Chinese and European American Undergraduates’ Perceptions of Maternal Warmth and Negativity as Predictors of Self-Esteem and Life Satisfaction

Abstract

This study examined whether culture moderated the associations between Chinese and European American undergraduates’ perceptions of parental warmth and negativity, and how these perceptions predicted undergraduates’ levels of self-esteem and subsequent life satisfaction. Participants consisted of 207 undergraduates (95 female, 112 male) from Wuhu, Anhui, China, and 129 (75 female, 54 male) European American undergraduates from Columbia, Missouri, the United States. Path analyses were conducted to examine the mediating role of self-esteem in the associations between parental warmth and negativity and young adults’ life satisfaction. Maternal warmth and negativity were more strongly negatively associated in the European American group than in the Chinese group; moreover, both warmth and negativity predicted self-esteem in the Chinese group, but only negativity predicted self-esteem in the European American group. Self-esteem significantly predicted life satisfaction in both groups, but the association was significantly stronger in the European American group. The results suggest cultural differences in the links between parental warmth and negativity, self-esteem, and subsequent life satisfaction. Therefore, it is important to consider parental warmth and negativity as separate constructs, because their associations with each other and with children’s levels of adjustment may vary across cultures.

Keywords

culture parent–child relations self-esteem life satisfaction

Research on parental warmth and negativity (parenting that displays negative regard toward children, such as anger) has tended to focus on how warmth is universally associated with better adjustment in children, and how parental negativity is universally associated with lower levels of adjustment. Typically, that research combines warmth and negativity, one of them reverse-scored, into a unidimensional measure of warmth and negativity. These may be at the item level, where items reflecting warmth are combined with items reflecting negativity, reverse-scored (e.g., Carlson & Harwood, 2003; Greenberger & Chen, 1996), or at the level of subscales where a warmth subscale is reverse-scored and combined with subscales reflecting rejection (e.g., Rohner, Khaleque, & Cournoyer, 2005; for a meta-analysis, see Khaleque & Rohner, 2012). Rohner’s research in particular is important, as it has established the universal importance of parental warmth and rejection as relevant to children’s adjustment. The present study sought to build on these findings by examining warmth and negativity as separate dimensions, as opposed to opposite ends of one continuum. We argue that perceptions of parental warmth and negativity might be strongly negatively associated in European American samples, but that the negative association between warmth and negativity should be weaker in East Asian cultures, including China. If this is the case, parental warmth and negativity should be more likely to predict unique variance in children’s adjustment in East Asian versus U.S., European American samples. We discuss these ideas below, and then consider the potential associations between parental warmth and negativity, children’s self-esteem, and children’s life satisfaction.

Cultural Variability in the Correlates of Parental Warmth and Negativity

Our first proposition has to do with the possibility that parental warmth and negativity may be differentially associated in groups of East Asian and European origin. This possibility is suggested by social-psychological research that has found that negative associations between positive and negative affect are stronger in groups of undergraduates from the United States as opposed to those from East Asian cultures. Bagozzi, Wong, and Yi (1999) argued that this is because in groups that emphasize independence, the behavior of oneself and others is thought to be caused by a self that has inner attributes, including affect-based motivational states that are consistent across situations. In groups that emphasize interdependence, however, the main influence of behavior has do with the rules for correct behavior in a given context, and less thought is devoted to considering the consistent, inner qualities of the self. Thus, positive and negative affect may be less strongly negatively associated and may co-occur more frequently. A number of studies, all regarding undergraduates reporting on their own affect, have found results consistent with this phenomenon (Bagozzi et al., 1999; Kitayama, Markus, & Kurokawa, 2000; Schimmack, Oishi, & Diener, 2002).

An alternative explanation of the findings of Bagozzi et al. (1999) is offered by Schimmack et al. (2002). They argue that differences in associations between positive and negative affect may be caused not by variations in independence and interdependence, but by variations in dialectical thinking. Dialectical thinking has to do with the idea that opposites are compatible and part of each other; thus in daily life, opposing ways of thinking and feeling may be more likely to coexist at any point in time (Spencer-Rodgers, Peng, Wang, & Hou, 2004). Whether due to independence-interdependence, or dialectical thinking, it is clear that negative and positive affect are less strongly negatively associated for East Asians than for European Americans.

The above research has important implications for parental warmth and negativity, although, to our knowledge, the issue has not yet been examined. Often, research on perceptions of parental warmth and negativity toward children combines the two constructs into one overall measure, with one construct reverse-scored. This has been done at the item level (Carlson & Harwood, 2003; Greenberger & Chen, 1996) as well as at the level of subscales (Khaleque & Rohner, 2012; Putnick et al., 2012). In this study, we examined parental warmth and negativity separately. We expected the associations between undergraduates’ reports of their parents’ warmth and negativity to be less strongly negative in a Chinese sample than in a sample of European Americans.

There are important ramifications of this possibility. First, many studies in the parenting literature assume that a measure of one dimension, reverse-scored, reflects the other dimension. Thus, low scores on a measure of parental negativity are sometimes considered to reflect higher levels of warmth (e.g., Luis, Varela, & Moore, 2008). This interpretation might be valid for European American samples, but may not hold for samples of Asian background. Second, if parental warmth and negativity are less strongly negatively associated in Asian groups, they might be more likely to account for unique variation in children’s outcomes when included in one model as predictors because they contain less shared variance. Because much cross-cultural research assesses parental warmth and negativity and combines them (with one dimension reverse-scored), the unique contributions of warmth and negativity may be obscured, especially in cultures where warmth and negativity are weakly associated. In this study, we sought to test this possibility by using undergraduates’ perceptions of parental warmth and negativity to predict their own perceptions of self-esteem, and subsequent life satisfaction.

Perceptions of Parental Warmth and Negativity as Predictors of Self-Esteem and Life Satisfaction

We expected parental warmth and negativity both to predict unique variation in self-esteem in the Chinese group, whereas in the European American group, we expected only one of these variables to predict self-esteem (when these variables are placed in the same model and are therefore controlling for one another). Again, this is because in the European American sample, perceptions of parental warmth and negativity should be strongly negatively associated and thus might not predict unique variance in children’s outcomes. Generally, evidence suggests that parental warmth and negativity are important predictors of self-esteem in both China and the United States (e.g., Farruggia, Chen, Greenberger, Dmitrieva, & Macek, 2004; Robertson & Simons, 1989; Yao, He, Ko, & Pang, 2014). Whether parental warmth and negativity might better predict self-esteem in China versus the United States is open to question. On one hand, there is research that suggests that self-esteem may be more likely to be promoted in the United States by parents who are warm. Chao (1995), in qualitative research, found that when asked what was important in raising children, both Chinese American and European American mothers were likely to state that love and warmth were important variables. European American mothers, however, were much more likely to state that the purpose of being loving was to help to develop a strong sense of self-confidence and self-esteem in their children. Chinese American mothers were more likely to state that the purpose of being loving was to establish a good relationship with their children. Thus, one might expect perceptions of parental warmth and negativity to predict levels of self-esteem more strongly in European American, as opposed to Chinese undergraduates.

On the other hand, there is evidence to suggest that parental warmth and negativity may be particularly relevant to groups of East Asian origin as opposed to groups that are primarily European American. Oishi et al. (2007), for example, found that Asian American participants were much more likely to stress the importance of parental approval than were European Americans. Asian Americans were also more likely to recall events that led to parental approval as opposed to events that were irrelevant to parental approval, whereas this distinction was not apparent for European Americans. This finding suggests that even though self-esteem may be more valued by European Americans, the relative importance of parental approval and warmth may be greater for Chinese undergraduates. Thus, the relationship between parental warmth and negativity and self-esteem may be of similar strength in both groups, but for different reasons—because self-esteem is relatively important for European Americans and because the parent–child relationship is relatively important for Chinese individuals.

We also sought to test whether self-esteem might play a mediating role between parental warmth/negativity and life satisfaction for both groups. We hypothesized that this indirect effect would be significant, based on previous findings that self-esteem is associated with life satisfaction in both Asian and European American samples, but that the association is typically stronger for European Americans (e.g., Diener & Diener, 1995; Kang, Shaver, Sue, Min, & Jing, 2003; Kwan, Bond, & Singelis, 1997). This is because in U.S. culture, an important route to life satisfaction is through the process of self-enhancement (Diener & Diener, 1995; Heine, 2015). On the contrary, in non–European American cultures, and in particular, Asian cultures, an important route to life satisfaction is to fit in harmoniously with others. Thus, Kwan et al. (1997) found that while both self-esteem and relationship harmony were significantly positively associated with life satisfaction for undergraduates from the United States and Hong Kong, the association between self-esteem and life satisfaction was significantly stronger in the U.S. sample than the Hong Kong sample, and the association between relationship harmony and life satisfaction was stronger for the Hong Kong sample than the U.S. sample.

Some studies have touched on the mediating role that self-esteem might play between parental warmth/negativity and life satisfaction. Stewart and colleagues (1998) found some evidence consistent with the possibility that self-esteem mediated the relationship between maternal warmth and life satisfaction in a sample of Chinese undergraduates; they found that the significant association between maternal warmth and life satisfaction became nonsignificant after controlling for levels of self-esteem. This approach, however, did not directly assess mediation, as was done in the present study. Kang et al. (2003) examined the general quality of interpersonal relationships rather than parental warmth and negativity. They argued that relationship qualities, including close relationships, should predict self-esteem, which in turn should predict higher levels of life satisfaction, particularly in cultures that emphasize collectivism. They found that relationship quality was significantly positively associated with self-esteem for Asian American, Korean, and Chinese undergraduates but not European American undergraduates. Self-esteem in turn was positively associated with life satisfaction for all groups, with the association significantly stronger for the European American group. While Kang and colleagues did not find a significant association between the general quality of interpersonal relationships and self-esteem in the European American group, research with more individualist samples, including the U.S. sample, has found parental warmth and negativity to be associated with self-esteem (e.g., Rudy & Grusec, 2006).

The Present Study

In the present study, we examined the ideas discussed above in two groups of undergraduates: one comprised of European American undergraduates and the other of undergraduates from China. We assessed participants’ perceptions of their parents’ warmth and negativity directed toward them as they grew up, as well as their current levels of self-esteem and life satisfaction. We assessed perceptions of parental warmth and negativity in the time growing up, as opposed to current emotions, because the nature of parent–child relationships changes as children transition to young adulthood (see Aquilino, 1997). However, parental emotions throughout childhood are important in shaping young adults’ outcomes (Aquilino, 1997; Khaleque & Rohner, 2012).

To test these ideas, we conducted a path analysis, as depicted in Figure 1. We used this model to test four hypotheses. Hypothesis 1 was that parental warmth and negativity would be more strongly negatively associated in the European American sample than in the Chinese sample. Hypothesis 2 was that, because we expected parental warmth and negativity to be less strongly negatively associated in the Chinese sample, we expected both variables to predict unique variance in self-esteem for that group. On the contrary, because we expected parental warmth and negativity to be strongly negatively associated in the European American group, we expected only one of the two variables to account for unique variation in self-esteem for that group. Hypothesis 3 was that self-esteem would mediate the associations between parental warmth/negativity and life satisfaction in both groups. We also tested for direct effects of parental warmth/negativity on life satisfaction, but did not make predictions as to whether self-esteem would fully or only partially mediate relations between parental warmth/negativity and life satisfaction. Hypothesis 4 was that self-esteem would more strongly predict life satisfaction in the European American group as opposed to the Chinese group.

Figure 1.

Relationships between participants’ perceptions of maternal warmth and negativity and their own levels of self-esteem and life satisfaction.

Method

Participants and Procedure

Participants were recruited from undergraduate classes in middle-class universities from their respective countries. Those included in the analyses reported on their age (30 European American and three Chinese participants did not report their age). Initial analyses revealed that participants older than 26 were outliers on age; these participants were dropped from the analyses due to their outlier status and the fact that while our definition of child had to do with relational status rather than age, these participants did not correspond well to the concept of child. Thirteen European American participants, ranging in age from 27 to 50, were excluded based on this criterion; no Chinese participants were excluded from the analyses. The remaining participants were 207 undergraduates (95 female, 112 male) from Wuhu, Anhui, China, and 129 (75 female, 54 male) European American undergraduates from Columbia, Missouri, the United States. Wuhu is an urban area in China, with approximately 3.8 million people; Columbia, Missouri, is a college town in the Midwest of approximately 120,000 people. Of this sample, the European American participants were significantly older than participants from China (for European Americans, M = 21.09, SD = 1.44; for Chinese participants, M = 18.86, SD = 1.130), F(1, 334) = 249.18, p < .001. Because of the age difference between the samples, age was statistically controlled in all analyses. However, the results were identical with and without this control. Other than age, missing data were dealt with using multiple imputation. There was very little missing data, however. Missing data for the variables ranged from 1 to 10 cases.

Participants also reported on their parents’ level of education. Mothers in the Chinese group were significantly less educated than their counterparts from the United States, χ²(4) = 202.09, p < .001. Seventy-one percent had less than a high school education, 22% had a high school degree, 2% had a degree from a technical school, 5% had a college degree, and none had a graduate degree. For the U.S. sample, 1% had less than a high school degree, 27% had a high school degree, 5% had a degree from a technical school, 47% had a college degree, and 20% had a graduate degree. Fathers from China were also less educated than their counterparts from the United States, χ²(4) = 142.71, p < .001. Forty-six percent had less than a high school degree, 40% had a high school degree, 6% had a degree from a technical school, 9% had a college degree, and none had a graduate degree. For the U.S. sample, 4% of fathers had less than a high school degree, 25% had a high school degree, 5% had a degree from a technical school, 48% had a college degree, and 19% had a graduate degree.

The Chinese sample was similarly educated to the average Chinese sample. UNESCO data using the International Standard Classification of Education showed that for men and women older than 25 in China, 78% of individuals had less than a high school equivalent degree, 14% had a high school degree, 8% had a college degree, and 1% had a graduate degree (United Nations Statistics Division, 2017; data regarding technical schools were not reported). The U.S. sample was more highly educated than the average American sample. U.S. census data showed that for men and women older than 25 in the United States, 11.6% had less than a high school degree, 46.1% had a high school degree, 9.8% had a degree from a technical school, 20.55% had a college degree, and 12% had a graduate degree (U.S. Census Bureau, 2016).

Measures

Maternal warmth and negativity

Participants reported on their perceptions of parental warmth and negativity by responding to items taken from the Child-Rearing Practices report (CRPR; Block, 1981). The items were adapted so that participants reported on their parents’ practices and responded to the items thinking about their years growing up (from Buri, 1991).

Negativity

The three-item measure from the CRPR negative affect toward child scale was used to assess parental negativity (e.g., “My parent often felt angry with me”), with one item added to improve reliability (“As I was growing up, my parent often became upset with me”). Participants rated each item on a scale from 1 = strongly disagree to 5 = strongly agree. One item loaded significantly differently for Chinese and European American participants when reporting on mothers and thus was not included in the scale (“There was often a good deal of conflict between my parent and me,” which loaded significantly only for European Americans). Cronbach’s alphas for this scale were .65 and .69 for the Chinese and European American groups, respectively. Items were averaged to create a score that could range from 1 to 5.

Warmth

Warmth was assessed with five items from the Open Expression of Affect Scale from the CRPR (e.g., “My parent felt I should be given understanding when I was scared or upset”). Participants rated each item on a scale from 1 = strongly disagree to 5 = strongly agree. One item was removed for conceptual reasons because it reflected the open expression of anger. For reports on maternal warmth, the chi-square values for models constraining loadings to be equal across groups were not significantly different from the unconstrained models; thus, all items were retained. Cronbach’s alphas for this scale were .59 and .82 for the Chinese and European American groups, respectively. Despite the low alpha for the Chinese group, the measure for this group was associated in expected ways with the other variables of interest for this group, as presented in the “Results” section, establishing its validity in that group. Items were averaged to create a score that ranged from 1 to 5.

Self-esteem

Self-esteem was assessed with Rosenberg’s (1989) 10-item measure. The scale contains a number of items regarding ideas about self-worth; half of the items are worded positively (e.g., “On the whole, I am satisfied with myself”) and half are worded negatively (e.g., “At times I think I am no good at all”). Participants completed each item on a scale from 1 = strongly disagree to 4 = strongly agree. Chi-square values for models constraining loadings to be equal across groups were not significantly different from the unconstrained models; thus, all items were retained. Alphas for this scale were .78 and .89 for the Chinese and European American groups, respectively. Items were averaged to create a mean score.

Life satisfaction

Life satisfaction was assessed with the Satisfaction With Life Scale (SWLS), consisting of five items (Diener, Emmons, Larsen, & Griffin, 1985). Participants rated each item on a scale from 1 = strongly disagree to 5 = strongly agree. Chi-square values for models constraining loadings to be equal across groups revealed that one item loaded significantly more strongly for European American than for Chinese participants when reporting on mothers (“In most ways, my life is close to ideal”); thus, this item was not retained. Alphas for the revised scale were .69 and .74 for the Chinese and European American groups, respectively. Items were averaged to create a mean score.

Translation

For the Chinese sample, questionnaire instructions were translated from English into Mandarin and then back-translated into English by the fourth author and a research assistant. Discrepancies in meaning were discussed and addressed, to arrive at questionnaires expressing the same meanings in both languages. Where measures already existed in Mandarin, we used the existing scales. The CRPR has been administered to Chinese samples (Chen, Dong, & Zhou, 1997), as have measures of life satisfaction and self-esteem (Kwan et al., 1997).

Metric invariance

Prior to the main analyses, a series of confirmatory factor analyses examining metric invariance were conducted on each subscale for all measures included in the study, with all items hypothesized to load on one main factor. Models with the items constrained to load equally on the scale in both groups were compared with models where this constraint was not specified. Each specific path was constrained to be equal across the two groups. When the difference in the χ² values for the two models were significant, items that loaded significantly differently were dropped from the analysis on an iterative basis, with the most significantly different items dropped first. We repeated this process until the chi-square values of the constrained and unconstrained models were not significantly different (Knight, Roosa, & Umaña-Taylor, 2009). Importantly, measurement equivalence could not be established for reports of paternal negativity. Thus, the analyses presented below are for perceptions of maternal warmth and negativity only. Because of the large number of parameters (as many as 11 indicators on one factor) and the multigroup design, the current study relied on observed variables and utilized path analysis as opposed to structural equation modeling (see Wolf, Harrington, Clark, & Miller, 2013).

Results

Preliminary Analyses

The means and standard deviations, as well as correlations for all variables, are presented in Table 1. The path analyses, presented in Figure 1, used multiple imputation for missing data and maximum likelihood estimation with robust standard errors in MPlus (MLR; Muthén & Muthén, 1998-2010). Age, gender, and parental education were initially entered as covariates but only gender was significantly associated with some variables predicted in the model. There was also a nonsignificant trend for age to be negatively associated with life satisfaction for the European American group and so this variable was included in the model. While there were significant correlations between parental education and some variables (see Table 1), parental education did not significantly predict any of the variables in the model and model fit was worse when this parental education was included as a covariate; thus, it was not included in the model.

Table 1.

Correlations, Means, and Standard Deviations of All Variables Used in Path Analysis.

	1.	2.	3.	4.	5.	6.	7.
WM		−.27***	.26***	.22	−.17*	−.09	.23***
NG	−.59***		−.27***	−.12	.01	−.04	.06
SE	.26	−.27		.42	.19	.12	.08
LS	.22	−.12	.60		.09	.06	.13
Gen	−.04	−.03	.05	.09		.21**	−.06
Age	−.02	−.05	.04	−.08	.14		−.11
ME	.12	−.06	−.03	.06	−.09	−.28**
Chinese sample, M (SD)	.230 (0.56)	3.69 (0.88)	4.10 (0.76)	1.95 (0.85)	0.41 (0.49)	18.86 (1.13)	1.40 (0.75)
U.S. sample, M (SD)	2.88 (0.69)	2.46 (0.73)	3.49 (0.67)	1.92 (0.83)	0.54 (0.50)	21.09 (1.44)	3.44 (1.15)

Note. Correlations below the diagonal are for European Americans, those with an absolute value of .18 or higher are significant; correlations above the diagonal are for Chinese undergraduates, those with an absolute value of .14 or higher are significant. Females are coded as 0, males as 1. All values are from the raw data rather than from the multiply imputed datasets. SE = Self-Esteem; LS = Life Satisfaction. For Gender, ME = maternal education; NG = Maternal negativity; WM = maternal warmth.

In our initial analyses, we compared a model that constrained the paths to be equal across groups with one where the paths were unconstrained. The Wald χ² test of parameter equalities revealed that, as expected, the constrained model was significantly different from the unconstrained model, χ²(6) = 34.75, p < .001. Thus, we tested a model that did not constrain the relationships between the variables to be equal across groups. The model fit in path analysis is considered good if the comparative fit index (CFI) is .95 or greater (fit is adequate at .90 or greater), the root mean square error of approximation (RMSEA) is less than or equal to .06 (fit is adequate at .08 or less), and the standardized root mean square residual (SRMR) value is close to .08 (Byrne, 2010; Hu & Bentler, 1999). Fit for the model was good, CFI = .992; RMSEA = .028; SRMR = .034, χ²(8) = 9.09, p = .33 (see Figure 1). We also ran a model with no covariates. All three models (with parental education, age, and gender as covariates, with only age and gender as covariates, and with no covariates added) yielded the same findings in terms of significant effects.

Main Analyses

Hypothesis 1 was that perceptions of maternal warmth and negativity would be more strongly negatively associated in the European American sample than the Chinese sample. In our path analysis, we tested for group differences in path loadings with the MODEL CONSTRAINT command in MPlus. The results showed that maternal warmth and negativity were significantly negatively associated in both samples, but that, as expected, the association was stronger for the European American group than the Chinese group, t(300) = 3.01, p = .009. Figure 1 shows that the standardized estimate for the association between these variables was −.27 for the Chinese group, and −.59 in the European group, both ps < .001.

As an alternative test of Hypothesis 1, we conducted exploratory factor analyses within each group on the items that reflected maternal warmth and negativity. We used the principal factor method to extract factors, followed by an oblique (promax) rotation. For the Chinese group, the scree plot suggested two meaningful factors. The first factor contained the three items reflecting negativity, all which loaded above .70 on the first factor and less than .40 in absolute value on the second factor. The second factor contained four of the five items reflecting warmth; these items loaded above .50 on the second factor and less than .40 in absolute value on the first factor. One item (My parent felt I should be given comfort and understanding when I was scared or upset) did not load above .40 on either factor. These two factors accounted for 49% of the common variance. For the European sample, the scree plot suggested one meaningful factor. All items loaded more than .50 in absolute value on this factor; the three items reflecting negativity had a negative loading on the factor and the five items reflecting warmth loaded positively. This factor accounted for 47% of the common variance. When we forced a two-factor solution with an oblique rotation, the results were not clearly interpretable. The results of the exploratory analysis are thus consistent with the first hypothesis.

Hypothesis 2 was that for the Chinese group, both maternal warmth and negativity would predict unique variance in self-esteem but that in the European group, only one of the measures of maternal warmth and negativity would predict unique variance in these variables. The results of the path analysis supported this hypothesis. In the Chinese sample, perceptions of maternal warmth were significantly positively associated with self-esteem and perceptions of maternal negativity were significantly negatively associated with self-esteem. For the European American sample, only perceptions of maternal negativity were significantly (negatively) associated with self-esteem; the association between perceptions of warmth and self-esteem was nonsignificant.

Hypothesis 3 was that self-esteem would play a significant mediating role between perceptions of maternal warmth/negativity and life satisfaction for both groups. In the path analyses, we conducted mediation tests in MPlus using MLR and found that for the Chinese sample, the indirect effect from warmth to life satisfaction via self-esteem was significant, b = .102, SE = .033, p < .001. Not surprisingly (given that perceptions of maternal warmth were not significantly associated with self-esteem for European Americans in the path analyses), the parallel indirect effect was not significant for European Americans, b = .008, SE = .067, ns. The indirect effect from perceptions of maternal negativity to life satisfaction via self-esteem was significant for the Chinese group, b = −.068, SE = .025, p < .001, and the European American group, b = −.162, SE = .064, p = .001. Thus, Hypothesis 3 was confirmed.

We also examined direct effects of maternal warmth and negativity on life satisfaction over and above the effects mediated by self-esteem, although we made no prediction as to whether these effects would be significant after accounting for the mediating effect of self-esteem. As can be seen in Figure 1, there remained a significant direct path from maternal warmth to life satisfaction for the Chinese group. The direct path from maternal negativity to life satisfaction approached significance for the European American group, though it was a nonsignificant trend.

Hypothesis 4 was that self-esteem would be more strongly associated with life satisfaction in the European American group than in the Chinese group. This hypothesis was confirmed. Figure 1 shows that the standardized estimate for the association between these variables was .38 for the Chinese group, and .56 in the European group, both ps < .001. The difference in the strength of the associations was significant, t(300) = 3.51, p < .01.

Discussion

The results from the present study showed clearly that maternal warmth and negativity were not as strongly negatively associated for the Chinese sample as compared with the European American sample. For European Americans, the association between the two variables was strongly negative, and it is possible that they could be considered to be part of one underlying factor. However, maternal warmth and negativity were much more weakly associated in the Chinese group, and were clearly separate constructs. Thus, it seems prudent, in cross-cultural research, to consider parental warmth and negativity as separate constructs with unique influence, with culture potentially moderating their negative associations with each other. There was also evidence to show that maternal warmth and negativity both predicted unique variance in self-esteem for the Chinese sample, but not the European American sample. This is not surprising given the fact that warmth and negativity were strongly negatively associated in the European American, but not the Chinese, sample. We expand on these points below.

Warmth and Negativity as Separate Constructs

Our first and second hypotheses had to do with the possibility that associations between parental warmth and negativity may vary in predictable ways as a function of culture. We found results consistent with these expectations. We could not examine associations between paternal warmth and negativity because we could not establish measurement invariance for the measure of paternal negativity. However, we did find that while maternal warmth and negativity were significantly negatively associated in both cultural groups, the association was much stronger in the European American group (−.59) than in the Chinese group (−.27). This is important, because there are numerous examples in the literature where negativity, reverse-scored, is considered to reflect warmth, or vice versa (e.g., Luis et al., 2008; Putnick et al., 2012), or where measures of warmth include items that reflect negativity (reverse-scored) along with items that reflect warmth (e.g., Carlson & Harwood, 2003; Greenberger & Chen, 1996). While these approaches are important, because they demonstrate that parental warmth is positively (and parental negativity negatively) associated with more desirable qualities in children across cultures, it seems appropriate to assess the two constructs separately. The data from the Chinese sample suggest that parental warmth and negativity are not always strongly opposed. Our results regarding maternal warmth and negativity are similar to findings in the social-psychology literature, which have examined associations between participants’ reports of their own levels of positive and negative affect, and found that associations between negative and positive affect are much more strongly negative in European as opposed to East Asian groups (Bagozzi et al., 1999; Kitayama et al., 2000; Schimmack et al., 2002).

Maternal Warmth and Negativity as Predictors of Self-Esteem and Life Satisfaction

Our second hypothesis followed from the first: that if negative associations between warmth and negativity were weaker for the Chinese participants, both warmth and negativity, in that group, should predict unique variance in self-esteem in the model that we tested. On the contrary, we expected only warmth or negativity to predict unique variance in self-esteem for European Americans. Our expectation held true; the measures of both warmth and negativity predicted unique variance in self-esteem for the Chinese group, but only negativity predicted variance in self-esteem for the European American group. Again, the results suggest the utility of separately modeling parental warmth and negativity, rather than combining them into one overall measure.

Our third hypothesis was that self-esteem would mediate associations between maternal warmth and negativity and life satisfaction. This was true for both groups. Self-esteem significantly mediated the negative associations between perceptions of maternal negativity and life satisfaction for both groups, and significantly mediated the positive association between perceptions of warmth and life satisfaction for the Chinese group. Past research has found self-esteem to be a significant predictor of life satisfaction in both European American and Asian samples (e.g., Diener & Diener, 1995; Kang et al., 2003; Kwan et al., 1997), and has found parental warmth to be a significant predictor of self-esteem (e.g., Farruggia et al., 2004) and life satisfaction (e.g., Stankov, 2013). However, to our knowledge, this is the first study to directly examine, using the products of the coefficients approach, the mediating role of self-esteem in the relations between parental warmth and negativity and subsequent life satisfaction.

We also examined direct effects from maternal warmth and negativity to life satisfaction, over and above the path mediated via self-esteem. There were significant direct effects from maternal warmth to life satisfaction for the Chinese group, and a nonsignificant trend from maternal negativity to life satisfaction for the European American group. Thus, there is some evidence that while self-esteem played a significant mediating role in the relationship between warmth/negativity and life satisfaction, it did not account entirely for the association between the two variables. It is easy to imagine other variables that might play a mediating role, such as relationship harmony. Kwan et al. (1997), for example, found that both self-esteem and relationship harmony mediated relationships between self-construal and life satisfaction in samples from Hong Kong and the United States, but that self-esteem played a stronger mediating role in the United States as compared with Hong Kong, whereas relationship harmony played a stronger mediating role in Hong Kong as compared with the United States.

Finally, we replicated the results of numerous previous studies that have found self-esteem to be more strongly associated with life satisfaction in groups that are European American, as opposed to East Asian (e.g., Diener & Diener, 1995; Kang et al., 2003; Kwan et al., 1997). This result appears to be quite robust, and fits with a long line of research that shows self-esteem to be part of a more general trend of self-enhancement in groups that are individualist in nature (e.g., Heine, 2015).

Limitations

The present study had a number of shortcomings that could be addressed in future research. First, we tested the hypotheses for mothers only and not fathers, as we could not establish measurement equivalence on negativity for fathers. Future research might assess parental negativity with a scale that contains more items reflecting parental negativity so that there would be a better chance of finding at least three items for which loadings on the principle factor are similar. We believe the Parental Acceptance–Rejection Questionnaire (PARQ; Rohner, 2005) to be a good candidate in this regard. In addition, future research should utilize latent factors to reduce measurement error.

The present study was also limited in terms of the sample, which was comprised of university undergraduates in provincial areas of two specific countries who reported on their parenting practices, as well as the measures used. It would be helpful to assess parental warmth and negativity from multiple sources (e.g., parental reports, behavioral observations), and to obtain results from samples from a broader range of age groups (e.g., younger children), socioeconomic levels, and countries. In addition, the European American sample was significantly older than the Chinese sample, and the European American sample also had mothers with higher education than the Chinese sample. While we found no differences in terms of the significance of effects in models that controlled for these variables versus models that did not control for them, ideally the samples would have been more similar, obviating the need to test the control variables. Future research could address this issue.

Furthermore, the present study also assumes a direction of causality as outlined in the figures, but all data were assessed at one point in time. Future research could assess these variables longitudinally, or examine specific links experimentally to establish direction of effect.

Another issue that we did not examine directly was the reasons why maternal warmth and negativity were more weakly associated in the Chinese group than in the European American group. Future research could assess parents’ and children’s levels of collectivism as well as dialectical thinking to determine whether collectivism and/or dialectical thinking interact with parental negativity to predict warmth (e.g, at low levels of collectivism, one would expect a stronger negative association between warmth and negativity than at high levels of collectivism).

Contributions

Regardless of these limitations and potential future directions, the results of the present study establish that differential associations between parental warmth and negativity, or parental acceptance–rejection, exist in East Asian and European American cultures. Currently, cross-cultural research on parental warmth and negativity tends to combine these two dimensions into one overall measure (with one dimension reverse-scored); this practice appears to be problematic given the results of the present study. Furthermore, it appears that culture may moderate the associations between maternal warmth and other variables, such as self-esteem. These findings are important for researchers to keep in mind as they conduct cross-cultural research on parental warmth and negativity and young adults’ well-being.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Alexandra N. Davis is an assistant professor of Family and Child Studies at the University of New Mexico in the Department of Individual, Family, and Community Education. Her research focuses on how stressors associated with context and culture predict parent–youth relationships and positive development among marginalized populations.

Duane Rudy is in the Department of Human Development and Family Science, University of Missouri. He is interested in parent–child relationships and how they manifest themselves as a function of culture.

Chang Su-Russell is a doctoral candidate in the Department of Human Development and Family Science at University of Missouri. She studies Chinese cultural values and cross-cultural parenting in early childhood.

Chenyi Zhang is an assistant professor at Georgia State University in the Department of Early Childhood and Elementary Education. His research examines the interplay between teachers’ classroom instruction and young children’s early literacy development with consideration of environmental factors in the classroom. His research aims to understand how high quality teacher–child interactions (i.e., emotionally and instructionally sensitive interactions) at school influence young children’s literacy, language, and socioemotional development. He completed his PhD in human development and family studies (emphasis: early childhood education) at Purdue University in 2013.

References

Aquilino

W. S.

(1997). From adolescent to young adult: A prospective study of parent-child relations during the transition to adulthood. Journal of Marriage and Family, 59, 670-686. doi:10.2307/353953

Bagozzi

R. P.

Wong

(1999). The role of culture and gender in the relationship between positive and negative affect. Cognition and Emotion, 13, 641-672. doi:10.1080/026999399379023

Block

J. H.

(1981). The Child-Rearing Practices Report (CRPR): A set of Q items for the description of parental socialization attitudes and values. Berkeley: University of California, Institute of Human Development.

Buri

J. R.

(1991). Parental Authority Questionnaire. Journal of Personality Assessment, 57, 110-119. doi:10.1207/s15327752jpa5701_13

Byrne

B. M.

(2010). Structural equation modeling with AMOS (2nd ed.). New York, NY: Routledge.

Carlson

V. J.

Harwood

R. L.

(2003). Attachment, culture, and the caregiving system: The cultural patterning of everyday experiences among Anglo and Puerto Rican mother-infant pairs. Infant Mental Health Journal, 24, 53-73. doi:10.1002/imhj.10043

Chao

R. K.

(1995). Chinese and European American cultural models of the self reflected in mothers’ childrearing beliefs. Ethos, 23, 328-354. doi:10.1525/eth.1995.23.3.02a00030

Chen

Dong

Zhou

(1997). Authoritative and Authoritarian parenting practices and social and school performance in Chinese children. International Journal of Behavioral Development, 21, 855-873. doi:10.1080/016502597384703

Diener

(1995). Cross-cultural correlates of life satisfaction and self-esteem. Journal of Personality and Social Psychology, 68, 653-663. doi:10.1037/0022-3514.68.4.653

10.

Diener

Emmons

R. A.

Larsen

R. J.

Griffin

(1985). The Satisfaction With Life scale. Journal of Personality Assessment, 49, 71-75. doi:10.1207/s15327752jpa4901_13

11.

Farruggia

S. P.

Chen

Greenberger

Dmitrieva

Macek

(2004). Adolescent self-esteem in cross-cultural perspective testing measurement equivalence and a mediation model. Journal of Cross-Cultural Psychology, 35, 719-733. doi:10.1177/0022022104270114

12.

Greenberger

Chen

(1996). Perceived family relationships and depressed mood in early and late adolescence: A comparison of European and Asian Americans. Developmental Psychology, 32, 707-716. doi:10.1037/0012-1649.32.4.707

13.

Heine

S. J.

(2015). Cultural psychology (3rd ed.). New York, NY: W.W. Norton.

14.

L.-T.

Bentler

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

15.

Kang

S.-M.

Shaver

P. R.

Sue

Min

K.-H.

Jing

(2003). Culture-specific patterns in the prediction of life satisfaction: Roles of emotion, relationship quality, and self-esteem. Personality and Social Psychology Bulletin, 29, 1596-1608. doi:10.1177/0146167203255986

16.

Khaleque

Rohner

R. P.

(2012). Transnational relations between perceived parental acceptance and personality dispositions of children and adults: A meta-analytic review. Personality and Social Psychology Review, 16, 103-115. doi:10.1177/1088868311418986

17.

Kitayama

Markus

H. R.

Kurokawa

(2000). Culture, emotion, and well-being: Good feelings in Japan and the United States. Cognition and Emotion, 14, 93-124. doi:10.1080/026999300379003

18.

Knight

G. P.

Roosa

M. W.

Umaña-Taylor

A. J.

(2009). Studying ethnic minority and economically disadvantaged populations: Methodological challenges and best practices. Washington, DC: American Psychological Association.

19.

Kwan

V. S. Y.

Bond

M. H.

Singelis

T. M.

(1997). Pancultural explanations for life satisfaction: Adding relationship harmony to self-esteem. Journal of Personality and Social Psychology, 73, 1038-1051. doi:10.1037/0022-3514.73.5.1038

20.

Luis

T. M.

Varela

R. E.

Moore

K. W.

(2008). Parenting practices and childhood anxiety reporting in Mexican, Mexican American, and European American families. Journal of Anxiety Disorders, 22, 1011-1020. doi:10.1016/j.janxdis.2007.11.001

21.

Muthén

L. K.

Muthén

B. O.

(1998-2010). Mplus user’s guide (6th ed.). Los Angeles, CA: Author.

22.

Oishi

Schimmack

Diener

Kim-Prieto

Scollon

C. N.

Choi

D.-W.

(2007). The value-congruence model of memory for emotional experiences: An explanation for cultural differences in emotional self-reports. Journal of Personality and Social Psychology, 93, 897-905. doi:10.1037/0022-3514.93.5.897

23.

Putnick

D. L.

Bornstein

M. H.

Lansford

J. E.

Chang

Deater-Deckard

Di Giunta

. . . Bombi

A. S.

(2012). Agreement in mother and father acceptance-rejection, warmth, and hostility/rejection/neglect of children across nine countries. Cross-Cultural Research: The Journal of Comparative Social Science, 46, 191-223. doi:10.1177/1069397112440931

24.

Robertson

J. F.

Simons

R. L.

(1989). Family factors, self-esteem, and adolescent depression. Journal of Marriage and Family, 51, 125-138. doi:10.2307/352374

25.

Rohner

R. P.

(2005). Parental Acceptance-Rejection/Control Questionnaire (PARQ/Control): Test manual. In Rohner

R. P.

Khaleque

(Eds.), Handbook for the study of parental acceptance and rejection (pp. 137-186). Storrs, CT: Rohner Research Publications.

26.

Rohner

R. P.

Khaleque

Cournoyer

D. E.

(2005). Parental acceptance-rejection: Theory, methods, cross-cultural evidence, and implications. Ethos, 33, 299-334. doi:10.1525/eth.2005.33.3.299

27.

Rosenberg

(1989). Society and the adolescent self-image (Rev. ed.). Middletown, CT: Wesleyan University Press.

28.

Rudy

Grusec

J. E.

(2006). Authoritarian parenting in individualist and collectivist groups: Associations with maternal emotion and cognition and children’s self-esteem. Journal of Family Psychology, 20, 68-78. doi:10.1037/0893-3200.20.1.68

29.

Schimmack

Oishi

Diener

(2002). Cultural influences on the relation between pleasant emotions and unpleasant emotions: Asian dialectic philosophies or individualism-collectivism? Cognition and Emotion, 16, 705-719. doi:10.1080/02699930143000590

30.

Spencer-Rodgers

Peng

Wang

Hou

(2004). Dialectical self-esteem and East-West differences in psychological well-being. Personality and Social Psychology Bulletin, 30, 1416-1432. doi:10.1177/0146167204264243

31.

Stankov

(2013). Depression and life satisfaction among European and Confucian adolescents. Psychological Assessment, 25, 1220-1234. doi:10.1037/a0033794

32.

Stewart

S. M.

Rao

Bond

M. H.

McBride-Chang

Fielding

Kennard

B. D.

(1998). Chinese dimensions of parenting: Broadening Western predictors and outcomes. International Journal of Psychology, 33, 345-358. doi:10.1080/002075998400231

33.

United Nations Statistics Division. (2017). Population 15 years of age and over, by educational attainment, age and sex. Retrieved from http://data.un.org/Data.aspx?d=POP&f=tableCode%3A30

34.

U.S. Census Bureau. (2016). Educational attainment in the United States: 2015. Retrieved from https://www.census.gov/content/dam/Census/library/publications/2016/demo/p20-578.pdf

35.

Wolf

E. J.

Harrington

K. M.

Clark

S. L.

Miller

M. W.

(2013). Sample size requirements for structural equation models an evaluation of power, bias, and solution propriety. Educational and Psychological Measurement, 73, 913-934. doi:10.1177/0013164413495237

36.

Yao

M. Z.

D. M.

Pang

(2014). The influence of personality, parental behaviors, and self-esteem on internet addiction: A study of Chinese college students. Cyberpsychology, Behavior, and Social Networking, 17, 104-110.