Parenting Attributions and Attitudes in Cross-Cultural Perspective

SYNOPSIS

Objective

This article used the Parenting Across Cultures Project to evaluate similarities and differences in mean levels and relative agreement between mothers’ and fathers’ attributions and attitudes in parenting in 9 countries.

Design

Mothers and fathers reported their perceptions of causes of successes and failures in caregiving and their progressive versus authoritarian childrearing attitudes. Gender and cultural similarities and differences in parents’ attributions and attitudes in 9 countries were analyzed: China, Colombia, Italy, Jordan, Kenya, the Philippines, Sweden, Thailand, or the United States.

Results

Although mothers and fathers did not differ in any attribution, mothers reported more progressive parenting attitudes and modernity of childrearing attitudes than did fathers, and fathers reported more authoritarian attitudes than did mothers. Country differences also emerged in all attributions and attitudes that were examined. Mothers’ and fathers’ attributions and their attitudes were moderately correlated, but parenting attitudes were more highly correlated in parents than were attributions.

Conclusions

We draw connections among the findings across the 9 countries and outline implications for understanding similarities and differences in mothers’ and fathers’ parenting attributions and attitudes.

INTRODUCTION

Culture and Parenting Science

Adults do not parent in isolation, but always do so in a social and cultural context. Parents and cultures are, therefore, intimately bundled because two intertwined major goals of parenting are to successfully transmit the prevailing culture across generations and effectively embed the next generation into the existing culture. Culture comprises the ways in which a collection of people process and make sense of their experiences and so shapes a wide array of functions, including cognitions and practices related to childrearing and child development.

The reasons to pursue cultural study in parenting science are many and compelling, and by now well known. One is description, and for this reason social commentary as a matter of course includes reports of family life. Insofar as cultural descriptions of parenting attempt to encompass the widest spectrum of human variation, they are also the most comprehensive. They are vital to delimiting the full range of human experience and so are also critical to establishing realistic and valid norms. Furthermore, our awareness of alternative modes of parenting sharpens our perceptions and enhances our understanding of the nature of childrearing in our own culture. Description is also prerequisite to other formal rationales, like explanation. This motive for submitting caregiving in different cultures to psychological scrutiny derives from the extraordinary and unique power that cultural comparisons furnish parenting science. Only the comparative view can expose variables that regulate care but may remain invisible from a monocultural perspective. The cultural approach to analysis helps to distinguish those parenting constructs, structures, functions, and processes that are community specific from those that transcend or are independent of culture, and this kind of analysis holds out the possibility of exposing how forces that vary globally (e.g., family structure, urbanization, nationality, religion, economics, and the like) differentially mold key features of human behavior.

This paper presents data from 9 countries that participated in the Parenting Across Cultures Project: China, Colombia, Italy, Jordan, Kenya, the Philippines, Sweden, Thailand, and the United States. When we refer to cultures and countries in this paper, we do not mean to imply that our samples are representative of the entire country. For example, the China sample is only representative of two-parent families of 7- to 10-year-old children in Jinan and Shanghai. For the sake of convenience, we refer to each sample according to the country in which it was collected. The paper uses data of a within-family framework to address two primary questions with mothers and fathers. First, we ask what are the similarities and differences in mean levels of two sorts of mothers’ and fathers’ parenting cognitions, namely attributions and attitudes? Second, we ask how highly are mothers’ parenting attributions and attitudes correlated with fathers’ parenting attributions and attitudes? We also explore whether these effects are moderated by culture. This paper employs samples that have been understudied to date from countries that are underrepresented in the psychological literature generally and the parenting literature specifically.

Cultural Variation in Parents’ Attributions and Attitudes about Parenting

Parenting is multidimensional, modular, and specific, and parenting is multiply determined (Bornstein, 2006). Here we are concerned with two sorts of parenting cognitions. Parental attributions and attitudes represent different constructs, and each is shaped by many factors, notably culture. Considerable attention has been devoted to trying to understand the intersection of culture and parents’ cognitions (Bornstein & Lansford, 2009). On the one hand, parents, regardless of culture, need and might share certain cognitions, and cognitions presumably serve comparable functions for everyone, irrespective of culture; on the other hand, community-specific cognitions about childrearing can be expected to arise and be compatible with each specific society’s setting and needs.

Parents’ attributions regarding successes and failures in caregiving

Parenting attributions consist of “a variety of judgments that parents make as they attempt to explain, evaluate, and predict their children’s behaviors” (Miller, 1995, p. 1558). Bugental and Shennum (1984) focused on the extent to which parents’ affective and behavioral responses to children’s behavior differ, based on their perception of whether parents’ own actions and those of their children are controllable. Thus, parents may attribute success or failure in caregiving to themselves, to their child, or to both themselves and their child. Dix and Grusec (1985) developed and tested an attributional model of parental cognitions wherein parents’ affective reactions to children’s behavior are asserted to vary based on their belief that such behaviors are intentional, controllable, or dispositional, as opposed to being determined by developmental limitations or situational constraints. Importantly, attributions for successes versus failures in caregiving are related to parents’ behaviors in actual caregiving situations. In one of the few cross-national studies of parents’ attributions regarding successes and failures in caregiving in Argentina, Belgium, France, Israel, Italy, Japan, and the United States, Bornstein et al. (1998) reported differences across countries with respect to the importance mothers placed on their own abilities and the child’s characteristics as reasons for parenting successes and failures. The present study expands the cross-cultural base of this research and extends it to older children and to fathers, allowing us to draw comparisons across countries and parents within families.

Parents’ progressive versus authoritarian childrearing attitudes

Attitudes constitute a unique and distinct type of social cognition. They represent the hypothetical construct of a “psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor” (Eagly, 1992, p. 693). Thus, attitudes are based on knowledge, but go beyond that information to evaluation. Childrearing attitudes have historically been a preeminent topic in parenting science. The first systematic study of parents published in North America assessed parental attitudes (Laws, 1927). Parents’ childrearing attitudes constitute important dimensions of caregiving because, like attributions, attitudes affect parents’ practices toward their children and the types of environments that parents create for their children.

A key domain of parents’ attitudes is the extent to which they hold progressive versus authoritarian childrearing views. Parents with progressive attitudes believe that children should be encouraged to think independently and verbalize their ideas, and these parents generally approach the parent-child relationship out of a democratic framework (Okagaki & Frensch, 1998). By contrast, parents with authoritarian attitudes emphasize strictness, respect for authority, and obedience (Dornbusch, Ritter, Liederman, Roberts, & Fraleigh, 1987). Both attitude orientations see child agency as important. Parents who hold progressive attitudes tend to grant children more agency than do parents who hold authoritarian attitudes. Depending on whether parents hold more progressive or more authoritarian attitudes about parenting, they presumably socialize their children differently and behave in ways that instill their attitudes in their children.

Similarities and Differences in Mothers’ and Fathers’ Parenting Attributions and Attitudes

Attributions regarding outcomes of caregiving and progressive versus authoritarian attitudes vary by characteristics of the parent, notably whether the parent is the mother or father. Two limitations of past research on parental attributions and attitudes have been their primary focus on mothers and on European Americans. The majority of children throughout the world grow up in family systems with more than one significant parenting figure. Until recently, however, most empirical research in parenting has focused on mothers alone, and virtually all guiding theories depict parenting and child development as unfolding within the context of the dyadic mother-child relationship. The parenting literature needs to look at both parenting figures. Researchers need to be attuned to the family’s “parenting map” when attempting to study family process (Demo & Cox, 2000).

Understanding the attributions and attitudes of both mothers (Barnard & Solchany, 2002) and fathers (Parke, 2002) is therefore important because families are interconnected social systems (Bornstein & Sawyer, 2005). The coparenting dynamic is co-constructed by both members of the family’s “executive subsystem” (Minuchin, 1974). Effectively functioning coparents collaborate to structure a family context that communicates to children solidarity and support between significant adult parenting figures, who project a consistent and predictable set of rules and standards and a safe and secure home (McHale, Lauretti, Talbot, & Pouquette, 2002). The nature of interadult coparenting dynamics within families is receiving attention as it becomes apparent that children are influenced, not just by discrete interactions with their mothers or with their fathers, but also by the parenting package of mothers and fathers (Feinberg, 2003; McHale, Khazan, Erera, Rotman, DeCourcey, & McConnell, 2002; Teubert & Pinquart, 2010).

Although children are socialized by both mothers and fathers, men’s reports of their parenting have historically been understudied (Parke, 2002). There is some evidence that fathers complement mothers in basic caregiving and that mothers and fathers divide the labors of child caregiving. Even when parents are equally engaged in parenting, mothers and fathers often differ in their styles of engagement (Dumas & Lechowicz, 1989). This paper focuses on understanding similarities and differences between mothers’ and fathers’ attributions for successes and failures in caregiving and in their progressive versus authoritarian attitudes.

Some differences have been found in mothers’ and fathers’ attributions regarding outcomes of caregiving, and parenting attributions appear to be associated with other constructs in different ways for women and men. For example, mothers’ attributions are more strongly related than are fathers’ attributions to sons’ and daughters’ own attributions (Bugental & Martorell, 1999). Still under-studied is the rather basic issue of whether mothers and fathers show similarities or differences in mean levels of, and degrees of agreement in, attributions regarding the causes of successes and failures in their caregiving and their progressive versus authoritarian attitudes. Of course, interparental levels and agreement might each be moderated by culture. For example, differences in attitudes between Asian fathers and mothers may be captured in the traditional adage of the “strict father, kind mother,” which describes the expectation that fathers exert authoritarian control whereas mothers exhibit warmth toward their children (Chao & Tseng, 2002). Exploring similarities and differences in mothers’ and fathers’ attributions and attitudes across cultures will contribute to a richer understanding of the joint socialization influences to which children are exposed in the family context.

Research Questions

Previous cross-national comparative studies have not examined similarities and differences or degrees of concordance in mothers’ and fathers’ parenting attributions and attitudes. Because strong cultural differences exist with respect to family roles of women and men in general (Best, 2010), and in their roles as parents in particular, the extent to which patterns of gender differences reported in one cultural context are found in other cultural contexts is not clear. Our analyses therefore address two primary research questions. First, what are the similarities and differences in mean levels of mothers’ and fathers’ attributions and attitudes across 9 countries? Second, how highly are mothers’ attributions and attitudes correlated with fathers’ attributions and attitudes across 9 countries? We also explored whether these effects were moderated by culture. Expanding research on parenting to include within-family analysis of previously underrepresented groups is important to advance our understanding of the extent to which parenting cognitions are community-specific or generalizable across cultural groups.

METHOD

Participants

Mother-father dyads (N = 1,133) of 7- to 10-year-old children from 9 countries provided data. Parents were recruited through schools that serve socioeconomically diverse populations in China (n = 239 from Jinan and Shanghai), Colombia (n = 108 from Medellín), Italy (n = 177 from Rome and Naples), Jordan (n = 112 from Zarqa), Kenya (n = 100 from Kisumu), the Philippines (n = 95 from Manila), Sweden (n = 77 from Trollhättan/Vänersborg), Thailand (n = 87 from Chiang Mai), and the United States (n = 138 European Americans, Latin Americans, and African Americans from Durham, NC). Parents of girls and boys were represented in approximately equal numbers in each country subsample. (Details about each sample appear in country-specific papers in this Special Issue). In China, Italy, and the United States, data from 2 or 3 geographic or cultural groups were collected. However, no differences were found between the subgroups in China, and despite the small differences found between groups in Italy and in the United States, we combined all subgroups to examine country-level differences.

Procedures

Detailed procedures for each country can be found in the Introduction (Lansford & Bornstein, 2012) and the individual country papers in this Special Issue. A procedure of forward- and back-translation ensured the linguistic and conceptual equivalence of all measures (Maxwell, 1996). Measures were translated and culturally adapted to ensure that the measures would be valid in all sites (Erkut, 2010; Peña, 2007). Interviews were conducted in participants’ homes, schools, or at another location chosen by the parents and used oral and written methods as appropriate. Parents completed a demographic questionnaire, a measure of social desirability bias (Reynolds, 1982), and two parenting measures.

The analyses in this paper focus on constructs from the measures of attributions and attitudes (see Lansford & Bornstein, 2012), the Parent Attribution Test (Bugental & Shennum, 1984) and Parental Modernity Inventory (Schaefer & Edgerton, 1985). The Parent Attribution Test measures parents’ perceptions of causes of successes and failures in hypothetical caregiving situations. Parents were presented with a hypothetical scenario that involved either a positive or negative interaction with a child (e.g., “Suppose you took care of a neighbor’s child one afternoon, and the two of you had a really good time together.”). Parents then rated on a 7-point scale (1 = not at all important, 7 = very important) how important factors such as the child’s disposition and the parent’s behavior were in determining the quality of the interaction. This measure yielded four variables: (1) attributions regarding uncontrollable success (6 items; e.g., how lucky you were in just having everything work out well); (2) attributions regarding adult-controlled failure (6 items; e.g., whether you used the wrong approach for this child); (3) attributions regarding child-controlled failure (6 items; e.g., the extent to which the child was stubborn and resisted your efforts); and (4) perceived control over failure (the difference between attributions regarding adult-controlled failure and attributions regarding child-controlled failure).

The Parental Modernity Inventory (Schaefer & Edgerton, 1985) assesses parents’ attitudes about childrearing and education. Each of 30 statements was rated on a 4-point scale (1 = strongly disagree, 4 = strongly agree). This instrument yielded three variables: (1) progressive attitudes (8 items; e.g., Children have a right to their own point of view and should be allowed to express it.); (2) authoritarian attitudes (22 items; e.g., The most important thing to teach children is absolute obedience to their parents.); and (3) modernity of childrearing attitudes (the difference between the progressive attitudes score and the authoritarian attitudes score). Internal consistency reliability for the four attribution and three attitude variables were all above .60 (range = .68–.88) except for mother and father progressive attitudes which were .58 and .56, respectively.

Finally, research has pointed to cross-cultural differences in self-serving bias (e.g., Chandler, Shama, Wolf, & Planchard, 1981; Markus & Kitayama, 1991). The 13-item Social Desirability Scale (SDS-SF; Reynolds, 1982) was used to assess parents’ possible social desirability bias. Statements like “I’m always willing to admit when I make a mistake.” were rated as True or False. Reliability of the SDS-SF was reported as .76, and the correlation with the full-length SDS .93 in a U.S. sample (Reynolds, 1982), but we acknowledge that the measure’s reliability across cultures may differ. The SDS-SF was used as a covariate for parents’ self-reported attributions and attitudes.

RESULTS

Analytic Plan

Analyses proceeded in two stages. First, repeated-measures linear mixed models with gender of parent as the within-subjects fixed factor tested for differences between mothers’ and fathers’ attributions for successes and failures in caregiving situations and progressive versus authoritarian childrearing attitudes. The covariance structure was modeled as heterogeneous compound symmetry, accounting for the likelihood that parents’ attributions and attitudes would be correlated, but allowing mothers’ and fathers’ variances to differ. Country and the interaction between Parent gender and Country were also included as fixed effects. In linear mixed models, fixed effects do not follow exact F distributions; therefore, the denominator degrees of freedom are estimated using the Satterthwaite (1946) approximation and are not necessarily integers. Test results are presented with and without controls for mothers’ and fathers’ ages, education, and possible social desirability bias.

Second, the equivalence of mother-father correlations across countries was tested using multiple group models in AMOS 17. Multiple group models in which covariances (i.e., correlations between mother and father scores) were constrained to be equal across the 9 countries were compared to models in which the covariances were free to vary across the 9 countries. To have the degrees of freedom to compute model fit statistics, we tested the 7 attribution and attitude measures in pairs, specifying no correlations across domains. If the differences in chi-square values for the unconstrained and constrained models were nonsignificant, we concluded that mother-father correlations were similar across countries. If the differences in chi-square values for the unconstrained and constrained models were significant, we attempted to improve the change in model fit by releasing the paths for one or more countries. Correlations were also computed between parents in the same family to assess similarity between mothers’ attributions and attitudes and fathers’ attributions and attitudes, respectively. Bivariate correlations as well as partial correlations, controlling for parents’ ages, levels of education, and possible social desirability bias, are presented. Correlations are interpreted following Cohen (1988) where r = .10 is interpreted as a small effect, r = .30 as a medium effect, and r = .50 as a large effect.

We controlled for parental age and education because parents who are older and those with higher levels of education are more likely to attribute outcomes of caregiving situations to both parents’ and children’s efforts, whereas younger parents and those with lower levels of education are more likely to attribute outcomes of caregiving situations to biological factors or chance (children “are the way they are”; Bugental & Happaney, 2002). Parental education also is related to parents’ attitudes regarding the advisability of particular parenting practices (Palacios & Moreno, 1996). We controlled for possible social desirability bias to account for parents’ tendency to present themselves and their children favorably (Bugental, Johnston, New, & Silvester, 1998; Holden & Edwards, 1989).

Country and Gender Similarities and Differences in Parents’ Attributions and Attitudes

Two significant Parent gender by Country interactions emerged (Table 1), controlling for parents’ age, education, and possible social desirability bias: uncontrollable success, F(8, 1109.72) = 2.29, p < .05, and adult-controlled failure, F(8, 1121.00) = 2.68, p < .01. Follow-up comparisons indicated that mothers scored higher than fathers on uncontrollable success attributions only in Italy, and mothers scored lower than fathers on adult-controlled failure attributions only in Sweden and the United States (mothers and fathers did not differ in all other countries). Given these significant interactions between Parent gender and Country, the main effects for uncontrollable success and adult-controlled failure attributions should be interpreted with caution.

TABLE 1.

Interaction Effects for Parent Gender X Country and Tests of the Equality of the Correlations between Mothers and Fathers across 9 Countries

	F	Fa	Δχ2	Δχ2b
Attributions
Uncontrollable success	2.39*	2.29*	33.70***	--
Adult-controlled failure	2.48*	2.68**	31.37***	8.51c
Child-controlled failure	1.86	1.82	20.34***	7.80c
Perceived control over failure	1.01	1.10	30.15***	6.00c
Attitudes
Progressive attitudes	1.06	.82	17.85*	11.85c
Authoritarian attitudes	2.04*	1.28	52.70***	11.63d
Modernity of attitudes	1.92	.92	34.51***	11.02e

-- = Releasing one or two path coefficients did not improve model fit.

^aControlling for parental age, education, and possible social desirability bias.

^bΔχ² after releasing the path coefficient for mother-father correlation in one or two countries.

^cReleasing the path coefficient for Jordan.

^dReleasing the path coefficient for the Philippines and United States.

^eReleasing the path coefficient for the United States.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

As shown in Table 2, across all 9 countries, there were no significant main effects of parent gender on the four attribution measures (but see above), but significant main effects were observed for all three attitude measures. Mothers reported more progressive attitudes and modernity of attitudes, and fathers reported more authoritarian attitudes overall. All of these differences remained significant after controlling for parents’ age, education, and possible social desirability bias. However, the significant differences in means between mothers and fathers were small, ranging from .12 to .18 standard deviations in the covariate controlled tests (see d^a column in Table 2).

TABLE 2.

Descriptive Statistics, Main Effects of Parent Gender, and Correlations between Mothers and Fathers across 9 Countries

	Mothers M (SD)	Fathers M (SD)	F	Fa	d	da	r	ra
Attributions
Uncontrollable success	5.02 (1.18)	4.96 (1.15)	2.13	2.27	.05	.05	.36***	.35***
Adult-controlled failure	4.27 (.77)	4.27 (.80)	.06	.00	.01	−.00	.17***	.17***
Child-controlled failure	3.97 (.68)	3.97 (.67)	.00	.00	.00	−.00	.14***	.14***
Perceived control over failure	.30 (1.00)	.30 (1.04)	.02	.00	.01	−.00	.16***	.16***
Attitudes
Progressive attitudes	3.14 (.35)	3.09 (.35)	16.84***	19.62***	.14	.16	.32***	.28***
Authoritarian attitudes	2.65 (.45)	2.68 (.44)	6.72**	15.38***	−.07	−.12	.59***	.48***
Modernity of attitudes	.50 (.63)	.41 (.60)	24.11***	37.12***	.13	.18	.59***	.49***

Note. Cohen’s d was computed using Equation 3 for paired samples in Dunlap, Cortina, Vaslow, and Burke (1996).

^aControlling for parental age, education, and possible social desirability bias.

^**p ≤ .01.

^***p ≤ .001.

As shown in Table 3, significant main effects of country emerged on all seven parenting constructs in uncontrolled and controlled analyses. Because we were not interested in specific country contrasts, we explored deviations from the grand mean of all countries as a method of investigating which countries had relatively more extreme scores. Results from uncontrolled and controlled analyses are reported in Table 3, but only covariate controlled analyses are detailed here and depicted in Figures 1 and 2. For uncontrollable success attributions, China, Italy, Kenya, and Sweden scored significantly lower than the grand mean, and Jordan, the Philippines, Thailand, and the United States scored significantly higher than the grand mean. For adult-controlled failure attributions, Colombia, Italy, and Jordan scored significantly lower than the grand mean, and China, Kenya, and Thailand scored significantly higher than the grand mean. For child-controlled failure attributions, Colombia, the Philippines, and the United States scored significantly lower than the grand mean, and Italy, Kenya, and Thailand scored significantly higher than the grand mean. For perceived control over failure, Colombia, Italy, and Jordan scored significantly lower than the grand mean, and China, the Philippines, and the United States scored significantly higher than the grand mean. For progressive attitudes, Kenya and the Philippines scored significantly lower than the grand mean, and China, Jordan, Sweden, and Thailand scored significantly higher than the grand mean.

TABLE 3.

Descriptive Statistics and Main Effects of Country

	China	Colombia	Italy	Jordan	Kenya	Philippines	Sweden	Thailand	United States	Grand Mean	F	Fa
Attributions
Uncontrollable success	4.27 (1.13)	5.25c (1.22)	4.87 (1.06)	5.51 (1.01)	4.73 (1.09)	5.68 (.97)	4.62 (.80)	5.32 (1.00)	5.49 (1.02)	5.08 (.48)	49.35	53.50
Adult-controlled failure	4.58 (.75)	3.96 (.75)	4.13 (.80)	3.84 (.93)	4.45 (.73)	4.34bc (.65)	4.16bc (.62)	4.42 (.78)	4.26bc (.66)	4.24 (.24)	26.51	24.67
Child-controlled failure	4.01bc (.78)	3.83 (.58)	4.11 (.60)	3.93bc (.77)	4.15 (.76)	3.80 (.47)	3.87bc (.51)	4.10 (.68)	3.80 (.59)	3.96 (.14)	9.39	9.51
Perceived control over failure	.57 (1.10)	.13 (.92)	.02 (1.00)	−.09 (1.22)	.30bc (.99)	.54 (.81)	.29bc (.72)	.32bc (.99)	.46 (.92)	.28 (.23)	13.62	12.23
Attitudes
Progressive attitudes	3.19 (.32)	3.11bc (.31)	3.10bc (.30)	3.18 (.37)	2.77 (.38)	3.04 (.33)	3.27 (.31)	3.22 (.35)	3.11bc (.33)	3.11 (.15)	32.30	30.99
Authoritarian attitudes	2.42 (.34)	2.87 (.38)	2.69bc (.42)	2.71bc (.33)	3.02 (.38)	2.96 (.42)	2.28 (.32)	2.73bc (.32)	2.56 (.54)	2.69 (.24)	59.10	68.23
Modernity of attitudes	.78 (.48)	.25c (.50)	.41bc (.53)	.47bc (.46)	−.25 (.57)	.08 (.52)	.99 (.47)	.49b (.52)	.55c (.68)	.42 (.36)	71.62	86.55

Note. M (SD). All F tests were significant at p < .001.

^aControlling for parental age, education, and possible social desirability bias.

^bCountry mean was not significantly different from the grand mean (p > .05) in tests without covariates.

^cCountry mean was not significantly different from the grand mean (p > .05) in tests controlling for parental age, education, and possible social desirability bias.

Figure 1.

Deviations from the grand mean for parental attributions.* p < .05.

Figure 2.

Deviations from the grand mean for parental attitudes. *p < .05.

For authoritarian attitudes, China, Sweden, and the United States scored significantly lower than the grand mean, and Colombia, Kenya, and the Philippines scored significantly higher than the grand mean. Finally, for modernity of childrearing attitudes, Kenya and the Philippines scored significantly lower than the grand mean, and China, Sweden, and Thailand scored significantly higher than the grand mean.

Within-Family Correlations between Parents’ Attributions and Attitudes across Country

The differences in chi-square values for the unconstrained and constrained multiple group models were significant for all measures, indicating that one or more countries had significantly different mother-father correlations (Table 1). Following the general tests, we attempted to improve the change in model fit by releasing the paths for one or more countries. As seen in Table 1, releasing the coefficient for Jordan significantly improved the model fit for adult- and child-controlled failure, perceived control over failure, and progressive attitudes. The correlation for Jordan was higher than the correlations in the other countries. Releasing the path coefficients for the Philippines and the United States significantly improved the model fit for authoritarian attitudes, and releasing the coefficient for the United States significantly improved the model fit for modernity of childrearing attitudes. The mother-father correlations in the Philippines and the United States were also higher than those in the other countries. Releasing the paths for three countries did not improve the model fit for uncontrollable success, indicating greater variability across countries in the mother-father correlation for uncontrollable success. Hence, the pooled correlations presented below should be interpreted with caution.

The final columns of Table 2 present bivariate correlations of mothers’ attributions and attitudes with fathers’ corresponding attributions and attitudes, respectively. As shown, all of the correlations were highly significant, but the size of the correlations differed for attributions and attitudes. Overall, mother-father partial correlations were significantly higher for attitudes (average r = .42, p < .001) than attributions (average r = .21, p < .001), z = 5.62, p < .001. Furthermore, each individual attitude correlation was higher than each attribution correlation, zs = 3.71 – 9.30, ps < .0001, with the exception of uncontrollable success and progressive attitudes, which differed only marginally, z = −1.83, p = .07.

DISCUSSION

The present study examined parenting attributions and attitudes among more than 1000 mothers and fathers of 7- to 10-year-olds in 9 countries. Country differences emerged in all attributions and attitudes we examined, and they remained significant after controlling for parents’ age, education, and possible social desirability of responding. Although mothers and fathers did not differ in any attribution overall, mothers reported more progressive parenting attitudes and modernity of childrearing attitudes than did fathers, and fathers reported more authoritarian attitudes than did mothers; these differences remained significant after controlling for parents’ age, education, and possible social desirability of responding. Mothers’ and fathers’ attributions and their attitudes were also moderately correlated, but parenting attitudes were more highly correlated in parents than were attributions. In sum, for attributions mothers and fathers were equal and somewhat concordant (with only 3 culturally moderated differences in means), and for attitudes mothers and fathers were unequal but concordant (and at a higher level than for attributions). Furthermore, on average mothers and fathers from all countries (except Kenya) held more progressive than authoritarian parenting attitudes.

Differences in Parenting Cognitions across Countries

During the lifecourse, individuals create ideas about parenting, children, and development that are adopted, interpreted, and transformed based on a variety of experiences. Although important sources of cognitions reside in the individual’s microsystem of personal experiences, cognitions must also be consistent with the culture because they are embedded in and are created within the macrosystem. Recognizing these multiple levels of influence fosters a more comprehensive understanding of the determinants of family life.

Country means of all seven parenting cognitions we investigated differed, even after controlling for parents’ age, education, and possible social desirability bias. These differences indicate that parents in different countries are unequally likely to attribute responsibility for caregiving successes and failures, as they are to profess progressive versus authoritarian parenting attitudes. These findings, first, fill a gap in knowledge regarding cultural variation in parents’ attributions of successes and failures in caregiving and parents’ progressive versus authoritarian childrearing attitudes, and, second, they point to the importance of acknowledging the significance of culture for parenting science (Bornstein, 1991, 2009). In the following discussion, we develop brief profiles of individual countries based on each country’s highs and lows relative to the grand means of countries we studied. We do not comment when a country’s mean responses do not differ from the grand mean.

In our across country analysis, we looked at deviations from the grand mean of countries that participated in the PAC (see Figures 1 and 2). So, for example, with respect to attributions for successes and failures and caregiving, Chinese parents rated uncontrollable success relatively low, whereas they rated adult-controlled failure and perceived control over failure relatively high. This suggests, with respect to parenting attributions, that Chinese parents see themselves as more than less responsible for caregiving successes and failures in this group of countries. Lower levels of attributions regarding uncontrollable success are consistent with higher levels of parenting efficacy. These findings accord with previous research showing that parents in China typically believe that children’s school performance is determined by effort--an attribution that assigns value to “trying harder” (rather than innate ability) as a means of doing better in school (Stevenson & Lee, 1990). In contrast, many parents in Western countries are more likely to believe that school performance is determined by ability—an attributional pattern that is less suggestive of ways that personal effort might improve performance. With respect to progressive versus authoritarian childrearing attitudes, Chinese parents rated themselves as relatively low in terms of authoritarianness and reciprocally high in progressive and modern childrearing attitudes. There has been some question of whether China’s one-child policy has rendered Chinese parents more lax and permissive; these findings regarding low levels of authoritarian attitudes support their modernization. Chang, Chen, and Ji (2012) have also pointed out that the parents who participated in the PAC represent a cohort that has experienced dynamic modernization.

In terms of attributions relative to other countries in the PAC, Colombian parents rated themselves low in adult-controlled, child-controlled, and perceived control over caregiving failures. This suggests that, with respect to failures in childrearing, Colombian parents find other sources more appropriate and responsible. Colombian parents were also higher than the grand mean of countries in endorsing an authoritarian stance to parenting, which is not atypical for South American countries. Latin American parents tend to display more authoritarian attitudes regarding parenting (Dornbusch et al., 1987; Livingston & McAdoo, 2007). Latina mothers tend to physically guide children’s actions, prefer discipline, and favor didactic teaching methods (Cardona, Nicholson, & Fox, 2000). Chaudhuri, Easterbrooks, and Davis (2009) studied emotional relationships in first-time (adolescent) mothers of toddlers and used cluster analysis to examine relations with mothers’ reports of parenting attitudes and behaviors; Latina mothers were more highly represented in a directive parenting group.

Italian parents rated themselves relatively low vis-à-vis parents in other PAC countries in terms of their attributions for uncontrollable success, but also rated adult-controlled failures and perceived control over failures relatively low with respect to the grand mean of all countries; reciprocally, Italian parents rated child-controlled failures relatively high. This pattern is consistent with the view that Italian parents see child development as more child-driven and as a naturally unfolding phenomenon and consider the effectiveness of adult intervention less requisite and robust (Bornstein et al., 1998; New, 1989).

Parents in Jordan rated adult-controlled failure and perceived control over failure in caregiving relatively low, whereas they rated uncontrollable success relatively high. Apparently, parents in Jordan do not see themselves as responsible for parenting failures, but they do not see themselves responsible for parenting successes either. One possibility is that, in the context of social change and recent public policies that are reforming the Jordanian education system and national policies related to children (Toukan, Alnoaimi, & Odibat, 2006), Jordanian parents perceive that many factors affecting their children’s success fall outside of parental control. That being said, Jordanian parents also profess relatively high progressive childrearing beliefs. These findings are consistent with several prior studies that have found Jordanian parents to hold progressive attitudes about parenting (Abu Aita, 2005; Sabri, 2002; Subehi, 1994).

Parents in Kenya rated uncontrollable success in caregiving as relatively low compared to parents in other PAC countries, and they rated adult controlled failures as well as child controlled failures relatively high. Childrearing ideology among the Kenyan Luos (the ethnic group included in our study) describes childrearing as a deliberate adult activity with the express aim of shaping children so that they show desirable outcomes; this ideology is illustrated by the common Luo saying that “A tree is shaped while young, or when it is grown up it breaks” (Oburu, 2004; Ocholla-Ayayo, 1976). Our findings suggest that Kenyan parents take both their children and themselves as responsible for failures of caregiving. With respect to their authoritarian versus progressive attitudes toward childrearing, Kenyan parents rated progressive attitudes and modern childrearing attitudes relatively low and reciprocally authoritarian attitudes relatively high, which is consistent with traditional structure in this social group where children occupy a relatively lower status in the social hierarchy than do their parents (Oburu & Palmerus, 2003).

In the Philippines, parents rated child-controlled failures as partly responsible as a source of attribution in parenting, whereas they rated uncontrollable successes and perceived control over failure relatively high. This profile suggests that Filipino parents do not see their children as responsible for failures in caregiving but tend to see themselves as responsible for failures, but not successes. In a qualitative study, Filipino mothers and fathers expressed the view that children do not have a “mind of their own,” have not yet developed reason and an understanding of reality, are impulsive, demand immediate gratification, and possess a natural penchant for mischief (De la Cruz, Protacio, Balanon, Yacat, & Francisco, 2001). Given these characterizations of children, it makes sense that Filipino parents hold themselves responsible for failures in caregiving. It is not clear why parents would not also hold themselves responsible for successes, but this might be part of a pervasive cultural pattern of modesty and not elevating oneself above others (Bulatao, 1992). With respect to their attitudes, like Kenyan parents, Filipino parents rated progressive and modern childrearing attitudes low relative to other countries and authoritarian attitudes high, consistent with authoritarian attitudes found among Filipino parents in previous research (De la Cruz et al., 2001; Hoffman, 1988).

Swedish parents rated uncontrollable success relatively low; that is, they see themselves and their children as responsible for successful caregiving. In the Swedish cultural context of emphasizing children’s agency and egalitarian relationships between children and parents (Carlson & Earls, 2001), these attributions regarding successful caregiving make sense. With respect to their authoritarian versus progressive childrearing attitudes, Swedish parents hold views that are the inverse of Kenyan and Filipino parents; that is, relative to other countries, Swedish parents rated authoritarian attitudes low, whereas they rated progressive and modern attitudes relatively high. In this sense, the pattern of Swedish childrearing attitudes parallels the pattern in China (Figure 2). Since the early 1900s, Sweden as a nation has emphasized children’s rights and equality between children and parents (Key, 1995). Sweden has been a leader in progressive social policies (e.g., with respect to generous parental leave and the aga law prohibiting corporal punishment), which are likely both a consequence and perpetuator of progressive attitudes.

In Thailand parents rated attributions about uncontrollable success, adult-controlled failure, and child-controlled failure high relative to the other 8 countries in the PAC. In other words, Thai parents see themselves and their children as responsible for negative childrearing outcomes but less so positive ones. In addition, Thai parents, like parents in Sweden, professed progressive and modern childrearing attitudes. Previous research has found that Thai parents stress the importance of children’s obedience (Cameron, Tapanya, & Gillen, 2006), but in this comparative analytic approach Thai parents appeared less oriented toward authoritarian attitudes than the average across the other PAC countries.

Finally, relative to other countries in the PAC, parents in the United States rated child-controlled failures relatively low, but they rated uncontrollable successes and perceived control over childrearing failures relatively high; that is, parents in the United States do not attribute caregiving failures to their children or successes to anyone, but assume responsibility for failures in caregiving. Finally, with respect to their progressive-authoritarian attitudes, parents in the United States rated authoritarian attitudes low relative to parents in other PAC countries. These findings accord with the U.S. American cultural emphasis on individualism (Tamis-LeMonda & McFadden, 2009), which focuses on freedom, choice, and autonomy. On average, parents’ attitudes about parenting seem to reflect cultural attitudes about granting children the freedom to make their own choices and express their own points of view.

Specific explanations for country-specific patterns are best left to ethnographies about the nature of children and the nature of parenting in specific countries, and many can also be found in individual country papers in this Special Issue. Folk theories and ethnotheories evolve within particular cultures for such purposes (Reid & Valsiner, 1986). Ethnotheories focus on particular childrearing cognitions that characterize peoples who share culture (Harkness & Super, 1992). It is assumed that members of a society employ childrearing methods that are derived from an underlying belief structure regarding the nature of child growth and development (LeVine, 1988). Every culture is in part instantiated in parent cognitions about the natures of parenting, childhood, and development. In some models cognitions are created through the transactions between culture and the individual, whereas in other models cognitions are created through the course of social interchanges between individuals. Of course such schema can be expected to vary among cultures and, to some extent, among individuals within the culture.

In general, country differences between attributions regarding adult-controlled and child-controlled failures of caregiving might be explained by the ways different peoples perceive control. Some may be more likely to think that they have control over outcomes in life, and this orientation could apply to parenting. Attributional patterns that reflect a generalized lack of perceived control have been found in Western samples to be associated with less effective parenting. For example, parents who believe that their newborns’ perinatal problems are outside their control tend not to adapt as well as do parents who see such problems as controllable (Tennen, Affleck, & Gershman, 1986). Later, too, parents who believe that they lack control in caregiving are more likely to engage in harsh parenting practices and show a lack of positive affect (Janssens, 1994). Parents’ beliefs in certain controllable causes, such as child effort, tend to predict their children’s academic success, whereas parents’ beliefs in uncontrollable causes, such as luck, are likely associated with children’s underachievement (O’Sullivan & Howe, 1996).

Similarly, country differences in progressive and authoritarian attitudes articulate with societal encouragement of child agency. Parents who hold more authoritarian attitudes may encourage less agency in their children than parents who hold more progressive attitudes. The effects of different parenting styles on children have been shown to vary within and across culture. For example, Leung, Lau, and Lam (1998) studied relations between parenting styles and academic achievement in Australia, China, Hong Kong, and the United States. Authoritarian parenting was associated with lower academic achievement in Australia, China, and the United States, but it was associated with higher academic achievement in Hong Kong. Furthermore, within Australia and the United States, authoritarian parenting was associated with higher academic achievement among parents with low education. Encouraging culturally appropriate child agency could lead to better outcomes for children.

Similarities and Differences between Mothers and Fathers

Overall, mothers and fathers did not differ in mean levels of any of four caregiving attributions, including uncontrollable success, adult or child controlled failure, and perceived control over failure (but see interactions for Italy, Sweden, and the United States). However, mothers and fathers differed in mean levels of all three caregiving attitudes, with mothers professing more progressive and modern childrearing attitudes and fathers more authoritarian ones. At the same time that mothers and fathers were similar in their mean levels of attributions and concordant in their attributions and attitudes, the two parents differed in their attitudes, and, though significant, mothers’ and fathers’ attributions and attitudes shared only 4% and 18% of their common variance. Medium effect sizes were found for concordances between parents in the same family for attributions and large effect sizes for attitudes, even after controlling for parents’ age, education, and possible social desirability bias.

What factors might explain similarities between parents? Both self-selection and mutual socialization could explicate why parents independently report possessing similar mean level attributions and relatively similar levels of attributions and attitudes regarding parenting. Through assortative mating, men and women who are similar on a number of dimensions (parenting cognitions perhaps being one) may be more likely to select into relationships and have children with one another than are men and women who hold divergent beliefs (Luo & Klohnen, 2005). Once they are in a relationship, men and women may influence one another’s attributions and (more) attitudes toward greater consonance. In some societies, too, culture instructs mothers and fathers alike in uniform conceptions of caregiving and child development (Durrant & Olsen, 1997), which may contribute to or reinforce discussions between parents about childrearing. Such discussions could result in more similar cognitions. Mother-father similarities could also be influenced by legislation that encourages parents to take equal caregiving responsibility (Haas, 1996). In Sweden, for example, about 80% of mothers and 90% of fathers work outside the home and are given equal opportunities to combine work and family (Allard, 2007). Swedish legislation, designed to encourage both parents to stay at home with their child, could help to render mothers’ and fathers’ attributions and attitudes relatively more similar to each other. Finally, once a child is born, responding to the same stimulus (“child effects”) could engender and maintain interparental similarity. Although they did covary, mothers’ and fathers’ attributions and attitudes were more different than the same, regardless of culture. One reason that mothers and fathers might differ in their parenting cognitions is that, on average, mothers spend more time with children and have more responsibility for their day-to-day care and well-being than do fathers (Day & Lamb, 2004). Moreover, mothers have been socialized into the parenting role more intensively than fathers, which may lead to more reflection and acceptance of alternative explanations of development on the part of mothers. Factors of motivation and experience in interactions may shape parents’ different cognitions. Of course, different explanations could obtain for attributions versus attitudes and mean level versus relative similarities and difference. However, the cross-cultural data undermine the plausibility of these general explanations because of the variation in mean level results between attributions and attitudes. Mothers and fathers did not differ in their mean levels of attributions regarding successes and failures in parenting, but did differ in mean levels of attitudes, and so explanations of differences in cognitions attributable to structural characteristics, like time with child, as well as explanations built on common cause, like child effects, lose force.

Whether the evidence is stronger for selection or socialization factors in explaining similarities between wives and husbands poses a question that defines further research. It is noteworthy in this regard that, on an assortative mating explanation, relatively strong similarities in childrearing cognitions would be hypothesized to come into play well before actual childrearing. Parents might opt into a relationship with someone who has similar cognitions about parenting. Explanations that appeal to parents electing into relationships with similar partners and shaping one another to become more similar over time are not mutually exclusive. Romantic partners might select each other because they share similar attitudes in general, but not specifically parenting attitudes. Once in a relationship, and with the advent of a child, discussions about parenting could eventually contribute to mother-father similarities. These considerations of the mother-father balance in parenting lead to our final point, about coparenting.

Coparenting

In most cultures, as in our samples, two collaborating persons, namely the child’s mother and father, coparent to meet the challenges of childrearing. Mutual support and cooperation contra antagonism and undermining are presumed to be important to succeed at this challenge (e.g., Teubert & Pinquart, 2010). Where parenting style may describe mothers’ and fathers’ parenting in their individual interactions with a child, the concept of coparenting describes collaboration in childrearing of two parental figures who share responsibilities for at least one child (Feinberg, 2003). Given that parents’ cognitions are multiply determined and based partly on their own families of origin, it is not surprising that coming to agreement on childrearing issues is an area of frequent difficulty, according to parents themselves (Feinberg, 2002). Children growing up in families that afford greater consistency will have an easier time divining and internalizing rules of comportment, achieving effective self-regulation, and coming to perceive a stable and trustworthy interadult alliance in the family than will children from families where the coparents are less consistent with one another.

The conceptual framework of coparenting operationalized by Feinberg (2003) includes four overlapping domains: agreement or disagreement on childrearing issues, division of child-related labor, support/undermining for the coparental role, and the joint management of family interactions. The first component of coparenting is the degree to which parental figures agree on child-related topics, such as attributions and attitudes about childrearing. We studied this first component of coparenting in mothers and fathers in the same family. Childrearing agreement, refers, not to interparental dynamics, but to whether and the degree to which parents’ views of how to rear a child are similar or not. This aspect can be seen as a fundamental platform on which coparenting relationships are built. If parents do not agree with each other about how to parent, the opportunity for conflict will be greater and coparenting supportively is likely to require substantial and ongoing negotiation and compromise.

The nature of coparenting in each family is likely affected by individual factors, but individual factors are formed in the context of larger cultural themes. Much, though not all, of the research on coparenting to date has been conducted with somewhat limited samples comprised of European descent, North American, middle-class families. Little research examines coparenting among more diverse family types, cultures, contexts, and stages (Hortacsu, 1999; McHale, Rao, & Krasnow, 2000). Yet, the form of the coparenting relationship might be shaped to a large extent by parents’ beliefs, values, desires, and expectations, which in turn are influenced by the dominant culture as well as associated subcultural themes within socioeconomic, ethnic, and religious groups.

A basic principle we would expect to obtain is that, regardless of who does what, socialization of young children should proceed most efficiently when there is functional agreement and coordination between parenting figures. Parents who are on the same page and who work in concert are in a better position to afford a stable, consistent, and predictable environment for children than are parents who work at cross-purposes. Preschool-aged children from families where larger discrepancies are found between mothers’ and fathers’ parenting cognitions exhibit more behavior problems than do children whose parents are more similar (Block, Block, & Morrison, 1981). These observations have been replicated by Deal, Halverson, and Wampler (1989), using an index of childrearing agreement-disagreement, and by Jouriles, Murphy, Farris, and Smith (1991), who employed a self-report estimate of coparental disagreement. However, Deal, Halverson, and Wampler (1999) hypothesized and found support for the view that what drives this relation is not parents’ attitudinal similarity per se, but rather the adherence of each parent to normative attitudes of the culture. Hence, any two parents sharing greater similarity in the standard cultural views of parenting or child development are socializing their children into socially accepted behavior. Not surprisingly, those children are considered to function more competently in the culture.

Limitations and Directions for Future Research

Certain characteristics of our samples limit the generalizability of our findings. Families are not static, but are relentlessly changing due in part to developmental processes occurring within all individuals. Thus, in addition to individual and between-family differences, coparenting changes over time (e.g., Kreppner, 1988). We studied mothers and fathers with children of a specific age (M ≈ 8 years). So, interparental agreement in terms of mean and relative levels could differ in parents of younger or older children as it would in parents married for different lengths of time. We also studied families from urban areas in each country. There may be within-country regional differences in parenting attributions and attitudes, with parents in urban areas perhaps holding more progressive and less authoritarian attitudes than parents in rural areas. For the purpose of comparing findings across countries, we analyzed data from all three U.S. ethnic groups as a single sample, but there are some ethnic group differences in attributions and attitudes within the United States (Lansford et al., 2012). Finally, we attempted to ensure that the measures used in this study were appropriately translated and adapted to each culture, but internal consistency was not ideal and we have no information about whether the scales were statistically invariant across countries.

In Western literature, lack of agreement between parents has been linked to child behavior problems and other negative outcomes (Deal et al., 1989; Vaughn, Block, & Block, 1988). It remains to be seen if such are the consequences in other contexts, where mothers and fathers are not similar in attributions but where delineations in parenting roles are culturally normative. It is possible that mothers’ attributions and attitudes may be more influential in determining children’s outcomes, and father’s cognitions, and the concordance between the two, may matter less. These issues are candidates for further study.

Recognition of these family relationships is important to foster a comprehensive understanding of the determinants of family life and to ensure that we do not neglect potential targets of intervention. What factors moderate the sizes of associations in coparenting and between coparenting and child adjustment? Other possible moderators include: children’s age, gender, clinical versus nonclinical status, family income, marriage status of parents, and design (cross-sectional versus longitudinal) and design biases. Children and adolescents of families residing at lower socioeconomic statuses are at elevated risk for developing problem behavior (Bradley & Corwyn, 2002). Under such adverse circumstances, positive coparenting (in terms of high agreement) may be of particular importance in fostering positive child development. Another sample limitation of this study is that, although we were able to examine mothers and fathers from 9 countries, we were not able to include others. Because of our focus on examining relations between mothers’ and fathers’ attributions and attitudes, our analyses were also limited to those families in which both a mother and father were available to respond to our measures. These included families in which the parents had never married (8%) or in which they had divorced or separated (3%), but there were still two parents to provide data. In addition, we assume greater disagreement in samples with separated or divorced parents because parental separation or divorce give more reasons for disagreements or conflicts about childrearing as well as triangulation and less opportunity for parental cooperation.

Children and youth do not only recognize and respond to striking aspects of parental behavior, like conflict and active coalition creation, but also to more subtle and perhaps hidden aspects, such as lack of agreement and absence of cooperation and support between parents. Grych, Seid, and Fincham (1992) showed that children recognize these subtle aspects of their parents’ interactions. Concerning the psychological mechanisms underlying associations between coparenting and child adjustment, positive agreement and cooperation patterns between parents may promote the development of social competence in children. Whether agreement is facilitating or protective across countries warrants further study.

Conclusions

The facet of parenting science concerned with cognitions attempts to understand what cognitions parents hold, why they hold specific cognitions, what functions those cognitions serve, how those cognitions are shared, and the effects of those cognitions on parents, children, and families. Across 9 countries, we found both country and parent gender variations in parents’ childrearing attitudes, but only country differences in attributions. Overall, moderate concordance also emerged between mothers’ and fathers’ attributions and more so attitudes. Country and parent gender differences suggest that likely powerful cultural processes help shape childrearing attitudes. Despite country and parent gender differences, however, the overall means revealed that, on average, mothers and fathers from all different countries held more progressive than authoritarian parenting attitudes. According to Norenzayan and Heine (2005), this kind of converging evidence across cultural groups is a key to understanding the generalizability of psychological processes. In conjunction with the other papers in this Special Issue, this work contributes to the literature a novel focus on ways that culture and parent gender relate to mothers’ and fathers’ attributions regarding successes and failures in caregiving and progressive versus authoritarian attitudes about childrearing.