Abstract
This study examined grandchildren’s perceptions of the quality of their relationship with their grandmothers and how these perceptions relate to psychosocial outcomes. Eighty-two youth from Mexico and 99 youth from the United States aged between 13 and 16 participated. Results suggested that both cultures benefit in unique ways from positive relationship with their grandmothers. Yet, there were also differences in the relational experience of grandmother–grandchild relationships across cultures. Specifically, grandchildren in the U.S. sample reported higher relationship quality, relational competency, and self-efficacy than the grandchildren in the Mexico sample. Within the U.S. sample, relationship quality was associated with grandchildren’s relational competence, while in the Mexico sample, relationship quality was associated with self-efficacy. Limitations of the present study include restricted generalizability to other age groups and to grandfathers and that only one element of the grandparent–grandchild dyad was sampled. Longitudinal research will improve our understanding of the causal dynamics of grandparent–grandchild relationships.
Relationships between grandchildren and their grandparents (about which we know more regarding grandmothers; see Hayslip & Page, 2012) are influenced by the interplay of many factors, that is, sociohistorical (e.g., economic pressures, technological changes), cultural factors (e.g., roles and values about grandparents’ influence on their grandchildren’s lives), family structure differences (e.g., marital status, relationship of parents with grandparents, custodial grandparenting), and individual differences (e.g., health, gender, grandchild’s birth order; Uhlenberg & Kirby, 1998). For example, one sociodemographic historical difference between current grandparent cohorts and those of the previous century is that of mortality. Although 99% to 100% of children in both 1900 and 2000 had at least one grandparent alive at birth, by age 40, 21% of adult grandchildren had one or more grandparents living in 2000, while only 1% did so in 1900. Thus, grandparents and grandchildren today may have longer relationships borne of the grandparent’s greater longevity. In addition, as fertility levels have leveled off relative to earlier in this century, the number of grandchildren has decreased, and grandchildren are more likely to have greater support from and contact with grandparents (Uhlenberg, 2009). Increasing health-care advances, and more years in comparatively better health, economic improvements, and higher average educational status in successive cohorts has also served to close the gap between grandparents and grandchildren (Uhlenberg, 2009; Uhlenberg & Kirby, 1998).
Further contributing to such complexity of grandparent–grandchild relationships is the reciprocal and dynamic nature of the relationships among grandparents, parents, and grandchildren, leading to elaborate and evolving patterns of interaction (Hayslip & Page, 2012). This reciprocity allows grandparents to have both a direct and indirect influence on their grandchildren (e.g., teaching values and skills to grandchild), impacting how the parent raises the child (Tomlin, 1998). Conversely, grandchildren may act to change grandparent behaviors or may elicit different grandparent responses in the context of their shared relationship. In addition, the nature of the grandparent–grandchild relationship is also likely to change as both individuals become older and accumulate more differential life experience and role change (Tomlin, 1998).
Culture and Grandparent–Grandchild Relationships
Culture is another fundamental factor in shaping the grandparent–grandchild relationship, and the grandparent–grandchild relationship is an integral part of intergenerational cultural transmission, the passing on of ideas and value for successful adaption (see, e.g., Kopera-Frye, 2009; Wiscott & Kopera-Frye, 2000). The daily family interactions, practices, and structure, which are heavily influenced and patterned by culture, shape the way developing children come to know and understand the world. Culture pervades all aspects of life; it organizes perceptions, cognitions, and behaviors (Cole, 1999). The relationship between grandchild and grandparent must therefore necessarily take place within a cultural context, as such present and future relationships are shaped by the culture in which they are embedded (Cole, 2005). Culture influences the way that grandparents experience aging and grandparenting, as well as the ways that grandchildren influence their grandparents and their own development.
Given the unique historical contexts and ecological challenges from which each culture emerges, we would expect there to be variation in the structure, roles, values, and practices of each culture. As such, there must be an appreciation for the relevant cultural context to fully understand the meaning and function of interactions in the grandparent–grandchild relationship, how each party is perceived by the other, and how this relationship affects each party. For example, Strom, Lee, Strom, Nakagawa, and Beckert (2008) have discussed how changes in China over the past 20 years (e.g., the end of martial law and media censorship) have dramatically changed Taiwanese culture and daily life (e.g., increases in women working and income, youth more immersed in Western culture), and thereby affected grandparent–grandchild relationships. Strom et al. explored how sociocultural changes put a strain on traditional cultural values (e.g., moving away from intergenerational cohabitation and hierarchical structure), intergenerational relationships, and how they influenced each generations’ perceptions of grandmothers’ role and functioning. Looking at cultures from within allows us to better understand the processes by which different variables within the respective ecological systems take on different meanings and lead to divergent outcomes for interpersonal relationships and individuals in that culture (e.g., the macrosystem affects the microsystem and the individual; Bronfenbrenner, 1986).
Sharing culture has been described as one of the primary functions of grandparenting (Wiscott & Kopera-Frye, 2000). In a study of ethnically diverse students, grandchildren in general felt that it was typical to engage in cultural activities with grandparents including sharing stories and family photos and learning about family traditions. They also reported a moderate influence of their grandparents, particularly on issues of religion, family, morality, and personal identity. While ethnicity was related to sharing everyday and cultural activities, as well as positive affect toward grandparents and culture, grandparent–grandchild relationship quality was described by students as moderately strong and did not differ by ethnicity (Wiscott & Kopera-Frye, 2000). Thus, to the extent that grandparents and grandchildren engage in shared activities, culture is transmitted, but this may not affect perception of relationship quality among grandchildren.
In another example, Pratt, Norris, Hebblethwaite, and Arnold (2008) found that adolescents who rated themselves as having more close relationships with their maternal grandparents were more likely to tell stories about experiences of the grandparents’ value teaching, and these stories were more likely to be interactive, specific, and caring. Telling stories with these characteristics was related to higher generativity scores for the adolescents, indicating higher concern about contributing to future generations. These results suggest that closeness to grandparents relates to more elaborate and positive perceptions of their value teaching, which would intuitively seem to be important for intergenerational cultural transmission.
Though the constructs of individualism and collectivism as they apply to cultures have been criticized as overly generalized in nature (Miller, Goyal, & Wice, 2015), American culture is typically considered more individualistic, and Hispanic culture is typically seen as more collectivistic in nature (see Greenfield, 2009; Manago, Guan, & Greenfield, 2015; Nisbett, 2003); it is reasonable to expect that differences in cultural values and associated familial structure and practices could be related to cultural differences in the experience of grandparent–grandchild relationships. One cultural value that is central to Hispanic culture and has received significant attention in the literature is familism, which is characterized by the desire to maintain strong family ties, value them over individual needs and desires, and the expectation that the family will be the primary source of instrumental and emotional support (Halgunseth, Ispa, & Rudy, 2006). Familism influences Hispanic family structure in many ways, including an emphasis on intergenerational relationships, frequent shared activities and high levels of interaction, and coresidency with older family members (Silverstein & Chen, 1999).
Among Mexican-American grandparents, 83% speak via telephone, and 68% in person, once a week or more with their young grandchildren (Schmidt & Padilla, 1983). Mexican-American grandparents frequently engaged in culture-transmitting activities during these times, such as teaching religion (85%), telling stories from the grandparent’s own childhood (77%), and teaching about family history (55%; Schmidt & Padilla, 1983). The amount of shared activity time with grandchildren relates to increased satisfaction, feelings of success in family role performance, less difficulty with grandparent obligations, and importantly greater teaching involvement with grandchildren among Mexican-American grandparents (Strom, Buki, & Strom, 1997). Both frequency and amount of contact therefore allow grandparents to teach and transmit culture to their grandchildren, as well as foster feelings that lead to affection.
In a study comparing Euro-American grandchildren and Mexican-American grandchildren, ratings of grandparent–grandchild affection were generally moderate to high; but ratings were relatively higher for Mexican-American grandchildren than Euro-American grandchildren (Giarusso, Feng, Silverstein, & Bengtson, 2001; Wiscott & Kopera-Frye, 2000). Euro-American grandparents report more affection for and similarity with their grandchildren than the grandchildren report, while gaps are much closer among Mexican-American dyads (Giarusso et al., 2001).
Because culture plays a significant role, it makes sense that as Mexican-American families acculturate to the United States, this would also impact the grandparent–grandchild relationship. A study of Mexican-American grandparents and grandchildren assessed in the early 1980s showed that, although grandchildren were more acculturated than their grandparents, both grandparents and grandchildren valued familism (Silverstein & Chen, 1999). On one hand, the difference in acculturation between grandchildren and grandparents, grandchildren’s value of family responsibility, and grandchildren’s ability to speak Spanish predicted more interactions between the dyad. On the other hand, grandparents’ level of education predicted more affection (Silverstein & Chen, 1999). Among Mexican-Americans, greater acculturation and grandchildren’s perception of similarity predict a larger gap between the affection reported by grandparent and grandchild toward each other (Giarusso et al., 2001). Thus, for both Mexican-American grandparents and grandchildren, collectivistic values and lack of acculturation predict more affectionate and interactive relationships. In turn, affectionate relationships with adult grandchildren predicted greater life satisfaction for older Mexican-American grandparents (Markides & Krause, 1985), highlighting the importance of grandparent–grandchild relationships among Mexican-Americans.
Despite the integral role of the grandparent–grandchild relationship in the grandchild’s development, the salience of cultural transmission of values from one generation to the next (see Hayslip, Maiden, Page, & Dolbin-MacNab, 2015; Hayslip & Page, 2012), and an ongoing interest in culture and grandparenting (Hayslip & Musil, 2017; Szinovacz, 1998), few studies have directly compared grandparent–grandchild relationships across cultures. Among those that have attempted to do so, these comparisons have been limited geographically, specifically within the United States, such that extraneous influences, for example, degree of acculturation, may affect results.
Given the cultural differences discussed earlier, and with the knowledge that grandparent–grandchild relationships exist in a cultural context, we expected differences in grandparent–grandmother relationships within and outside of the United States. Thus, the purposes of this study were (a) to examine grandchildren’s perceptions of their relationship quality with their maternal grandmothers in American and Mexican samples and (b) to explore how these perceptions relate to psychosocial outcomes of the grandchildren as a function of culture.
Method
Participants and Procedure
Mexican sample
Eighty-two youth aged between 13 and 16 (44 male and 38 female) whose grandmothers were living at the time the study was conducted were recruited from a high school and middle school in Jalisco, Mexico. This high school was selected via professional contacts between the school’s administration and both the second and sixth authors; school administrators agreed to contact parents about the project with an information sheet about the survey provided to each parent. After having first obtained written parental consent via the schools, the researchers read and explained each section of the questionnaire to a classroom of students one at a time, and the students wishing to participate and from whom parental consent had been obtained each completed the forms by themselves. The surveys were presented to each adolescent in his or her native language, having been first translated into Spanish and back-translated into English by the second and sixth authors, each of whom was born and raised in Mexico.
U.S. sample
For the purpose of this study, a subset of participants was taken from a larger U.S. sample (n = 324) collected by the first and third authors and matched with the Mexican sample primarily based on age and secondarily on gender (the gender distributions did not vary across samples, χ2 = .311, ns). This subset consisted of 99 youth aged between 13 and 16 (49 male, 50 female). The initial sample was recruited from local high schools and included youth whose maternal grandmothers were living at the time the study was conducted. Meetings were arranged with principals at local high schools to secure their agreement to announce the study in classes and at parent–teacher association meetings. In compliance with research ethical standards, written parent permission was first obtained for students under age 18, and these students also provided their informed assent to participate in the research; students each completed the forms by themselves.
In both cases, surveys were completed independently of parents so as to avoid any bias via influence by the parent regarding the child’s perceptions of his or her relationship to the maternal grandmother. Adolescents’ enrollment in the study was voluntary; they were not compensated in any form for their participation.
Measures
Participants were asked to complete questionnaires beginning with a short demographic section (e.g., requesting information on participants’ demographic backgrounds, family structure, the number of years since divorce if one had occurred) and a variety of measures targeting the participants’ relationship to their grandmothers and participants’ psychological adjustment.
Inventory of Parent and Peer Attachment
The Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987) is a 53-item self-report inventory measuring adolescents’ and young adults’ relationship quality with parents and peers. For the current study, we used the trust, communication, and alienation subscales for parents, modified to be appropriate for grandmothers. The alienation subscale was reverse scored here to make the interpretation consistent with the other subscales, where higher alienation scores indexed greater degrees of attachment. Samples of modified items include the following: “My grandmother [formerly parent] respects my feelings” and “I feel my grandmother [formerly parent] is successful as a grandparent [formerly parent].” Participants completed this measure with respect to each grandparent (maternal grandmother and grandfather and paternal grandmother and grandfather) although we used data only for maternal grandmothers for the reasons described previously. Further, participants reported many different combinations of living grandparents, making consistent comparisons unwieldy from an analytic standpoint. Previous validity studies conducted by Armsden and Greenberg (1987) have shown that parent attachment scores correlated significantly with family support as well as with a tendency to seek out parents in times of need. In the present study, in both the U.S. and Mexican samples, coefficient alphas exceeded .84 for each subscale.
Contact with grandmothers
Extent of contact between grandchildren and maternal grandmothers at the time the study was conducted was assessed by two items, one indicating the current frequency of visits and the other the current frequency of phone conversations with their maternal grandmothers. Response options included “daily,” “at least once per week,” “at least once per month,” and “less than once per month.” These items were summed to provide an index of contact the participants had with their grandmothers, with larger values indicating more contact (alpha = .70) for participants’ contact with their maternal grandmothers.
Relationship Competence Scale
The Relationship Competence Scale (Hansson, Jones, & Carpenter, 1984) is a 100-item measure of competence in interpersonal skills. Relational competence measures abilities that serve to maintain and enhance relationships and is composed of the intimacy, trust, interpersonal sensitivity, empathic concern, and perspective-taking subscales. Published internal consistency coefficients for the subscales range from .76 to .90. Responses range on a scale from 1 to 4 with 4 meaning “strongly agree or very much like me” and 1 meaning “strongly disagree or very much unlike me.” Sample items include “People are usually very dependable and trustworthy” and “I really am thoughtful and considerate of others.” The interpersonal sensitivity and trust subscales were used in the current study, and these scales showed adequate reliability here, with alphas exceeding .80 in both the U.S. and Mexican samples.
Self-Efficacy Scale
The Self-Efficacy Scale (Sherer et al., 1982) is a 23-item scale that was used to measure an individual’s estimate of his or her ability to cope with a situation. Woodruff and Cashman (1993) found that a general self-efficacy subscale could be divided into three subscales (general efficacy magnitude, strength, and competence). Concurrent validity estimates are reported by Woodruff and Cashman (1993) who used Pearson correlations to demonstrate moderate correlations between the Self-Efficacy Scale and the Mastery Scale (.54; Pearlin, Lieberman, Menaghan, & Mullan, 1981) and the Rosenberg Self-Esteem Scale (.54; Rosenberg, 1965). The magnitude, strength, and competence subscales of general self-efficacy were used in the current study. Alphas for these scales exceeded .86 in the U.S. sample; in the Mexican sample, they exceeded .74. Higher scores indexed greater self-efficacy.
Hopkins Symptom Checklist
The Hopkins Symptom Checklist (HSCL; Derogatis, 1974; Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974) is a 58-item self-report symptom inventory using a 5-point anchored rating scale and covering five symptom dimensions: somatization, obsessive-compulsive, interpersonal sensitivity, anxiety, and depression. It is the precursor, and therefore very similar in terms of item content and targeted assessment areas, to the Symptom Checklist 90 (Derogatis, Lipman, & Covi, 1973) and the Brief Symptom Inventory (Derogatis & Melisaratos, 1983), which have been used in numerous studies. Because the participants were not sampled from a clinical population, subscales were selected on the basis of their relevance to nonclinical adolescent and young adult populations and included depression, anxiety, and somatization. Derogatis et al. (1974) report alpha coefficients from .84 to .87 for individual HSCL subscale scores. Several studies have supported the criterion-related validity by demonstrating the sensitivity of this measure to a variety of treatment effects (see Derogatis et al., 1974 for a review). The interpersonal sensitivity subscale was not used due to its overlap with the Relationship Competence Scale, nor was the obsessive-compulsive subscale due to the low probability of occurrence in the targeted population (see Derogatis et al., 1974). Coefficient alphas for the depression, anxiety, and somatization subscales exceeded .84 in both the U.S. and Mexican samples. Importantly, HSCL subscales were rescaled here so that higher scores indicate better adjustment.
Data Analysis
The objectives of this study were to compare youth living in the United States and Mexico on the extent to which (a) the groups differed on the means for grandmother–grandchild contact, relationship quality, self-efficacy, relational competence, and psychological symptoms and (b) the groups differed in the strength of the relationships between grandmother relationship quality and youths’ psychological adjustment.
Results
To explore the extent to which there were cultural differences in a variety of aspects of the grandparent–grandchild relationship, as well as in indicators of grandchild adjustment, we conducted a series of multivariate analyses of variance (listwise deletion) of grandchild closeness to/contact with the maternal grandmother, grandchild self-efficacy, as well as grandchild relational competence, attachment, and psychological symptomology.
Contact With Grandmothers
Analysis of contact with grandmothers indicated that grandchildren in the United States interacted with their grandmothers more frequently (M = 6.81, SD = 2.18) than did grandchildren in Mexico (M = 5.98, SD = 2.61; F(1, 79) = 4.90, p < .03).
Grandparent Attachment/Relationship Quality
This analysis revealed a reliable cultural difference for indicators of grandparent attachment (trust, communication, alienation) as a set, F(3, 77) = 47.63, p < .01, eta2 = .650, with univariate comparisons being statistically significant for trust, F(1, 79) = 99.15, p < .01, eta2 = .557, and communication, F(1, 79) = 14.85, p < .01, eta2 = .158; each favoring the U.S. sample (see Table 1). No cultural differences were obtained for alienation.
Descriptive Statistics for Primary Variables and Results of Statistical Tests of Mean Comparisons for Youth Living in the United States and Mexico.
Note. Higher scores on alienation index greater attachment; higher HSCL scores index more adequate adjustment.
p < .01.
Relational Competence
For indicators of relational competence (trust, interpersonal sensitivity), main effects of the variables as a set were statistically reliable, F(2, 137) = 6.50, p < .01, eta2 = .087. While interpersonal sensitivity scores favored the U.S. sample, F(1, 138) = 9.26, p < .01, eta2 = .063 (see Table 1), trust scores did not vary across cultures.
Self-efficacy
For indicators of self-efficacy (magnitude, strength, competence), cultural differences for the variables as a set were found, F(3, 136) = 53.92, p < .01, eta2 = .543, with scores for magnitude, F(1, 138) = 18.48, p < .01, eta2 = .118; strength, F(1, 138) = 160.96, p < .01, eta2 = .538; and competence, F(1, 138) = 23.67, p < .01, eta2 = .146, all being higher in the U.S. sample.
Psychological Symptomology
For indicators of psychological symptomology, cultural effects were significant for the variables as a set, F(3, 103) = 52.19, p < .01, eta2 = .603, with scores for somatization, F(1, 105) = 156.83, p < .01, eta2 = .599; depression, F(1, 105) = 69.22, p < .01, eta2 = .397; and anxiety, F(1, 105) = 124.10, p < .01, eta2 = .542, all being higher (indicating better adjustment) in the U.S. sample (see Table 1). Table 2 presents the correlations among study variables by nationality.
Bivariate Correlations Among Relationship Quality, Influence, and Adjustment Variables by Nationality.
Note. Correlations among variables reported by youth living in the United States appear below the diagonal. Correlations among variables reported by youth living in Mexico appear above the diagonal. The sample covariance matrix is available from the authors by request.
Comm = Communication; Comp = Competence; Som = Somatization.
p < .05. **p < .01.
Thus, youth in the U.S. sample, who reported greater contact with their grandmothers, also reported higher levels of closeness to their grandmothers, self-efficacy, interpersonal sensitivity, as well as more adequate psychological functioning than youth in the Mexican sample.
Grandmother Relationship Quality as a Predictor of Grandchild Functioning
A multivariate general linear modeling approach was used to test the hypothesis that better grandmother–grandchild relationship quality (as indexed by trust, communication, and alienation) was associated with better relational competence and self-efficacy as well as fewer psychological symptoms and that the relations between relationship quality and these domains differed by sample (i.e., U.S. and Mexican grandchildren). As these research questions are both novel and exploratory, we relaxed the typical progression of multivariate modeling, which requires a significant multivariate test to proceed to univariate tests. In the event that the multivariate effect was not statistically significant, we report the post hoc results of the univariate tests to explore the nature of the interaction with culture regarding the association between indicators of grandmother–grandchild relationship quality and measures of relational competence, self-efficacy, and psychological symptoms. Given a significant univariate interaction (post hoc or otherwise), we then followed up with within-sample regressions to probe its nature within cultures. Only such relationships are discussed here.
Predicting Relational Competence
In the first model, we examined relationships between grandmother relationship quality (trust, communication, alienation) and relational competence. The multivariate test for the interaction between grandparent trust and sample was significant for relational competence, F(4, 118) = 2.81, p = .029, eta2 = .09. This effect was specific to interpersonal sensitivity, F(2, 60) = 5.53, p = .006, eta2 = .16, with the relationship between the variables being positive and significant in the U.S. sample (b = 0.23, β = .35, SE = 0.08, t = 2.95, p = .005), but a nonsignificant relation being shown in the Mexican sample (b = 0.02, β = .03, SE = 0.13, t = 0.18, ns).
The multivariate test for the interaction between grandmother alienation and sample was marginal, F(4, 118) = 2.20, p = .073, eta2 = .07, with, however, a significant univariate post hoc relation shown for interpersonal trust, F(2, 60) = 3.70, p = .030, eta2 = .11. Examining these relationships within each sample revealed that the effects were significant for both U.S. (b = 0.28, β = .43, SE = 0.08, t = 3.44, p = .001) and Mexican (b = 0.29, β = .42, SE = 0.13, t = 2.14, p = .043) samples, but the effects were somewhat stronger (higher t value) in the U.S. sample (note that the grandmother alienation measure was rescaled so that higher values reflected less alienation).
There was no interaction with sample at either the multivariate or univariate levels for communication.
Predicting Self-efficacy
For the next set of models, we examined relations with respect to self-efficacy (magnitude, trust, competence). There were no significant multivariate effects; however, there was a significant post hoc univariate effect for the interaction between sample and grandmother alienation on self-efficacy magnitude, F(2, 62) = 3.19, p = .048, eta2 = .09. Within-sample regressions revealed that this effect was significant in the Mexican sample (b = 0.32, β = .57, SE = 0.10, t = 3.05, p = .006), but not in the U.S. sample (b = 0.10, β = .21, SE = 0.08, t = 1.17, ns). Similarly, there was a significant univariate interaction effect between sample and grandmother trust, F(2, 62) = 3.83, p = .027, eta2 = .11; however, post hoc testing indicated that the relation was not significant in either sample (United States: b = 0.08, β = .19, SE = 0.08, t = 1.05, ns; Mexican: b = −0.02, β = .21, SE = 0.13, t = −0.13, ns).
Predicting Psychological Symptoms
Finally, we examined relations between indicators of grandmother relationship quality and psychological symptoms (somatization, depression, anxiety). There were no significant multivariate effects; however, there was a significant univariate effect for the interaction between sample and grandmother alienation for depression ratings, F(2, 48) = 2.00, p = .026, eta2 = .16. Post hoc testing revealed that this effect was shown in the U.S. sample (b = 0.63, β = .59, SE = 0.12, t = 5.37, p < .001), but not in the Mexican sample (b = −0.08, β = −.07, SE = 0.25, t = −0.33, ns).
Discussion
Cultural Differences
This study found statistically reliable differences in the relational experience of grandmothers and grandchildren across the U.S. and Mexican samples. Importantly, different conclusions could be drawn when comparing the U.S. and Mexican samples based on the two levels of analysis used in this study—the comparisons across culture in measured scale means versus moderation by sample with respect to the grandmother relationship quality-psychological adjustment relation variables. In this respect, as Cole (1999, 2005) has pointed out, cultural differences questionably presume that culture is causal in nature in explaining such differences. In contrast, we argue that regression findings are more capable of speaking to process, that is, treating culture as a moderator. Such findings speak to the cultural lens through which intergenerational relationships may be seen and therefore can potentially address the why of cultural differences.
On one level, our findings suggest that grandchildren in the U.S. sample reported greater grandmother contact, higher relationship quality/attachment, higher relational competence, and greater self-efficacy than did grandchildren in the Mexican sample (see Table 1). This suggests that higher levels of such indicators of grandchild functioning are paralleled by greater grandparent relationship quality among U.S. participants. The finding that Mexican grandchildren reported less relationship quality is surprising in light of previous work finding higher ratings of grandparent grandchild affection and more interaction and shared activities in Mexican-American samples when compared with Euro-American samples (Giarusso et al., 2001; Schmidt & Padilla, 1983; Wiscott & Kopera-Frye, 2000). This pattern is on the surface, also inconsistent with the greater salience of familism as a value presumably shared by grandchildren in Mexico versus values stressing individualism in the United States (Greenfield, 2009; Manago et al., 2015; Nisbett, 2003). It may also be that in American culture, the nature of the grandparental role is more volitional, and therefore creating meaningful interactions with a grandchild in the American cultural context is more effortful and intentional. The marked investment required to build grandparent–grandchild relationship quality in American culture may account for the differential interpersonal skills and self-confidence among American grandchildren found here versus Mexican grandchildren.
In explaining these differences, it may also be that the measures of interpersonal competence used here are simply too generalized to apply to a specific grandmother–grandchild relationship that is symbolic in nature in Mexico. It is also possible that Mexican grandchildren are indeed impacted negatively by cultural forces (e.g., crime, drug use, poverty, peer relationships) in that the potential benefits of (a) contact with a grandmother to whom one is securely attached, (b) having shared values stressing collectivism and familism bringing the generations closer, and (c) both interpersonal facility and personal adjustment borne of one’s relationship to a grandmother are all undermined. Clearly, these explanations are speculative and underscore the fact that the basis for these findings needs to be further explored, perhaps using observational measures of grandmother–grandchild interactions to more fully understand them, especially among Mexican grandchildren.
Another potential explanation for the cultural differences for here is offered by Ruiz, Roosa, and Gonzalez (2002), which is that in Mexican culture, where there are more people/caregivers included in a given family system, the impact of each person in that system is likely to be less than in a family system with fewer people/caregivers. In addition, as across cultures there is varying meaning ascribed to grandparent–grandchild relationship quality, different cultures may attribute different meaning to socialization or attachment-related behaviors and that these behaviors can be differentially related to adaptive functioning in a given culture (Cole, 1999). For example, in American culture, democratic interaction is highly valued even between adults and children, and this type of interaction is presumed to lead to adaptive functioning. Conversely, in Mexican culture, there is less relationship placed on mutuality and more emphasis placed on respect for adults (respeto; Halgunseth et al., 2006), which helps to maintain familial solidarity. These differing cultural values teach children to have different expectations of what is important and influential in the grandparent–grandchild relationship; these differing expectations may help explain the cultural differences found here. Although it could not be explored here, the potentially culture-specific mediating role of the parent on grandchild-grandmother interactions may also be influential.
That such differences are an artifact of sample-specific scale reliabilities is not supported by the psychometric data for the measures used here, and all grandmother–grandchild related measures correlated positively with one another, to include HSCL subscales. It is worth noting that though they were not statistically significant, correlations between the aforementioned measures and degree of contact were consistently negative, that is, less contact correlated with poorer relationship quality, less relational competence, less self-efficacy, and less adequate psychological functioning in the U.S. sample. In contrast, such relationships were negative only for alienation, that is, more alienation (greater relationship quality) was related more contact in the Mexican sample. Moreover, more contact was correlated positively (though not significantly) with somatization (more adequate functioning) among Mexican youth.
Within-Sample Relationships
Regarding within-sample regression findings, the trust (an indicator of grandmother relationship quality) by sample interaction was significant for relational competence, with higher interpersonal sensitivity being linked to greater relationship quality in the U.S. sample, but not so in the Mexican sample. Likewise, greater interpersonal trust was linked to higher alienation (greater grandmother relationship quality) in both samples, with effects being somewhat stronger in U.S. grandchildren.
Collectively, these findings generally suggest that irrespective of culture, a grandchild’s interpersonal skill development is facilitated to a greater extent by a strong attachment to a grandmother, paralleling the literature on grandparenting (see, e.g., Giarusso et al., 2001; Hayslip & Page, 2012; Hayslip, Shore, & Henderson, 2000; Hayslip et al., 2015; Levitt, Guacci, & Weber, 1992; McGuinn & Mosher-Ashley, 2002; Tomlin, 1998). Together these findings therefore provide strong support for fostering and protecting the quality of the grandparent–grandchild relationship.
We found that regarding indicators of adjustment (HSCL), greater grandmother alienation (better relationship quality) was linked, surprisingly, to higher self-rated depression among U.S. grandchildren, but not among Mexican youth. This might suggest that U.S. grandchildren, despite reporting more positive grandmother relationships, nevertheless report less positive affect exchange. This might be due to intergenerational discord regarding any number of personal or contemporary issues (e.g., drug use, sexuality, violence, peer relationships, experiences at school), or it simply could suggest that the two (grandmother relationship quality and depression) coexist in parallel, in that grandchildren who are depressed may rely on their grandmothers as sounding boards or supports to enable them to cope or express their feelings.
These findings for depressive symptoms among U.S. grandchildren qualify the conclusion that grandparents’ influence on their grandchildren is necessarily positive and reinforce the observation that grandparents can interact with their children in ways or model behaviors that are contrary to a parent’s wishes, intensify adolescent–parent conflicts, and are otherwise counterproductive (see Bailey, Hill, Oesterle, & Hawkins, 2009; Hayslip et al., 2015; Silverstein & Ruiz, 2006).
Regarding self-efficacy, greater alienation (a component of better relationship quality) was associated with higher self-efficacy magnitude in the Mexican sample, but not in the U.S. sample. Thus, at least one aspect of self-efficacy is enhanced via greater attachment to a grandmother among Mexican grandchildren, and in this case, having a strong relationship with a grandmother seems to be a psychologically protective factor for a grandchild in Mexico. These results underscore the importance of the grandmother’s role in Mexico and thus reinforce the findings of Ruiz et al. (2002) who found that parental acceptance, rejection, and inconsistency each had less influence on the self-esteem of children in the Mexican-American sample than children in the European example (where parents may have more influence and indeed, serve as gatekeeper in modulating a grandparent’s impact on the grandchild). That in Mexican culture the grandparent role is vital and is seamlessly embedded in cultural structure (Schmidt & Padilla, 1983; Silverstein & Chen, 1999; Strom et al., 1997) is consistent with this positive influence.
Limitations of the Present Study
A number of limitations must be recognized in interpreting the cross-cultural findings here. In this study, we used the trust, communication, and alienation subscales of the Inventory of Parent and Peer Attachment (Armsden & Greenberg, 1987) to measure relationship quality. Beyond such measures, there may be other characteristics of grandparent–grandchild relationship that would be more salient and culturally relevant to Mexican grandchildren, and these may have greater influence on the development of their relational competence and self-efficacy. As noted earlier, such measures may not reflect the familial nature of grandparenting that may reflect Mexican cultural values, instead focusing upon those that are volitional. Thus, it may be that items in the scales used here presume or tap adjustment and relational processes that more adequately reflect American norms stressing individual choice and control and that values reflecting familism are better assessed by measures that are necessarily dyadic, process-oriented, or interpersonal/interactional in nature. Thus, it may be that Mexican youths’ relationships with their grandmothers transcend measured grandmother relationship quality, interpersonal competence, and self-efficacy in being symbolically interactional in nature.
The findings of this study are limited in generalizability to the ages of the grandchildren in each sample (13–16 years old) and are particular to the regional areas from which the samples were drawn. This investigation was based on self-report data from grandchildren, one element of the grandmother–grandchild dyad. They are also limited in not targeting grandchildren’s perceptions of their grandfathers. Consequently, due to the complex and bidirectional nature of the grandparent–grandchild relationship, this data could be enriched and corroborated by including the grandparent perspective and relying upon observational and qualitative methodology.
As these data raise as many questions as they answer and the earlier explanations for the cultural differences found here are speculative, future research that is both qualitative/observational, emphasizing dyadic interactions is needed to improve our limited knowledge about the nature of adult grandchildren’s relationships with their grandparents (Hayslip, Fruhauf, & Dolbin-MacNab, 2017; Hayslip & Page, 2012) as well as the long-term effects of grandparent–grandchild relationship quality on grandchild development. In this respect, inferences regarding the causal nature of grandmother relationship quality may be alternatively interpreted as bidirectional in nature. Consequently, grandchildren with greater relationship competence or more self-efficacy may be more successful in bonding with a grandmother. Longitudinal work has the potential to establish the nature of such causality.
Perhaps the most important implication for clinicians and researchers is that the dynamics of grandparent–grandchild relationships indeed do vary by culture (see Dolbin-MacNab & Yancura, 2017), wherein distinctions between mean level findings and moderator relationships need to be made (see Cole, 1999, 2005). Not only is there cultural specificity in the impact of grandmother relationship quality on grandchildren, but the nature of that influence likely varies along a continuum of facilitating to detrimental regarding a child’s developing social self. As there is likely not one universal or ideal model to understand such relationships, we cannot assume that similar interventions or interpersonal behaviors will have the same impact in different cultural contexts (McLoyd, Cauce, Takeuchi, & Wilson, 2000). Although the importance of the grandparent–grandchild relationship and the need for interventions targeting the grandparent–grandchild relationship has been acknowledged for several decades (see Hayslip & Page, 2012; Strom et al., 1997, 2008; Szinovacz, 1998), work based upon an empirically grounded knowledge of grandparents’ impact on their grandchildren’s development is needed so that optimal outcomes in grandparent–grandchild relationships in various cultural contexts can be achieved.
Footnotes
Acknowledgments
Yunuen Zihuatlililli Valencia significantly contributed to this article. Yunuen Zihuatlililli Valencia was a graduate student at the time we did this work and we gratefully acknowledge his contribution.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
