externalising problems in Chinese and Anglo‐Celtic children in Australia

Abstract

The Child Behaviour Checklist (CBCL) and Youth Self‐Report (YSR) are widely used measures of internalising and externalising problems in children and adolescents, but cross‐cultural differences in these problems have not been examined in Australian children of Chinese and Anglo‐Celtic ancestry. Fifty‐nine Chinese (27 boys and 32 girls) and 65 Anglo‐Celtic Australian (35 boys and 30 girls) children aged 10–13 years completed the YSR, and one of their parents completed the CBCL. Cross‐informant agreement between these two measures was also examined. No differences were found between the Chinese and Anglo‐Celtic groups on parents' and children's ratings on any syndrome or overall scales. The overall level of agreement between YSR and CBCL scores was poor in the Anglo‐Celtic group (intraclass correlation coefficient for total problems = 0.33), but there was no agreement between these scores in the Chinese group. Results highlight the need to obtain multiple sources of information in assessments of mental health problems in children from different cultural backgrounds.

Key words:

• cross‐cultural differences
• cross‐informant agreement
• externalising problems
• internalising problems

There is abundant research on the Child Behaviour Checklist (CBCL; Achenbach & Rescorla, 2001) and Youth Self‐Report (YSR; Achenbach & Rescorla, 2001), a parent‐report and self‐report of mental health problems in children, respectively. Yet differences in the extent to which these problems are evident in Australian children from different cultural backgrounds have been largely neglected in the research literature, apart from a few studies that have compared children of Australian‐born mothers and children of overseas‐born mothers (e.g., Alati, Najman, Shuttlewood, Williams, & Bor, 2003; Bond, Nolan, Adler, & Robertson, 1994; Hensley, 1988). The majority of children in Australia are of Australian (39%) or Anglo‐Celtic (English—21.3%, Irish—5%, Scottish—3.8%, or Welsh—<1%) ancestry, but there are also many children of Chinese (2.4%) ancestry in Australia (Australian Bureau of Statistics (ABS), 2008a). The present study examined mental health problems in children of Australian/Anglo‐Celtic ancestry and children of Chinese ancestry, and the level of agreement between parents' reports and children's self‐reports of such problems.

CROSS‐CULTURAL DIFFERENCES IN INTERNALISING (INT) AND EXTERNALISING (EXT) PROBLEMS

Mental health problems in children are often classified as INT, where there are disturbances in children's mood of emotions (e.g., anxiety), or EXT, where there are disturbances in children's behaviours that are disruptive and harmful to others (e.g., aggression). The prevalence of INT or EXT problems in different cultural contexts may depend on the degree to which these behaviours are accepted or discouraged in the broader sociocultural milieu (Zahn‐Waxler, Klimes‐Dougan, & Slattery, 2000). Chinese children may be more likely to express psychological distress with INT problems, particularly of the psychosomatic type, because EXT problems are more visible, disruptive, and undermining of traditional Chinese values of behavioural restraint, harmony, and order (Chen, 2000). Additionally, EXT problems often entail non‐conformity to rules set by authority figures; such non‐conformity may be discouraged in a society that fosters reverence and respect by children towards parents and teachers (Leung & Fan, 1996). A recent review by Anderson and Mayes (2010) also highlighted biological and social factors and other factors like parenting and acculturation processes that might contribute to cross‐cultural differences in the prevalence of INT disorders, such as depressive and anxiety disorders, in adolescents. Empirically, the idea of culturally bound expressions of psychological distress has received equivocal support.

In two separate reviews, Rescorla and her colleagues (Rescorla, Achenbach, Ivanova, Dumenci, Almqvist, Bilenberg, Bird, Broberg, et al., 2007; Rescorla, Achenbach, Ivanova, Dumenci, Almqvist, Bilenberg, Bird, Chen, et al., 2007) compared CBCL and YSR scores with studies in 31 and 24 countries, respectively, including Australia, the USA, People's Republic of China (PRC), Hong Kong Special Administrative Region (HK), and Taiwan. CBCL Total Problems scores for these countries were ranked, in order from lowest to highest, PRC (4th), Australia (9th), the USA (20th), Taiwan (23rd), and HK (26th), and for YSR Total Problems scores, Australia (12th), the USA (13th), and HK (21st), out of all the countries surveyed. Researchers have found lower EXT problems in Chinese children and adolescents in PRC and Taiwan, compared with children and adolescents in the general American population, but findings for INT problems have been more mixed depending on the informant (parent, teacher, or self) and type of INT problems (anxious‐depressed, somatic, or total) examined (Verhulst et al., 2003; Weine, Phillips, & Achenbach, 1995; Yang, Soong, Chiang, & Chen, 2000; Zhou, Lengua, & Wang, 2009). There have been fewer studies of children or adolescents in immigrant Chinese families, but Wu and Chao (2005) found that Chinese American adolescents reported fewer EXT problems than European American adolescents on the YSR with no difference for EXT problems. In contrast, Chang, Morrissey, and Koplewicz (1995) found lower INT and EXT problems in Chinese American children than in the normative US sample for the CBCL.

Although studies of Chinese children in PRC, HK, and Taiwan are typically considered together, socio‐political and historical differences among these Chinese societies should not be overlooked. Such differences may explain other findings of varying levels of INT and EXT problems reported in children in different Chinese societies (see, for example, Guo et al., 2000). Moreover, differences may exist between Chinese children who live in societies with predominantly Chinese populations (e.g., children in PRC) and those from Chinese families living in another cultural setting where the Chinese community is an ethnic minority group (e.g., Chinese American children). Even among ethnic minority groups, there are differences between, for instance, Chinese children in the USA and those in Australia (see, for example, Chiu, Feldman, & Rosenthal, 1992). Studies involving children in the USA may also overlook within‐group variation, which may mask potential between‐group differences when the US group is compared with other groups (see, for example, Greenberger, Chen, Tally, & Dong, 2000). Although Chinese societies may share many similarities (Ho, 1986), research specific to each cultural setting is needed rather than relying on previous findings for groups that appear to be culturally similar. In Australia, the CBCL and YSR have not been used to compare INT and EXT problems in children of Chinese and Australian or Anglo‐Celtic ancestry. In one study of Australian‐born children, Alati et al. (2003) found no difference in INT or EXT problems among children of Australian‐born and overseas‐born mothers (UK, Ireland, or New Zealand, other Europe, or Southeast Asia). However, in the overseas‐born mothers group, the level of EXT problems was higher in children of mothers who had lived in Australia for more than 12 years, compared with children of mothers who had lived in Australia for fewer than 5 years, and this level was comparable with the level of EXT problems evident in children of Australian‐born mothers. Second‐generation Australian children who are more acculturated may resemble third‐generation Australian children in the way they express psychological distress, but this similarity may not be observed in second‐generation Australian children whose families recently settled in Australia, as they are likely to be less acculturated (Alati et al., 2003).

AGREEMENT BETWEEN PARENTS' REPORTS AND CHILDREN'S SELF‐REPORTS

One of the advantages of the CBCL and YSR is that these parallel instruments permit comparisons of parents' reports and children's self‐reports of INT and EXT problems. Researchers have generally found moderate levels of agreement between parents' ratings on the CBCL and their children's ratings on the YSR in samples from the USA, Australia, and PRC (e.g., Achenbach & Rescorla, 2001; Rey, Schrader, & Morris‐Yates, 1992; Wang, Zhang, & Leung, 2005). Adolescents themselves typically report higher levels of problems on the YSR than their parents on the CBCL (Rescorla, Achenbach, Ivanova, Dumenci, Almqvist, Bilenberg, Bird, Broberg, et al., 2007). These trends have been interpreted as variability of behavioural ratings by different informants, which may be partly attributed to variability in children's behaviours across different social contexts (Achenbach, McConaughy, & Howell, 1987). The level of agreement between parents and their children on reports of mental health problems in children depends on several factors, including children's age and sex, parents' mental health, aspects of the parent–child relationship, and children's and parents' cultural backgrounds (see De Los Reyes & Kazdin, 2005, for a review). Lau et al. (2004) examined parents' and adolescents' ratings of INT and EXT problems among adolescents in the USA who self‐identified as Caucasian, African American, Hispanic, or Asian/Pacific Islander. There were smaller discrepancies and stronger associations between parents and adolescents' reports in the Caucasian American group than in all of the other groups. Different rates of acculturation in parents and children of immigrant backgrounds may increase the level of parent–child disagreement on reports of INT and EXT problems in children, as the expression and acknowledgement of such problems are likely to be culturally bound (Fung & Lau, 2010; Lau et al., 2004). The study by Lau et al. involved adolescents recruited from five systems of care (e.g., juvenile justice system and mental health services) in the USA; it remains to be investigated whether the level of parent–child agreement varies among parents and children in the general Australian population and who are from different cultural backgrounds.

THE CURRENT STUDY

The aim of the current study was to explore INT and EXT problems in Australian children from Chinese and Australian/Anglo‐Celtic cultural backgrounds, on the basis of parent‐reports on the CBCL and children's self‐reports on the YSR. The level of agreement between parents' reports and children's self‐reports of these problems were also examined. Although not a main aim of the current study, sex differences in INT and EXT problems were also explored, given the well documentation of such differences in the research literature (Rescorla, Achenbach, Ivanova, Dumenci, Almqvist, Bilenberg, Bird, Chen, et al., 2007) and the potential for interaction effects between children's cultural background and sex. Differences between the Chinese and the Anglo‐Celtic group and between boys and girls in INT and EXT (and total) problems were explored using both raw scores and standardised T scores on these scales. While higher raw scores reflect higher frequencies of problems, higher T scores reflect levels of problems that are more likely to be of clinical concern based on normative scores for the YSR and CBCL (that are stratified by children's age and sex) and, thus, provide another dimension into understanding potential cross‐cultural or sex differences in INT and EXT problems.

METHOD

Participants

The sample comprised 59 Chinese Australian (27 boys and 32 girls (M age: 11.03, standard deviation (SD) = 0.81)) and 65 Anglo‐Celtic Australian (35 boys and 30 girls (M age: 11.06, SD = 0.58)) children. A larger proportion of parents in the Chinese group (15%) did not return the questionnaires for parents, as compared with parents in the Anglo‐Celtic group (5%). A larger proportion of parents in the Chinese group (22%) were also fathers, as compared with parents in the Anglo‐Celtic (13%) group. Participating parents provided demographic information for themselves and for their partners. Chinese Australian children were defined as those who themselves or their parents were born in Australia, PRC, HK, Taiwan, or Southeast Asia, who themselves or their parents identified with Chinese or Chinese and Australian ancestry, and whose home language was Chinese and/or English. Anglo‐Celtic Australian children were defined as those who themselves or their parents were born in Australia or the UK, who themselves or their parents identified with Australian and/or another Anglo‐Celtic ancestry (i.e., English, Irish, Scottish, or Welsh), and whose home language was exclusively English.

Chinese Australian children

The majority of children in the Chinese group were second‐generation Chinese Australians (75%). Those who were born overseas (PRC, n = 5; HK, n = 5; Singapore, n = 2; Malaysia/South Korea/New Zealand, n = 3) had lived in Australia for an average of 4.72 years (SD = 3.21). All mothers and all but two fathers were first‐generation Chinese Australians, having lived in Australia for an average of 13.28 years (SD = 7.07) for mothers and 13.74 years (SD = 8.44) for fathers. Most mothers were born in PRC (n = 24), followed by Southeast Asia (n = 20), Hong Kong (n = 11), and Taiwan (n = 3). Most fathers were also born in PRC (n = 24), followed by Southeast Asia (n = 19), Hong Kong (n = 9), Taiwan (n = 3), Australia (n = 2), and Macau (n = 1). Twenty children were recruited from Chinese‐language weekend schools, and an additional 22 children reported that they also attended Chinese‐language weekend schools. The majority of children lived in intact families (85%), with the remaining in single‐parent families (12%) and step/blended families (3%).

Anglo‐Celtic Australian children

The majority of children in the Anglo‐Celtic group were third‐generation Anglo‐Celtic Australians (83%). Those born overseas (UK, n = 1; USA, n = 2; HK, n = 1) lived in Australia for an average of 8.50 years (SD = 1.29). For these five children, their parents were born in Australia and identified with Australian or Anglo‐Celtic ancestry. The majority of mothers (89%) and fathers (94%) in the Anglo‐Celtic group were also at least second‐generation Anglo‐Celtic Australians. Parents who were born in the UK had lived in Australia for an average period of 34.80 years (SD = 9.83) for mothers (n = 5) and 35.50 years (SD = 2.12) for fathers (n = 2). The majority of children lived in intact families (91%), with the remaining in single‐parent families (6%) and step/blended families (3%).

Parents' occupation and educational attainment for the Chinese and Anglo‐Celtic groups are displayed in Table-1. A larger proportion of parents in the Anglo‐Celtic group were employed in professional positions and held postgraduate qualifications, as compared with parents in the Chinese group. Compared with the overall Australian population, parents in both groups had higher educational qualifications, and more were engaged in professional work (ABS, 2008b, 2008c).

Table 1 Parents' occupation and educational attainment

–	Chinese	Anglo‐Celtic
Mother's occupation	(n = 49)	(n = 61)
Professional	33%	59%
Non‐professional	27%	15%
Own business	16%	6%
Not employed	24%	20%
Father's occupation	(n = 46)	(n = 60)
Professional	46%	75%
Non‐professional	26%	12%
Own business	24%	13%
Not employed	4%	0%
Mother's education	(n = 50)	(n = 61)
Postgraduate	8%	18%
Undergraduate	42%	33%
TAFE/High school	36%	38%
Did not complete high school	14%	12%
Father's education	(n = 47)	(n = 60)
Postgraduate	17%	33%
Undergraduate	40%	38%
TAFE/High school	19%	17%
Did not complete high school	23%	12%

TAFE = Technical and Further Education institutions.

Measures

The YSR for ages 11–18 (Achenbach & Rescorla, 2001) and the CBCL for ages 6–18 (Achenbach & Rescorla, 2001) consist of 112 and 113 items, respectively, and each item is rated on a 3‐point scale, where (0) = not true, (1) = somewhat/sometimes true, and (2) = very true or often true. Scores on all items on the YSR and CBCL are summed to form the total problems score, which consists of the INT problem scale (anxious/depressed, withdrawn/depressed, and somatic complaints syndrome scales), the EXT problem scale (rule‐breaking behaviour and aggressive behaviour syndrome scales), and four other syndrome scales (social problems, thought problems, attention problems, and other problems). Norms are available to convert raw scores on the INT, EXT, and total problems scales to standardised T scores, according to children's sex and age, and T scores can be classified in the normal range (<60), borderline clinical range (60–64), and clinical range (≥65). The YSR and CBCL have very good psychometric properties and are widely used measures of INT and EXT problems in children and adolescents (Achenbach & Rescorla, 2001). Both measures have been translated into the Chinese language and satisfactory reliability, and validity indices have been reported for the translated version (Leung et al., 2006).

Procedure

Approval to conduct research was obtained from the authors' university ethics committee, the Department of Education and Early Childhood Development, Victoria, Australia, and the Catholic Education Office, Melbourne, Victoria, Australia. Based on 2001 Census data (ABS, 2006), eastern and western metropolitan regions in Victoria were identified as areas where the Chinese language was commonly spoken among residents, and 250 primary and Chinese‐language weekend schools in these areas were contacted in 2006–2007. Thirty‐two primary school principals (132 declined, and 62 did not respond) and 12 weekend school principals (12 declined, and 12 did not respond) agreed to allow their schools participate in this research. Within schools that participated, all children who were 10–13 years were invited to take part in the research. The number of Chinese students in primary schools could not be determined as many school principals provided estimated numbers only. The overall response rate from parents in primary schools was 8.1% and in weekend schools, 6%. Data for the current study were collected as part of a larger study on children's temperament, family functioning, and mental health problems in Australian children from different cultural backgrounds. All materials for parents, apart from the CBCL (an official Chinese version was already available), were professionally translated from English to the traditional Chinese script, and letters to parents were sent in both English and Chinese. Out of all Chinese parents who took part in the research, 19 parents (32%) indicated on the consent forms of their preference to receive all materials in Chinese. Children completed a demographic questionnaire and the YSR in groups of four at their regular day school or language school. Children took home the demographic questionnaire for parents and the CBCL for their parents. Parents retuned both questionnaires to the authors via reply‐paid envelopes.

RESULTS

Plan for statistical analyses

The data were collated and analysed using IBM SPSS Statistics (Version 19.0), New York, USA. Treatment of missing items on the YSR and CBCL were in accordance to guidelines set out in the test manual (Achenbach & Rescorla, 2001). YSR and CBCL scores were examined in two ways for cross‐cultural and sex differences. First, raw scores were used in 2 × 2 between‐subjects analyses of variance (ANOVAs) that investigated main effects of sex and ancestry, and interaction effects between sex and ancestry on all syndrome, and INT, EXT, and total problems scale scores. Second, standardised T scores for INT, EXT, and total problems scales were analysed in 2 × 2 ANOVAs for main and interaction effects of sex and ancestry. Although standardised T scores were also available for syndrome scales, Achenbach and Rescorla (2001) recommended the use of raw scores for these syndrome scales because T scores were truncated at 50 for these scales. T scores were also used to identify the proportion of children who obtained scores in the clinical range on the INT, EXT, and total problems scales. To investigate the level of agreement between YSR and CBCL scores, raw scores on 105 items common to both the YSR and CBCL were used to compute intraclass correlation coefficients (ICC; McGraw & Wong, 1996). A two‐way mixed effects model investigating absolute agreement was specified in SPSS to calculate the single measures intraclass correlation for all YSR and CBCL scale scores. Paired‐samples t‐tests were also computed to investigate the magnitude of difference between YSR and CBCL scale scores. There was a potential for inflated type I error because of the number of analyses conducted for a total of 12 scale scores, and a Bonferroni adjustment was applied, which resulted in an adjusted alpha level of .004.

YSR and CBCL scores by children's sex and ancestry group

Descriptive statistics for the various syndrome, and INT, EXT, and total problems scales are presented in Tables-2 and 3.

Table 2 Means and standard deviations (in parentheses) for YSR syndrome and internalizing, externalizing, and total problem scales by children's ancestry group and sex

–	Chinese			Anglo‐Celtic
Scale	Boys	Girls	Total	Boys	Girls	Total
Anxious‐depressed	5.00(3.97)	4.38(2.69)	4.66(3.32)	3.91(3.36)	3.73(2.77)	3.83(3.07)
Withdrawn‐depressed	3.81(2.77)	3.28(2.14)	3.52(2.44)	3.03(2.32)	2.07(1.86)	2.58(2.16)
Somatic problems	4.15(2.74)	2.75(2.23)	3.39(2.55)	3.31(2.39)	3.33(3.04)	3.32(2.69)
Social problems	3.30(2.83)	3.09(2.05)	3.19(2.42)	3.46(2.74)	2.33(2.62)	2.94(2.72)
Thought problems	4.37(3.58)	2.88(2.98)	3.56(3.32)	3.57(3.81)	3.33(3.90)	3.46(3.82)
Attention problems	4.37(2.31)	3.31(2.16)	3.80(2.27)	4.94(3.36)	3.07(2.42)	4.08(3.09)
Rule‐breaking behaviour	3.63(4.13)	1.66(1.60)	2.56(3.16)	3.23(2.33)	1.67(1.73)	2.51(2.20)
Aggressive behaviour	6.37(5.10)	4.41(3.72)	5.31(4.48)	6.57(3.73)	5.40(4.45)	6.03(4.09)
Other problems	4.15(2.27)	3.03(2.35)	3.54(2.35)	4.97(3.03)	4.10(2.68)	4.57(2.89)
Internalizing Problems	12.96(8.49)	10.41(5.76)	11.58(7.19)	10.26(7.13)	9.13(6.21)	9.74(6.70)
Externalizing Problems	10.00(8.70)	6.06(4.76)	7.86(7.06)	9.80(5.64)	7.07(5.85)	8.54(5.86)
Total Problems	39.15(23.62)	28.78(15.80)	33.53(20.27)	37.00(20.54)	29.03(19.06)	33.32(20.12)
n	27	32	59	35	30	65

YSR = Youth Self‐Report.

Table 3 Means and standard deviations (in parentheses) for CBCL syndrome and internalizing, externalizing, and total problem scales by children's ancestry group and sex

–	Chinese			Anglo‐Celtic
Scale	Boys	Girls	Total	Boys	Girls	Total
Anxious‐depressed	2.43(2.48)	2.67(2.92)	2.56(2.70)	2.70(2.77)	2.59(2.53)	2.65(2.64)
Withdrawn‐depressed	1.22(1.81)	1.52(2.08)	1.38(1.95)	1.61(2.37)	1.07(1.58)	1.35(2.04)
Somatic problems	1.00(1.57)	1.04(1.37)	1.02(1.45)	1.21(1.47)	1.62(1.93)	1.40(1.70)
Social problems	2.04(2.27)	1.89(2.22)	1.96(2.22)	1.82(2.69)	1.48(3.00)	1.66(2.82)
Thought problems	1.74(2.36)	1.52(1.89)	1.62(2.10)	1.36(1.73)	1.03(1.15)	1.21(1.48)
Attention problems	3.91(3.64)	2.52(2.74)	3.16(3.23)	3.45(3.04)	1.93(2.02)	2.74(2.70)
Rule‐breaking behaviour	1.35(1.37)	0.89(1.12)	1.10(1.25)	1.39(1.64)	1.07(2.63)	1.24(2.15)
Aggressive behaviour	3.09(2.41)	2.59(3.30)	2.82(2.90)	4.21(3.50)	4.34(6.14)	4.27(4.87)
Other problems	2.17(2.23)	1.70(1.73)	1.92(1.97)	3.00(2.76)	2.38(2.54)	2.71(2.66)
Internalizing problems	4.65(4.71)	5.22(5.84)	4.96(5.30)	5.52(5.48)	5.28(5.10)	5.40(5.26)
Externalizing problems	4.43(3.38)	3.48(3.84)	3.92(3.63)	5.61(4.78)	5.41(8.62)	5.52(6.79)
Total problems	18.96(14.36)	16.33(15.41)	17.54(14.84)	20.76(16.48)	17.52(19.04)	19.24(17.65)
n	23	27	50	33	29	62

CBCL = Child Behaviour Checklist.

Results of ANOVAs were boys reported higher raw scores than girls on the attention problems (M difference = 1.50, 95% CI = 0.57 to 2.43, F(1, 120) = 9.48, p = .003, partial η² = .07), rule‐breaking behaviour (M difference = 1.74, 95% CI = 0.83 to 2.65, F(1, 120) = 14.50, p < .001, partial η² = .11), EXT problems (M difference = 3.34, 95% CI = 1.12–5.56, F(1, 120) = 8.65, p = .004, partial η² = .07) scales. Boys also obtained higher T scores than girls on INT (M difference = 6.39, 95% CI = 2.90–9.88, F(1, 120) = 14.32, p < .001, partial η² = .11), EXT (M difference = 5.03, 95% CI = 1.76–8.30, F(1, 120) = 8.87, p = .004, partial η² = .07), and total problems (M difference = 6.10, 95% CI = 2.58–9.61, F(1, 120) = 11.75, p = .001, partial η² = .09) scales. There were no main effects of sex on CBCL scale scores (raw or T scores), nor were there main effects of ancestry or interaction effects between sex and ancestry on YSR or CBCL scale scores (raw or T scores). The proportion of children scoring in the clinical range of T scores (i.e., ≥65) on INT, EXT, and total problems scales are reported in Table-4.

Table 4 Proportion of children in clinical range of scores (T scores ≥65) on internalizing (INT), externalizing (EXT), and total problem scales on YSR and CBCL

–	Chinese			Anglo‐Celtic
	Boys	Girls	Total	Boys	Girls	Total
YSR
INT problems	18.5%	6.3%	11.9%	14.3%	0%	7.7%
EXT problems	14.8%	3.1%	8.5%	2.9%	3.3%	3.1%
Total problems	22.2%	3.1%	11.9%	17.1%	3.3%	10.8%
n	(27)	(32)	(59)	(35)	(30)	(65)
CBCL
INT problems	8.7%	11.1%	10.0%	9.1%	13.8%	11.3%
EXT problems	0%	0%	0%	3.0%	6.9%	4.8%
Total problems	4.3%	3.7%	4.0%	9.1%	3.4%	6.5%
n	(23)	(27)	(50)	(33)	(29)	(62)

CBCL = Child Behaviour Checklist; EXT = externalizing; INT = internalizing; YSR = Youth Self‐Report.

Comparisons of YSR and CBCL scores

Table-5 presents ICC and results from paired‐samples t‐tests comparing YSR and CBCL scale scores. There was no agreement in the Chinese group between YSR and CBCL scores on all scales, whereas six scales showed poor to moderate agreement with an ICC of 0.33 for total problems in the Anglo‐Celtic group. In the Chinese group, YSR scores on all scales but social problems and attention problems were significantly higher than their counterparts on the CBCL. In the Anglo‐Celtic group, YSR scores on all scales, but anxious‐depressed and aggressive behaviours were significantly higher than CBCL scores on the corresponding scales. Additional Pearson's r coefficients showed significant associations (p < .001) between INT and EXT problems raw scores on the YSR (Chinese group: r = .58, Anglo‐Celtic group: r = .59) and CBCL (Chinese group: r = .52, Anglo‐Celtic group: r = .51).

Table 5 Intraclass correlation coefficients (ICC), mean difference, and 95% range for raw scores on YSR and CBCL scales

–	Chinese (n = 50)			Anglo‐Celtic (n = 62)
	ICC	M diff	95% range	ICC	M diff	95% range
Anxious‐depressed	.07	2.38***	−7.73to10.73	.11	1.18	−8.43to8.85
Withdrawn‐depressed	−.01	2.30***	−5.18to8.73	.31**	1.21***	−4.43to6.42
Somatic problems	.01	2.62***	−3.00to9.90	.11	2.03***	−3.85to8.00
Social problems	.09	1.20	−7.00to6.73	.50***	1.27***	−5.85to7.00
Thought problems	.10	2.28***	−4.45to12.00	.14	2.40***	−2.00to13.42
Attention problems	.17	0.78	−8.90to6.00	.37***	1.29**	−7.00to7.42
Rule‐breaking behaviour	−.08	1.62^†	−3.00to15.73	.24	1.26***	−6.97to6.85
Aggressive behaviour	.01	2.80**	−9.63to16.73	.36**	1.61	−13.25to11.42
Other problems	.18	1.98***	−4.00to7.73	.40***	2.15***	−2.00to7.85
Internalizing problems	.01	8.20***	−15.18to25.63	.21	4.21***	−12.00to20.27
Externalizing problems	−.04	4.42**	−11.63to30.80	.33**	2.87**	−17.92to17.00
Total problems	.01	17.96***	−45.93to85.23	.33***	14.40***	−33.85to56.12

M diff = Mean difference.

*** p < .001, **p < .004,

† p = .004.

DISCUSSION

Cross‐cultural differences in INT and EXT problems

The finding of no significant differences in INT, EXT, and total problems between Chinese and Anglo‐Celtic groups is inconsistent with the findings of other studies that have used the CBCL and YSR with Chinese American and European American children (Chang et al., 1995; Wu & Chao, 2005) but are in accord with the findings of no difference between Australian children whose mothers were born in Australia and those whose mothers were born overseas in the study by Alati et al. (2003). Most children in the Chinese group were second‐generation Australians, and results might have been different for first‐generation children. When children from immigrant families become more acculturated, they may become more like to other children in the host culture in the manner in which psychological distress is manifested (Alati et al., 2003; Chiu et al., 1992). It has been suggested that Chinese children may be more likely to internalise than to externalise psychological distress, given Chinese socialisation patterns that promote behavioural restraint and social harmony (Chen, 2000). Such socialisation patterns probably exist to some degree in other cultural settings, which may explain why both Chinese and Anglo‐Celtic children in the current study exhibited similar levels of INT and EXT problems, when compared with each other.

Cross‐cultural differences in levels of parent‐child agreement

In both the Chinese and Anglo‐Celtic groups, children reported having higher levels of most types of problems than their parents reported for them, which is consistent with previous findings (Rescorla, Achenbach, Ivanova, Dumenci, Almqvist, Bilenberg, Bird, Broberg, et al., 2007). One interpretation of these results is that parents' reports of such problems are limited to their observations within home and family settings and to those children's behaviours that are more overt (Achenbach et al., 1987). Current results also highlighted stronger agreement and smaller differences between parents' ratings and children's self‐ratings on most types of problems in the Anglo‐Celtic than in the Chinese group. The level of parent–child agreement found in the Anglo‐Celtic group was poor to moderate, as was found in the normative samples for the CBCL and YSR in the USA (Achenbach & Rescorla, 2001). In contrast, there was no agreement between children's and parents' reports of any type of problem in the Chinese group, a finding that differs from the moderate level of agreement reported in other studies of Chinese children in PRC and the USA (Fung & Lau, 2010; Wang et al., 2005). Weaker parent–child agreement in the Chinese than Anglo‐Celtic group was somewhat expected because there may be an acculturation gap between parents and children in the Chinese group, which could have led to different beliefs regarding mental health problems (Fung & Lau, 2010; Lau et al., 2004). However, the complete lack of agreement between parents and their children in the Chinese group was unexpected, and other factors such as the quality of the parent–child relationship or parental mental health status might have also contributed to this level of disagreement (De Los Reyes & Kazdin, 2005).

Methodological limitations

Limitations of the current study were a low response rate and consequently, a small sample size, which might not be representative of Chinese or Anglo‐Celtic children in Australia in general. Due to the sample size, it was not statistically feasible to examine further variations within the Chinese or Anglo‐Celtic group, which may have arose from differences in children's or parents' self‐identified ancestry, country of birth, and length of stay in Australia, or families' overall socioeconomic status. Although separate response rates for the Chinese and Anglo‐Celtic groups could not be determined, the response rate for the Chinese group seemed lower than the Anglo‐Celtic group. Recruitment and data collection for Chinese participants took twice as long as that for Anglo‐Celtic participants (24 months vs 12 months) despite recruiting Chinese participants from two sources (primary schools and weekend language schools) and in areas with large Chinese populations. Even among parents who gave consent for their child and themselves to participate in the current research, there was a poorer return rate from Chinese than Anglo‐Celtic parents. Chinese parents may have been reluctant to allow their child to participate because of a perceived stigma associated with mental health issues, which may have important ramifications for Chinese parents' willingness to engage with mental health‐care services for themselves and for their children (Chan, 2009).

Clinical implications and recommendations for future research

There were no differences in the level of INT or EXT problems between Chinese and Anglo‐Celtic children; nevertheless, the close association between INT and EXT problems in both groups highlights the frequent co‐occurrence of these problems in children regardless of their cultural background. The level of parent–child agreement on ratings of mental health problems appeared poorer in the Chinese than in the Anglo‐Celtic group, but the level of agreement in the latter group was not very high either. Current findings suggest that children may be more reliable informants than their parents regarding INT and EXT problems in children, but such a proposition would require further corroboration from other informants (e.g., teachers) and clinical assessments. These findings also underscore the importance of multiple informants in clinical assessments, especially for clinicians working with Chinese Australian children, as well as the need for further research into the factors that influence the level of parent–child agreement on ratings of mental health problems in children from different cultural backgrounds.

CONCLUSIONS

Main findings of the present study were no difference between the Chinese and Anglo‐Celtic groups in the type of mental health problems experienced by children, but poorer parent–child agreement on ratings of such problems in the Chinese than Anglo‐Celtic group. The next step for researchers is to investigate whether there are differences in the aetiology of mental health problems in Australian children from different cultural backgrounds. Findings from these future studies will have further ramifications as to how clinicians assess, treat, and prevent mental health problems in an Australian context.