Responsiveness in Mother-Infant Social Interactions among Immigrant and Nonmigrant Families: Japanese, South Korean, South American, and European American

Abstract

A culture learning perspective motivated the present study of the acculturation of responsiveness in mother-infant interactions. Several conceptual and analytic features of responsiveness in mother-infant social interactions were examined: Temporal contingency, mean differences in responsiveness among and within dyads, attunement of mother and infant responsiveness withing dyads, and the influence of acculturation on individual responsiveness. Methodologically, acculturation was assessed at group and individual levels in immigrant Japanese, South Korean, and South American dyads in comparison with nonmigrant dyads in their respective cultures of origin (Japan, South Korea, and South America) and their single common culture of destination (United States). In total, 408 mothers and their 5½-month-old infants were observed in the naturalistic setting of the home, and observations were coded for mothers’ speech to infant, social play, and encouraging her infant to look at her, and infants’ looking at mother and nondistress vocalizations. Odds ratios were then generated for mother and infant responsiveness in four types of social interactions: Mother speaks to infant and infant looks at mother (Mother Speak/Infant Attend), mother plays with infant and infant looks at mother (Mother Play/Infant Attend), mother plays with infant and infant vocalizes (Mother Play/Infant Vocalize), and mother encourages infant to look at her and infant vocalizes (Mother Encourage/Infant Vocalize). Five key findings emerged. Specifically, mother and infant responsiveness in Mother Speak/Infant Attend interactions were temporally contingent in all cultures. Mean differences in responsiveness among cultures emerged, and within dyads infants were more responsive than their mothers in Mother Speak/Infant Attend interactions. Mother and infant responsiveness in Mother Speak/Infant Attend interactions were attuned in all cultures. Responsiveness in Mother Play/Infant Vocalize interactions showed acculturation effects at the individual level. Implications of these findings for understanding the development of responsiveness in social interactions and acculturation in immigrant families are discussed.

In the first few months of life, infants display social sensitivities and communicative expectations of their mothers when they engage in dyadic interactions. A basis of these sensitivities and expectations is a partner who will respond contingently to an infant’s social behavior (Legerstee, 2009; Mcquaid et al., 2009; Soussignan et al., 2006), and this dynamic is illustrated dramatically in the infant’s response to “still face” paradigm. Moreover, mothers’ abilities to maintain attention, show warmth and sensitivity, and respond to her infant have repercussions for infant social development (Markova & Legerstee, 2006). However, parent-child relationships and social interactions are known to differ cross-culturally (e.g., Bornstein & Esposito, 2020; Bornstein & Lansford, 2019; Parke & Buriel 2006; Shwalb et al., 2010) and vary with immigration experience (Bornstein & Cote, 2019). Still relatively little is known about young children’s development in immigrant families, despite the fact that their numbers are large and growing (Bhabha & Abel, 2019). Social interactions among immigrant families are therefore a matter of contemporary concern. This study explored and compared several types of social interactions among immigrant families and their nonmigrant counterparts in their countries of origin and destination to better understand maternal and infant responsiveness in mother-infant social interactions among immigrant families.

Cultural Variability in Socialization Goals

Cultural ideas about relationships and interactions vary, and as relevant here include the extent to which mothers and infants are expected to lead and respond in interpersonal interactions. In the United States, European Americans tend to believe that allowing infants to lead interactions with their mothers responding will foster a sense of agency and, ultimately, independence and assertiveness in children (Tamis-LeMonda & McFadden, 2010); these are valued character traits in more individualist and autonomous cultures. However, other cultures, particularly those that favor collectivist and relational characteristics, place more emphasis on teaching children to respect and obey their parents and other authority figures (Mesman et al., 2018), often in order to maintain harmony in interpersonal relationships. Germane to the present study, Japanese parental goals include fostering amae (emotional dependency; Azuma, 2005; Doi, 2002; Ng & Wang, 2019) in the mother-infant relationship and shitsuke (teaching children proper social behavior; Holloway, 2010). Korean familial virtues include zeung-zee (respect for parents’ teachings and authority) and jull-ze (controlling personal desires to maintain harmony; Kim, 2006). South American parental goals include teaching children to be bien educado (well-mannered and behave properly in social situations; Bornstein et al., 1999; de Castro Ribas, 2010), and to achieve this goal parents tend to be more directive and authoritarian rather than autonomy-supportive in interactions with their children (Halgunseth, 2019). Cultural variation in the relative importance of autonomy-promotion and harmony-maintenance in interpersonal relationships may lead to differences in mother-infant social interactions in terms of whether responses are temporally contingent, whether there are mean differences in responsiveness among cultural groups or within dyads, and whether mother and infant responsiveness is attuned (correlated). When socialization goals manifest differently in social responsiveness in the countries of origin and destination, this may present challenges for immigrant families in the process of behavioral acculturation (Bornstein & Cote, 2001; Cote & Bornstein, 2021b).

Theoretical and Methodological Approaches to Acculturation

Theoretically, Ward’s (Masgoret & Ward, 2012) culture learning approach motivated the present investigation of responsiveness in mother-infant social interactions among immigrant families. The culture learning approach investigates immigrants’ behavioral adaptation when the behaviors under study are known to differ between the immigrants’ cultures of origin and destination. Specifically, mother and infant responsiveness in interactions involving three maternal behaviors (speech to infants, social play, encouragement of infant attention to the mother herself) and two infant behaviors (look at mother, nondistress vocalization) were studied. These social behaviors were chosen because vocalizations, eye contact, facial expression, and gestures are species-universal and characteristic of mother-infant interactions during early infancy (Legerstee, 2009), but at the same time their prevalence varies cross-culturally, including among the cultural groups in this investigation (Bornstein, 2022; Caudill & Frost, 1972; Cote & Bornstein, 2021b; LeVine, 1991). Specifically, cross-culturation examination of parenting behaviors among middle-class mothers has revealed that Japanese mothers encourage their infants to look at them more than European American mothers; South Korean mothers engage in more social play and child-directed speech than European American mothers; and Argentine mothers encourage their infants to look at them, vocalize child-directed speech, and engage in social play more than European American mothers (Cote & Bornstein, 2021b). Differential engagement in the three types of maternal behaviors studied among mothers in the cultures of origin and destination may indicate preferences for some types of social interactions over others, leading to differential maternal responsiveness to infants’ social behavior across cultural groups.

Methodologically, we adopted both comparative (group) and individual-differences approaches to analyzing behavioral acculturation (Berry et al., 1987). Namely, the comparative (group-level) approach to acculturation compared mother and infant responsiveness in social interactions among immigrant families from three cultural groups to nonmigrant mother-infant dyads in their respective cultures of origin and one common culture of destination. Specifically, mother-infant dyads who immigrated from Japan were compared to nonmigrant dyads in Japan and European American dyads in the United States (called the Japanese comparison), mother-infant dyads who immigrated from South Korea were compared to nonmigrant dyads in South Korea and European American dyads in the United States (called the South Korean comparison), and mother-infant dyads who immigrated from South America were compared to nonmigrant dyads in Argentina and European American dyads in the United States (called the South American comparison). European American dyads were selected as the culture of destination because they are the majority cultural group at present in the United States and they provide a familiar reference point since most research on parenting and children’s development has been conducted with European American samples (e.g., Parke, 2000). However, European Americanism represents neither the desirable nor probable acculturation goal for immigrants to the United States (e.g., Ferguson & Bornstein, 2012; Masgoret, & Ward, 2012). In this study, acculturation at the group level was assessed by comparing mean levels of responsiveness among acculturation groups (culture of origin, immigrant, culture of destination), examining patterns of mean levels of responsiveness within dyads (between mother and infant) across acculturation groups, and comparing differences in magnitudes of attunement of mother and infant responsiveness within dyads (differences between correlation coefficients) across acculturation groups. The individual approach to acculturation examined correlations between immigrant mothers’ psychological acculturation and their responsiveness and initiation in social interactions with their infants. Observing mother-infant dyads from three separate cultural groups permitted investigation of the specificity and generality of group and individual acculturation across cultures.

The investigation of responsiveness in social interactions among immigrant mother-infant dyads is a matter of pressing contemporary concern for developmental and cultural reasons. Specifically, maternal responsiveness in mother-infant social interactions promotes the development of children’s social competence (e.g., Bozicevic et al., 2021; Legerstee & Varghese, 2001; Mcquaid et al., 2009). Although less well studied, infant responsiveness influences children’s development among U.S. immigrant families. For example, Japanese immigrant infants who were more responsive to their mothers in social interactions had mothers who were more likely to engage in physical affection with them when they were toddlers, and in turn preschoolers whose mothers engaged in more physical affection with them during toddlerhood showed less hostility as preschoolers (Cote & Bornstein, 2018). Despite the large and growing numbers of immigrant families (Urban Institute, 2019), relatively little is still known about their family processes. In the balance of this Introduction, we review the literatures on the four sets of analyses that constitute the foci of this study – temporal contingency, mean differences in responsiveness among and within dyads, attunement of mother and infant responsiveness within dyads, and the influence of acculturation on individual responsiveness.

Responsiveness as Temporal Contingency

Behavior is considered responsive when it depends on an interactional partner’s previous activity (Bornstein et al., 2021). In this study we define responsiveness as temporal contingency; that is, a target behavior occurring within a specified time frame following an initiating behavior. Temporal contingency in social interactions is important to infant development because contingency focuses infant attention and thereby fosters infants to develop turn-taking and perhaps an understanding of the intentions of their communicative partner. These achievements in turn serve as a catalyst for infant language learning (Kuhl et al., 2003; Kuchirko et al., 2018). Temporal contingency in mother-infant interactions has been observed in a variety of behavioral domains and across cultures. For example, across a number of cultures, mothers respond to their infants’ person-directed behaviors (Bornstein et al., 2012; Bornstein et al., 2019; Cote et al., 2008), object-directed behaviors (Bornstein et al., 2012; Bornstein et al., 2019; Cote et al., 2008), and vocalizations (Bornstein et al., 2015; Cote & Bornstein, 2021a) with complementary behaviors that are temporally contingent. By contrast, research findings for infant responsiveness are more variable. Although infants have been found to respond to person-directed behaviors (as when mothers encourage their infant to engage in face-to-face interaction with her and the infant responds by looking at mother’s face in a temporally contingent way; e.g., Bornstein et al., 2012; Bornstein et al., 2019; Cote et al., 2008), infants’ responses are not uniformly temporally contingent for object-directed behaviors or vocalizations. Thus, mother and infant (temporally contingent) responsiveness seem to be more common for social interactions than for other kinds of interactions.

Responsiveness in social interactions also varies within cultures depending on the specific social behaviors studied. For example, in the domain of social interactions, Van Egeren et al. (2001), who also defined responsiveness as temporal contingency, found that mothers in the United States do not reliably respond to their 4-month-old infants’ looking at mother with social play but instead do so by vocalizing, and that mothers do not reliably respond to infant vocalization with social play. Van Egeren et al. (2001) also found that infants do not reliably respond to their mothers’ social play by looking at them or vocalizing, but respond to their mothers’ speech by looking at them. Suwalsky et al. (2012) examined the same six social interactions and defined responsiveness in the same way as Van Egeren et al. (2001), and like Van Egeren et al. found that only interactions in which the mother spoke and the infant looked at mother (Mother Speak/Infant Attend interactions) were temporally contingent, regardless of which partner initiated the interaction.

The present study investigated all four types of social interactions (Mother Speak/Infant Attend, Mother Play/Infant Attend, Mother Play/Infant Vocalize, and Mother Encourage/Infant Vocalize). Given the potential importance of Mother Speak/Infant Attend interactions for children’s language development (Masek et al., 2021a), we hypothesized that Mother Speak/Infant Attend interactions would be temporally contingent regardless of which partner was responding, and we hypothesized that this contingency would hold across cultural groups. Previous researchers found that Mother Play/Infant Attend and Mother Play/Infant Vocalize interactions were not temporally contingent (Van Egeren et al., 2001; Suwalsky et al., 2012). However, in both of those studies nonmigrant mother-infant dyads in the United States were sampled; Van Egeren et al.’s (2001) sample was reported as 97% Caucasian with an average age of 21.8 years and 12 years of education and could be considered working class socioeconomically, and Suwalsky et al.’s (2012) sample was reported as European American with an average age of 35 years and a college degree, and could be considered upper middle class socioeconomically. Given the differential emphasis on autonomy-promotion among European Americans compared to the emphasis on harmonious social interactions among Japanese, Korean, and South American cultures (Bornstein et al., 1999; de Castro Ribas, 2010; Holloway, 2010; Kim, 2006), and because Mother Play/Infant Attend, Mother Play/Infant Vocalize, and Mother Encourage/Infant Vocalize interactions have been less well studied, we made no specific hypotheses about temporal contingency for these types of social interactions in the culturally diverse middle-class immigrant samples studied here.

Mean Differences in Responsiveness Among and Within Dyads

Cultural differences in expectations for many human behaviors are common, including responsiveness in mother-infant social interactions. Immigrant mothers who are in the process of acculturating may be particularly sensitive to these differences and may or may not adapt their behavioral expectations to those of the culture of destination. Deeply personal interactions, such as those that occur in private spaces and within the family, are especially likely to be retained despite acculturation (Arends-Toth & van de Vijver, 2003). For example, middle-class European American mothers are more responsive in person-directed interactions than middle-class South American immigrant mothers (who were primarily from Argentina, Colombia, and Peru; Cote et al., 2008), and, although middle-class Japanese immigrant mothers’ responsiveness in person-directed interactions do not differ from middle-class mothers in Japan and the United States, Japanese immigrant mothers are more responsive than mothers in Japan and less responsive than European American mothers (Bornstein et al, 2012). However, middle-class South Korean immigrant mothers are more responsive to person-directed social interactions than either middle-class South Korean or European American mothers (Bornstein et al., 2019). Previous research has not revealed differences in infant responsiveness in person-directed interactions among immigrant infants and infants in cultures of origin and destination living in middle-class families (Bornstein et al., 2019; Bornstein et al., 2012; Cote et al., 2008). In consequence, we hypothesized that European American mothers would be more responsive than immigrant mothers and mothers in the cultures of origin (Japan, South Korea, Argentina) in social interactions, and we did not expect cultural differences in infant responsiveness. This study extends previous research on person-directed social interactions among immigrant families by examining cross-cultural differences in maternal and infant responsiveness to other types of social interactions that have not been studied in immigrant families before (i.e., Speak/Attend, Play/Attend, Play/Vocalize, Encourage/Vocalize).

Although previous research examining responsiveness in parent-infant interactions with primarily European American samples and conducted by researchers with European origins in the United States, Canada, or Europe has reported that responsiveness in dyadic interactions is driven primarily by parents compared to child-driven or bidirectional (Kochanska & Aksan, 2004), it is unknown whether the same is true in other cultural contexts. The cultures studied here differ in valuing autonomy-promotion and harmony-maintenance in interpersonal interactions, which can manifest in mean differences in dyads in responsiveness and lead-taking in social interactions. Specifically, in cultures that value autonomy-promotion such as the United States, mothers may begin to foster culturally valued traits of independent action and assertiveness from a young age (Tamis-LeMonda & McFadden, 2010) by responding to infants’ lead in dyadic interactions. In contrast, in cultures that value harmony-maintenance, such as Japan and South Korea (Mesman et al., 2018 and Kim, 2006, respectively), and cultures that are more authoritarian and where parents are directive, such as Argentina (Halgunseth, 2019), adults may lead in dyadic interactions and emphasize teaching children early how to be responsive partners. Previous research has supported this expectation in part. For example, middle-class European American, Japanese immigrant, South Korean immigrant, and South American immigrant mothers are more responsive in person-directed social interactions than their infants (Bornstein et al., 2012, Bornstein et al., 2019; Cote et al., 2008), suggesting that these forms of behavioral interaction acculturate quickly. However, no differences in responsiveness between mothers and infants emerged for dyads in Japan or South Korea, and unexpectedly mothers in Argentina were more responsive than their infants in person-directed interactions. On this basis, we hypothesized that within European American and immigrant dyads mothers would be more responsive than infants in social interactions, and that for dyads in Japan, South Korea, and Argentina, infants would be more responsive than their mothers. This study extends research on nonmigrant Caucasian and European American families (by Van Egeren et al., 2001, and Suwalsky et al., 2012, respectively) by comparing maternal to infant responsiveness within dyads for social interactions that have not been studied previously among immigrant families.

Attunement of Mother and Infant Responsiveness Within Dyads

Mother-infant attunement can be operationally defined as the correlation between mother and infant levels of responsiveness, such that infants who respond more frequently to mothers have mothers who respond more frequently to them (Bornstein et al., 2021). Mothers and infants in different cultures respond to each other in an attuned fashion across a number of behavioral domains (Bornstein, 2012, 2013; Kochanska & Aksan, 2004). Moreover, behavioral attunements appear to be a common and perhaps a necessary feature of mother-infant social interactions in that, for example, they promote children’s social well-being (Bornstein, 2013; Field et al., 2007; Stern, 1985). High degrees of attunement in mother and infant responsiveness characterize middle-class Japanese immigrant, South Korean immigrant, South American immigrant (primarily from Argentina, Colombia, and Peru), and European American dyads for one type of social interaction, person-directed behaviors (in which mother encourages her infant to look at her and her infant looks at the mother; Cote & Bornstein, 2018). With respect to the specific behavioral interactions studied here, Van Egeren et al. (2001) found attunement for interactions in which the mother vocalizes and the infant looks at the mother (Mother Speak/Infant Attend), but not for interactions in which the mother engages in social play and the infant vocalizes (Mother Play/Infant Vocalize), nor for interactions in which the mother engages in social play and the infant looks at mother (Mother Play/Infant Attend) among Caucasian working-class families; interactions in which the mother encourages her infant to look at her and the infant vocalizes (Mother Encourage/Infant Vocalize) were not examined.

In the present study, attunement was operationalized as the correlation between mother and infant responsiveness (Bornstein et al., 2021; Karger, 1979), and attunement was explored separately in each cultural group and all four types of social interactions. Moreover, because previous research has indicated that mother-infant interactions may be disrupted in immigrant families (e.g., Gratier, 2003), acculturation differences in attunement were explored directly (using z-tests). Because previous research had revealed that interactions in which mother spoke and infant looked at mother were attuned, and because they may be important for infants’ language development (Masek et al., 2021a), we hypothesized that Mother Speak/Infant Attend interactions would be attuned in the cultural groups studied here. Attunement has not been found for Mother Play/Infant Vocalize and Mother Play/Infant Attend interactions in European American mother-infant dyads; however, because attunement of mother and infant responsiveness in Mother Play/Infant Vocalize, Mother Play/Infant Attend, and Mother Encourage/Infant Vocalize interactions have not been studied in Japanese, South Korean, or South American samples, we had no specific predictions.

The Influence of Acculturation on Individual Responsiveness

In addition to group-level differences in acculturation, individuals experience acculturation uniquely (Bornstein, 2017; Bornstein & Cote, 2019). Previous research on parenting and children’s development among middle-class immigrant families has reported associations between parents’ beliefs and behaviors and their acculturation at the individual level (e.g., Bornstein & Cote, 2006; Farver & Lee-Shin, 2000; Jain & Belsky, 1997). To measure acculturation at the individual level, we used a psychological measure of acculturation that reflects involvement in the new culture (e.g., identity, values; Güngör et al., 2013; Phinney, 2006). Because European American mothers promote culturally valued autonomy in their infants and young children by being responsive to them (Tamis-LeMonda & McFadden, 2010), it follows that mothers who were more acculturated to U.S. American culture would display higher levels of responsiveness to their infants in social interactions. Previous research has supported this view. For example, college educated South Korean immigrant mothers who are more acculturated to U.S. American culture encourage autonomy and show more warmth toward their children than less acculturated mothers (Shin et al., 2010). Middle-class South Korean immigrant mothers who are more acculturated to South Korean culture are less responsive to their infants in person-directed social interactions, whilst those who are more acculturated to European American culture are more responsive in object-directed interactions (Bornstein et al., 2019). Mexican and Puerto Rican mothers living in low-income settings who are more acculturated to U.S. American culture engage in more verbal inquiry and structuring during parent-child interactions than those who are less acculturated (Howes & Obregon, 2009; Teichman & Contreras-Grau, 2006). In consequence, we hypothesized that immigrant mothers who are more acculturated to U.S. American culture would display higher levels of responsiveness in their social interactions with their infants and lower levels of initiation, and mothers who are more acculturated to their culture of origin would display lower levels of responsiveness and higher levels of initiation.

Research Goals

This study sought to investigate: 1) temporal contingency in mother-infant social interactions as a measure of responsiveness, 2) acculturation of responsiveness at the group level by comparing mean differences in responsiveness among and within dyads, 3) attunement of mother and infant responsiveness within dyads (including whether attunement differs across acculturation groups), and 4) the influence of acculturation on individual responsiveness. The specific hypotheses associated with these research goals were, first, that Mother Speak/Infant Attend interactions would be temporally contingent regardless of which partner was responding in every cultural group. Our second hypothesis exploring the acculturation of responsiveness at the group level had multiple parts: European American mothers would be more responsive than immigrant mothers and mothers in the cultures of origin, and there would be no cultural differences in responsiveness among infants. Within dyads, European American and immigrant mothers were expected to be more responsive than their infants, and in the cultures of origin, infants were expected to be more responsive than their mothers. Third, we hypothesized that Mother Speak/Infant Attend interactions would be attuned in every cultural group. Fourth, we hypothesized that immigrant mothers who are more acculturated to U.S. American culture would display higher levels of responsiveness in their social interactions with their infants and lower levels of initiation, and mothers who are more acculturated to their culture of origin would display lower levels of responsiveness and higher levels of initiation.

Method

Participants

In total 408 mothers and their infants participated: Japanese (n = 46), South Korean (n = 56), Argentine (n = 48), Japanese immigrant (n = 32), South Korean immigrant (n = 59), South American immigrant (n = 32), and European American (n = 135). All mothers lived in major metropolitan regions of their respective countries (i.e., Japanese dyads lived in Tokyo, South Korean dyads lived in Seoul, and Argentine dyads lived in Buenos Aires); immigrant and European American mothers lived in the Washington, D.C metropolitan area in the United States. All immigrant mothers were born in their respective country of origin and had immigrated to the United States during their lifetimes, and all their infants were born in the United States. Japanese-origin mothers were ethnically Japanese, South Korean-origin mothers were ethnically Korean, and European American mothers in the United States were 4^th or 5^th generation descendants of European immigrants. Argentine and South American immigrant mothers were native Spanish speakers of European descent. South American immigrant mothers self-identified as a shared ethnicity by responding to an advertisement for “familias de origen Sud Americano”/“South American families” (Marín & Marín, 1991). In areas of the United States such as the location for this research, where there is not a large concentration of a single Latin American group, self-identification by regional affiliation, rather than specific country of origin, is common (Winn, 1992), and even within South America it is common for adults from Spanish-speaking countries to share supranational identities (Salazar & Villegas, 1999), lending cultural validity to this grouping. South American immigrant mothers were primarily from Argentina (n = 6), Colombia (n = 6), and Peru (n = 8), with smaller numbers from Bolivia (n = 3), Chile (n = 3), Ecuador (n = 3), Venezuela (n = 2), and Paraguay (n = 1); empirically there were no differences in this group on any dependent variables (when Argentine, Colombian, Peruvian, and other South American countries were compared).

Immigrant mothers in the study had been in the United States an average of 12.85 years (SD = 10.27). South Korean immigrant mothers were significantly younger when they immigrated and had lived in the United States significantly longer than either Japanese immigrant or South American immigrant mothers: age at immigration, F(2, 109) = 24.89, p < .001, n²_p = .31, and number of years lived in the United States, F(2, 110) = 13.71, p < .001, n²_p = .20, respectively. Descriptive statistics for the immigrant mothers were: Japanese immigrant age at immigration, M = 27.92, SD = 3.66, and years in U.S., M = 5.48, SD = 3.13; South Korean immigrant age at immigration, M = 14.23, SD = 10.38, and years in U.S., M = 16.79, SD = 11.49; South American immigrant age at immigration, M = 23.31, SD = 8.13, and years in U.S., M = 11.18, SD = 7.48.

Three separate European American samples were drawn from a larger study of infant development and parenting in the United States so that the European American samples would be comparable to the culture of origin and immigrant samples within each acculturation comparison (Japanese, South Korean, or South American) with respect to sample size, infant gender distribution, maternal age and education level, percent of mothers who were employed, number of hours per week mothers worked outside of the home, and percent of nuclear families (Table 1). The European American comparison groups were ns = 40 each for the Japanese and South American acculturation comparisons and n = 55 for the South Korean acculturation comparison. There were no differences among the European American groups on the dependent variables (DVs).

Table 1.

Sociodemographic Characteristics of Participants

Acculturation Group:
Variable	Acculturation Comparison	Origin	Immigrant	Destination	F or χ2
Infant age (days)	Japanese	161.89a (9.16)	171.84b (13.60)	165.15a (8.01)	F(2, 115) = 9.03, p <.001, η2p = .14
	South Korean	162.16a (4.89)	172.63b (15.91)	162.33 (5.81)	F(2, 167) = 19.46, p < .001, η2p = .19
	South American	167.08 (6.63)	168.91a (13.19)	163.25b (5.38)	F(2, 117) = 4.22, p = .02, η2p = .07
Infant gender (girls:boys)	Japanese	23:23	15:17	18:22	χ2(2, N = 118) = 0.22, p = .90
	South Korean	27:29	30:29	28:27	χ2(2, N = 170) = 0.11, p = .95
	South American	18:30	14:18	17:23	χ2(2, N = 120) = 0.38, p = .83
Infant awake	Japanese	2777.91a (346.93)	2818.01 (271.89)	2948.68b (131.00)	F(2, 115) = 4.53, p = .013, η2p = .07
	South Korean	2855.36a (203.87)	2840.58a (305.12)	2965.08b (80.92)	F(2, 165) = 5.35, p = .006, η2p = .06
	South American	2871.80 (185.86)	2910.95 (160.67)	2817.36 (285.77)	F(2, 117) = 1.67, p = .192, η2p = .03
Mother’s age	Japanese	29.03a (2.90)	31.95b (3.22)	30.06 (5.34)	F(2, 109) = 4.75, p = .01, η2p = .08
	South Korean	29.06a (2.17)	31.74b (3.65)	31.07b (3.33)	F(2, 166) = 11.22, p < .001, η2p = .12
	South American	27.02a (3.82)	31.74b (4.87)	29.43b (4.95)	F(2, 117) = 10.73, p < .001, η2p = .16
Mother’s education level (Hollingshead index) a	Japanese	5.58 (0.88)	5.75 (0.76)	6.00 (1.04)	F(2, 112) = 2.22, p = .11, η2p = .04
	South Korean	5.69 (0.86)	6.12 (1.20)	6.00 (0.84)	F(2, 166)= 2.82, p = .06, η2p = .03
	South American	5.60 (1.63)	5.94 (0.88)	5.73 (0.99)	F(2, 117) = 0.67, p = .52, η2p = .01
Employed mothers: Not employed	Japanese	8:9	9:23	21:19	χ2(2, N =89) = 4.48, p = .11
	South Korean	25:31	38:21	29:26	χ2(2, N = 170) = 4.58, p = .10
	South American	17:31	17:15	20:20	χ2(2, N = 120) = 3.04, p = .22
Hours per week mothers works outside the home (all mothers)	Japanese	14.30 (20.57)	7.06 (14.51)	16.60 (17.79)	F(2, 84) = 2.82, p = .07, η2p = .06
	South Korean	17.54 (22.47)	20.14 (18.84)	16.65 (19.13)	F(2, 167) = 0.46, p = .63, η2p = .01
	South American	6.98a (12.01)	14.66 (16.37)	16.93b (17.99)	F(2, 115) = 4.89, p = .01, η2p = .08
% Nuclear families	Japanese	94.1%	93.8%	100%	χ2(2, N = 63) = 0.90, p = .64
	South Korean	48.2%	81.4%	85.8%	χ2(2, N = 170) = 23.09, p < .001
	South American	79.2%	75.0%	75.0%	χ2(2, N = 120) = 0.28, p = .87
Time (in s) mother was in view of infant during observation period	Japanese	2728.26 (549.14)	2864.62 (283.65)	2781.77 (366.25)	F(2, 115) = 0.95, p = .31, η2p = .02
	South Korean	2878.14 (264.84)	2811.76 (266.31)	2879.38 (196.81)	F(2, 167) = 1.44, p = .24, η2p = .02
	South American	2760.89 (251.62)	2852.28 (264.94)	2817.20 (236.95)	F(2, 117) = 1.36, p = .26, η2p = .02

Note. M (SD) unless specified. ANOVAs with one between-subjects factor (Acculturation group) were performed separately for each sociodemographic characteristic, followed by t-tests with Bonferroni’s correction; different alphabetic subscripts in the same row indicate significant group differences.

^aMaternal education was scored using the 7-point Hollingshead (2011) Four Factor Index of Social Status, where 5 indicates some college or additional training beyond high school and 6 indicates a baccalaureate degree. Because differences exist between countries in the frequency, quality, and content of schooling, bicultural researchers adjusted mothers’ years of schooling so that the scales were equivalent to the Hollingshead index.

Inclusion criteria were as follows: Mothers’ ethnic identity as described above; infants were firstborn full term, healthy (≥ 2300g at birth) singletons; and dyads needed to have at least 2400 s (40 min) of videorecorded interaction. The gender distributions for infants were approximately equally divided. Mothers were living with their infant’s father, and 99.5% were married to their infant’s father. Table 1 presents participant sociodemographic information.

Procedure

All participants were treated in accordance with the ethical standards of the American Psychological Association and the Declaration of Helsinki. This research study was approved by the National Institutes of Health’s IRB. Written informed consent was obtained from mothers for their and their infants’ participation in the study prior to data collection. All consent forms and sociodemographic questionnaires were written in English and translated into Japanese, Korean, and Spanish by bicultural bilingual developmental scientists who were natives of Japan, South Korea, and Argentina, respectively, following standard forward- and back-translation procedures recommended by Brislin (1986) and Peña (2007). Immigrant mothers completed consent forms and questionnaires in the language of their choice. Our goal was to recruit firstborn singleton infants of a particular age, so mothers in each sample and culture were recruited in multiple ways to maximize the likelihood of finding participants who met participation criteria; mothers were recruited from government vital records, hospital birth notifications, medical group patient lists, newspaper advertisements, newspaper birth announcements, and mass mailings. The same recruitment techniques were used in each country.

Mothers and their 5½-month-old infants were visited in their homes by a female researcher native to the country. Mothers were instructed to ignore the researcher insofar as possible and behave as they typically would with their infant. After a recommended period of acclimation (McCune-Nicolich & Fenson, 1984; Stevenson et al., 1986), dyads were videorecorded consecutively for 50 min (3000 s), which is an optimal window for capturing mother-child interaction (Holden & Miller, 1999). Infants were awake for nearly the entire observation (see Table 1). Afterwards mothers completed an evaluation of the visit, an acculturation questionnaire, and a sociodemographic questionnaire.

Behavioral Coding

Coders made 7 coding passes through the data – 5 to observe mother and infant behaviors that would be used to generate dependent variables, and 2 that would be used as covariates. These 7 coding modes and the specific behaviors that made up each mode are described in detail in Table 2 along with information about coding reliability. Within each coding mode, behaviors were mutually exclusive and exhaustive, and onset times for behaviors were coded to the nearest 0.1 sec for the duration of the videorecording (50 min). Behaviors coded in this way can be used to generate duration and frequency information as well as timed event-sequential data (Bakeman & Quera, 2011). Timed event sequences allow for examination of temporal relations between mother and infant behaviors in dyadic interaction (Bakeman & Quera, 2011; Cote & Bornstein, 2021a). Strengths of this approach to measuring interaction include: Elimination of coder bias because mother and infant behaviors are coded separately by coders who are unaware of study hypotheses (Leclère et al., 2014); it is easy to distinguish which partner initiates the interaction and which responds; and it is flexible because it can be used to identify unexpected interaction patterns.

Table 2.

Behavioral Coding

Coding mode	Mutually exclusive behaviors that comprised the coding mode	Behaviors used in sequential analysis
Maternal affective behavior	Onset times of the following mutually exclusive maternal behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) social play: The mother directs verbal or physical behavior to the infant, the purpose of which appears to be to amuse the infant (i.e., to elicit smiles, positive vocalizations, laughter, or motoric excitement in the context of a primarily social dyadic interaction). The types of exchanges coded as social play are: (a) physical contact with a fun-like quality (e.g., tickling); (b) introducing the element of surprise, suspense, or quick release of stimuli (e.g., peek-a-boo); (c) singing to the infant; and (d) playing a game that involves physical manipulation of the infant’s body (e.g., pattycake). b) expressing affection: The mother expresses affection or positive evaluation to the infant either physically (e.g., kissing, patting, stroking, or caressing) or verbally (using explicit phrases denoting praise or endearment). c) negative affect: The mother directs negatively-toned verbal or physical behavior to the infant. (e.g., using an irritated or angry tone of voice or handling the baby harshly) d) none of the above. Behavior is coded as “none of the above” when the mother has not interacted for 3 seconds and when she is no longer oriented to the infant and poised to continue the exchange; pauses of any duration when the mother clearly remains poised to continue are coded as part of the play sequence.	Social play (code a), referred to in this paper as Play
Maternal encouragement of infant attention	Maternal encouragement could be physical or verbal; these were not distinguished because they tend to co-occur. Examples of encouragement of infant attention include mother saying, “Look at me,” pointing to herself, or directing the infant to look at a specific object by pointing to it or saying, “Look at [specific object].” Encouragement had to be direct; simply speaking to the infant or calling the infant’s name or naming an object was not coded as encouragement. Pauses of 2 seconds or longer were coded as terminations of an ongoing behavior. Onset times of the following mutually exclusive maternal behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) mother encourages attention to herself: The mother attempts to draw the infant into face-to-face social interaction with herself. Physical attempts include intentionally moving her face toward the infant or moving the infant toward her face. Verbal attempts include making very specific comments about herself that are clearly designed to capture the infant’s interest. Pauses of 2 seconds or longer are coded as terminations of an ongoing behavior. b) mother encourages attention to another person (e.g., the researcher) c) mother encourages attention to object: The mother physically moves the infant or an object so that the infant can see or touch it, or the mother verbally refers to an object or an object-related event or activity that is no more than 12 feet from the infant. d) none of the above.	Mother encourages attention to herself (code a), called Encourage
Infant visual attention	For each of these codes listed below, focused fixation must be evident. An active behavioral component often accompanies clear and focused fixation (e.g., brightening of the face, widening of the eyes, stilling, increased motor excitement, positive vocalizations, or reaching). A change in fixation is coded after the infant has looked away from target for 1 second. Onset times of the following mutually exclusive infant behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) face-to-face with caregiver: The faces of both the infant and caregiver are simultaneously on film and enface gazing is evident. This is not simply the opportunity for enface, it must actually occur. b) looks at mother: The infant looks at the mother’s face or head. The mother need not return the infant’s gaze. Focused fixation must be evident. An active behavior component often accompanies clear and focused fixation (e.g., brightening of the face, widening of the eyes, stilling, increased motor excitement, positive vocalizations, or reaching). A change in fixation is coded after the infant has looked away from target for 1 second. c) looks at familiar person: The infant looks at the face of a familiar person, child or adult, other than the primary caregiver. The person need not return the infant’s gaze. The person need not respond to the infant’s gaze. d) looks at stranger: The infant looks at the face of an unfamiliar person, child or adult. The person need not respond to the infant’s gaze. e) looks at object: The infant looks at any discrete object or body part other than a face that is within a radius of 15 feet. f) none of the above: The visual focus of the infant is not clear, the infant is off-camera, or the infant’s eyes are closed.	Two behaviors, face-to-face with caregiver (code a) and looks at mother (code b) were combined to form infant Attend
Maternal vocalization	The minimum duration for a vocalization was set to 0.30 s, and there had to be a gap at least 1.00 s between vocalizations for a second vocalization to be counted as a new vocalization. Onset times of the following mutually exclusive maternal behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) mother talks to infant using adult conversational tones (adult-like speech): Words and speech-like sounds directed by the mother to the infant that are characterized by normal intonation patterns. Included are syllable sounds, parts of words, single words, conversations, and singing. b) mother talks to infant using infant-directed speech (child-like speech): The special speech register used by mother when talking to her infant, including short sentences, greater repetition and questioning, and higher and more variable intonation than that of speech addressed to adults. Changes and pauses in vocalization lasting less than 1 second are not recorded. c) vocal imitation: Mother attempts to imitate the infant’s vocalization. Vocalizations may be of any type (e.g., babbling, distress, nondistress, vegetative sounds). The mother’s vocalization must be temporally contiguous (i.e., occur during or within 5 sec) of the infant’s vocalization in order to be considered imitation. d) talks to other people: Mother talks to another person such as the researcher or someone on the telephone (note that in this study, mothers were instructed to be alone with their infants during the observation) e) none of the above (e.g., mother not speaking).	Two behaviors, mother talks to infant using adult conversational tones (code a) and mother talks to infant using infant-directed speech (code b) were combined to form mother Speak
Infant vocalization	The minimum duration for a vocalization was set to 0.30 s, and there had to be a gap at least 1.00 s between vocalizations for a second vocalization to be counted as a new vocalization. Onset times of the following mutually exclusive infant behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) nondistress vocalization: Any positively or neutrally toned infant vocalization that is clearly audible. Included are babbling, cooing, laughing, vocal play, shrieking, and sighs or grunts not indicative of distress. b) distress vocalization: Vocalizations produced by the infant that indicate protest, complaint, anger, or upset, as indicated by vocal quality, facial expression, or other negative behaviors (e.g., intense squirming, back arching). c) vegetative sounds (e.g., yawning, burping, sneezing, coughing, raspberries, tongue clicks) d) silence/none of the above.	Nondistress vocalization (code a) is called Vocalize
Coding mode	Mutually exclusive behaviors that comprised the coding mode	Behavioral covariates
Infant State	Coding for infant state of arousal was adapted from Brazelton’s (1973) conception of infant states of arousal. Onset times of the following mutually exclusive infant behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) asleep (eyes closed, regular breathing, little or no spontaneous activity), b) drowsy (i.e., eyes closed or open but glassy, eyelids fluttering, fussing), c) awake (i.e., infant’s eyes are open and look bright and focused, may or may not be accompanied by motor or vocal activity), d) crying (i.e., infant exhibits full-blown distress including sobbing, kicking, facial expressions such as open mouth and tightly closed eyes for 5 sec or more).	Proportion of time infant was awake was computed by summing all durations for code c (awake) then dividing this sum by the total duration of the observation period (3000 sec = 50 min). This is called infant awake.
Mother’s activity	Onset times of the following mutually exclusive maternal behaviors were recorded (the onset of a behavior indicated the offset of the previous behavior): a) Feed: The mother attempts to give the infant liquid or solid foods by cup, bottle, breast, or spoon. b) Burp/Wipe face or hands: The mother attempts to burp the infant in connection with a feeding, or the mother wipes the infant’s face, hands, or clothing at any time. c) Bathe: The mother washes and dries the infant’s body and/or hair. d) Check/Change diaper: The mother checks to see if the infant needs a diaper change or changes the diaper. e) Dress: The mother removes or puts an article of clothing on the infant f) Groom: The mother engages in behavior designed to enhance the infant’s appearance (e.g., combs hair). g) Meet other health needs: The mother attends to other health needs of the infant (e.g., wipes or suctions the infant’s nose; gives medicine from a dropper or medicine spoon). h) Infant in caregiver’s view but none of the activities listed is occurring (e.g., mother is doing something without the infant such as washing her hands) i) Infant is out of caregiver’s view.	Proportion of time mother was in view of the infant was computed by subtracting the duration of time the mother’s activity was coded as i (infant out of caregiver’s view) from the duration of the observation period (3000 sec) and then dividing that number by the total duration of the observation period (3000 sec). This is called mother in view.

Note. Coders were bicultural natives of the participants’ culture of origin and unaware of study goals. Coders were trained to reliability on a standard set of digital records (Cohen’s, 1960, kappa: κ > .60). Interrater agreement was calculated using ~15% of each sample (range = 10–30%). Coder reliability was checked every 10 records to guard against coding drift. The Kappa Acc program (available at http://bakeman.gsucreate.org/kappaacc/) was used to calculate estimated observer accuracy (Bakeman, 2018). Cohen’s κs were above .60 and estimated observer accuracy was above 94% across codes and cultural groups, with three exceptions: infant awake for the Argentine (κ = .56), South American immigrant (κ = .29), and Japanese (κ = .32) samples; to achieve these kappas, observers were at least 98% accurate (Bakeman, 2018; Bakeman & Quera, 2011).

Behavioral Dependent Variables

Timed event sequences were used to measure interactions, temporally and sequentially, and they were generated via a multi-step process. From the coded videorecords, the following 5 behaviors (which were coded independently in 5 separate coding passes) were the focus of this study: mother social play, mother and infant attention, and mother and infant vocalization.

The next step in generating the dependent variables was to program the Generalized Sequential Querier program (GSEQ version 5.1.23, https://www.mangold-international.com/en/products/software/gseq) to identify the following 8 behavioral sequences in the data: Mother Speak/Infant Attend interactions encompassed both 1) Speak → Attend (the odds that an infant responds by looking at mother within 3 s of the onset of mother speaking to infant), and 2) Attend → Speak (the odds that a mother responds by talking to the infant within 3 s of the onset of the infant looking at her); Mother Play/Infant Attend interactions included both 3) Play → Attend (i.e., the odds that the infant responds by looking at mother within 3 s of the onset of maternal social play), and 4) Attend → Play (i.e., the odds that mother responds by engaging in social play within 3 s of the onset of the infant looking at her); Mother Play/Infant Vocalize interactions included both 5) Play → Vocalize (the odds that an infant responds by vocalizing nondistress within 3 s of the onset of maternal social play), and 6) Vocalize → Play (the odds that a mother responds by engaging in social play within 3 s of the onset of the infant vocalizing nondistress); and Mother Encourage/Infant Vocalize interactions included both 7) Encourage → Vocalize (the odds that an infant responds by vocalizing nondistress within 3 s of the onset of mother encouraging the infant’s attention to her), and 8) Vocalize → Encourage (the odds that a mother responds by encouraging the infant to look at her within 3 s of the onset of the infant vocalizing nondistress). Separately for each dyad, the GSEQ program tallied time units into 2 by 2 tables (covering all 4 possible combinations of signal and response), and from those tables odds ratios (OR) were computed for the 8 behavioral sequences. OR statistics were used as the dependent variables in this study. A strength of this approach is that ORs compare the number of successful interaction attempts to unsuccessful attempts (as suggested in Leclère et al., 2014) rather than the raw number of mother-infant interactions. The OR is a descriptive measure of effect size (Bakeman et al., 2005) and is easy to interpret (e.g., an OR of 3.00 for Speak → Attend means that the odds of an infant looking at the mother within 3 s of the onset of mother speaking to infant are 3.00 times greater when the mother is speaking to the infant than when the mother is not speaking to the infant).

Acculturation Scales

The acculturation scales completed by Japanese immigrant and South American immigrant mothers were Cote & Bornstein’s (1996a, 1996b) Japanese American Acculturation Scale (JAAS) and South American Acculturation Scale (SAAS), respectively. These 21-item scales measured participants’ identification with U.S. American culture and Japanese or South American culture (respectively). Participants rated items using 5-point scales, with higher mean ratings indicating greater identification with U.S. American culture and lower mean ratings indicating greater identification with Japanese or South American culture. Each acculturation scale has demonstrated high levels of test-retest reliability (rs ≥ .82), internal consistency (Cronbach’s αs ≥ .91), and construct validity (i.e., correlated positively with generation level and number of years mother lived in the U.S., and correlated negatively with mothers’ age at immigration, rs ≥ │.33 – .76│).

Data collection for the South Korean immigrant families occurred after data collection for the other immigrant families, over which time advances in measurement of acculturation had occurred; therefore, a multidimensional rather than unidimensional measure of acculturation was used. Specifically, the Korean American Acculturation Scale 2 (KAAS2; Cote & Bornstein, 2004) contains one 17-item subscale that measured Korean identification and a separate 13-item subscale that measured U.S. American identification using 5-point rating scales (with higher scores indicating greater identification). The Korean and U.S. American subscales each demonstrated good test-retest reliability (rs ≥ .88), internal consistency (Cronbach’s αs ≥ .88), and construct validity (i.e., Korean identification correlated positively with age at immigration and negatively with number of years in the United States, whereas U.S. American identification correlated negatively with age at immigration and positively with number of years in the United States, rs ≥ │.59│).

Immigrant mothers were bicultural, as indicated by the fact that Japanese immigrant and South American immigrant mothers scored midway on their acculturation scales (Ms = 2.02 and 2.20, and SDs = 0.38 and 0.28, respectively). South Korean immigrant mothers scored midway on both the Korean identification (M = 3.48, SD = 0.71) and U.S. American identification subscales (M = 3.61, SD = 0.81), and there were no differences between their scores on these two subscales, t(51) = −0.66, p = .51, d = 0.09.

Preliminary Analyses

Missing Data

Less than 1% of the data was missing. Of the 408 participants in the original sample, the sound on the digital records was poor for two South Korean dyads; thus, their vocalizations could not be coded. Therefore, 6 dependent variables could not be computed for 2 of the 408 participants (0.37%). Bakeman et al. (2005) advised that the signal (initiating) behavior had to occur at least 5 times for there to be sufficient data to draw reliable conclusions about sequences of behaviors. Using this criterion, for 6 DVs 93%–99% of participants had sufficient data; however, for maternal social play 82% of participants had sufficient data because this behavior occurred less frequently than the other behaviors studied, which means fewer participants had data for Play → Attend and Play → Vocalize interactions than for the other DVs. Appendix A in Supplementary Materials details the ratio of participants for whom there was insufficient:sufficient data separately for each DV and cultural group. Because such a small amount of data was missing, no attempt was made to estimate missing data. Sample sizes for some analyses reported in Tables 1 and 3–7 are smaller than the ns listed due to missing data.

Table 3.

Temporal Contingency and Mean Differences in Responsiveness Among and Within Dyads for the Japanese Acculturation Comparison

Acculturation Group
Dependent Variables	Japan (Origin) (n = 46)	Japanese Immigrant (n = 32)	European American (Destination) (n = 40)	AN(C)OVA
Mother Speak/Infant Attend b
Speak → Attend	3.12* (2.81)	2.48* (2.53)	3.27* (4.21)	Respondent within-subjects effect: F(1, 109) = 3.85, p = .05, η2p = .03 Acculturation Group main effect: F(2, 109) = 0.72, p = .49, η2p = .01 Respondent × Acculturation Group interaction: F(2, 109) = 2.64, p = .08, η2p = .05
Attend → Speak	2.00* (1.28)	2.06* (1.31)	2.49* (1.92)
Mother Play/Infant Attend
Play → Attend	3.69* (3.99)	3.15 (4.68)	3.36* (2.72)	Respondent within-subjects effect: F(1, 84) = 0.88, p = .35, η2p = .01 Acculturation Group main effect: F(2, 84) = 6.82, p = .002, η2p = .14 Respondent × Acculturation Group interaction: F(2, 84) = 3.04, p = .05, η2p = .07
Attend → Play	3.12 (4.07)	1.46 (2.17)	4.60* (3.79)
Mother Play/Infant Vocalize ab
Play → Vocalize	1.02 (0.93)	1.57 (1.50)	1.26 (1.40)	Respondent within-subjects effect: F(1, 83) = 5.04, p = .03, η2p = .06 Acculturation Group main effect: F(2, 83) = 0.04, p = .96, η2p = .00 Respondent × Acculturation Group interaction: F(2, 83) = 0.49, p = .62, η2p = .01
Vocalize → Play	1.43 (1.55)	1.27 (1.76)	1.24 (1.56)
Mother Encourage/Infant Vocalize
Encourage → Vocalize	1.41 (1.28)	1.71 (2.07)	1.46 (1.41)	Respondent within-subjects effect: F(1, 104) = 0.04, p = .84, η2p = .00 Acculturation Group main effect: F(2, 104) = 0.48, p = .62, η2p = .01 Respondent × Acculturation Group interaction: F(2,104) = 0.93, p = .40, η2p = .02
Vocalize → Encourage	1.54 (1.10)	1.26 (1.07)	2.01 (2.72)

Note. Data (M, SD) are untransformed ORs (ORs) not controlled for covariates. Asterisks (*) indicate temporal contingency in mother-infant interactions, specifically, that the OR was significantly greater than 1 (as indicated by a one-sample t-test where p ≤ .05). AN(C)OVAs examined acculturation of responsiveness at the group level. Superscripts indicate the specific covariates that were used in the mixed-design AN(C)OVAs:

^ainfant age

^bmother in view.

Table 7.

The Influence of Acculturation on Individual Responsiveness

	Immigrant Families
Dependent Variable	Japanese immigrant (n = 32)	South Korean immigrant (n = 59)		South American immigrant (n = 32)
	Acculturation	European American Acculturation	South Korean Acculturation	Acculturation
Speak → Attend	−.02	.06 be	.05 be	.00 c
Attend → Speak	.04	−.16 bce	.17 bce	.12 c
Play → Attend	−.01 a	.06 b	.00 b	.03 c
Attend → Play	−.13	−.20 bd	.22 bd	−.11 cd
Play → Vocalize	−.22	−.28 ab	.37 ab	.50** c
Vocalize → Play	−.06 c	−.32* b	.33* b	−.03 c
Encourage → Vocalize	.04	−.13 b	.17 b	−.07 c
Vocalize → Encourage	.08 e	.12 b	.05 b	−.02 c

Note. Statistics are two-tailed Pearson correlations. Higher acculturation scores for Japanese immigrant and South American immigrant mothers mean that they are more acculturated to European American culture, and lower scores mean that they are more acculturated to their cultures of origin. For the South Korean immigrant sample, European American and South Korean acculturation were measured using two different scales in which a higher score means higher levels of acculturation. Superscripts in each cell indicate the specific covariates that were used in the analysis in that particular cell:

^ainfant age

^bmother age

^cmother education

^dmother in view

^einfant awake.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Data Screening

Data were screened to assess assumptions associated with parametric statistical testing (Tabachnick & Fidell, 2018). Tables 3–5 report descriptive statistics (Ms, SDs) for untransformed ORs, uncorrected for covariates and excluding outliers; these are presented for ease of interpretation and to allow comparisons across research studies. For some samples and DVs, data were not normally distributed or contained statistical outliers; thus, transformations were applied (see Appendix B for details). In Tables 3–7 all analyses were performed with transformed data, excluding outliers (if there were any after transformation) and controlling covariates; r and F statistics reported in Tables 3–7 derive from those analyses. The specific covariates used for each analysis can be found in the Tables.

Table 5.

Temporal Contingency and Mean Differences in Responsiveness Among and Within Dyads for the South American Acculturation Comparison

Acculturation Group:
Dependent Variables	Argentine (Origin) (n = 48)	South American Immigrant (n = 32)	European American (Destination) (n = 40)	AN(C)OVA
Mother Speak/Infant Attend ac
Speak → Attend	2.03* (1.57)	1.91* (1.07)	2.32* (1.34)	Respondent within-subjects effect: F(1, 114) = 4.35, p = .04, η2p = .04 Acculturation Group main effect: F(2, 114) = 0.76, p = .47, η2p = .01 Respondent × Acculturation Group interaction: F(2, 114) = 0.11, p = .90, η2p = .00
Attend → Speak	1.66* (0.60)	1.83* (0.89)	2.00* (1.10)
Mother Play/Infant Attend c
Play → Attend	2.72* (2.21)	2.75* (2.26)	3.17* (2.96)	Respondent within-subjects effect: F(1, 99) = 0.02, p = .88, η2p = .00 Acculturation Group main effect: F(2, 99) = 0.71, p = .49, η2p = .01 Respondent × Acculturation Group interaction: F(2, 99) = 0.18, p = .83, η2p = .00
Attend → Play	3.32* (2.10)	3.80* (3.01)	4.57* (4.01)
Mother Play/Infant Vocalize b
Play → Vocalize	1.11 (0.95)	0.64 (0.88)	1.50 (1.09)	Respondent within-subjects effect: F(1, 99) = 0.002, p = .96, η2p = .00 Acculturation Group main effect: F(2, 99) = 5.94, p = .004, η2p = .11 Respondent × Acculturation Group interaction: F(2, 99) = 1.04, p = .36, η2p = .02
Vocalize → Play	1.10 (0.86)	0.89 (1.08)	1.47 (1.45)
Mother Encourage/Infant Vocalize
Encourage → Vocalize	1.27* (0.71)	1.25 (1.57)	1.97* (1.42)	Respondent within-subjects effect: F(1, 109) = 0.07, p = .79, η2p = .00 Acculturation Group main effect: F(2, 109) = 6.26, p = .003, η2p = .10 Respondent × Acculturation Group interaction: F(2, 109) = 0.17, p = .85, η2p = .00
Vocalize → Encourage	1.33* (0.88)	1.19 (1.40)	1.74* (1.18)

Note. Data (M, SD) are untransformed ORs (ORs) not controlled for covariates. Asterisks (*) indicate temporal contingency in mother-infant interactions, specifically, that the OR was significantly greater than 1 (as indicated by a one-sample t-test where p ≤ .05). AN(C)OVAs examined acculturation of responsiveness at the group level.

Superscripts indicate the specific covariates that were used in the mixed-design AN(C)OVAs:

^ainfant age

^bmother age

^cmother in view.

Sociodemographic and Behavioral Covariates

Infant age and gender and maternal age and education were screened as potential sociodemographic covariates if they correlated significantly (p < .05) with the dependent variables and for ANCOVAS if they had a significant independent effect. Infant gender had neither an independent nor an interaction effect with Acculturation Group; therefore, it was not used in any analyses.

Two variables were generated from the observational data as candidate behavioral covariates: duration of the time the infant was awake and duration of the time mother was in view of the infant during the observation period (see Table 2 for details). Covariates used in each analysis (if any) are reported in the Tables.

Power Analysis

Post-hoc power analyses were computed using GPower 3.0 (available for download at https://www.psychologie.hhu.de/arbeitsgruppen/allgemeine-psychologie-und-arbeitspsychologie/gpower.html) to determine whether the sample sizes for analyses provided sufficient power to detect a medium-sized effect in a mixed-design ANOVA with one between-subjects factor with 3 levels and two dependent variables. Power estimates ranged from .81 – 1.00, indicating sufficient power to detect medium or large (between × within) interaction effects (for α = .05, effect size f = .25 as recommended by Faul et al., 2007, and Ns between 87 – 160) for analyses that investigated differences among acculturation groups and within dyads.

Results

Results are reported following the 4 research goals and tied to hypotheses stated in the Introduction.

Responsiveness as Temporal Contingency

Responsiveness in mother-infant interactions was assessed using one-sample t-tests. As mentioned previously, in this study responsiveness was defined as temporal contingency, which was operationalized as the likelihood of one partner responding with a target behavior to the signal (i.e., initiating) behavior of another partner within 3 sec of the onset of that signal behavior. ORs greater than 1 indicate that the target behavior is more likely to begin within 3 sec of the onset of the signal behavior than at other times (Bakeman & Quera, 2011), thus one-sample t-tests compared mean ORs for each sample with 1.00 (α = .05). ORs > 1 are identified in Tables 3–5 with an asterisk (*); t-test statistics appear in Appendix C in the Supplementary Materials. Significant results summarized here reflect small-to-large effect sizes; thresholds for small, medium, and large effects for Cohen’s (1988) d are 0.20, 0.50, and 0.80, respectively.

Our hypothesis that Speak → Attend and Attend → Speak interactions would be temporally contingent in every cultural group was supported. No other hypotheses were made; however, other behaviors were temporally contingent. Specifically, Play → Attend interactions were temporally contingent for all except Japanese immigrant dyads, and Attend → Play interactions were temporally contingent for all except Japanese and Japanese immigrant dyads. Play → Vocalize and Vocalize → Play interactions were not temporally contingent in any cultural group. Encourage → Vocalize and Vocalize → Encourage interactions were temporally contingent only for Argentine and European American dyads (although not indicated with an asterisk in Table 3 for European American dyads, this was due to large intragroup variability for these DVs for the European American sample in the Japanese comparison; when all three European American samples were combined, the ORs significantly exceeded 1.00: Encourage → Vocalize M = 1.56, SD = 1.21, and Vocalize → Encourage M = 1.82, SD = 1.86).

Mean Differences in Responsiveness Among and Within Dyads

Acculturation of responsiveness at the group level was assessed by exploring mean differences in responsiveness among and within dyads with AN(C)OVAs. Specifically, one mixed-design AN(C)OVA with one between-subjects factor (Acculturation Group with 3 levels: Country of Origin, Immigrant, Country of Destination) and one within-subjects factor (Respondent with two levels: infant responds, mother responds) was performed separately for each interaction type (Mother Speak/Infant Attend, Mother Play/Infant Attend, Mother Play/Infant Vocalize, and Mother Encourage/Infant Vocalize) and acculturation comparison (Japanese, South Korean, South American), leading to 12 mixed design AN(C)OVAs. Full F, covariate information, and descriptive statistics for the Japanese comparison are reported in Table 3, the South Korean comparison in Table 4, and the South American comparison in Table 5. When significant main effects or interactions were found, pairwise t-tests with Bonferroni’s correction were used to decompose those effects and identify mean differences. Interaction effects supersede main effects, so when interaction effects were found, only they are discussed. Only statistically significant results (α = .05) are described below, and they reflect small-to-large effect sizes (η²_p of .01 is considered a small, .06 a medium, and .14 a large effect size; Cohen, 1988). As a reminder, the hypotheses were that among cultural groups, European American mothers would be more responsive than immigrant mothers and mothers in the culture of origin, and there would be no cultural differences in responsiveness among infants. Within dyads, European American and immigrant mothers were expected to be more responsive than their infants, and infants were expected to be more responsive than their mothers in the cultures of origin.

Table 4.

Temporal Contingency and Mean Differences in Responsiveness Among and Within Dyads for the South Korean Acculturation Comparison

Acculturation Group
Dependent Variables	South Korean (Origin) (n = 56)	South Korean Immigrant (n = 59)	European American (Destination) (n = 55)	AN(C)OVA
Mother Speak/Infant Attend bcd
Speak → Attend	2.24* (1.61)	2.08* (1.11)	3.11* (3.38)	Respondent within-subjects effect: F(1, 150) = 24.14, p < .001, η2p = .14 Acculturation Group main effect: F(2, 150) = 9.73, p < .001, η2p = .12 Respondent × Acculturation Group interaction: F(2, 150) = 1.10, p = .34, η2p = .01
Attend → Speak	1.46* (0.58)	1.73* (1.12)	2.02* (0.98)
Mother Play/Infant Attend
Play → Attend	3.90* (5.17)	3.00* (3.69)	4.28* (4.58)	Respondent within-subjects effect: F(1, 138) = 2.72, p = .10, η2p = .02 Acculturation Group main effect: F(2, 138) = 2.12, p = .13, η2p = .03 Respondent × Acculturation Group interaction: F(2, 138) = 0.22, p = .80, η2p = .00
Attend → Play	4.50* (5.09)	3.79* (3.91)	4.67* (5.05)
Mother Play/Infant Vocalize c
Play → Vocalize	1.31 (1.77)	1.52 (2.21)	1.14 (1.01)	Respondent within-subjects effect: F(1, 138) = 0.07, p = .80, η2p = .00 Acculturation Group main effect: F(2, 138) = 0.94, p = .39, η2p = .01 Respondent × Acculturation Group interaction: F(2, 138) = 1.59, p = .21, η2p = .02
Vocalize → Play	1.13 (1.57)	1.31 (1.83)	1.49 (1.45)
Mother Encourage/Infant Vocalize a
Encourage → Vocalize	1.53 (1.52)	1.71 (1.58)	1.38* (0.82)	Respondent within-subjects effect: F(1, 156) = 1.53, p = .22, η2p = .01 Acculturation Group main effect: F(2, 156) = 1.26, p = .29, η2p = .02 Respondent × Acculturation Group interaction: F(2, 156) = 0.19, p = .83, η2p = .00
Vocalize → Encourage	1.46 (1.57)	1.49 (1.81)	1.74* (1.51)

Note. Data (M, SD) are untransformed ORs (ORs) not controlled for covariates. Asterisks (*) indicate temporal contingency in mother-infant interactions, specifically, that the OR was significantly greater than 1 (as indicated by a one-sample t-test where p ≤ .05). AN(C)OVAs examined acculturation of responsiveness at the group level. Superscripts indicate the specific covariates that were used in the mixed-design AN(C)OVAs:

^ainfant age

^bmother age

^cmother education

^dmother in view.

Japanese Comparison.

Two of the four AN(C)OVAs yielded significant results. First, when the significant Respondent within-subjects effect for the Mother Speak/Infant Attend ANCOVA was decomposed, infants were significantly more likely to respond to their mothers (M = 3.00, SD = 3.29) than their mothers were to them (M = 2.18, SD = 1.54), regardless of cultural group; this pattern of results was expected only for dyads in the culture of origin (Japanese dyads). Second, in support of our hypothesis when the significant Respondent × Acculturation group interaction for Mother Play/Infant Attend ANOVA was decomposed, results indicated that European American mothers were significantly more likely to respond to their infants looking at them by engaging in social play than either Japanese or Japanese immigrant dyads. Unexpectedly, Japanese immigrant infants were significantly more responsive than their mothers in Mother Play/Infant Attend interactions (i.e., M for Play → Attend > M for Attend → Play). Although the Respondent within-subjects effect for the Mother Play/Infant Vocalize ANCOVA was statistically significant, pairwise comparisons were not significant.

South Korean Comparison.

For the South Korean comparison, one of the four AN(C)OVAs yielded significant results. The Mother Speak/Infant Attend ANCOVA showed a Respondent within-subjects effect: Infants were more responsive than their mothers (infant responds M = 2.48, SD = 2.29, mother responds M = 1.75, SD = 0.96); that this would be the case was hypothesized for dyads in the culture of origin (South Korean) only. Additionally, an Acculturation group main effect emerged: European American dyads were more mutually responsive than either South Korean or South Korean immigrant dyads in Mother Speak/Infant Attend interactions (estimated marginal mean [EMM] = 2.56, SD = 1.37, for European American dyads, EMM = 1.85, SD = 1.38, for South Korean dyads, and EMM = 1.91, SD = 1.37, for South Korean immigrant dyads); that this pattern of results would be found was hypothesized only for maternal responsiveness (and not responsiveness of the dyad).

South American Comparison.

For the South American comparison, three of the four AN(C)OVAs yielded significant results. First, a Respondent within-subjects effect for the Mother Speak/Infant Attend ANCOVA emerged: Infants were more responsive than their mothers (infants respond M = 2.10, SD = 1.37; mothers respond M = 1.82, SD = 0.88); this was expected to be the case for dyads in the culture of origin (Argentine dyads) only. Second, a main effect of Acculturation Group emerged in the Mother Play/Infant Vocalize ANCOVA: European American dyads (EMM = 1.48, SD = 0.87) were more responsive to their partners in Mother Play/Infant Vocalize interactions than South American immigrant dyads (EMM = 0.76, SD = 0.87). Third, an Acculturation Group main effect emerged in the Mother Encourage/Infant Vocalize ANOVA: European American dyads (EMM = 1.86, SD = 0.89) were more responsive to their partners in Mother Encourage/Infant Vocalize interactions than South American immigrant dyads (EMM = 1.22, SD = 0.89). The second and third results partially support hypotheses because, although Ms were expected for culture of destination > immigrant (in this case, European American > South American immigrant), culture of destination > culture of origin was also expected (for this comparison, European American > Argentine) and this result was expected only for maternal responsiveness.

Attunement of Mother and Infant Responsiveness Within Dyads

To investigate the attunement of mother and infant responsiveness within dyads, one-tailed Pearson correlations were performed with transformed data controlling for covariates and excluding outliers as necessary (Table 6). Attunement was indicated by significant positive correlations between pairs of ORs; for example, a correlation for Mother Speak/Infant Attend in Table 6 reflects the correlation between infant looks at mother then mother responds by speaking to infant within 3 sec of the onset of infant looking (Attend → Speak), and mother speaks to infant then infant looks at mother in response within 3 sec of the onset of mother speaking (Speak → Attend). Statistically significant results (α = .05) represented medium (r = .30) to large (r = .50) effects (Cohen, 1988) and were as follows: As hypothesized, Mother Speak/Infant Attend interactions were attuned for dyads in every cultural group. No other hypotheses were made, but Mother Play/Infant Attend interactions were attuned for South Korean immigrant and South American immigrant dyads only. Mother Play/Infant Vocalize interactions were attuned for South Korean, South Korean immigrant, Argentine, and European American dyads. Argentine and South American immigrant dyads were attuned for Mother Encourage/Infant Vocalize interactions.

Table 6.

Attunement of Mother and Infant Responsiveness Within Dyads

	Countries of origin				Immigrant dyads		Destination
Interaction types	Japan (n = 46)	South Korea (n = 56)	Argentina (n = 48)	Japanese immigrant (n = 32)	South Korean immigrant (n = 59)	South American immigrant (n = 32)	European American (n = 135)
Mother Speak/Infant Attend	.35* cd	.59*** cd	.63*** d	.62***	.31** abce	.63***	.67*** abcd
Mother Play/Infant Attend	.01	.20 d	−.11ad	.15	.30*	.42*	−.06
Mother Play/Infant Vocalize	.32	.36**	.64*** b	.07 c	.49*** c	.30	.42*** a
Mother Encourage/Infant Vocalize	.13	.19	.44**	.09	.05	.49** e	.11

Note. Statistics are one-tailed Pearson correlations. Data for the three European American samples were combined for these analyses. Superscripts in each cell indicate specific covariates that were used in the analysis in that particular cell:

^ainfant age

^bmother age

^cmother education

^dmother in view

^einfant awake.

^*p ≤ .05.

^**p ≤ .01.

^***p ≤ .001.

Two-tailed z-tests were computed to determine whether the magnitude of the correlations for the immigrant dyads differed from those of dyads in their respective cultures of origin or destination, as another indicator of group-level acculturation. No specific hypotheses were made. For Mother Speak/Infant Attend interactions, the correlation for European American dyads was larger than for South Korean immigrant dyads, z = 3.04, p = .002, although both were significant positive relations. For Mother Play/Infant Attend interactions, the correlations for South American immigrant dyads were larger than those for either Argentine, z = 2.17, p = .03, or European American dyads, z = 2.20, p = .03, and the correlation for South Korean immigrant dyads was larger than for European American dyads, z = 2.00, p = .046. No differences were found for Mother Play/Infant Vocalize interactions. For Mother Encourage/Infant Vocalize interactions, the correlation for South American immigrant dyads was larger than for European American dyads, z = 2.00, p = .046.

The Influence of Acculturation on Individual Responsiveness

To examine acculturation of mother-infant social interactions at the individual level for immigrant dyads, two-tailed Pearson correlations were performed between mothers’ acculturation level and each DV, controlling covariates as necessary (see Table 7). The hypotheses were general and predicted that mothers who were more acculturated to the U.S. would be more responsive and that mothers who were more acculturated to the culture of origin would have infants who were more responsive. The few significant relations that were found involved Mother Play/Infant Vocalize interactions: South Korean immigrant mothers who were more acculturated to South Korean culture were more likely to have infants who responded to their social play by vocalizing, supporting the hypothesis. Contrary to hypotheses, South Korean immigrant mothers who were more acculturated to South Korean culture and less acculturated to European American culture were more likely to respond to their infants’ vocalizations with social play. Contrary to hypotheses, South American immigrant mothers who were more acculturated to European American culture were more likely to have infants who responded to their social play by vocalizing. No relations obtained for Japanese immigrant mothers. The effect sizes for these significant correlations were medium to large (rs = .30 to .50, respectively; Cohen, 1988).

Discussion

This study examined four types of social interactions to learn the extent to which mother-infant interactions are temporally contingent and influenced by acculturation at the group and individual levels. The results expand our understanding of parenting and infant development by exploring: Four understudied types of mother-infant social interactions, infant responsiveness in addition to maternal responsiveness, and the extent to which these interactions characterize immigrant dyads from three cultural groups. As hypothesized, mother and infant responsiveness in Mother Speak/Infant Attend interactions were temporally contingent and attuned in all cultures. Contrary to expectations, infants were more responsive to their mothers than their mothers were to them in Mother Speak/Infant Attend interactions across cultures (i.e., infants responded by attending to mothers’ speech more often than mothers responded by speaking when their infants looked at them). Some cultural variability in the acculturation of responsiveness and attunement of responsiveness in social interactions at the group level was found, and responsiveness in Mother Play/Infant Vocalize interactions was influenced by acculturation at the individual level, although few of those findings were as expected. In the balance of the Discussion, the developmental significance of these findings is discussed.

Responsiveness as Temporal Contingency

As hypothesized, Mother Speak/Infant Attend interactions were temporally contingent in every acculturation group (as evidenced by ORs significantly greater than 1.00; Bakeman & Quera, 2011), suggesting that the behaviors that compose these interactions tend to co-occur. Mother Speak/Infant Attend interactions were the only social behaviors temporally contingent in every acculturation group in this study. Previous research has shown that mothers are likely to respond contingently to their infants’ social (person-directed) and didactic (object-directed) behaviors and their infants’ vocalizations, but that infants only respond contingently to their mothers in person-directed social behaviors at 5½ months of age, and do not yet do so for object-directed behaviors (Bakeman & Adamson, 1984; Cote et al., 2008) or vocalizations (Bornstein et al., 2015; Cote & Bornstein, 2021a; Kuckirko et al., 2018). This pattern of responsiveness reveals a developmental component to the temporal contingency of infant responses, namely that temporal contingency in interactional sequences follows a developmental timetable with interactional sequences that are most salient to optimal development of the human infant likely acquired earlier. Infant responsiveness in social interactions has been found to increase over time with advances in development, for example as infants become more responsive to their parents’ bids for social interaction between 7 and 15 months of age (Kochanska & Aksan, 2004). Among the types of social interactions examined here, Mother Speak/Infant Attend interactions were temporally contingent in all cultural groups explored, even when infants were the responding partner, suggesting that Mother Speak/Infant Attend interactions may precede developmentally the other types of social interactions studied here, with other types of social interactions becoming more prevalent later in infant development. Because Mother Play/Infant Vocalize interactions were not temporally contingent in any acculturation group (for mothers or infants), they may trail other types of social interactions developmentally. Mother Encourage/Infant Vocalize interactions were temporally contingent only for Argentine and European American dyads, suggesting that this type of interaction is characteristic of European-origin parenting. It should be noted that infant responses were more likely to be temporally contingent when they consisted of looking at mother (Speak → Attend and Play → Attend) rather than vocalizing (Encourage → Vocalize and Play → Vocalize), and developmentally infants attend to people earlier than they can intentionally vocalize nondistress; moreover, attending is a more passive and thus “easier” response to make than intentional nondistress vocalization.

Mean Differences in Responsiveness Among and Within Dyads

When social interactions of mother-infant dyads in the cultures of origin, destination, and immigrant samples were compared, cultural differences by interaction type obtained. Specifically, for the Japanese comparison a difference was found for Mother Play/Infant Attend interactions: European American mothers were more likely to respond to their infant looking at them by engaging in social play than either Japanese or Japanese immigrant dyads. For the South Korean comparison, European American dyads were more responsive to their partners than either South Korean or South Korean immigrant dyads in Mother Speak/Infant Attend interactions (regardless of whether mother or infant initiated). For the South American comparison, European American dyads were more likely than South American immigrant dyads to respond to their partners in both Mother Encourage/Infant Vocalize and Mother Play/Infant Vocalize interactions (regardless of who initiated the interaction). Across all three acculturation comparisons (Japanese, South Korean, South American), when differences in an acculturation group were found, immigrant dyads were less likely to show temporally contingent responsiveness than European American dyads (and not different from dyads in the cultures of origin). The immigrant groups were similar in terms of mothers’ education and all at least middle class, suggesting that these demographic factors and immigrant status may be more important determinants of behavioral interactions than cultural group.

When group-level acculturation differences in interactions were found, immigrant dyads engaged in fewer temporally contingently responsive interactions than their nonmigrant peers, which may indicate adaptation issues for immigrant families. In addition to being new parents, even if the migration was not recent, immigrant mothers in this study have experienced a major life transition and likely do not enjoy the extended support networks that most new mothers have (viz., parents or other close relatives). Other research with immigrant samples supports this view; for example, Gratier (2003) found that Indian immigrant mother-infant dyads engaged in lower levels of interactional synchrony in vocal interactions than nonmigrant Indian and U.S. American dyads. However, in the current study, Japanese immigrant and South Korean immigrant dyads and dyads in their respective cultures of origin both displayed lower levels of responsiveness than European American dyads (although for different types of social interactions), perhaps stemming from differential valuing of autonomy-promotion and harmony-maintenance in interpersonal relationships. Practitioners should be aware of such differences in interactions among immigrant families to more effectively serve them.

One pervasive within-dyad difference in responsiveness was found across all acculturation comparisons: Infants were more likely to respond to their mothers in Mother Speak/Infant Attend interactions than their mothers were to them. These results suggest that infants learn early to pay attention and look at their mothers when their mothers are speaking, further supporting the idea that this type of interaction is developmentally significant. In accord with this view, researchers suggests that infants’ attention to faces in temporally contingent interactions facilitates language development (Masek et al., 2021a).

Attunement of Mother and Infant Responsiveness Within Dyads

It is likely that attunement in social interactions emerged between mothers and infants because they are developmentally significant or culturally important. In this study, Mother Speak/Infant Attend interactions were attuned for all cultural groups. Mother Speak/Infant Attend interactions were the only type of social interaction that showed a high degree of both responsiveness (temporal contingency) and attunement in every cultural group, further suggesting that this type of interaction is developmentally significant. Van Egeren et al. (2001) found that infants were equally likely to respond to maternal speech by looking at their mothers as by vocalizing. Although maternal speech and infant vocalizations have received attention as interactional precursors to language development, infants’ attention to maternal speech is also important. Infants clearly do a lot of listening (have “exposure”) to language before they produce language on their own; that is, receptive language develops well before expressive language. By looking at mother’s face when she is speaking, infants gain information about how to articulate certain sounds (phonemes; Meltzoff & Kuhl, 1994; Patterson & Werker, 1999), which is an important component of learning to speak a language. The processing of speech by infants is intermodal and biologically relevant (Meltzoff & Kuhl, 1994). Earlier attunement between mothers’ and infants’ responsiveness may be greater for behaviors that are developmentally salient and important, such as speech (for hearing infants).

The other social interactions examined here tend to conform to cultural specificities in attunement (Bornstein, 2022). For example, the finding that Mother Encourage/Infant Vocalize interactions were attuned only for Argentine and South American immigrant dyads, coupled with the fact that these interactions were temporally contingent only for Argentine and European American dyads, suggests that Mother Encourage/Infant Vocalize interactions may be a particular style of social interaction for European-origin families. Additionally, if these analyses had only examined mean differences and not attunement, South American immigrant families may have appeared deficient because their mean levels of responsiveness in Mother Encourage/Infant Vocalize interactions were lower than European American dyads for Mother Encourage/Infant Vocalize and Mother Play/Infant Vocalize interactions. However, behavioral attunement was found for ¾ of the interactions studied for South American immigrant dyads and only ½ for European American dyads, and for two of these interactions the relation was stronger for South American immigrant than European dyads. Thus, attunement constitutes another way to construe social interactions beyond statistical contingency and mean level differences among acculturation groups. Attunement may be more important than mean differences in responsiveness because it appears to be a necessary feature of mother-infant social interactions and promotes children’s social well-being (Bornstein, 2013; Field et al., 2007; Stern, 1985).

The Influence of Acculturation on Individual Responsiveness

When culturally valued behavioral domains are examined, for example object mastery, acculturative differences are found at the individual level. For example, a positive relation emerged between South Korean American mothers’ responsiveness in object-directed interactions with their infants and acculturation to European American culture (Bornstein et al., 2019). However, in this study, few significant relations between acculturation and mother-infant interactions were found. Mothers who were more acculturated to U.S. culture were expected to be more responsive in interactions than their infants, and mothers were more acculturated to the culture of origin were expected to be less responsive in interactions with their infants (mothers would initiate interactions). This hypothesis was supported only for Play/Vocalize interactions among South Korean immigrants (i.e., mothers who were more acculturated to South Korean culture were more likely to have responsive infants). Contrary to hypotheses, South Korean immigrant mothers who were more acculturated to South Korean culture and less acculturated to European American culture were more likely to respond to their infants’ vocalizations with social play. Contrary to hypotheses, South American immigrant mothers who were more acculturated to European American culture were more likely to have infants who responded to their social play by vocalizing. No relations obtained for Japanese immigrant mothers. Acculturation at the individual level may be more elusive when samples of mothers and infants are relatively homogenous in terms of generation level (all immigrant mothers in this study were born in their cultures of origin and moved to the United States and their infants were born in the United States and would be considered first generation). Additionally, these results suggest that multidimensional measures of acculturation, such as was used for the South Korean immigrant sample, are more likely to yield findings than when unidimensional measures of acculturation are used (as were used for the South American immigrant and Japanese immigrant samples).

Limitations and Future Directions

Limitations of this study include the fact that this study defined responsiveness as temporal contingency, and as such could not assess maternal sensitivity in interactions. However, as Keller (2018) pointed out, sensitive responsiveness is subjective and culture-bound, and thus as a construct is particularly risky to assess in cross-cultural research. Defining responsiveness as temporal contingency eliminates such coding bias (Leclère et al., 2014).

Additionally, the results of this study generalize only to middle-class urbanites in the particular cultural groups studied. It may be that few acculturation differences at the group level were found because the immigrant dyads and those in the cultures of origin and destination were relatively socioeconomically homogenous (middle-class two-parent families where parents had some college education on average; Hollingshead, 2011). Cross cultural researchers (Lin et al., 2023) have found variability in ideal parenting beliefs due to parental education level in Asian (Japan), Hispanic-Italian (Chile, Peru, Ecuador, Uruguay), and Western English-speaking countries (e.g., the United States). Among these samples of immigrants in the United States, socioeconomic status may have had a homogenizing effect on parenting, because few differences among immigrant samples have been found in previous work (when immigrant mothers’ nurturing, physical, and didactic behavior, and provision of materials in the environment were compared,Bornstein & Cote, 2001; and when immigrant mothers’ play was compared, Cote & Bornstein, 2009). Moreover, Masek et al. (2021b) reported that mother-toddler interactions differentially predict children’s language outcomes based on family socioeconomic status even among families that are sociodemographically homogenous in other ways (i.e., all lived in the United States, were monolingual English-speaking, and the majority were White).

However, one type of behavioral interaction, Mother Speak/Infant Attend (mothers speak to their infants and infants look at their mothers), was temporally contingent across samples, was attuned in mother and infant dyads across samples, and showed acculturation differences. This dynamic raises compelling questions about the developmental importance of this type of interaction for children in immigrant families. Previous research has suggested that infants’ attention to faces in temporally contingent interactions facilitates language development (Masek et al., 2021a), however, most research in this area has been conducted with monolingual English-speaking White middle-class families, and it is unknown whether these relations might obtain among bilingual immigrant families from diverse cultural contexts. Future research should investigate the developmental significance of these interactions and their variability for infants’ later language and social development, because the developmental sequelae of parent-infant interactions likely vary based on cultural context.

Conclusion

By investigating social interactions among nonEuropean origin samples, this study expands the current literature and contributes to a more inclusive and global understanding of early human development (Barbot et al., 2020). Furthermore, by sampling multiple immigrant groups, this research highlights commonalities and specificities in mother-infant social interactions among immigrant and nonmigrant families. Research on immigrant families is timely and necessary, because most developmental research has been conducted on European-origin children, and it is erroneous to assume that theory and research findings from one cultural context automatically extend to all children everywhere. Because the numbers of immigrant children around the world are growing, testing the extent to which theories based on White middle-class samples extend to other cultural contexts is an urgent contemporary concern for developmental science. For example, one common estimate suggests that 1 in 4 children under the age of 5 in the United States has at least one foreign-born parent (Urban Institute, 2019). Research on the acculturation of social interactions among immigrant families can reward professionals with a deeper understanding of culturally typical mother-infant interactions, which can help healthcare, human services, and educational professionals interact with their clients with improved understanding and appreciation of cultural differences.

Highlights.

• Mother Speak/Infant Attend interactions were contingent in every culture

• Mother and infant responsiveness in Mother Speak/Infant Attend interactions was attuned in all cultures

• Infants were more responsive than their mothers in Mother Speak/Infant Attend interactions

• Cultural variability in the acculturation of responsiveness in social interactions at the group level emerged

• Responsiveness in Mother Play/Infant Vocalize interactions were amenable to acculturation at the individual level

Data Availability Statement

Anonymized data are available upon request from the authors. Authors do not have permission from participants to publicly share the data.