CULTURAL SOCIALIZATION IN CHILDHOOD: ANALYSIS OF PARENT-CHILD CONVERSATIONS WITH A DIRECT OBSERVATION MEASURE

Abstract

Cultural factors influence the development of all children. Yet, current knowledge of explicit cultural socialization processes in childhood remains limited, mainly by failing to incorporate the experiences of young children. To address this critical gap, the authors introduce the OMERS-Peds task, an observational measurement designed to systematically identify and compare the content of cultural messages passed down from caregivers to offspring during early school age years. The OMERS-Peds was administered to mothers and children (n= 275) from three diverse racial/ethnic backgrounds (African American (n= 153), Hispanic (n=61), and non-Hispanic White (n= 61)) within the longitudinal Multidimensional Assessment of Preschoolers (MAPS) Study. The OMERS-Peds coding system was used to rate how strongly families endorsed five key constructs: family culture, religion, identity, ethnicity, and race. A series of χ² statistic tests were used to compare scores across racial/ethnic backgrounds, and within families (between children and their mothers). Analyses revealed that in the cultural socialization conversations occurring in early childhood, parents and children prioritize talking about their family’s culture and religion. Independent of their racial/ ethnic backgrounds, mothers and children seldom discussed race and ethnicity. Contrary to research with older children, differences were mainly identified within families, rather than across racial/ethnic groups. Findings support the need to include children’s perspectives in the assessments of cultural socialization, as opposed to relying primarily on parent reports, and highlight the importance of having an observational methodology that allows researchers to examine parent-child bidirectional interactions during early school age years in a systematic manner.

INTRODUCTION

Racial and ethnic disparities across cognitive, social, behavioral, and health outcomes have their origins in the early years of life (Dixon, Peña, & Taveras, 2012). For example, racial and ethnic differences in age- and sex-adjusted body mass index (BMI) are present by age 6 years (Cunningham, Kramer, & Narayan, 2014). Elucidation of early-life risk and protective factors associated with the development of racial and ethnic disparities is a national priority (Koh, Piotrowski, Kumanyika & Fielding, 2017). Although early childhood is increasingly being recognized as a critical period in the development of racial and ethnic disparities, limited understanding of the underlying causes of disparities precludes the development of effective strategies to abate inequalities. At the root of developing effective prevention and treatment strategies to address the early-life factors that explain disparities in childhood and adulthood, is the recognition of the importance of culture (Koh et al., 2017).

Cultural factors influence the health of all children, but culture is particularly salient for children in racial and ethnic minority families (Power, O’Connor, Orlet-Fisher & Hughes, 2015; Hughes et al., 2006; Yasui & Dishion, 2007). Children from racial and ethnic minority families experience cultural socialization practices that diverge from those of non-Hispanic White majority families (Cabrera, Kuhns, Malin, & Aldoney, 2016; Hughes et al., 2006; Priest et al., 2014; Yasui & Dishion, 2007). Nonetheless, despite the significance of culture in children’s health, and overall development, and the potential implications of cultural socialization in childhood for adult health, our understanding of parent-child cultural socialization processes remains limited, particularly at younger ages (Cabrera et al., 2016; Hughes et al., 2006; Priest et al., 2014; Yasui, 2015). A better understanding of cultural socialization in early life can advance understanding about the ways in which the formation of ethnic identity is associated with mental and physical health, as well as educational attainment.

Cultural Socialization

Cultural socialization is a multi-faceted empirically defined component of parental socialization processes that are key to the development of children’s sense of values, priorities, and morals for interacting with the world around them (Lee, 2003; Maccoby, 1992). Cultural socialization includes all processes whereby children develop into competent and functional members of culturally diverse societies (Lee, 2003). In the context of the family, cultural socialization refers to implicit and explicit practices that parents use to communicate or promote cultural pride, teach cultural knowledge, beliefs, and behaviors, and pass down cultural norms, expectations, and traditions (Hughes et al., 2006; Lee, 2003; Umaña-Taylor & Fine, 2004).

Studies of explicit parent-child cultural socialization in young children are very limited (Caughy, Randolph, & O’Campo., 2002; Loyd & Gaither, 2018; Priest et al., 2014). Despite the shifting attention to bidirectional processes in the field of child socialization, current understanding of cultural socialization in early ages continues to be defined by reports from parents or other adults who are prominent socializing agents (Cabrera et al., 2016; Loyd & Gaither, 2018; Maccoby, 1992; Priest et al., 2014). Investigations of cultural socialization of young children have mainly focused on measures that examine parent ethnic-racial socialization and different demographic factors or agents of influence associated with parental engagement and avoidance in cultural socialization practices (i.e., family, peers, media, community) (Priest et al., 2014).

Current knowledge of cultural socialization in early childhood is mostly defined by measurements of parents’ behaviors, experiences and attitudes, particularly by the well-documented parental reluctance towards the introduction of ethnic-racial socialization messages to young children (Priest et al., 2014). Studies have documented that ethnic-racial socialization avoidance in early childhood is most often attributable to parents’ beliefs that acknowledging differences promotes prejudice (Cabrera et al., 2016; Loyd & Gaither, 2018; Priest et al., 2014). Parents’ reluctance towards the introduction of cultural and ethnic-racial socialization messages to young children is also linked to the belief that children are not developmentally ready (Loyd & Gaither, 2018; Priest et al., 2014; Yasui, 2015). By age 12 years and older, more is known about the specific content of explicit ethnic-racial socialization processes of children in racial and ethnic majority and minority families (e.g., cultural socialization, preparation for bias, cultural mistrust) (Hughes et al., 2006; Loyd & Gaither, 2018; Priest et al., 2014; Yasui, 2015; Yasui & Dishion, 2007). Findings suggest that in some instances, engagement in explicit cultural socialization increases with child’s age (Loyd & Gaither, 2018; Priest et al., 2014; Yasui, 2015). However, to our knowledge, conversations between parents and their early school age children have not been examined within this context.

Although most studies of cultural socialization have focused on adolescence, there is increasing evidence from implicit cultural socialization assessments suggesting that the roots of these processes occur much earlier (Rogoff et al., 2007). Contrary to what parents believe, research on implicit cultural socialization practices recognizes young children are constantly learning culturally relevant messages through observation of their parents and participation in routine activities (Rogoff et al., 2007). Therefore, early school age represents a key period in the development of cultural socialization (Anderson, Hughes, Fisher, & Nicklas, 2005; Cabrera et al., 2016). As children transition to school in multicultural societies, they must learn about and acquire the skills to negotiate racial, ethnic, and cultural diversity within their everyday lives (Hughes et al., 2006; Priest et al., 2014). Transitions to school coincide with changes in explicit parent-child interactions, potentially in response to demands introduced by children’s frequent encounters in multicultural settings (Anderson, Hughes, Fisher, & Nicklas, 2005; Cabrera et al., 2016).

Children’s cultural socialization during early school age has important implications for children’s psychological, social, and emotional development (Cabrera et al., 2016). Early school age is a critical period for children’s future health when most children adopt health-related attitudes and behaviors that will last a lifetime (Harrison et al., 2011; Hughes et al., 2006; Priest et al., 2014; Yasui, 2015). Since culture permeates every aspect of children’s socioecological system, it is imperative that researchers further examine children’s socio-cultural interactions with their parents using a bidirectional framework that enables them to understand how early school age children develop and lay the foundation to navigate an increasingly diverse world (Cabrera et al., 2016). Deeper explorations are needed to identify communication patterns and cultural constructs discussed across racial and ethnic diverse families, and to better understand the key differences in explicit messages parents and children exchange as part of the cultural socialization process.

Characterizing how school age children are socialized and make meaning of messages in their social environments is challenging. Children’s limited capacity to self-report on complex and highly sensitive issues introduces a myriad of methodological difficulties inherent to examining this bidirectional, interactive process with developmentally appropriate means in culturally diverse contexts (Carter, Briggs-Gowan, & Davis, 2004; Wakschlag et al., 2005; Yasui, 2015). To address some of the methodological challenges, previous research studies relied on parent-reports to examine how young children are culturally socialized (Carter et al., 2004). Parent-reports of the communications exchanged during cultural socialization processes provide critical information from the parent perspective but fail to capture the bidirectional context of the child-caregiver relationship, the children’s experience, and the unique cultural messages introduced during parenting of which parents may or may not be aware (Yasui, 2015; Yasui & Dishion, 2007). In addition, parent-reports have been shown to be highly sensitive to social desirability, and interviewer or respondent bias, such as concerns about political correctness, and being viewed as racist or prejudiced (Loyd & Gaither, 2018). Thus, scholars have highlighted the need to identify novel methodologies that can overcome these measurement issues and facilitate the study of the messages exchanged within the cultural socialization processes during key developmental stages. Studies of cultural communications during key developmental stages, like early school age, provide the opportunity to gain a deeper understanding of how cultural socialization processes introduce lasting effects upon children’s mental, and potentially physical health (Loyd & Gaither, 2018).

The present study

This study was embedded in the second wave of the Multidimensional Assessment of Preschoolers Study (MAPS). MAPS is a comprehensive longitudinal study designed to characterize neurodevelopmental profiles of diverse young children (Briggs-Gowan et al., 2019; Wakschlag et al., 2015). The MAPS cohort recruited parent-child dyads from an urban pediatric clinic setting and conducted assessments at preschool, early school age, and pre-adolescence (Briggs-Gowan et al., 2019; Wakschlag et al., 2015). The focus of this study is on the early school age wave, where we employed a developmental adaptation of the Observational Measure of Ethnic-Racial Socialization (OMERS) task: the OMERS-Peds (Yasui, 2008; Yasui, Dishion, Stormshak, & Ball, 2015). The OMERS-Peds is a standardized semi-structured direct observation paradigm developed to examine the variability in the cultural socialization processes in parents’ communications with their young children (Yasui, 2008). For the purpose of this study, we operationalized the concept of cultural socialization to focus on the messages parents use to transmit their values, beliefs, customs, and behaviors to the child and the extent to which the child internalizes these messages, adopts them, and acquires the skills and competence to integrate these messages into his/her/their discourse when discussing five specific constructs: (1) Family culture, (2) Religion or Spirituality, (3) Identity, (4) Ethnicity, and (5) Race. We employed the OMERS-Peds direct observation method to enable a deeper examination of the construct endorsement and message content of young children’s cultural socialization processes within, and across families from diverse racial/ethnic backgrounds. To our knowledge, this is the first study to directly observe cultural socialization processes at this young age, and in a large racial/ethnically diverse sample.

One of the most important attributes of the MAPS cohort is the diverse multicultural constitution of the sample, which includes African Americans, Hispanics, and non-Hispanic White mother-child dyads. Furthermore, the MAPS study specifically sought to reduce the confounding effects of poverty on racial/ethnic diversity by including socioeconomic diversity among minority participants (Wakschlag et al., 2014). Because of these attributes, this racially/ethnically and socioeconomically diverse cohort introduces a unique opportunity to expand the literature of explicit cultural socialization processes in childhood and address the two aims of this study. The first aim is to identify what constructs are endorsed during the cultural socialization process of early school age children and to compare construct endorsement across diverse racial/ethnic families. Our second aim is to examine the content of the messages exchanged in the socialization process in depth, by contrasting across diverse racial/ethnic families how each cultural construct is discussed by mothers and children in terms of promotion of cultural affiliation and involvement. We hypothesized that when comparing the content (what) and the quality (how) of the messages exchanged between children and their mothers we will document key differences across racial-ethnic mothers and children, and similarities within families (when comparing the constructs endorsed and messages of children with those from their mothers). Considering the centrality of race and ethnicity in the lives of families from minority backgrounds, we expect that when compared with non-Hispanic White mothers, African American and Hispanic mothers will place greater emphasis on racial and ethnic constructs (Hughes et al. 2006).

METHODS

Participants

Table 1 introduces the demographic characteristics of the children and mothers (n=275) recruited from the longitudinal Multidimensional Assessment of Preschoolers (MAPS) Study (for a detailed description see Briggs-Gowan et al., 2019; Wakschlag et al., 2015) to participate in the current study. Preschoolers and their mothers were recruited from five pediatric clinic waiting rooms in the Chicago area. The sample was broadly stratified by child age, sex, race/ethnicity, and poverty. Poverty was defined based on census guidelines (by household size and income). In addition, participation in TANF (Temporary Assistance for Needy Families) and participation in food stamps or the Link card program were included in the criteria for poverty, but only among families who were borderline poor based on the poverty line calculated by household size and income.

Table 1.

Child and mother’s demographic characteristics

	Overall (n=275)		African American (n=153)		Hispanic (n=61)		non-Hispanic White (n=61)		χ2/ f	df	p
Child Characteristics	n	%	n	%	n	%	n	%
Child sex									2.6	2	.276
Male	128	46.5	74	48.4	23	37.7	31	50.8
Female	147	53.5	79	51.6	38	62.3	30	49.2
Child age (years)
mean, sd	7	1.0	7.0	0.9	7.1	1.02	7.0	1.18	0.4	2,262	.684
Mother Characteristics
Age in years
mean, sd	33	6.3	32.1	5.9	31.0	6.3	37.6	5.1	22.9	2,262	<.001
Educational attainment									74.6	4	<.001
High school or less	59	22.5	30	20.1	24	41.4	5	9.1
Some college (less than bachelor’s)	132	50.4	98	65.8	22	37.9	12	21.8
Bachelor’s degree or higher	71	27.1	21	14.1	12	20.7	38	69.1
Employment status									6.7	4	.151
Employed full time	114	43.7	68	45.6	19	33.3	27	49.1
Employed part time	61	23.4	34	22.8	12	21.1	15	27.3
Not employed	86	33.0	47	31.5	26	45.6	13	23.6
Marital status									76.4	2	<.001
Married	113	44.8	31	21.7	35	64.8	47	85.5
Not married	139	55.2	112	78.3	19	35.2	8	14.5
Family Poverty status									77.1	4	<.001
Not poor	106	40.6	35	23.5	21	36.8	50	90.9
Borderline poor	47	18.0	34	22.8	13	22.8	0	0.0
Poor	108	41.4	80	53.7	23	40.4	5	9.1

Procedure

Upon arrival, mothers and their children were consented and assented, respectively. Participants completed the OMERS-Peds task as part of a battery of assessments during the early school age phase of the longitudinal study. The OMERS-Peds was embedded in a video-recorded, interactive 30-minute parent-child observational assessment. During the OMERS-Peds, parent-child dyads were left alone in a research room and given instructions for six tasks in a numbered flipbook. The beginning and end of each timed activity was signaled to the family using a bell, which was heard through an intercom system. Child and family demographic characteristics were collected by maternal report. At the end of the visit, families were compensated and thanked for their time. All study procedures received Institutional Review Board (IRB) approval from the Northwestern University IRB.

Measures

The OMERS-Peds.

To assess cultural socialization, participants completed a video-recorded behavioral observational task modeled after the Observational Measure for Ethnic-Racial Socialization (OMERS) (Yasui, 2008; Yasui, et al., 2015). OMERS was developed to determine if cultural socialization bidirectional processes in adolescence can be reliably observed (Yasui, 2008; Yasui, et al., 2015). Observations made with the OMERS were used to identify variability in the cultural socialization messages of adolescents from American Indian, African American, and European American families (Yasui, 2008; Yasui, et al., 2015). The OMERS operationalizes the family culture paradigm by coding the communications from each informant within key cultural socialization constructs (Yasui, 2008; Yasui, et al., 2015).

Because the OMERS was designed for adolescent samples, developmental modifications were made to facilitate its use with younger children and examine constructs relevant to their cultural socialization processes. To achieve this, three researchers on the team collaboratively modified the OMERS Family Culture Paradigm coding scheme. Specifically, an array of videos was randomly selected to assess the overall variability in the types of responses observed in both the mothers and the children. Based on these assessments, several changes were implemented. First, questions related to the mothers’ and children’s approach to the task were added. These questions formed the Parent and Child Approach to Task component which intended to gauge whether the parent-child dyads considered the task and their participation to be valuable and important to them. Second, various similar questions within the Family Culture Paradigm were combined and organized into three major components: Parent and Child Approach to Task, Family Structure and Relationships, and Cultural Socialization Messages, the focus of this study.

Similarly to the OMERS, the OMERS-Peds relied on a prompt and examples to ask families to have a conversation about values and traditions that are important to their family or cultural group. The instructions were designed to allow participants to discuss one or two topics in detail or to more broadly discuss several topics in the allotted time (~4-minutes). Mothers were instructed to:

“Please have a conversation with your child about values and traditions that are most important to you, your family, and your cultural group. Talk about why these values are important to your family and how they may be alike or different from other families or cultural groups. By culture we mean traditions and practices, how you celebrate holidays, family traditions and celebrations, history of your family, and/or culture, beliefs, role of the family, how family members treat each other, importance or religion or spirituality within your family. Talk to your child about values and traditions.”

Coding.

As with the prompt, the coding scheme was also modeled after the OMERS, but modified for use with conversations between early school age children (ages 5 to 8 years old) and their mothers (Yasui, 2008; Yasui, et al., 2015). The coding was revised to capture the types of conversations that children and their mothers were having. The research team modified the codes according to both the types of topics covered but also according to the age of the child. While the coding scheme for mothers remained constant, a separate coding scheme was developed for children between the ages of 5 to <7 years and children between the ages of 7 to 8 years. The different schema were intended to account for developmental differences between the contributions of younger versus older children. To capture the salient ideas brought into the conversation by children and the contributions of each conversational partner to the communication to greatest extent possible, coders were asked to focus on one coding scheme at the time (i.e., focus on children or mothers only). As a result of this approach, the bidirectionality was captured only in the nature of the exchanges, which were most often parent initiated. Lastly, the coding was simplified by changing the scale from 1 to 9 to a 1 to 4 scale. Table S1 introduces the different items and scores included in the coding scheme. Once the coding scheme was finalized, a researcher and two research assistants, who were present during the development of the coding scheme, independently watched and coded the cultural socialization task. The conversations from each member (child or mother) were coded separately using the corresponding coding scheme. The research assistants overlapped in 20% of their coding and had a 78% overall agreement. Table S2 introduces the observed inter-rated agreement weighted Cohen’s kappa, which was examined separately for each of cultural construct. Weighted Cohen’s kappa ranged from 0.64 to 1.00.

Modifications to the final coding scheme included dividing it into the three major components: Parent and Child Approach to Task, Family Structure and Relationships, and Cultural Socialization Messages. Cultural Socialization Messages were coded in relationship to five key constructs: (1) Family culture, (2) Religion or Spirituality, (3) Identity, (4) Ethnicity, and (5) Race. Trained coders assessed how families talked about each construct following the OMERS-Peds coding system that included six items for each construct. Briefly, the construct of family culture, referred to mothers or children conversations about instilling appreciation of their families’ traditions, and discussions of their cultural heritage in a broader manner, either by sharing knowledge about their cultural beliefs, and activities, social events, traditions, customs, or talking about practices related to the cultural group to which they belong. Conversations about religion or spirituality mainly referred to transmitting knowledge about their religious or spiritual traditions, customs and religious heritage. For the construct of identity, coders identified instances where mothers or children shared their knowledge about different representations of their identity, and their family identity (i.e. features that distinguished their family from others). Lastly, ethnicity referred to specific conversations about the family’s ethnic and cultural backgrounds or ancestry, and race referred to messages that convey information about having different individual and social affiliations, relationships, and social position because of their identification with a social group. Since coders assessed one construct at the time following the OMERS-Peds coding system (detailed in Supplement 1), it was possible that if a statement endorsed more than one cultural construct (i.e. religion and identity), both constructs were coded based on the same statement.

The OMERS-Peds coding system guided coders as they rated how strongly families endorsed the five key constructs examined: family culture, religion, identity, ethnicity, and race. The coding system specifically assessed these six items: (1) parent and child overall appreciation for the cultural construct discussed; (2) their knowledge about the topic they discussed; (3) their participation in activities or social events related to the topic; (4) the importance of passing down their knowledge about the topic; (5) the extent to which their religion, ethnicity, etc. influenced the way they live their lives; and (6) whether the family differentiated themselves from other religious, ethnic, etc., groups. All questions were rated on a scale from 1 to 4, with a 4 indicating that the specific belief or behavior was supported, or that the question was thoroughly discussed. Constructs that were not discussed were coded with a categorical value, as opposed to a missing value. We collapsed codes by construct by taking the maximum value from the 6-items scored. Using the maximum value limited the focus on the constructs that were not discussed, which accounted for the fact that only one or a few items within each construct could be covered in the 4-minute session.

All cultural constructs (family culture, religion, identity, ethnicity, and race) were assigned two scores. First, dichotomous scores were calculated to identify what did children and mothers talk about during conversations about cultural socialization. This approach was used to determine the cultural constructs that were endorsed, and the variability in the endorsement each cultural construct received from mothers and children across the different racial/ethnic groups. Participants received a score of 1 if they discussed a specific construct, and 0 if the construct was not discussed at all.

The second score was based on the categorical rating scale from 1 to 4 previously described. The collapsed score derived from the 6-items examined was assigned to each construct. This coding scheme classified low and high endorsement in any of the five cultural constructs coded (religion, ethnicity, culture, identity and race), and was used to determine how mothers and children talked about the different cultural constructs. Scores of 1 and 2 were associated with lower promotion of cultural pride, or discouraging involvement in each cultural construct, and scores of 3 and 4 were classified as higher endorsement, and promotion of cultural pride.

Statistical Analysis

A series of χ² statistic tests were used to address the first aim and compare the specific construct endorsement by each racial/ethnic group, and within families (parent vs. child) using the dichotomous scores. In lieu of post-hoc analysis, we conducted pairwise comparisons to further define the source of any significant differences identified. To assess our second aim, χ² statistic tests were used to examine differences in the categorical distribution of each cultural construct’s scores by racial/ethnic groups. Children and parent categorical distribution of scores per construct were examined separately.

RESULTS

Table 1 displays the sociodemographic characteristics for the sample overall, and by each racial/ethnic group. Children’s ages ranged from 5 to 8 years (mean age 7±1.0 years), with a higher proportion of children in the 6–7 group. Overall, 22% (61/275) of children were non-Hispanic White, 56% non-Hispanic Black or African American (153/275) and 22% Hispanic (61/275). About half of the children in the sample were female (54%, N=147), with a higher proportion of females in the Hispanic group (62.3%, N=38). Mother’s ages ranged from 22 to 50 years old, and mean mother’s age significantly varied by race and ethnicity (p<0.001). It was observed that the Hispanic mothers were the youngest (M=31, SD= 6.3), and the non-Hispanic White mothers were the oldest group (M=37.6, SD=5.1).

There were significant racial/ethnic differences in mothers’ educational attainment, marriage status and poverty status (p<0.001), but not in mothers’ employment status. It was observed that most African American mothers had some college education (65.8%, 98/153), but not a bachelor’s degree. Most non-Hispanic White mothers completed a bachelor’s degree or higher (69%, 38/61), and most Hispanic mothers completed high-school or less (41.4%, 24/61). Regarding marital status, 64.8% (35/61) of Hispanics, and 85.5% (47/61) of non-Hispanic White were married, which contrasted with 21.7% (31/61) of African American mothers. Lastly, based on the US Census definition of poverty, and our criteria to identify borderline poverty (participation in TANF, food-stamps or the link card program), it was observed that 58.6% (153/275) of families were considered non-poor. The distribution of poor, and non-poor families significantly varied across the different racial/ethnic groups, 46.3% (69/153) of the Black or African American families, 90.9% (50/61) of the non-Hispanic White families, and 59.6% (34/61) of the Hispanic families were non-poor.

What did children and mothers talk about during conversations about cultural socialization?

The modified OMERS-Peds methodology evidenced similarities and differences in the cultural constructs mothers and children endorsed and omitted during their conversations about culture. As shown in Table 2A, it was observed that the attention paid to each construct among mothers overall, and within each racial/ethnic group significantly varied χ²(4) =370.02, p <0.001. Of the five constructs of culture assessed, broad mentions of family culture, religion/spirituality, and identity, were included in the discussions of all families, independently of their race and ethnicity. In contrast, comments about ethnicity, and race were scarce among all racial and ethnic groups. Tables 2a, 2b, and 2c introduce a summary of what constructs were endorsed by mothers (2a) and children (2b) during the cultural socialization communications, and a comparison within mother-child dyads (2c) by race/ethnicity. As shown in Table 2c, there was high agreement between mothers and children in the constructs endorsed and omitted in their communications, and the overall attention paid to discussing the different constructs.

Table 2A.

Distribution of cultural constructs endorsed by mothers, overall, and by families race and ethnicity

	Overall n, %	African-American (n=153) n, %	Hispanic (n=61) n, %	non-Hispanic White (n=61) n, %	χ2(2)	P
Family Culture	235, 85.5	127, 83.0	54, 88.5	54, 88.5	1.66	.435
Religion/spirituality	189, 68.7	109, 71.2	39, 63.9	41, 67.2	1.17	.558
Identity	151, 54.9	84, 54.9	26, 42.6	41, 67.2	7.45	.024
Ethnicity	32, 11.6	9, 5.9	14, 23.0	9 ,14.8	13.10	.001
Race	13, 4.7	11, 7.2	0, .0	2, 3.3	5.37	.068
χ2(4)	370.02	184.90	118.61	105.63
P	<.001	<.001	<.001	<.001

Table 2B.

Distribution of cultural constructs endorsed by children, overall, and by families race and ethnicity

	Overall	African-American (n=153) n, %	Hispanic (n=61) n, %	non-Hispanic White (n=61) n, %	χ2(2)	P
Family Culture	234, 85.1	126, 82.4	54, 88.5	54, 88.5	2.04	.361
Religion/spirituality	186, 67.6	107, 69.9	38, 62.3	41, 67.2	1.17	.557
Identity	151, 54.9	84, 54.9	26, 42.6	41, 67.2	7.45	.024
Ethnicity	30, 10.9	8, 5.2	13, 21.3	9, 14.8	12.80	.002
Race	11. 4.0	10, 6.5	0, .0	1, 1.6	5.99	.050
χ2(4)	381.83	192.59	118.61	97.42
P	<.001	<.001	<.001	<.001

Table 2C.

Agreement between mothers vs. children (kappa statistics)

	Overall	African-American (n=153)	Hispanic (n=61)	non-Hispanic White (n=61)
Family Culture	.986	.977	1.000	1.000
Religion/spirituality	.975	.969	.965	1.000
Identity	1.000	1.000	1.000	1.000
Ethnicity	.964	.938	.953	1.000
Race	.913	.949	-	.659

All racial/ethnic groups showed greater interest in discussing their family’s culture and religion/spirituality, compared to the attention paid to race or ethnicity. Overall, 85.7% (235/275) of mothers, and 85.1% (234/275) of children, talked about family culture, in a broader sense, with no significant differences in endorsement by racial/ethnic groups. Religion/spirituality was the second most frequently endorsed construct by all racial ethnic groups. Overall, 68.7% (189/ 275) of mothers, and 67.6% (186/275) of children talked about religion and spirituality.

Endorsement of identity in the conversations was significantly different across the diverse racial/ethnic parent-child dyads. Of the three racial/ethnic groups, non-Hispanic White mothers and children spoke about identity the most (67.2%, 41/275 for parent and children) compared with 54.9% of African American (84/153), and 42.6% (26/61) of Hispanic parent-child dyads. Post-hoc pairwise comparisons revealed the significant differences between the three groups were mostly attributable to differences between non-Hispanic parent child-dyads and Hispanic parent-child dyads (χ² (2) = 12.84, p = 0.002 for pairwise comparisons of Hispanic and non-Hispanic white mothers, and χ² (2) = 7.99, p = 0.018 for pairwise comparisons of their children).

Across all groups, focus on ethnicity was limited, and significantly differed by race and ethnicity. Overall, only 11.6% of mothers (32/275) and 10.9% (30/275) of children talked about ethnicity. Ethnicity was mostly endorsed by the Hispanic group, and least endorsed among African Americans. Twenty-three percent (14/61) of Hispanic mothers, and 21.3% (13/61) of Hispanic children talked about ethnicity, compared with 5.9% (9/153) of African American mothers and 5.2% (8/153) of African American children. In contrast, endorsement of messages about race was the greatest among African Americans. Seven percent (11/153) of African American mothers, and 6.5% (10/153) of African American children talked about race.

Still, discussion specific to race across all racial/ethnic groups, were minimal. Hispanics did not talk about race at all, and among non-Hispanic Whites, only 2 of 61 mothers, and 1 of 61 children discussed race. Given the limited mentions of race we combined responses from race and ethnicity for pairwise post-hoc comparisons. Results showed the significant differences across the three racial/ethnic groups were attributable to the contrasting emphasis placed in these constructs among the African American and Hispanic mothers χ²(2) = 6.74, p = 0.034.

How did mothers and children talk about the different cultural constructs?

Within each domain, mothers endorsed different aspects of each cultural construct when talking with their young children. For example, within the family culture domain, it was noticed that most mothers talked about valuing one’s culture, whereas it was less common for mothers to focus on encouraging children to rely on their family culture in challenging times. Tables S3a and S3b introduce examples of the different messages mothers and children exchanged when talking about each cultural domain based on the high and low categorical scores previously described. For several constructs, no examples with low scores were discussed, thus more examples of high instead of low scores are presented.

In addition to assessing the arguments in depth, the coding scheme was used to examine the quality of the conversations. Scores assigned to the content of the messages revealed that variations in the quality of the messages across the different racial/ethnic groups were limited to the least discussed constructs. Namely, variations were found in conversations about identity, ethnicity and race. In contrast, most differences were found between children and their mothers within racial/ethnic groups. Mothers’ messages were scored as mostly positive or higher, compared with children messages, which were less positive than their mothers. Tables 3a, and 3b introduce the distribution of the scores used to examine the quality of the messages, and how the messages compared by race/ethnicity.

Table 3.a.

Distribution of how mothers discussed each construct by race and ethnicity

		Overall	African-American (n=153)	Hispanic (n=61)	non-Hispanic White (n=61)	χ 2	df	P
Construct	Score	n (%)	n (%)	n (%)	n (%)
Family Culture	Low (1–2)	85 (31%)	52 (34%)	15 (25%)	18 (30%)	4.78	4	.310
	High (3–4)	150 (55%)	75 (49%)	39 (64%)	36 (59%)
	Did not discuss (5)	40 (14.5%)	26 (17%)	7 (11%)	7 (11%)
	Mean/ sd	3.36 (1.16)	3.22 (1.17)	3.51 (1.10)	3.56 (1.16)
Religion/Spirituality	Low (1–2)	69 (25%)	41 (27%)	14 (23%)	14 (23%)	1.33	4	.856
	High (3–4)	120 (44%)	68 (44%)	25 (41%)	27 (44%)
	Did not discuss (5)	86 (31%)	44 (29%)	22 (36%)	20 (33%)
	Mean/ sd	2.91 (1.44)	2.92 (1.35)	2.84 (1.52)	2.97 (1.57)
Identity	Low (1–2)	52 (19%)	30 (20%)	13 (21%)	9 (15%)	13.14	4	.011
	High (3–4)	99 (36%)	54 (35%)	13 (21%)	32 (52%)
	Did not discuss (5)	124 (45%)	69 (45%)	35 (57%)	20 (33%)
	Mean/ sd	2.52 (1.46)	2.51 (1.46)	2.08 (1.32)	2.98 (1.49)
Ethnicity	Low (1–2)	14 (5%)	5 (3%)	5 (8%)	4 (7%)	14.08	4	.007
	High (3–4)	18 (7%)	4 (3%)	9 (15%)	5 (8%)
	Did not discuss (5)	243 (88%)	144 (94%)	47 (77%)	52 (85%)
	Mean/ sd	1.32 (0.93)	1.15 (0.63)	1.67 (1.30)	1.39 (1.02)
Race	Low (1–2)	9 (3%)	8 (5%)	0 (0%)	1 (2%)	5.71	4	.222
	High (3–4)	4 (1%)	3 (2%)	0 (0%)	1 (2%)
	Did not discuss (5)	262 (95%)	142 (93%)	61 (100%)	59 (97%)
	Mean/ sd	1.11 (0.49)	1.16 (0.60)	1.00 (0.0)	1.07 (0.40)

Table 3.b.

Distribution of how children discussed each construct.

		Overall	African-American (n=153)	Hispanic (n=61)	non-Hispanic White (n=61)	χ 2	df	P
Construct	Score	n (%)	n (%)	n (%)	n (%)
Family Culture	Low (1–2)	146 (53)	84 (55)	29 (48)	33 (54)	4.83	4	.305
	High (3–4)	88 (32)	42 (27)	25 (41)	21 (34)
	Did not discuss (5)	41 (15)	27 (18)	7 (11)	7 (11)
	Mean/ sd	2.99 (1.07)	2.88 (1.10)	3.18 (0.99)	3.10 (1.04)
Religion/Spirituality	Low (1–2)	109 (40)	67 (44)	21 (34)	21 (34)	2.98	4	.561
	High (3–4)	77 (28)	40 (26)	17 (28)	20 (33)
	Did not discuss (5)	89 (32)	46 (30)	23 (38)	20 (33)
	Mean/ sd	2.60 (1.30)	2.59 (1.23)	2.52 (1.39)	2.67 (1.40)
Identity	Low (1–2)	98 (36)	58 (38)	17 (28)	23 (38)	9.79	4	.044
	High (3–4)	53 (19)	26 (17)	9 (15)	18 (30)
	Did not discuss (5)	124 (45)	69 (45)	35 (57)	20 (33)
	Mean/ sd	2.26 (1.27)	2.22 (1.24)	1.92 (1.20)	2.70 (1.32)
Ethnicity	Low (1–2)	21 (8)	6 (4)	10 (16)	5 (8)	14.63	4	.006
	High (3–4)	9 (3)	2 (1)	3 (5)	4 (7)
	Did not discuss (5)	245 (89)	145 (95)	48 (79)	52 (85)
	Mean/ sd	1.25 (0.75)	1.11 (0.49)	1.46 (0.96)	1.38 (0.95)
Race	Low (1–2)	8 (3)	8 (5)	0 (0)	0 (0)	7.52	4	.111
	High (3–4)	3 (1)	2 (1)	0 (0)	1 (2)
	Did not discuss (5)	264 (96)	143 (93)	61 (100)	60 (98)
	Mean/ sd	1.08 (0.41)	1.12 (0.49)	1.00 (0.00)	1.05 (0.38)

Family Culture.

There were no statistically significant differences in how mothers and children from different races and ethnicities talked about their family’s culture (χ²(4) = 4.78 p=.31 for mothers, and χ²(4) = 4.83 p=.31 for children). However, it was observed that when mothers talked about this construct, they promoted cultural engagement and shared their knowledge about their family’s culture in positive terms, while children messages, were less positive when compared with their mothers.

Religion/ Spirituality.

Tables S3a and S3b demonstrate how conversations introduced different activities, social events, traditions, customs, or practices in the context of affiliation with religious or spiritual groups. No statistically significant differences were found in how mothers and children from different races and ethnicities discussed religion and spirituality (χ²(4) =1.33, p =0.86, for mothers, and χ²(4) =2.98, p =0.56 for children). Overall, of the mothers who discussed religion or spirituality with their children, 44% (120/275) promoted or discussed religion in positive or high terms, while 25% (69/275) discouraged religious or spiritual engagement or talked about it in an ambivalent or less positive manner. In contrast, children’s endorsement of religious messages was low, or more ambivalent. Forty percent of children (109/275) talked about culture in lower or less positive terms, compared with 28% (77/275) who talked about it in high terms, or a more supportive manner.

Identity.

The main difference in how mothers and children communicated was found in their exchanges about identity. This construct included elements from their knowledge of traditions, customs and history, as well as participation in activities, social events, traditions, customs, or practices related to their identity group. This construct was mostly endorsed by non-Hispanic White mothers, of whom 52% (32/61) spoke about it in high or positive terms, and 15% (9/61) in lower, or less supportive manner. Notably, endorsement observed among non-Hispanic White children was not as prevalent as the endorsement found in mothers. Of the 61 non-Hispanic White children in our sample, 38% (23/61) discussed identity in low terms, and 30% (18/61) in high terms. Among both mothers and children, the way identity was talked about, was significantly different when compared by race or ethnicity (χ²(4) =13.14, p =0.011 for mothers, and χ²(4) =9.79, p =0.044 for children).

Ethnicity.

Overall, few mothers and children talked about their ethnicity. Among mothers, of the 12% (32/275) who talked about it, most (7%, 18/275) did so in high terms. In contrast, among children, of the 11% (30/275) whom talked about ethnicity, most (8%, 21/275) did so in low terms.

There were significant differences in how mothers and children spoke about ethnicity (χ²(4) =14.08, p =0.007 for mothers, and χ²(4) =14.63, p =0.006 for children), with this construct being spoken most often by Hispanic mothers and children.

Race.

Overall, race was the least discussed construct. It was spoken by 4% of mothers (13/275), and children (11/275). Although there were no statistically significant differences in how race was talked about among mothers or children, it was observed that Hispanics did not talk about race at all, and the few mentions were mostly among African American mothers and children.

DISCUSSION

The primary purpose of this study was to conduct a deeper examination of the construct endorsement and message content of parent-child cultural socialization processes during early school age. Using the OMERS-Peds, a novel direct observation paradigm which examined both parent and child contributions to the cultural socialization process in early school age, we found that of the five cultural constructs examined, family culture, and religion or spirituality, were the focus of the conversation amongst most families, independent of their race or ethnicity. Differences across the different racial/ethnic backgrounds were mainly found in the endorsement mothers and children gave to the construct of identity. While messages of ethnicity and race also significantly differed across families’ racial/ethnic backgrounds, contrary to our expectation, the differences were mainly associated with the limited focus placed on this construct. It was also observed that the differences in messages of ethnicity and race documented across all racial/ethnic child-mother dyads, resulted mostly, from the different emphasis African American and Hispanic mothers placed on both constructs. In depth analysis of the content of messages showed that in the three racial/ethnic groups included, the quality of the messages exchanged was different when comparing endorsements between mothers and children. Mothers talked mostly in a positive fashion about their culture, whereas children messages were scored slightly lower and considered less supportive.

This in-depth characterization of how school age children are socialized was accomplished by introducing the OMERS-Peds, a new observational methodology designed to systematically compare explicit socialization communications within, and across families from three diverse racial/ethnic backgrounds: African American, non-Hispanic White, and Hispanic. Existing studies have largely centered on the socialization of race and ethnicity (Yasui & Dishion, 2007; Priest et al., 2014). The OMERS-Peds advances the literature by allowing us to conduct a deeper analysis of the multidimensional nature of cultural socialization –one that it is inclusive of the transmission of identity, family specific practices and values, and religion. Using the OMERS-Peds this study identified few differences, mostly in the attention paid to the less spoken constructs. Cultural socialization practices among minority families are believed to diverge from those of the non-Hispanic White or dominant culture (Yasui & Dishion, 2007). However, contrary to what the literature from studies with older children has suggested (Yasui, et al., 2015), using this novel methodology we documented mostly similarities between the culturally socialization practices of racially and ethnically diverse families during early school ages.

Previous studies have suggested that ethnic minority families share similar ecological systems, derived from their historical, economical, and political disadvantaged positions (Harrison et al., 1990). A history of discrimination and limited access to societal resources led them to develop similar adaptive observable social behaviors, which became embedded into their culture (Harrison et al., 1990; Loyd & Gaither, 2018). Although it can be presumed that African-American and Hispanic families in our study shared minority status and social ecologies, it is important to observe that families differed significantly in mother’s age, educational attainment, marital status, and poverty status. Notably, these differences in demographic characteristics did not translate into differences in cultural socialization messages. Recent research suggests that for African Americans and Hispanics, gains in socioeconomic status may result in more exposure to discrimination and unfair treatment, as opposed to less (Colen et al., 2018). Thus, despite the diverse demographic characteristics of our sample, and the inclusion of both poor and non-poor African American and Hispanic families, it is likely that social ecologies did not change for the upwardly mobile minority families. The lack of differences in social ecologies may explain the similarities in cultural socialization processes. Still, while the similarities in social ecologies may explain the similarities in the cultural socialization processes between Hispanic and African families, similarities seen with non-Hispanic White families are still unclear.

We speculated that the similarities in mothers and children endorsement to messages about family culture from a broad perspective, and religion, is related to children’s age and parental perceptions of developmentally-appropriate conversations. In a previous study, Suizzo et al., (2008) found that African American mothers believed it was more developmentally appropriate to focus on cultural socialization messages with younger children and introduce other types of ethnic-racial socialization messages when children are older (Suizzo, Robinson, & Pahlke, 2008). This seems to suggest that while cultural socialization processes in early school ages have many common features, differences likely increase with children’s age. Future longitudinal studies tracing changes over time in these processes will be important for elucidating the developmental unfolding of these important culturally salient conversations.

In a review of predictors of racial and ethnic parental socialization practices, authors observed that parental perceptions of children’s cognitive maturity were associated with parents’ likelihood to discuss race, ethnicity, and discrimination with their children (Hughes, 2006). Mean age for children in all racial/ethnic groups in our sample was 7-years old. At these early ages, it is possible that the low frequency of ethnic and racial messages is explained by mother’s assessment of their children’s cognitive developmental maturity. Although it was not directly examined, it may be postulated that mothers’ tendency no to discuss racial and ethnic issues with their children suggests they perceived the children were not developmentally prepared for these conversations. A detailed analysis of mother’s motivations to talk about, or avoid discussing racial and ethnic issues with their children is beyond the scope of this study, but it merits further investigation.

It has been documented that among adolescents, the experience of discrimination enhance children’s response to racial socialization (Yasui & Dishion, 2007). In contrast, among young children the awareness of being treated differently due to race may not be as well developed and may be less frequent. Correspondingly, we speculate that in our study, parents and children were able to focus their cultural socialization conversations around constructs they prioritized instead of having to discuss race and ethnicity in response to potential experiences with discrimination.

Contrary to our expectations, differences were mainly identified within families rather than across racial ethnic groups. It is believed that mothers are the primary, and most important source of cultural socialization for children (Maccoby, 1992). Thus, we expected few differences would emerge when comparing the quality of the messages expressed by mothers with the messages from their children. Finding that the quality of messages differed between mothers and children at this early age supports the need to include the children’s perspective in the assessments of cultural socialization, and the relevance of having a methodology that allows researchers to examine parent-child interactions in a systematic manner.

Further examinations of these differences revealed that mothers and children tended to talk about the different cultural constructs in a neutral manner. In a scale of 1 to 4, where 1 represents a low, or negative endorsement, and a 4 is a high or positive endorsement, it was observed that mothers scored moderately high (e.g. mostly 3’s), whereas children scored moderately low (e.g. mostly 2’s). Another possible explanation to the differences in the quality of the messages within families, may be attributable to children’s age and developmental differences. It is possible that older children may have shown a deeper appreciation or at least a higher level of engagement during the cultural socialization task. While the OMERS-Peds adaptation developed different coding schemes that accounted for age so that an 8-year-old that scored a 4 could theoretically be similar to a 5-year-old that also scored a 4, developmental differences may still be salient.

To our knowledge, this is the first study to implement an observational methodology to examine the cultural socialization practices of families from different racial and ethnic groups in early childhood. Previous studies that used direct observational measurements to examine the explicit socialization communications in depth have targeted older children (e.g. OMERS) (Yasui, 2008). We identified two previous studies that used observational assessments to examine cultural socialization experiences of preschool age children. Yet, these previous observational assessments focused primarily on the parents’ socialization strategy, and on the home as the environment where cultural socialization takes place (Caughy, Randolph, & O’Campo, 2002; Pahlke, Bigler, & Suizzo, 2012). In addition, both of these previous studies limited their analysis to one racial ethnic group, which precluded the assessment of cross-cultural comparisons (Caughy et al., 2002; Pahlke et al., 2012).

The OMERS-Peds introduces an opportunity for researchers to learn about what mothers and children are talking about, and how they are discussing different constructs in their specific cultural socialization interactions. Previous research examining what mothers talked about and constructs endorsed has mostly relied on self-reported instruments administered to mothers only (Yasui, 2008; Yasui & Dishion, 2007). While these previous studies are important, the present study expands the literature by documenting differences within families, which can only be elucidated with a methodology that includes children’s insights. The OMERS-Peds documented the constructs endorsed not only by mothers but also by children and enable us to examine how these compared across different racial and ethnic groups. This novel methodology provided evidence of similarities in the cultural socialization conversations held by families from different racial and ethnic backgrounds, which were otherwise presumed to primarily diverge (Yasui & Dishion, 2007).

While the OMERS-Peds provides important insights, it is important to acknowledge the limitations of this study, and how these affect the generalizability of our findings. First, the conversations analyzed were obtained in a laboratory setting. Actual cultural socialization conversations between mothers and children may place emphasis in different constructs, may be discussed in a different way, or may not have happened at all. While the OMERS-Peds provides the opportunity to expand our approach to research on cultural socialization, there is still a need to identify novel methodologies that can facilitate the study of real-life cultural socialization processes among diverse racial/ethnic groups during key developmental stages and the role of race/ethnicity in culturally-framed understanding of behavior (Shaunfield, et al., (under review)). An adaptation of the OMERS-Peds to be implemented in the home-, school, or even social-media context could potentially satisfy this methodological need. Second, it is important to acknowledge that measurements were collected only once. The extent to which these processes are stable over time It has been suggested that cultural socialization changes over time with respect to both the parents and children (Loyd & Gaither, 2018; Priest et al., 2014). Future studies should consider using repeated observational measurements, and longitudinal data to examine the stability and the variations in the constructs discussed, and changes in the quality of the messages exchanged during explicit cultural socialization communications between mothers and their children. Third, although we employed our extensive developmental adaptation knowledge to modify an observational methodology previously validated to examine cultural socialization processes among racially and ethnically diverse families and young children, its predictive utility is not yet known for children’s subsequent racial identify, self-esteem and functioning. Fourth, this study was conducted within an ethnically and socioeconomically diverse sample of African American, and Hispanic urban families. Future research is needed to examine the cultural socialization processes among racially and ethnically diverse rural, and multi-racial families during children’s early school age. Fifth, findings about the frequency and relevance of messages shared or omitted were drawn from examining explicit cultural socialization practices and do not take into account the corresponding implicit messages. As such, our findings and conclusions should be interpreted in this contexts, as implicit messages are highly relevant during children’s early school ages (Rogoff et al., 2007).

Lastly, we acknowledge that currently, non-parental cultural socialization sources, such as peers, mass-media, and other interactions, may also exert a strong influence on young children’s cultural socialization. In fact, these influences may have led children to diverge from their mothers’ views. Further examination in non-parental explicit and implicit influences of cultural socialization are warranted.

CONCLUSION

This study documented important similarities in the cultural socialization communications across diverse racial/ethnic families, and differences when comparing the content of the cultural messages from early school age children, with those from their mothers. Findings were elucidated using an observational approach, the OMERS-Peds introduced to systematically examine the content of explicit cultural socialization messages exchanged between mothers and children. Namely, insights revealed what cultural messages are being transmitted, and how cultural information is passed down from caregivers to offspring during early school age. Considering the importance of culture to physical and mental health, and the increasingly diverse landscape, we encourage researchers to continue exploring cultural socialization processes during key developmental ages among minority and majority families. Future efforts are needed to explore how to translate research about explicit cultural messages exchanged between children and mothers into programs that promote positive multicultural relationships. Focusing on cultural socialization during early childhood may be an important approach to promote positive multicultural relationships, limit the harmful effects of racism and prejudice, and narrow disparities in physical and mental health among racial and ethnic minority families.