Faculty of Social and Behavioral Sciences
Graduate School of Child Development and Education
Understanding Affective Teacher-Student Relationships in a Cross-Cultural Context: A Comparison Between Dutch and Chinese Primary School Teachers
and Students
Mengdi Chen (student number: 11328215) University of Amsterdam
Research Master of Child Development and Education Supervisor: dr. D. L., Roorda
Abstract
The affective quality of teacher student relationships (TSRs) has mostly been studied in Western contexts and little is known about whether previous findings also apply to Eastern contexts. Therefore, the present study examined measurement invariance, mean differences, and agreement in teachers’ and students’ perceptions of relationship quality between the Netherlands and China. The sample consisted of 789 primary school students (403 girls) and 35 teachers from the Netherlands, and 587 primary school students (309 girls) and 14 teachers from Zhejiang, China. Both teachers and students reported about the quality of their mutual relationships. Structural equation modeling showed that teachers’ perceptions of Closeness, Conflict and Dependency, as well as students’ perceptions of Closeness and Conflict, reached partial strong invariance. Chinese students perceived more Closeness and less Conflict in their relationships with teachers than Dutch students. Chinese teachers also reported less Conflict in their relationships with students but comparable levels of Closeness and Dependency as Dutch teachers did. Chinese teacher-student dyads had higher
agreement on Closeness but lower agreement on Conflict than Dutch teacher-student dyads. Possible explanations are provided for the measurement non-invariance and cultural difference in reporting TSRs. More research is needed to explore cultural differences in TSRs while also taking the improvement of relationship questionnaires into account.
Keywords: cross-cultural comparisons, individualism and collectivism, student
Understanding Affective Teacher-Student Relationships in a Cross-Cultural Context: A Comparison Between Dutch and Chinese Primary School Teachers and
Students
The affective quality of teacher-student relationships (TSRs), which is the emotional bond between teachers and students, has frequently been found to be important for children’s social-emotional and academic development (for meta-analytic overviews, see Roorda, Koomen, Spilt, Oort, 2011; Roorda, Jak, Zee, Oort & Koomen, 2017). Despite the importance of TSRs for children’s school functioning, they have been primarily studied in Western contexts and not in Eastern contexts. Nevertheless, cultural contexts may have an important influence on the development of TSRs (Pianta, Hamre & Stuhlman, 2003). Furthermore, findings of studies on Western, individualistic countries (i.e., countries in which presenting oneself as an independent individual is valued; Markus & Kitayama, 1991) do not necessarily fit to Eastern, collectivistic countries (i.e., countries that stress the fundamental
connectedness to each other; Markus & Kitayama, 1991; Oyserman, Coon, Kemmelmeier, 2002). In addition, teachers and students may develop different
perceptions of their relationship (Pianta et al., 2003), and the agreement between their perceptions is found to be low in Western contexts (Hughes, 2011; Hughes & Cavell, 1999; Rey, Smith, Yoon, Somers & Barnett, 2007). However, it remains unknown whether this is also the case in Eastern contexts. Therefore, the present study included both teachers’ and students’ perceptions of their relationships. As the instruments measuring both perceptions are not used in a cross-cultural context before, we first investigated the measurement invariance of the instruments across the Netherlands (an individualistic country) and China (a collectivistic country), and thereafter the cultural differences in both teachers’ and students’ perceptions of their relationships.
Finally, we investigated the cultural differences in the agreement between the two perceptions.
Culture and Interpersonal Relationships
The impact of culture on children’s development is acknowledged by different theories, for example, ecological system theory. According to the ecological system theory (Bronfenbrenner, 1979), children develop within a series of nested
environmental systems (i.e., microsystem, mesosystem, exosystem and macrosystem) where factors (e.g., families, schools, or culture) interact with each other and affect children’s development. Pertinent with this study, the macro system consists of cultural values, customs and laws, which have impact on children and the relationships they have with their surroundings (e.g., children’s interaction with parents; Berk, 2000; Bronfenbrenner, 1979). The impact of culture on the development of affective teacher-student relationships is specified in the
developmental systems model developed by Pianta et al. (2003). According to this model, TSRs is embedded in a system where there are external influences such as cultural values and norms (Pianta et al., 2003). Cultural values such as general beliefs about children, shapes individuals’ expectation about each other (e.g., students’ performances) and therefore affect the way students and teachers relate to one another (Pianta et al., 2003).
With regard to cultural values, a distinction can be made between individualistic and collectivistic cultures (Hofstede, 1980; Oyserman et al., 2002; Voronov & Singer, 2002). By definition, individualism refers to a culture that focuses on expressing ones’ unique attributes and construing oneself as an independent individual, whereas
collectivism refers to a culture that focuses more on the fundamental connectedness to each other and linking individuals as parts of a whole (Markus & Kitayama, 1991). In
an individualistic culture, people tend to be independent from their in-groups (i.e., family, tribe and nation) and guide their behaviors on the basis of their own goals (Triandis, 2001; Triandis, Bontempo & Villareal, 1988). On the contrary, in a collectivistic culture, people tend to be dependent on their in-groups and guide their behaviors on the basis of in-group norms (Triandis, 2001; Triandis, Bontempo & Villareal, 1988). As a result, the interpersonal relationships in collectivistic cultures are more intensive and the interdependence is higher than in an individualistic culture, where there are more distance and detachment (Mark & Kitayama, 1991; Triandis et al., 1988). Furthermore, collectivists and individualists differ in their tolerance towards conflict in interpersonal relationships and the mechanisms to deal with conflict (Chen & French, 2007). In collectivistic cultures, individual tends to avoid conflict whereas in individualistic cultures one tends to bring conflict to open
(Triandis et al., 1988). These cultural differences in interpersonal relationships may as well apply to teacher-student relationships. As the Netherlands is generally considered individualistic and scored high in individualism (Hofstede, 1986), and China is
generally considered collectivistic and scored high in collectivism (Oyserman et al., 2002), this study aimed to compare the quality of teacher-student relationships in the Netherlands and China.
Attachment Theory and Cross-Cultural Differences
Other than ecological system theory, attachment theory (Bowlby, 1980)
provided a more nuanced theoretical framework to understand affective relationships in cross-cultural contexts. It was initially used to explain how parent-child
relationships serve as a context for children’s development. According to attachment theory (Bowlby, 1980), children seek proximity with their mothers when they feel threatened or hurt. Mothers’ sensitive responses (i.e., perceiving children’s needs
accurately and responding appropriately and timely towards them) will serve as a secure base for children to explore the environment (Ainsworth, Blehar, Waters, & Wall, 1978). As a result, children who are securely attached become more socially and emotionally competent than children who are insecurely attached (Ainsworth et al., 1978). These statements are known as the three core hypotheses in attachment theory, namely, sensitivity, secure base and competence hypotheses (Rothbaum, Weisz, Pott, Miyake & Morelli, 2000).
Attachment may be defined or understood differently in different cultural context (Beyazuk & Kesner, 2005; Keller, 2013; Mizuta, Zahn-Waxler, Cole & Hiruma, 1996; Oyserman et al., 2002). For example, Rothbaum et al. (2000) argued that the three core hypotheses of attachment were constructed differently in the United States (i.e., individualistic culture) and Japan (i.e., collectivistic culture). That is, Japanese mothers perceived sensitivity as being responsive to infants’ social engagement, which served for exploration of interdependence and entailed being competent in preserving social harmony (Rothbaum et al., 2000). On the contrary, American mothers perceived sensitivity as being responsive to infants’ need for individuation, which served for autonomous exploration and entailed being competent in obtaining autonomy (Rothbaum et al., 2000).
Since the 1990s, it has been argued that attachment theory could also be used to conceptualize teacher-student relationships (Pianta, 1992; Verschueren, 2015; Verschueren & Koomen, 2012). Sensitive teachers may be regarded as attachment figures as well and, hence, enable children to function adequately within the school environment (Pianta et al., 2003; Verschueren & Koomen, 2012). More specifically, sensitive teachers are more likely to form positive TSRs with students and provide them with a secure base to explore the environment whereas less sensitive teachers
are more likely to form negative TSRs with students and interfere their attempt to explore the school environment (Pianta et al., 2003; Verschueren & Koomen, 2012). Based on attachment theory, Pianta (2001) defined three dimensions for measuring TSRs, namely Closeness (i.e., the degree of warmth and openness in the relationship), Conflict (i.e., the discordance in the relationship) and Dependency (i.e., a child’s overreliance and possessiveness in the relationship). As attachment theory also applies to TSRs, the cultural differences found in mother-child attachment patterns can also be found in TSRs.
Attachment and Interpersonal Relationships in Cross-Cultural Contexts
In line with the theories, some empirical studies found evidences on cultural differences in attachment patterns between collectivistic and individualistic cultures. Mizuta et al. (1996) compared Japanese and American mother-preschooler dyads in a separation-reunion task and found that Japanese mothers showed more physical presence and Japanese children showed more amae (i.e., desire for bodily closeness) after reunion than their American counterparts. Furthermore, in Japan (Takahashi, 1986), Indonesia (Zevalkink, Riksen-Walraven, Van Lieshout, 1999) and Korea (Jin, Jacobvitz, Hazen & Jung, 2012), the percentage of infants classified as avoidant (i.e., around 1%) is significantly lower than in an American sample (i.e., around 21% ; Ainsworth et al., 1978); this might be due to the encouragement of close infant-mother relationships in Eastern countries(Jin et al., 2012).
Furthermore, some empirical studies have found evidence on the differences in interpersonal relationships that are theorized to exist between individualistic and collectivistic countries. In terms of Closeness, Dwairy and Achoui (2010) investigated the connectedness between children and their families in eight countries and found that family-child connectedness was lower in Western countries than in Eastern
countries. In addition, Indonesian youth reported more companionship with mothers and more intimacy with siblings compared to American youth (French, Rianasari, Pidada, Nelwan & Buhrmester, 2001). Besides, American mothers were found to maintain more physical distance when playing with their pre-school children than Japanese mothers (Dennis, Cole, Zahn-Waxler & Mizuta, 2002). With regard to Conflict, youth in the United States reported more Conflict with mothers, fathers, siblings, and friends than Indonesian youth (French et al., 2001). Students in Taiwan were found to report less Conflict in their relationships with peers compared to their counterparts in Canada (Benjamin, Schneider, Greenman, Hum, 2001). Lastly, regarding Dependency, Dennis et al. (2002) found that American mother-preschooler dyads emphasized more individuality in their conversations whereas Japanese dyads emphasized more mutuality.
Teacher-Student Relationships in Cross-Cultural Contexts
As far as we know, there are only five published studies that examined whether cross-cultural differences also exist in affective teacher-student relationships. With regards to positive dimensions of TSRs, previous studies found higher relationship quality in collectivistic cultures from both teachers and students’ perceptions (Beyazkurk & Kesner, 2005; Jia et al., 2009; Yang et al.,2013). Yang et al. (2013) investigated 3 to 12 grade students in China and the United States, and found that Chinese students perceived better relationships with teachers compared to American students. Beyazkurk and Kesner (2005) found that Turkish teachers in kindergarten and primary schools reported more Closeness with their students than elementary school teachers in the United States. In addition, Jia et al. (2009) found that Chinese students in urban middle schools perceived higher levels of teacher support compared to their counterparts in America. However, two other studies found contradictory
results (Aldridge & Fraser, 2000; Bear et al., 2014). Bear et al. (2014) included both teachers’ and students’ perceptions of the relationship quality in their cross-cultural comparison between China and the United States and found that in elementary schools, American teachers reported better TSRs than Chinese teachers, while the students reported equally favorable relationships with teachers across countries. Aldridge and Fraser (2000) found no difference in the degree of student-reported teacher support between junior high school students (13-15 years old) from Australia and from Taiwan. In terms of negative dimensions of the TSRs, Beyazkurk and Kesner (2005) also investigated teachers’ perceptions of Conflict in Turkey and the United States, and found no significant difference.
Considering the limited number of empirical studies available, teacher-student relationships in cross-cultural contexts is a rather unstudied field. Specifically, previous studies rarely included negative dimensions of TSRs such as conflict. Therefore, more study is needed to compare teacher-student relationships in cross-cultural contexts and fill in the gaps.
Agreement in Teachers’ and Students’ Relationship Perceptions
According to attachment theory, mental representation (i.e., the participants’ feelings, emotions, and thoughts about themselves, the other, and their mutual relationships; Pianta et al., 2003) forms one of the key elements of TSRs. Mental representations are formed based on attachment histories with parents, previous teachers or students (Pianta et al., 2003). Therefore, teachers and students may develop different mental representations of their relationships, and their perceptions may not always agree with each other (Pianta et al., 2003;). In line with attachment theory, empirical studies have found low agreement between teacher-reported and student-reported relationship quality, with the correlations ranging from .08 to .43
(Hughes, 2011; Hughes & Cavell, 1999; Rey et al., 2007). However, these findings were largely found under Western contexts and we do not know whether they also apply in Eastern contexts. Benjamin et al. (2001) did find higher agreement among friends in Taiwan in reporting the conflict in their friendships compared to friends in Canada. Therefore, we also examined whether the cultural differences on the
agreement of reporting relationships with friends also exist in the context of TSRs.
The Present Study
The present study aimed to compare Dutch and Chinese third to sixth graders’ affective relationships with teachers, and included both student’s and teacher’s perceptions of the relationship quality. As the instruments measuring TSRs had not been used in a cross-cultural context before, we first tested the measurement
invariance. Second, we investigated cultural differences between the Netherlands and China in both teachers’ and students’ perceptions of their relationships. Based on previous studies (Beyazkurk & Kesner, 2005; Jia et al., 2009; Yang et al., 2013), we hypothesized that Chinese students and teachers would report more Closeness than their Dutch counterparts. Enlightened by previous studies about interpersonal relationships (Benjamin et al., 2001; Dennis et al., 2002; French et al., 2001), we expected to find less Conflict and more Dependency between Chinese teachers and students compared to their Dutch counterparts. Third, we examined whether there were cross-cultural differences in the degree of agreement in relationship perceptions between teacher-student dyads from the Netherlands and China. Based on Benjamin et al. (2001), we hypothesized that the Chinese teacher-student dyads had higher agreement in relationship perceptions compared to Dutch dyads.
Method Participants
Dutch sample. The Dutch sample consisted of 789 students (including 386 boys
and 403 girls) and 35 teachers (including 28 females and 7 males) from 11 elementary schools in the Netherlands. The Dutch students were third to sixth graders and had a mean age of 9.99 years (SD = 1.23), ranging from 7 to 13 years old. Most of the students identified themselves (71.7%) and their parents (66.2% of the mothers and 64.3% fathers) as belonging to the Dutch ethnic group. Other students frequently identified themselves as belonging to the Moroccan (8.6%), Turkish (6.3%), and Surinamese (2.2%) ethnic groups, as well as identifying their parents as belonging to the Moroccan (12.2% of fathers and 11.4% of mothers), Turkish (8.2% of fathers and 7.2% of mother) and Surinamese (4.2 % of fathers and 3.04% of mothers) ethnic groups. Teacher reports were available for 269 students (132 boys and 137 girls). The teachers of the Dutch sample had on average 14.01 years of teaching experience (SD = 11.68). Among all teachers, eight teachers worked full time (i.e., more than 32 hours a week), nine teachers worked four days a week and 18 teachers worked three days a week.
Chinese sample. The Chinese sample consisted of 587 students (including 278
boys and 309 girls), and 14 teachers (including 13 females and one male) from three elementary schools in Zhejiang, China. The Chinese students were third to sixth graders and had a mean age of 11.49 years (SD = 1.29), ranging from 9 to 14 years old. Chinese students’ ethnicity was roughly divided into two categories, namely Han (i.e., the major ethnic group in China) and minority (including 55 minor ethnic groups in China, such as She Zu). Most of the students identified themselves (94.38%) and their parents (96.25% of the fathers and 94.72% of the mothers) as belonging to the Han group. In addition, 5.3% of students identified themselves and their parents (3.6% of the fathers and 4.1% of the mothers) as belonging to the minority group. The
teachers reported their affective relationship quality with 112 students (56 boys and 56 girls). They had on average 9.65 years of teaching experience (SD = 6.60). Twelve teachers worked full time, one teacher worked four days a week and one teacher worked two days a week.
Procedure
Dutch sample. We received approval from the ethical committee of the
University of Amsterdam (file number: 2016-CDE-7243) to carry out the data collection in the Netherlands. Data from the Dutch sample were collected by master thesis students under the supervision of the second and last author. The students contacted schools through their own networks via telephone and email. In case schools agreed to participate, teachers received information letters explaining the aim of this study as well as consent forms. Meanwhile, the parents of the students also received an information letter and could express their disagreement for their children’ participation in this study.
During school visits in February and March 2017, all students in the classroom were asked to fill out a questionnaire about their background information and their relationships with their teachers. In case students had multiple part-time teachers, they were instructed to report on their relationships with one of these teachers. It took the students around 30 minutes to finish the questionnaire and while students were filling out the questionnaire, teachers were asked to leave the classroom to encourage
students’ honest responses. Meanwhile, a researcher explained the procedure and answered students’ questions. The total response rate was 99% and nonparticipation was mainly due to students’ absence.
Around the same time of the school visits, the teachers were emailed and asked to complete a digital questionnaire, measuring their relationship quality with eight
randomly selected students (four boys and four girls for each teacher). It took the teachers around 40 minutes to complete the questionnaire and the teachers were asked to finish the questionnaire within two weeks.
Chinese sample. Although there is no ethical committee in China to approve this
study, the ethical committee of the University of Amsterdam stated that they would approve if the data were collected in the Netherlands. Meanwhile, a senior researcher in China was invited to review the proposal and drew a letter stating that the data collection procedure complied with the laws and rules in China. Data from the
Chinese sample were collected by the first author for the research master thesis, under the supervision of the second author. Schools were contacted by telephone. The principals of all contacted schools gave consent for the schools to participate in this study. Subsequently, within each school, classes from grade 3 to grade 6 were randomly selected and all students from the selected classes were invited to participate in this study. Different from the Netherlands, where primary school teachers teach all the subjects, in China, primary school teachers often teach one of the subjects, for example Chinese language. Compared to Dutch teachers, Chinese teachers might therefore spend less time with students. However, in China, a head teacher is in charge of a class and is responsible for the daily affairs and activities of the class (Chen, Rubin & Li, 1997). Since head teachers’ responsibilities are most comparable to those of Dutch teachers, they were invited to participate in this study. The head teachers received information letters about the purpose of this study and consent forms. The parents were also informed of the aim of this study and could express their disagreement for their children’s participation.
During school visits in March 2018, all students were asked to complete a questionnaire about their background information and quality of their relationship
with head teachers. The teachers were either asked to leave the classroom or not to intervene with students if they insisted to stay. The first author was present to explain the procedure and answer students’ questions. Among the sample, 99% of the students completed the questionnaire and nonparticipation was mainly due to students’ absence at the time of the data collection.
During the school visits, the teachers were asked to complete a questionnaire about their relationship quality with eight students (four boys and four girls per classroom). Most of the teachers completed the questionnaire during the school visits and two teachers, who were absent at the time of data collection, were emailed and asked to complete a digital questionnaire within two weeks. All invited teachers finished the questionnaires.
Measurements
Students’ perceived relationship quality. Third to sixth graders’ perceptions of
their relationship with their teacher was measured with the Student Perception of Affective Relationship with Teacher Scale (SPARTS; Koomen & Jellesma, 2015), which includes three subscales: Closeness, Conflict, and Negative Expectations. The Closeness subscale consists of 8 items and taps students’ positive feelings and reliance on teachers when they feel stressed, for example I feel relaxed with my
teacher. The Conflict subscale includes 10 items and describes students’ perceptions
of negative behaviours and distrustful feelings from teachers, for example I easily
have quarrels with my teacher. The Negative Expectations subscale consists of 10
items and taps students’ uncertain feelings about the relationship with the teacher, for example When I am with my teacher, I feel nervous. Students rated the extent to which these statements applied to their relationship with the teacher on a 5-point Likert scale, ranging from 1 (No, that is not true) to 5 (Yes, that is true).
The Closeness (Cronbach’s alpha ranging from .74 to .80), Conflict (Cronbach’s alpha ranging from .79 to .80) and Negative Expectation subscales (Cronbach’s alpha is .70) showed adequate reliability in previous studies (Koomen & Jellesma, 2015; Zee & de Bree, 2017; Zee & Koomen, 2017). The factorial, convergent, and concurrent validity of the SPARTS has also been supported in previous researches (Koomen & Jellesma, 2015; Zee & de Bree, 2017). The SPARTS was originally developed in Dutch. To measure the perceptions of the Chinese students, the English version was translated into Chinese by a researcher who is fluent in both languages. A back-translation procedure was applied to examine the precision of wording in the translated version. In our study, the Dutch SPARTS showed satisfactory reliability, with the Cronbach’s alphas being .84, .84 and .68 for the Closeness, Conflict and Negative Expectations respectively. In the Chinese SPARTS, the Closeness (! = .83) and Conflict (! = .72) subscales showed adequate reliability, whereas the Negative Expectations subscale showed inadequate reliability (! = .36). Therefore, the Negative Expectation subscale was omitted from further analyses.
Teachers’ perceived relationship quality. Teachers’ perceptions of their
relationship with the eight selected students were measured with the Dutch adaption of the Student Teacher Relationship Scale (STRS, Pianta, 2001; Koomen,
Verschueren, & Pianta, 2007; Koomen, Verschueren, van Schooten, Jak, & Pianta, 2012). The STRS also includes three dimensions to measure teachers’ affective relationships with individual students: Closeness, Conflict, and Dependency. The Closeness subscale taps the degree of warmth, trust and open communications in the relationship, for instance This child openly shares his/her feelings and experiences
with me. The Conflict subscale measures the degree of anger, negativity and
seems to be struggling with each other. The Dependency subscale measures the extent
to which the teacher experiences clingy and demanding behaviors from the child, for example This child reacts strongly to separation from me. The Closeness and Conflict subscales of the TSRS are comparable with the Closeness and Conflict subscale of the SPARTS whereas the Dependency subscale of STRS and the Negative Expectation subscale of the SPARTS tapped somewhat different constructs. Teachers indicated the degree to which each of the statements applied to their relationship with a particular child on a 5-point Likert scale, from 1 (Definitely does not apply) to 5 (Definitely
applies).
In the present sample, the shortened version of the Dutch STRS was used, which includes five items per dimension with the highest factor loadings reported in
previous researches (Koomen et al., 2007; Koomen et al., 2012). For the Chinese sample, we selected the same items as in the shortened Dutch version from the validated Chinese version of the STRS (Zhang, 2010). As three of these items are not available in the Chinese version, we translated them into Chinese with a
back-translation procedure.
The shortened and adapted Dutch version of STRS has showed satisfactory reliability in previous studies (Cronbach's alphas ranging between .85 and .93 for Closeness, .88 and .93 for Conflict, and .77 and .91 for Dependency; Koomen et al., 2007; Koomen et al., 2012; Zee & Koomen, 2017; Zee, Koomen & van der veen, 2013). The validity of the shortened version of the STRS has also been supported in previous research (Zee et al., 2013; Zee & Koomen, 2017). The complete Chinese version of the STRS showed satisfactory reliability (Cronbach’s alphas being .81 for Closeness, .83 for Conflict, and .67 for Dependency) and support for the validity was also provided (Zhang, 2010). In the present study, the Dutch STRS showed high
reliability, as the Cronbach’s alphas were .86 for Closeness, .89 for Conflict, and .82 for Dependency. The Chinese STRS showed adequate reliability, with the Cronbach’s alphas being .83 for Closeness, .67 for Conflict and .78 for Dependency.
Analyses
Analyses were conducted with the lavaan package in R version 3.3.1(Rosseel, 2012). To answer the research questions, structural equation modeling (SEM) was used. Model building consisted of three steps. First, we built multigroup models to test measurement invariance of both teachers’ and students’ perceptions of the
relationships across countries. Second, we compared the latent means of teachers’ and students’ perceptions of the relationships between the two samples. Third, we
estimated the correlations between the teacher reports and student reports and examined the cultural differences in the agreement. The details of each step are described in the following.
Measurement invariance. Separate measurement models with mean structures
were built for students’ and teachers’ perceptions. The proposed measurement model for students’ perceptions consisted of two latent factors, Closeness and Conflict. Given the small sample size for the teachers’ perceptions, we fitted three separate models to increase statistical power (for Closeness, Conflict and Dependency). To test weak invariance, we then constrained the factor loadings to be equal across groups and freely estimated the variance(s) of the common factor(s) in one group. To test strong invariance, we added equality constraints across groups to the intercepts and freely estimated the mean(s) of the common factor(s) in one group.
Cultural differences in relationship quality. By means of the models with
strong invariance, we could test the differences in latent means between the Dutch and Chinese samples. The means of the latent factors in the Dutch sample were fixed to zero and Z tests were used to examine whether mean differences between the Dutch sample and Chinese sample were different from zero. We standardized the latent means to ease interpretation of the magnitude of the mean group differences, using 0.2, 0.5 and 0.8 as rule of thumb for small, medium and large effects respectively (Cohen, 1988).
Cultural differences in agreement. Based on the models with (partial) strong
invariance, we extracted the factor scores and estimated the correlations between teachers’ and students’ perceptions in the Dutch sample and in the Chinese sample. Because Dependency was not included in the students’ perceptions, only Closeness and Conflict were considered. To compare correlation coefficients, we used the Fisher’s r to z transformation (Fisher, 1915; Fisher, 1921) and used a one-tailed Z-test to examine cultural differences in teacher-student agreement.
Model estimation and model comparison. Listwise deletion was used in all
models to deal with missing values as percentage of missingness is acceptable (9.23% of the students and 1.79% of the teachers missed information about at least one of the questions; Bennett, 2001; Schafer, 1999). To control for the nested structure of our data (i.e., students were nested within classes), we adjusted the standard errors and Chi-square values (Oberski, 2014). For all statistical tests, .05 was taken as the significance level.
To test exact model fit, Chi-square test was used. To test the overall goodness-of-fit, Root mean square error of approximation (RMSEA), Comparative Fit Index (CFI) and Standardized Root Mean Square Residual (SRMR) were used. RMSEA values lower than .08 (Browne & Cudeck, 1992; Hu & Bentler, 1999), CFI values higher than .90 (Bentler, 1992), and SRMR values lower than .10 (Hu & Bentler, 1999) were used as indicators of satisfactory fit. To compare nested models, we used Chi-square difference tests and included the CFI difference (∆CFI), RMSEA differences
(∆RMSEA) and SRMR difference (∆SRMR) as alternative indicators. Two models were considered equivalent when the Chi-square difference was not significant, when ∆CFI did not exceed .02 (Cheung & Rensvold, 1999; Rutkowski & Svetina, 2014), when ∆RMSEA did not exceed .03 (Rutkowski & Svetina, 2014) and when ∆SRMR did not exceed .03 for weak invariance and .01 for strong invariance (Chen, 2007). When three out of the four indicators indicated satisfactory fit or model equivalence, we proceeded to the next step.
Results Measurement Invariance in Student Perceptions
The model fit and model comparison statistics are provided in Table 1. The measurement model had a satisfactory fit and the model with weak invariance did not fit significantly worse than the measurement model, ∆c2 (16) = 89.32, p < .001,
∆CFI = .010, ∆RMSEA = .001, ∆SRMR =.014. However, the model with strong invariance fitted the data significantly worse than the model with weak invariance, ∆c2 (16) = 879.9, p < .001, ∆CFI = .121, ∆RMSEA = .028, ∆SRMR =.029.
Therefore, we checked the standardized mean residuals in both groups and
modification indices (MI) to locate the items showing non-invariance (Bentler, 1995; Bentler & Chou, 1993). As a result, we decided to free the equality constraints on the intercepts of item 1, 3, 5, 8, 10, 15, 20, and 23 (see Appendix). Item 1, 10, 15, 20 and 23 had a higher intercept in the Dutch group, indicating that given the same level of the latent constructs, Dutch students scored higher on these items. On the contrary, item 3, 5 and 8 had a higher intercept in the Chinese group, indicating that given the same level of Closeness or Conflict, Chinese students scored higher on these items. Freeing these parameters resulted in a model with partial strong invariance, and this model did not differ significantly from the model with weak invariance, ∆c2 (8) =
13.75, p = .089, ∆CFI = .011, ∆RMSEA = .001, ∆SRMR =.001. Therefore, the partial strong invariance did hold and this model was taken as the final model (see Figure 1 for the estimated coefficients).
Figure 1. The estimated coefficients of the partial strong invariance model for students’
perceptions
Note. ***: p <.001. The coefficients between brackets were standardized coefficients. Measurement Invariance in Teacher Perceptions
fit, c2 (10) = 27.96, p =.002, RMSEA = .130, CFI =.961, SRMR = .039. To improve
the model even further, we checked the correlation residuals and MI, and decided to add residual correlations between item 1 (i.e., This child openly shares his/her
feelings and experiences with me) and item 3 (i.e., If upset, this child will seek comfort from me) because these two items both tapped a child’s specific behavior in seeking
open communications with the teacher. The adjusted model fitted the data
satisfactorily, c2 (8) = 12.36, p =.136, RMSEA = .054, CFI =.993, SRMR = .021. The
model with weak invariance did not differ significantly from the adjusted measurement model, ∆c2 (4) = 3.21, p =.524, ∆CFI = .002, ∆RMSEA
= .013, ∆SRMR =.025. However, fit of the model with strong invariance deteriorated significantly compared to the model with weak invariance, ∆c2 (4) = 72.51, p < .001,
∆CFI = .112, ∆RMSEA = .136, ∆SRMR =.047. Checking the mean residuals and MI, we released the constraints on item 9, 12 and 15 (see Appendix). The model with partial strong invariance reached good fit, c2 (13) = 15.36, p =.286, RMSEA = .040,
CFI =.995, SRMR = .047, and did not significantly differ from the model with weak invariance, ∆c2 (1) =1.00, p = .318, ∆CFI = .000, ∆RMSEA = .001, ∆SRMR =.001.
The model with partial strong invariance was the final model for further analyses and the estimated coefficients were provided in Figure 2 (as well as the coefficients of Conflict and Dependency models).
The Conflict model. The measurement model of Conflict had a satisfactory fit
and the model with weak invariance did not differ significantly from the measurement model, ∆c2 (4) = 6.54, p =.163 , ∆CFI = .009, ∆RMSEA = .016, ∆SRMR =.019. The
model with strong invariance fitted the data significantly worse than the model with weak invariance, ∆c2 (4) = 20.87, p < .001, ∆CFI = .050, ∆RMSEA = .058, ∆SRMR
=.029. Checking the mean residuals and MI, we released the constraints on item 7 and 11 (see Appendix). The model with partial strong invariance had satisfactory fit, c2
(16) = 19.33, p = .252, CFI =.991, SRMR = .048, RMSEA = .050 and did not
deteriorate significantly from the model with weak invariance, ∆c2 (2) =.40, p = .820,
∆CFI = .006, ∆RMSEA = .019, ∆SRMR =.002. The model with partial strong invariance was taken as the final model for further analyses.
The Dependency model. The measurement model of Dependency did not reach
satisfactory fit, c2 (10) =32.04, p <.001, RMSEA = .149, CFI =.939, SRMR = .044.
Checking the correlation residuals and MI, we decided to add residual correlation between item 8 (i.e., This child reacts strongly to separation from me) and 10 (i.e.,
This child is overly dependent on me) as they both directly described a child’s
dependence/unwillingness of separation to the teacher. Adding the residual correlation led to a model with reasonable fit, c2 (8) =17.37, p =.030, CFI =.975, SRMR = .035,
RMSEA = .106. The model with weak invariance did not significantly differ from the adjusted measurement model, ∆c2 (4) = 4.05, p = .399, ∆CFI = .000, ∆RMSEA
= .019, ∆SRMR =.020. The model with strong invariance deteriorated significantly compared to the model with weak invariance, ∆c2 (4) = 23.54, p < .001, ∆CFI
= .061, ∆RMSEA = .052, ∆SRMR =.030. Checking the mean residuals and MI, we decided to release equal constraints on item 4 and 6 (see Appendix). The model with partial strong invariance reached satisfactory fit, c2 (14) =22.67, p = .066, RMSEA
= .084, CFI =.973, SRMR = .058, and did not significantly differ from the model with weak invariance, ∆c2 (2) = 2.60, p = .272, ∆CFI = .002, ∆RMSEA = .003, ∆SRMR
=.003, so it was taken as the final model for further analyses.
Figure 2. The estimated coefficients of partial strong invariance models for teachers’
perceptions
Table 1
Model Fit and Model Comparison Statistics for Model
c2 df p RMSEA 90% CI CFI SRMR ∆c2 ∆df p ∆CFI ∆RMSEA ∆SRMR Students’ Perceptions
Configural invariance 801.74 268 <.001 .055 [.050, .058] .925 .046
Weak invariance 891.05 284 <.001 .056 [.052, .060] .915 .060 89.32 16 <.001 .010 .001 .014 Strong invariance 1770.96 300 <.001 .084 [.081, .088] .794 .089 879.9 16 <.001 .121 .028 .029 Partial strong invariance 904.80 292 <.001 .055 [.051, .059] .904 .061 13.75 8 .089 .011 .001 .001
Teachers’ Perceptions - Closeness
Configural invariance 27.96 10 .002 .130 [.074, .188] .961 .039 Configural invariance.2 12.36 8 .136 .054 [.000, .125] .993 .021
Weak invariance 14.40 12 .276 .041 [.000, .108] .995 .046 3.21 4 .524 .002 .013 .025 Strong invariance 77.22 16 <.001 .177 [.139, .218] .883 .093 72.51 4 <.001 .112 .136 .047 Partial strong invariance 15.36 13 .286 .040 [.000, .105] .995 .047 1.00 1 .318 .000 .001 .001
Teachers’ Perception - Conflict
Configural invariance 12.76 10 .238 .053 [.000, .129] .994 .027
Weak invariance 19.69 14 .140 .069 [.000, .134] .985 .046 6.54 4 .163 .009 .016 .019 Strong invariance 42.07 18 .001 .127 [.077, .177] .935 .075 20.87 4 <.001 .050 .058 .029 Partial strong invariance 19.33 16 .252 .050 [.000, .119] .991 .048 .40 2 .820 .006 .019 .002
Teachers’ Perception - Dependency
Configural invariance 32.04 10 <.001 .149 [.093, .208] .939 .044 Configural invariance.2 17.37 8 .030 .106 [.035, .175] .975 .035
Weak invariance 20.04 12 .070 .087 [.000, .152] .975 .055 4.05 4 .399 .000 .019 .020 Strong invariance 43.46 16 <.001 .139 [.091, .190] .914 .085 23.54 4 <.001 .061 .052 .030 Partial Strong invariance 22.67 14 .066 .084 [.000, .144] .973 .058 2.60 2 .272 .002 .003 .003
Cross-Cultural Differences in Relationship Quality
Students’ perceptions. Based on the final model for students’ relationship
perceptions with partial strong invariance, we tested the mean differences of student-reported Closeness and Conflict across the Netherlands and China. For Closeness, the mean difference between the Chinese sample and Dutch sample was .69 and it was significant different from zero, ! = 3.98, ) < .001, 95% /0 = [ .56, .83], with a large effect (Cohen’s d = .83). In terms of Conflict, the mean difference between the two samples was -.42 and it significantly differed from zero, Z = -3.51, p < .001, 95% CI = [-.44, -.20], with a medium effect size (Cohen’s d = -.71). The results implied that Chinese students perceived higher level of Closeness and lower level of Conflict with teachers compared to Dutch students.
Teachers’ perceptions. Based on the models for teachers’ relationship
perceptions with partial strong invariance, we tested the mean differences of teacher-reported Closeness, Conflict and Dependency across the two countries. With regards to Closeness, the mean difference between the two samples was -.21 and it was not significantly different from zero, ! = 4.16, ) = .490, 95% /0 = [ .81, .39]. In terms of Conflict, the mean difference between the two samples was -.31 and it was significantly different from zero, ! = −1.04, ) = .047, 95% /0 = [−.61, −.04], with a medium effect size (Cohen’s d = -.57). With regards to Dependency, the mean difference was .39 and it did not significantly differ from zero, ! = 1.45, ) = .147, 95% /0 = [ −.14, .92]. The results indicated that Chinese teachers perceived lower level of Conflict and comparable levels of Closeness and Dependency with
students compared to Dutch teachers.
Agreement in Student and Teacher Perceptions Across Cultures
The estimated correlations between students’ and teachers’ perceptions towards Closeness and Conflict are provided in Table 2. The positive correlation between teacher-reported and student-reported Closeness was significantly higher in the Chinese group than the correlation in the Dutch group, Z = 1.98, p =.022. On the contrary, the correlation between teacher-reported and student-reported Conflict was significantly higher in the Dutch group compared to the Chinese group, Z = -2.43, p = .007.
Table 2
The Correlations Between the Students’ and Teachers’ Perceptions Towards Closeness and Conflict in the Netherlands and China
Closeness-T Conflict-T Closeness-S Conflict-S
Closeness-T * .06 .36 -.24
Conflict-T -.40 * -.11 .17
Closeness-S .16 -.32 * -.73
Conflict-S -.17 .42 -.83 *
Note. Closeness-T and Conflict-T refer to the teachers’ perceptions while Closeness-S
and Conflict-S refer to the students’ perceptions. The correlations in the Dutch sample are presented in the lower-left corner and the correlations in the Chinese sample are presented in the upper-right corner.
Discussion
In this study, we investigated the affective quality of teacher-student
relationships in the Netherlands, a Western, individualistic country and in China, an Eastern, collectivistic country. We examined the measurement invariance of teachers’ and students’ perceptions of the relationships across countries, and then tested the cultural differences in the teacher-reported and student-reported TSRs, as well as the
cultural difference in the agreement between the two reports.
Measurement Non-Invariance
First, we only found partial strong invariance for both students’ and teachers’ perceptions of the relationship, indicating that the items might be biased by cultural values and measured somewhat different constructs in Eastern and Western contexts. This is not surprising though, since previous cross-cultural studies about affective TSRs also failed to establish strong measurement invariance (Jia et al., 2009; Yang et al., 2013). Still, it might be relevant to take a closer look at the items showing
measurement non-invariance to investigate where cultural values might bias them. A plausible explanation is that collectivistic and individualistic cultures hold different expectations for interpersonal interactions. For example, in collectivistic cultures, individuals are expected to suppress emotions in interpersonal relationships (Matsumoto et al., 2008; Rubin et al., 2006; Wu et al., 2002) whereas in individualistic cultures, open and free expression of emotions are valued and encouraged (Matsumoto et al., 2008). This difference might account for the lower scores of the Chinese sample in the items related to directly sharing experiences and feelings (e.g., If I have a problem, I can share it with my teacher), and items with strong emotional tendency (e.g., I can be very angry with my teacher). Furthermore, compared to individualistic culture, collectivistic cultures emphasize more on
children’s respect and obedience to teachers (Bear et al., 2014; Triandis, 1989; Yang et al., 2013) and therefore Chinese teachers tend to be more critical towards children’s deviant behaviors in teacher-student interactions (Yang et al., 2013). This might
explain the higher score of the Chinese sample in items related to students’
misbehaviors (e.g., When this child is in a bad mood, I know we are in for a long and difficult day). These findings can serve as a basis for researchers who aim to develop
cross-cultural questionnaires. Our findings thereby suggest that special attention needs to be paid to cultural expectations when developing questionnaires to assess relationship quality.
Cross-Cultural Differences in Relationship Quality
Second, in contrast to our expectations of finding higher Closeness in China for both perceptions (Beyazkurk & Kesner, 2005; Jia et al., 2009; Yang et al., 2013), different results were found for students’ and teachers’ perceptions of Closeness. Chinese students perceived a higher level of Closeness with teachers than their Dutch counterparts. However, Chinese and Dutch teachers hold equally favorable
perceptions of Closeness with students. The result from the students’ reports was generally consistent with what Jia et al. (2009) and Yang et al. (2013) found in their studies, namely that students in China perceived better relationships with their teachers compared to American students. However, our finding with regard to teachers’ perceptions contradicted with findings from two previous studies, where Turkish kindergarten teachers reported more Closeness than their American
counterparts (Beyazkurk & Kesner, 2005) and American elementary school teachers reported better TSRs than their Chinese counterparts (Bear et al., 2014). Nevertheless, these studies included different instruments and different educational levels compared to our study and the results might therefore not be directly comparable. A plausible
reason for the different results we found regarding students’ and teachers’ perceptions might be the time that Dutch and Chinese teacher-student dyads spent with each other. Compared to teachers in the Western society, it seems that Chinese teachers spend on average more time with their students. While a normal school day in China lasts for eight hours (i.e. from 8 am to 4 pm) with additional tutoring after school, a school day in the Netherlands only lasts for about six hours (i.e. from 9 am to 3 pm), normally without further tutoring. However, Chinese teachers also need to deal with larger classes compared to Dutch teachers. While classes in China mostly comprise 40 to 50 students, Dutch classes are mostly only half the size, comprising between 20 to 30 students on average. Therefore, from students’ perspectives, Chinese students may spend more time with teachers and thus receive more support compared to Dutch students. From the teachers’ perspectives, Chinese and Dutch teachers may spend a similar amount of time with individual students, and thus have no difference in their perceptions of Closeness with a specific student.
As expected (Benjamin et al., 2001; Dennis et al., 2002), both Chinese students and teachers perceived less Conflict in their mutual relationships compared to the Dutch sample. Found in both perceptions, the cultural difference in Conflict is rather strong and might be due to the cultural contexts that teachers and students were embedded to. More specifically, the traditional Chinese culture values filial piety (i.e., respect of parents and family members), which puts an emphasis on children’s
compliance and obedience to parents (Bear et al., 2014; Yang et al., 2013). This value is also reflected in school settings, where students are expected to respect their
teachers and avoid deviant behaviors that might raise conflict (Bear et al., 2014; Yang et al., 2013). On the contrary, Western cultures often emphasize an egalitarian point of view with a lower power distance between teachers and students (Triandis, 1989; den Brok & Levy, 2006). As a result, teachers grant students a considerable freedom of action and students are more critical of the curriculum and teaching methods
(Boekaerts, 2003). Therefore, compared to Dutch students, Chinese students may (1) display more socially desirable behaviors (Yang et al., 2013), and (2) be less critical about their teachers (Jia et al., 2009), which might in turn contribute to less
observable conflict with teachers.
Different from our expectation of finding higher Dependency in Chinese teachers’ relationship perceptions (French et al., 2001), there was no significant cultural difference in teachers’ perceptions of Dependency. A plausible reason might be that cultural values simultaneously affect students’ dependent behaviors and teachers’ understanding about Dependency. For example, Chinese students may actually display more dependent behaviors but the emphasis of interdependence in collectivistic cultures (i.e. demanding more dependency from students) (Markus & Kitayama, 1991; Triandis et al., 1988) also increases Chinese teachers’ tolerance towards these behaviors. The two effects might cancel out and as a result there is no cultural difference in teachers’ perceptions towards Dependency. Another plausible explanation is that the larger class size in China makes it less likely for students to be overly dependent on teachers and therefore there is no cultural difference despite the emphasis of interdependence in collectivistic cultures. Future study is needed to look
into this dimension in cross-cultural contexts and preferably also include students’ perceptions of Dependency to have more information.
Our findings about cultural differences in relationship quality are profitable for both researchers and practitioners. For example, future studies could further explore the mechanisms behind the cultural differences by including behavioral measures such as students’ externalizing behaviors. Furthermore, teachers and students in both contexts may adjust their interaction patterns, for example, being more open and less critical in relationships, to improve the quality of mutual relationships.
Cross-Cultural Difference in Agreement
Third, partly different from our expectation of finding higher agreement in Chinese sample for both dimensions (Benjamin et al.,2001), Chinese teacher-student dyads displayed higher agreement on reporting Closeness and lower agreement on reporting Conflict compared to Dutch dyads. A plausible reason might be the emphasis on social harmony in collectivistic cultures. Maintaining a harmonious social interaction requires a full understanding of how others are feeling and thinking in the relationships (Markus & Kitayama, 1991), and relationship dyads in
collectivistic cultures might therefore be more sensitive about each other’s perceptions of the relationships. As a result, it might be more likely for Chinese teacher-student dyads to reach consensus about Closeness in their relationships. In addition, as individuals in collectivistic cultures tend to avoid bringing up conflict and there is a larger power distance between teachers and students (Triandis et al., 1988), it might be more difficult for teachers and students to learn about each other’s
dissatisfaction about the relationship. For example, a Chinese teacher may never know that the students felt unfairly treated as they would not say it out loud. On the contrary, the power distance between teachers and students is smaller in Western countries like the Netherlands (den Brok & Levy, 2006), and individuals tends to bring conflict to the open (Triandis et al., 1988). The conflict may be more salient in the Dutch context and therefore the agreement might be higher. As an inspiration, teacher-student dyads from both contexts may learn from each other and adjust their interacting strategies accordingly (e.g., attaching more importance to connectedness or being more open to Conflict) to reach a higher level of mutual understanding.
Limitations and Suggestions for Future Research
Some limitations need to be taken into account when interpreting the results of the present study. First, the teacher sample is somewhat smaller than the student sample and we might therefore have affected power to detect effects in the multigroup models (Chen, 2007). Future research including larger teacher sample is therefore needed. Second, the reliability of the Negative Expectations turned out to be inadequate in the Chinese sample and we were not able to make meaningful cross-cultural comparisons on this dimension. In addition, teachers’ and students’ perceptions on the other dimensions reached partial strong invariance, with some items showing measurement non-invariance. As a result, future studies could take a closer look at the scales so that they can be improved to function better in a cross-cultural context. Third, although we already included both teachers’ and students’ perceptions, they primarily reflected the mental representation of the relationships and
not necessarily the real time teacher-student interactions. Therefore, observational measures and qualitative data (e.g., in depth interviews) may be included in future studies to provide more information about the teacher-student interactions and help addressing how culture casts influences on the interactions. Forth, in this study, we mainly focused on students in middle childhood and future studies could further explore the cross-cultural difference across different levels of education, and favorably include younger children’s perceptions (e.g., kindergarteners), as little is known about their perceptions of the relationship quality with teachers.
Conclusion
The teacher’s perception towards Closeness, Conflict and Dependency, as well as students’ perceptions towards Closeness and Conflict reached partial strong
invariance across the Netherlands and China. In this study, Chinese primary school students perceived higher levels of Closeness and lower levels of Conflict with teachers compared to their Dutch counterparts. The Chinese teachers reported less Conflict and comparable levels of Closeness and Dependency as Dutch teachers did. Chinese teacher-student dyads had higher agreement in reporting Closeness and lower agreement in reporting Conflict than the Dutch dyads. The cultural differences we found might be due to the different interpersonal interaction patterns in collectivistic and individualistic cultures. Accordingly, teachers and students are advised to adjust their interaction patterns to improve the relationship quality. Future research is needed to further explore the cultural difference in TSRs, specifically regarding the
References
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment. Hillsdale, NJ: Erlbaum.
Aldridge, J., & Fraser, B. (2000). A Cross-cultural Study of Classroom Learning Environments in Australia and Taiwan. Learning Environments Research, 3, 101-134. doi: 10.1023/A:1026599727439
Bear. G. G., Yang., C., Glutting, J., Huang., X., He., X., Zhang., W., & Chen., D. (2014). Understanding Teacher-Student Relationships, Student-Student Relationships, and Conduct Problems in China and the United States. International Journal of
School & Educational Psychology, 2, 247-260. doi:
10.1080/21683603.2014.883342
Benjamin, W. J., Schneider, B. H., Greenman, P. S., & Hum, M. (2001). Conflict and childhood friendship in Taiwan and Canada. Canadian. Journal of Behavioural
Science, 33, 203-211. doi: 10.1037/h0087142
Bennett, D. A. (2009). How can I deal with missing data in my study? Australian and New Zealand Journal of Public Health, 25,464-468. doi:10.1111/j.1467-842X.2001.tb00294.x
Bentler, P. M. (1992). On the fit of models to covariances and methodology to the Bulletin. Psychological Bulletin, 112, 400–404.
Bentler, P. M., & Chou, C. P. (1993). Some new covariance structure model improvement statistics. Sage Focus Editions, 154, 235–255.
Berk, L. E. (2000). Child Development (5th ed.). Boston: Allyn and Bacon. 23-38 Beyazkurk, D., & Kesner, J. E. (2005). Teacher-Child Relationships in Turkish and
United States Schools: A Cross-Cultural Study, International Education Journal,
Boekaerts, M. (2003). Adolescence in Dutch culture: A self-regulation perspective. In F. Pajares & T. Urdan (Eds.), Adolescenc and education: Vol. 3. International
perspectives on adolescence (pp. 101–124). Greenwich, CT: Information Age.
Bowlby, J. (1980). Attachment and loss: Vol.3. Loss: sadness and depression. NewYork: BasicBooks.
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. London: Harvard University Press.
Browne, M. W. & Cudeck, R. (1992). Alternative ways of assessing model fit,
Sociological Methods & Research, 21, 230-258. doi: 10.1177/0049124192021002005
Chen, F. (2007), Sensitivity of goodness of fir indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14, 464-504. doi:10.1080/10705510701301834
Chen, X., Rubin, K. H. & Li, D. (1997), Relation Between Academic Achievement and Social Adjustment: Evidence from Chinese Children. Developmental Psychology,
33, 518-525.
Cheung., G. W., & Rensvold, R. B. (1999). Testing factorial invariance across groups: A reconceptualization and proposed new method. Journal of Management, 25, 1-27. doi:10.1177/014920639902500101
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Dennis, T. A., Cole, P. M., Zahn-Waxler, C., & Mizuta, I. (2002), Self in context: Autonomy and relatedness in Japanese and U.S. mother–preschooler dyads. Child
Development, 73, 1803- 1817. doi: 10.1111/1467-8624.00507
Cross-cultural research on parenting and psychological adjustment of children. Journal
of Child and Family Studies, 19, 8-15. doi: 10.1007/s10826-009-9335-1
den Brok, P., & Levy, J. (2006), Teacher-student relationships in multicultural classes: Reviewing the past, preparing the future. International Journal of Educational
Research, 43. 72-88. doi: 10.1016/j.ijer.2006.03.007
Fisher, R. A. (1915). Frequency Distribution of the Values of the Correlation Coefficient in Samples from an Indefinitely Large Population. Biometrika, 10, 507-521. doi:10.2307/2331838
Fisher, R. A. (1921). On the probable error of a coefficient of correlation deduced from a small sample. Metron, 1, 1–32.
French, D. C., Rianasari, M., Pidada, S., Nelwan, P., & Buhrmester, D. (2001). Social support of Indonesian and U.S. children and adolescents by family members and friends. Merrill-Palmer Quarterly, 47, 377–394. doi: 10.1353/mpq.2001.0015 Hofstede, G. (1980). Culture and Organizations. International Studies of Management
& Organization, 10, 15-41. doi: 10.1080/00208825.1980.11656300
Hofstede, G. (1986). Cultural Differences in Teaching and Learning. International
Journal of Intercultural Relations, 10, 301-320. doi:
10.1016/0147-1767(86)90015-5
Hu, L., & Bentler., P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation
Modeling: A Multidisciplinary Journal, 6, 1-55. doi:
10.1080/10705519909540118
Hughes, J. N. (2011). Longitudinal effects of teacher and student perceptions of teacher-student relationship qualities on academic adjustment. The Elementary School
Hughes, J. N., & Cavell, T. A. (1999). Influence of the teacher-student relationship in childhood conduct problems: A prospective study. Journal of Clinical Child
Psychology, 28, 173-184. doi: 10.1207/s15374424jccp2802_5
Jia, Y., Way, N., Liang, G., Yoshikawa, H., Chen, X., Hughes., D., Ke, X., & Lu, Z. (2009). The Influence of Student Perceptions of School Climate on Socioemotional and Academic Adjustment: A Comparison of Chinese and American Adolescents. Child Development, 80, 1514-1530. doi: 10.1111/j.1467-8624.2009.01348.x
Jin, M. K., Jacobvitz, D., Hazen, N., & Jung, S.H. (2012). Maternal sensitivity and infant attachment security in Korea: Cross-cultural validation of the Strange Situation. Attachment & Human Development, 14, 33-44. doi: 10.1080/14616734.2012.636656
Keller, H. (2013). Attachment and culture. Journal of Cross-Cultural Psychology, 44, 175-194. doi: 10.1177/0022022112472253.
Koomen, H. M. Y., & Jellesma, F. C. (2015). Can closeness, conflict, and dependency be used to characterize students’ perceptions of the affective relationship with their teacher? Testing a new child measure in middle childhood. British Journal of
Educational Psychology, 85, 479-497. doi: 10.1111/bjep.12094
Koomen, H., Verschueren, K., & Pianta, R. (2007). Leerling Leerkracht Relatie Vragenlijst (LLRV).
Koomen, H. M. Y., Verschueren, K., van Schooten, E., Jak, S., & Pianta, R. C. (2012). Validating the Student-Teacher Relationship Scale: Testing factor structure and measurement invariance across child gender and age in a Dutch sample. Journal
of School Psychology, 50, 215–234. doi:10.1016/j.jsp.2011.09.001
Emotion, and Motivation. Psychological Review, 98, 224-253. doi: 10.1037/0033-295X.98.2.224
Matsumoto, D., Yoo, S. H., Nakagawa, S., & Multinational Study of Cultural Display Rules. (2008). Culture, emotion regulation, and adjustment. Journal of
Personalibity and Social Psychology, 94, 925-937. doi:
10.1037/0022-3514.94.6.925
Mizuta, I., Zahn-Waxler, C., Cole, M., & Hiruma, N. (1996). Cross-cultural Study of Preschoolers' Attachment: Security and Sensitivity in Japanese and US Dyads.
International Journal of Behavioral Development, 19, 141-160. Doi:
10.1080/016502596385992
Oyserman, D., Coon, H. M., & Kemmelmeier, M. (2002). Rethinking individualism and collectivism: Evaluation of theoretical assumptions and meta-analyses.
Psychological Bulletin, 128, 3-72. doi: 10.1037/0033-2909.128.1.3
Oberski, D. (2014), lavaan.survey: A R Package for Complex Survey Analysis of Structural Equation Models. Journal of Statistical Software, 57, 1-22.
Pianta, R.C. (Ed.) (1992). New directions for child development: Vol. 57. Beyond the
parent: The role of other adults in children’s lives. San Francisco, CA: Jossey Bass.
Pianta, R. C. (2001). STRS Student Teacher Relationship Scale. Odessa, FL: Psychological Assessment Resources.
Pianta, R. C., Hamre, B., & Stuhlman, M. (2003). Relationships between teachers and children. In W. M. Reynolds, G. E. Miller & I. B. Weiner (Eds.), Handbook of
psychology: Educational psychology, 7, 199-234. doi:
10.1002/0471264385.wei0710
Rey, B., Smith, A. L., Yoon, J., Somers C., & Barnett, D. (2007), Relationships between teachers and urban African American children: The role of informant. School
Psychology International, 28, 346-364. doi: 10.1177/0143034307078545
Roorda, D. L., Jak, S., Zee, M., Oort, F. J. & Koomen, H.M.Y. (2017), Affective Teacher–Student Relationships and Students’ Engagement and Achievement: A Meta-Analytic Update and Test of the Mediating Role of Engagement. School
Psychology Review, 46, 239-261. doi: 10.17105/SPR-2017-0035.V46-3
Roorda, D. L., Koomen, H. M. Y., Spilt J. L., Oort, F. J. (2011). The Influence of Affective Teacher–Student Relationships on Students’ School Engagement and Achievement: A Meta-Analytic Approach, Review of Educational Research, 81, 493-529. doi: 10.3102/0034654311421793
Rosseel., Y. (2012). Lavaan: A R package for structural equation modeling and more.
Journal of Statistical Software, 48, 1-36.
Rothbaum, F., Weisz, J., Pott, M., Miyake, K., & Morelli, G. (2000). Attachment and culture: Security in the United States and Japan. American Psychologist, 55, 1093-1104. doi: 10.1037/0003-066X.55.10.1093
Rubin, K. H. (1990). Introduction: Peer relationships and social skills in childhood – an international perspective. Human Development, 33, 221-224.
Rutkowski, L., & Svetina, D. (2014), Assessing the hypothesis of measurement invariance in the context of large-scale international surveys. Educational and
Psychological Measurement, 71, 31-57. doi: 10.1177/0013164413498257
Schafer, J. L. (1999). Multiple imputation: a primer. Statistical Methods in Medical
Research. 8, 3–15. doi: 10.1191/096228099671525676.
Takahashi, K. (1986), Examining the strange-situation procedure with Japanese mothers and 12-month-old infants. Developmental Psychology, 22, 265-270. doi: 10.1037/0012-1649.22.2.265
Personality, 69. 907-924. doi: 10.1111/1467-6494.696169
Triandis, H. C., Bontempo, R., Villareal, R., Marcelo., J., Masaaki., A., & Nydia., L.(1988). Individualism and collectivism: Cross-cultural perspectives on self-ingroup relationships. Journal of Personality and Social Psychology, 54, 323-338. Verschueren, K. (2015). Middle childhood teacher–child relationships: Insights from
an at- tachment perspective and remaining challenges. In G. Bosmans & K. A. Kerns (Eds.), Attachment in middle childhood: Theoretical advances and new
directions in an emerging field. New Directions for Child and Adolescent Development, 148, 77–91.
Verschueren, K., & Koomen, H.M.Y. (2012). Teacher-child relationships from an attachment perspective, Attachment & Human Development, 14, 205-211.doi: 10.1080/14616734.2012.672260
Voronov, M., & Singer, J. A., (2002). The Myth of Individualism-Collectivism: A Critical Review, The Journal of Social Psychology, 142, 461-480. doi: 10.1080/00224540209603912
Wu, P., Robinson, C. C., Yang, C., Hart, C. H., Olsen, S. F., Porter., C. L., Jin, S., Wo, J., & Wu, X. (2002), Similarities and differences in mothers’ parenting of preschoolers in China and the United States. International Journal of Behavioral
Development, 26, 481-491. doi: 10.1080/01650250143000436
Yang., C., Bear., G. G., Chen., F., Zhang., W., Blank., J. C., & Huang, X. (2013), Students’ Perceptions of School Climate in the U.S. and China. School Psychology
Quarterly, 28, 7-24. doi: 10.1037/spq0000002
Zee., M., & de Bree., E. (2017), Students’ self-regulation and achievement in basic reading and math skills: the role of student-teacher relationship in middle childhood, European Journal of Development Psychology, 14, 265-280. doi:
10.1080/17405629.2016.1196587
Zee., M., & Koomen., H.M.Y. (2017), Similarities and dissimilarities between teachers' and students' relationship views in upper elementary school: The role of personal teacher and student attributes, Journal of School Psychology, 64, 43-60. doi: 10.1016/j.jsp.2017.04.007
Zee, M., Koomen., H.M.Y., van der Veen, I. (2013). Student-teacher relationship quality and academic adjustment in upper elementary school: The role of student personality. Journal of School Psychology, 51, 517-533. doi: /10.1016/j.jsp.2013.05.003
Zevalkink, J., Riksen-Walraven, J. M., Van Lieshout, C.F.M. (1999), Attachment in the Indonesian Caregiving Context. Social Development, 8, 21-40. doi: 10.1111/1467-9507.00078
Zhang, X. (2010), [Reliability and Validity of Teacher-Child Relationship Scale], Chinese Journal of Clinical Psychology, 18, 582-583. doi: 1005-3611(2010)05-0582-02
