Cultural variations in maternal cognitive demand elicitation and child shyness during emergent literacy among Mexican American and Chinese American immigrant families

Shared book reading is a culturally situated activity where maternal scaffolding may be adapted to a child's temperament. This study is grounded in Sigel's Psychological Distancing Model and Gutiérrez and Rogoff's framework of repertoires of practice to examine the relations among maternal elicitation, characterized by cognitive demand, child narrative production, and child shyness in low-income immigrant families. Participants were 147 mother-child dyads, split between Mexican American (n = 64) and Chinese American (n = 84) families, recruited from preschools in Northern California, USA. The parent-child interaction was captured during a 5-min audio-recorded shared reading session in their home language (Spanish or Cantonese). Every maternal utterance was coded into four levels of cognitive demand (yes/no, referential, behavioral, or inferential). Results showed that although Chinese American mothers utilized more questions across all cognitive demand levels, the proportional distribution of these levels and their positive associations with child utterances were largely consistent across both groups. Chinese American children produced more total utterances, but no group differences were found in mean length of utterance or total tokens, suggesting these differences were interactional rather than structural. A variable-centered analytic approach utilizing a two-way Multivariate Analysis of Variance and partial correlations showed that shy children navigate distinct communicative environments. Specifically, maternal inferential questions were negatively associated with child shyness among Mexican American dyads after controlling for demographic covariates, whereas maternal yes/no questions were positively associated with child shyness among Chinese American dyads. These findings underscore how universal cognitive scaffolding is functionally adapted to child temperament within specific cultural repertoires of practice and socialization goals.

Shared book reading is recognized as a crucial activity for supporting children's emergent literacy and cognitive development (e.g., Deckner et al., 2006; Lenhart et al., 2022; Shahaeian et al., 2018; Sim et al., 2014). In addition to the quantity of talk, the quality of maternal interaction, specifically the use of cognitively demanding language, serves as a robust predictor of children's linguistic outcomes (Bojczyk et al., 2016; Luo and Tamis-LeMonda, 2017). Applying Sigel's (2002) Psychological Distancing Model, maternal elicitation can range from low-demand labeling to high-demand inferential questioning, which scaffolds a child's ability to move beyond the immediate reality toward abstract and representational thinking (Walsh et al., 2016). Additionally, these parent-child conversations during shared book reading serve as culturally embedded family narratives and discourse practices shaped by parents' socialization goals and sociocultural contexts (Escobar et al., 2017; Huang et al., 2026; Luo et al., 2014).

Temperamental shyness, characterized by social wariness and inhibited verbal performance (Kagan et al., 1988; Reynolds and Evans, 2009), poses a unique challenge for a child in social-evaluative tasks such as book reading (Crozier and Perkins, 2002; Reynolds and Evans, 2009). At the same time, shyness is also interpreted differently across cultural contexts through values such as 做人 (zuò rén; becoming a proper person) in Chinese culture and bien educado (being well-mannered) in Hispanic culture, where it may be viewed as a sign of maturity, respect, or modesty (Chen, 2019; Gudiño and Lau, 2010). Despite evidence that shy children may struggle to demonstrate their true linguistic proficiency in social-evaluative tasks, research on maternal scaffolding and child shyness is limited, particularly among diverse low-income immigrant families. The present study addresses this gap by examining how the cognitive demand of maternal elicitation relates to child shyness and narrative production among low-income, home-language-dominant Mexican American and Chinese American dyads, moving beyond the samples from individualistic cultural contexts in the literature.

Cognitive demand in maternal elicitation

The concept of cognitive demand, which categorizes utterances by the mental effort required of the child, is grounded in Sigel's (2002) Psychological Distancing Model. Sigel (2002) posits that environmental stimuli, such as parental questions, require a child to mentally “distance” themselves from the immediate reality, thus fostering their cognitive development. The greater the distance between a concept and a child's immediate context, the more mental effort is required for the child to respond to or engage with that concept (Mazachowsky et al., 2019). This model provides a hierarchy of maternal language input during shared book reading, in which the level of cognitive distancing corresponds to the cognitive demand placed on the child (Sigel, 2002). These conversations stimulate the representational thinking necessary to foster various academic and developmental skills, including language, literacy, executive function, and mathematics (e.g., Maier et al., 2022; Mak et al., 2025; Mauer et al., 2024; Sigel, 2002).

Shared book reading is one of the most consistently supported and powerful contributors to children's emergent literacy, as it facilitates inferential talk, which involves conversation beyond the literal reading of the text (Anderson et al., 2012). A large body of literature has often focused on the quantity of talk (e.g., word count and utterance frequency), but contemporary studies have begun to shift the focus to examining the quality of language input and have shown it to be a more robust predictor of child language outcomes (e.g., Bojczyk et al., 2016; Luo and Tamis-LeMonda, 2017). The cognitive demand of parental talk has been associated with children's language output and vocabulary in the home language, particularly among bilingual children who are acquiring two languages simultaneously (Mak et al., 2025; Walsh et al., 2016).

Applying the Psychological Distancing Model (Sigel, 2002), maternal elicitation can be categorized into four levels, ranging from low to high demand (Luo and Tamis-LeMonda, 2017). The low-demand elicitation involves minimal cognitive distancing from the book's illustrations, such as yes/no questions seeking confirmation or referential questions seeking identification or simple recall (e.g., “What is this?”). These questions provide the foundation for joint attention and literal comprehension and remain relevant to the book content (Atim and Azahar, 2012; Mak et al., 2025; Zucker et al., 2010). At the mid-level of the cognitive demand spectrum, behavioral questions require the child to infer unobserved actions of the story characters (e.g., “What is the boy doing?”). This mid-level talk can be positively associated with children's executive function skills, like cognitive flexibility (Mauer et al., 2024). On the highest end of the cognitive demand spectrum, inferential questions require the child to engage in representational thinking that is beyond the visible content of the book, such as prediction, causal explanation, and character analysis (e.g., “Why is the boy mad?”). These questions scaffold the zone of proximal development and prompt the child to bridge the gap between what is explicitly presented in the book and what is implied by the content or illustrations (McGinty et al., 2012; Van Steensel et al., 2024). It is important to note that mothers may decrease these basic referential questions in favor of more behavioral or inferential questions as children mature, specifically between ages 3 and 5 years (Kuchirko et al., 2016).

Shyness among Chinese American and Mexican American children

Shyness is a well-established construct in developmental science, rooted in temperament theory (e.g., Chen, 2019; Coplan and Rubin, 2010). Temperamental shyness refers to a biologically based disposition to react with fear or wariness when encountering unfamiliar people or situations involving social evaluation, a pattern that tends to remain relatively stable over time (Kagan et al., 1988; Pérez-Edgar and Fox, 2005). While shyness is based in early neurobiology and temperament, culture and socioecological conditions may influence how shyness in children is interpreted and the degree to which shy behaviors are approved, supported, or valued by others.

According to the contextual development perspective (Chen, 2012), cultural values shape the extent to which shyness is viewed as a desirable characteristic in children. Research has documented cultural variations in childhood shyness, including among Chinese American and Mexican or Hispanic American children (e.g., Balkaya et al., 2018; Gudiño and Lau, 2010). Studies of Mainland Chinese mothers in the early twenty first century show that shyness can be understood not only as a reactive and anxiety based form of social inhibition, referred to as anxious shyness, but also as a self-regulated and intentional form of restraint, referred to as regulated shyness, that supports fitting in with others and maintaining social harmony (e.g., Liu et al., 2020; Xu et al., 2007; Xu and Zhang, 2025). However, it is important to distinguish an individual's traits (e.g., temperament) from their repertoire of practice, the culturally patterned ways of participating, learning, and engaging in sociocultural activities that develop through histories of involvement in family routines, community roles, and schooling practices (e.g., Gutiérrez and Rogoff, 2003). This distinction helps avoid mislabeling individuals' culturally shaped participation histories as inherent temperament traits.

Within traditional Chinese culture, parents who endorse traditional Chinese cultural values tend to guide their children toward 做人 (zuò rén), which is an indigenous concept referring to becoming a proper person who embodies moral character and respect for others, especially elders. These qualities can manifest as deference to elders, restraint, humility, and modesty that could easily be interpreted by observers as shyness, which may help explain why studies from the late 1990s to early 2000s found positive associations between shyness and developmental outcomes in Mainland China, particularly among rural migrant children whose families more strongly adhered to traditional cultural values (e.g., Chen et al., 2009). Although Confucian and collectivistic traditions historically regarded shy children as mature, well-behaved, and socially competent leaders, recent meta-analytic evidence indicates a cultural shift in which shyness among contemporary Chinese children is increasingly associated with negative social-emotional and developmental outcomes (Zhang et al., 2021).

Like traditional Chinese cultural values, Hispanic cultural values are often considered collectivistic. Parents who endorse traditional Hispanic cultural values tend to guide their children toward bien educado, which is an indigenous concept translated as being “well-educated” or well-mannered and embodying proper courtesy, politeness, and respect toward others, especially elders or authority figures. Similar to 做人 (zuò rén), these qualities of bien educado can manifest as deference to elders, restraint, humility, and modesty that could easily be interpreted by observers as shyness. In their study of Hispanic American children's shyness, Gudiño and Lau (2010) reported that children's shyness varies with the extent to which parents have adopted United States values, attitudes, and behaviors, such that parents with more collectivistic orientations aligned with traditional Hispanic cultural values more strongly endorsed shy child behavior. Overall, while bien educado primarily reflects expectations for respectful and prosocial behavior, some of its associated values may manifest in ways that resemble shyness in certain social or educational contexts.

Shared book reading among Mexican American and Chinese American families

Shared book reading is a culturally situated activity in which the manner and style of how parents engage with their child vary across cultures (e.g., Caspe, 2009; Leyva et al., 2021; Luo and Tamis-LeMonda, 2017). Research with ethnically diverse and low-income families in the United States has identified distinct patterns in the shared reading experiences of Mexican American and Chinese American dyads, which reflect different cultural values and socialization goals within these communities (e.g., Escobar et al., 2017; Luo et al., 2014). Mexican American mothers often prioritize narrative coherence and the expansion of the story's descriptive and emotional aspects (Escobar et al., 2017; Luo et al., 2014; Mak et al., 2025), while Chinese American mothers emphasize linguistic precision, factual knowledge, and didactic instruction (Huang et al., 2026; Luo et al., 2014). These differences are more apparent in how mothers choose to expand upon the story content, with Mexican American mothers typically referring to more story components, including story setting, character traits, and internal states, and Chinese American mothers focusing on labeling illustrations, describing actions visible in the book, and reinforcing and teaching vocabulary (Huang et al., 2026; Luo et al., 2014; Mak et al., 2025).

The structure of a shared reading task plays a significant role in modulating these interactions and the child's resulting narrative production, specifically through the use of wordless picture books. These wordless picture books lack predetermined texts and thus create a negotiation setting in which dyads co-construct the narrative (Devescovi and Baumgartner, 1993). Studies have found that using wordless picture books significantly increases children's spontaneous language production by prompting mothers to provide higher levels of instructional support and elicitation (Chaparro-Moreno et al., 2017; Gámez et al., 2017).

The level of cognitive demand placed on the child during shared book reading also follows cultural norms. In Mexican American families, the mother-child conversations are usually characterized by embellished elaborations, in which mothers provide evaluative or descriptive information beyond the literal text (Escobar et al., 2017). Studies suggest that these specific elaborations are highly functional and that the embellished elaborations used by Latino mothers positively predict children's language and storytelling skills 6 months later (Schick et al., 2017). Similarly, in a more recent study comparing low-income immigrant families, Mak et al. (2025) found that Mexican American mothers utilized more high-demand inferential statements, such as explaining a character's motives, which exposed their children to high levels of narrative complexity. This story-building style is further supported by the reciprocity in responding to their child's cognitive demands, where Mexican American mothers are more likely to match their child's cognitive demand contribution level, thus maintaining a balanced dialogic interaction (Luo and Tamis-LeMonda, 2017).

In contrast, the shared book reading style in Chinese American families has traditionally been described as more didactic and less dialogic, with mothers providing less storyline information and asking fewer open-ended questions than their Mexican American or African American counterparts (Luo et al., 2014). However, recent literature suggests that this style is related to a specific cultural preference for instruction rather than a lack of engagement. For instance, Huang et al. (2026) found that Chinese mothers may produce more overall extra-textual talk than Western mothers, even if the talk is directed toward labeling or moral lessons. Furthermore, these patterns are not immutable and are highly responsive to environmental changes. An intervention study shows that a 30-min training session on dialogic reading strategies significantly shifted maternal behavior in Chinese dyads and increased their frequency of open-ended questions and child-initiated talk (Zhao et al., 2021). Overall, both Mexican and Chinese mothers utilize the shared reading context to scaffold development, but they may do so through different linguistic practices, with one favoring co-building the story arc and the other prioritizing factual and instructional knowledge.

Although cultural values like 做人 (zuò rén) and bien educado profoundly shape maternal socialization goals (Chen, 2019; Gudiño and Lau, 2010; Valdés, 1996), operationalizing these broad, multifaceted socioemotional orientations during shared book reading is complex. The present study focuses on the structural form of maternal elicitation through the well-established lens of cognitive demand as an observable proxy (Luo and Tamis-LeMonda, 2017; Sigel, 2002). Importantly, while the highest cognitive demand level in our coding scheme, inferential questions, captures discussions of internal states and social knowledge, it also aggregates these socioemotional prompts with other complex cognitive leaps, such as plot prediction. Therefore, rather than claiming to isolate the specific thematic content of 做人 (zuò rén) and bien educado directly, we use cognitive demand to objectively assess the proximal scaffolding styles that these cultural orientations foster. For instance, a didactic focus on factual recall (lower cognitive demand) aligns with the aspects of pedagogical authority and deference in 做人 (zuò rén), whereas co-constructing narratives about characters' intentions (higher cognitive demand) facilitates the interpersonal awareness central to bien educado. By focusing on these structural linguistic mechanisms, this study addresses a critical gap in measuring how culturally patterned instructional and scaffolding styles operate distinctly across groups (Escobar et al., 2017; Luo et al., 2014) and examining how these styles relate to children's temperamental expression, such as shyness, during shared book reading (Chen, 2012; Reynolds and Evans, 2009). While this structural focus precludes claims about the specific culturally-relevant content of the mother-child conversations, it provides a cultural and quantifiable metric of the linguistic environments these Chinese American and Mexican American children navigate.

Shared book reading and child shyness

The cultural interpretations of temperament have significant implications for how children navigate the communicative demands of shared book reading. Shy children may face unique challenges in demonstrating their true linguistic competence during shared book reading, as it is a social and verbal task (Crozier and Perkins, 2002; Reynolds and Evans, 2009). However, the literature examining the bidirectional relationship between maternal scaffolding or elicitation and child shyness during these interactions is limited. Baardstu et al.'s (2023) study suggests that this relationship may be transactional in that early shyness predicts later language difficulties, but having strong early language skills may act as a protective factor against the development of future internalizing problems.

A main concern in the studies examining child shyness is whether shy children lack specific linguistic skills or if their linguistic performance is suppressed by the social and verbal nature of the assessment or interaction (Crozier and Perkins, 2002). Research indicates that shy children tend to speak less, use shorter utterances, and show a longer latency period before responding to questions (Crozier and Perkins, 2002). Additionally, Spere et al. (2004) found that shy preschoolers scored lower on both receptive and expressive language measures than their non-shy counterparts. However, they still performed at their age-equivalent level. These findings suggest that shyness does not necessarily compromise the developmental trajectory of language skills, but rather that shy children may lack the performance boost that non-shy children exhibit. These differences in speech often persist even when vocabulary scores are controlled for, suggesting that the reticence or quietness observed during shared book reading may not reflect the child's true knowledge and proficiency but rather a manifestation of social anxiety (Crozier and Perkins, 2002; Spere and Evans, 2009).

The quality of shared book reading is often determined by how caregivers adjust their language to align with the child's communicative style. However, existing research demonstrates a complex picture regarding how these adjustments manifest in the presence of child temperament. For instance, Laible (2004) found that mothers did not adjust their speech to a child's temperament during shared book reading, as maternal elaborations were not related to child temperamental traits. Similarly, regarding child shyness specifically, Reynolds and Evans (2009) observed that the parents of shy children did not differ from those of non-shy children in the amount of semantic information or scaffolding they provided, despite the shy children contributing less during a shared reading task with a wordless picture book. This suggests that caregivers' language input may remain stable regardless of the child's level of verbal participation during shared book reading. A recent meta-analysis concludes that parents often prioritize maintaining a consistent instructional or narrative flow as they view shared reading as a goal-oriented educational task (Nan and Tian, 2025). Although some studies suggest that a lack of adjustment could lead to missed opportunities for verbal engagement and undermine these children's confidence or language proficiency (Spere et al., 2004), these patterns may instead reflect the parent's focus on providing a consistent linguistic environment to complete the activity.

The present study

The present study builds on the existing literature by examining the relations between maternal elicitation and child shyness during shared book reading among low-income immigrant families from two cultural groups. Extensive research has established that shared book reading is a critical activity for supporting children's language and literacy development (e.g., Deckner et al., 2006; Lenhart et al., 2022; Sim et al., 2014). However, there remains a lack of consensus on how these interactions vary across cultural groups and how they relate to children's temperamental traits, such as shyness (Luo et al., 2014; Reynolds and Evans, 2009), particularly when shyness may be viewed as a desirable characteristic in certain cultures rather than social inhibition (Chen, 2019).

This study contributes to the field by shifting the focus toward low-income, home-language-dominant Mexican American and Chinese American dyads, who share similar socioeconomic status but often hold different cultural socialization goals regarding book reading practices and shyness (Chen, 2019; Leyva et al., 2021). By moving beyond the predominantly individualistic cultural contexts that have historically shaped much of the literature (Evans and Ennis, 2017), this study offers a more nuanced comparison of how culturally patterned practices shape children's linguistic environments. Specifically, we examine the cognitive demand of maternal questions to determine how these styles relate to child shyness across cultural contexts. While our theoretical framework posits that mothers intuitively adapt their shared reading and scaffolding styles to their child's temperament, mother-child interactions are fundamentally bidirectional. Recognizing that the cross-sectional design of the present study cannot establish causality, we address the associations among variables to describe the culturally distinct maternal language environments in which shy children are embedded.

The following research questions and hypotheses guide the present study: Are there significant differences between Mexican American and Chinese American mother-child dyads in the frequency of maternal elicitation at varying levels of cognitive demand, child narrative production, and child shyness? It is hypothesized that Chinese American mothers will utilize more questions across all cognitive demand levels compared to Mexican American mothers, which reflects the cultural emphasis on didactic instruction and academic preparation during reading (Luo et al., 2014; Wang, 2001), and that these differences will also be detected in child narrative production as the hypothesized frequency of Chinese American maternal questions will be positively associated with Chinese American child output (Mak et al., 2025). It is also hypothesized that Chinese American children are expected to have higher reported levels of shyness compared with Mexican American children (Balkaya et al., 2018; Gudiño and Lau, 2010).

What are the bivariate associations between child shyness, the cognitive demand of maternal elicitation, and child narrative production within each cultural group? It is hypothesized that higher child shyness will be correlated with lower narrative production across both groups (Crozier and Perkins, 2002; Reynolds and Evans, 2009) and that shy children may receive different patterns of maternal elicitation, such as more frequent low cognitive demand questions to facilitate verbal engagement (McGinty et al., 2012).

What are the associations between the cognitive demands of maternal elicitation and child shyness in Mexican American and Chinese American families after controlling for relevant demographic covariates? Given the cross-sectional nature of the data, it is hypothesized that specific cognitive demand levels of maternal elicitation will be uniquely associated with child shyness across cultures. For Mexican American dyads, it is hypothesized that high-demand inferential questions will be negatively associated with child shyness, controlling for demographic covariates. This may reflect an environment in which mothers adopt a more supportive, lower-demand narrative flow to accommodate a shy child and fulfill the cultural expectation of being respectful during storytelling (Schick et al., 2017). For Chinese American dyads, it is hypothesized that low-demand yes/no or referential questions will be positively associated with child shyness after controlling for demographic covariates. This may reflect maternal responses to linguistic accuracy and regulated shyness (e.g., Xu and Zhang, 2025), in which mothers use simpler prompts to maintain social harmony and ensure the child provides correct responses (Chen, 2019; Luo and Tamis-LeMonda, 2017).

Participants

This study utilized data from a larger longitudinal project investigating the language and socio-emotional development of children from low-income immigrant families. Participants were 147 low-income immigrant mother-child dyads, including 64 Mexican American dyads and 84 Chinese American dyads. Families were recruited from federally funded Head Start programs and state-subsidized preschools in Northern California. Children in the sample (79 boys, 69 girls) were preschool-aged at the time of the home visit (M age = 48.33 months, SD = 6.43). Their mothers were first-generation immigrants to the United States and identified as the primary caregivers.

In the Mexican American subsample (n = 64, 33 boys, 31 girls), the mean child age was 48.98 months (SD = 6.45), and mothers primarily spoke Spanish during shared reading. In the Chinese American subsample (n = 84, 46 boys, 38 girls), the average child age was 47.68 months (SD = 6.36), and mothers primarily spoke Cantonese. All families met the federal poverty guidelines for Head Start enrollment to ensure a relatively homogeneous socioeconomic background across both cultural groups. The Mann-Whitney U test showed that the groups did not differ significantly in child age or per capita income. However, a significant difference was observed for maternal education (U = 1,973.5, p < 0.05), with Chinese American mothers reporting more years of formal education (M = 12.62, SD = 2.53) than Mexican American mothers (M = 11.31, SD = 2.53). Additionally, a significant difference was observed in mothers' length of residence in the United States (U = 2,574, p < 0.001), with Mexican American mothers reporting a longer average residence (M = 15.8, SD = 6.84) than Chinese American mothers (M = 8.54; SD = 7.26).

Procedure

Data collection took place during home visits, with each lasting 3–4 h. Breaks were provided as needed for the child or to allow caregivers to attend to parenting responsibilities, thereby ensuring participants' comfort. Each visit was staffed by three trained bilingual research assistants. Two research assistants focused on engaging the child in a series of bilingual assessments and activities; however, only data from the mother-child shared book reading task were used in the present study. Concurrently, the third research assistant assisted the mother in completing a comprehensive questionnaire packet that included parent-reported demographic information and child shyness. At the conclusion of the visit, families were compensated for their time, receiving US$100 for one parent's participation or US$150 for both parents' participation.

Measures

Cognitive demand of maternal elicitation

The shared book reading interaction was video- and audio-recorded to capture the mother-child dyadic exchange. A wordless picture book, Frog, Where Are You? (Meyer, 1969), was used to minimize the influence of varying literacy levels and to standardize the task across languages. Mothers were instructed to tell the story to their child in their preferred language, as they would typically do at home. The book reading interactions lasted between 5 and 10 min.

The recordings of the shared book reading sessions were transcribed verbatim following the Codes for Human Analysis of Transcripts (CHAT) conventions from the Child Language Data Exchange System (CHILDES; MacWhinney, 2000) by bilingual research assistants fluent in the language of the interaction (Spanish or Cantonese). Only transcriptions in which the mother used the home language for at least 80% of total words were included in the present study to ensure maternal dominance in the home language (Bedore et al., 2012). Two trained bilingual research assistants (Spanish-English and Cantonese-English) coded each utterance using a coding scheme adapted from Luo and Tamis-LeMonda (2017), which categorizes verbal adult input into four mutually exclusive levels of cognitive demand. At the lower levels, yes/no utterances consisted of simple affirmation or negation (e.g., “Do you see the frog?”). Referential utterances focused on labeling or identifying visible objects in the illustrations (e.g., “Where is the frog?”). Behavioral utterances, which are at the mid-level, focus on inferring unobserved actions of the characters (e.g., “What is the boy doing?). Inferential utterances involve high-level processing, requiring the speaker to go beyond literal images to discuss internal states, social knowledge, predictions, or narrative interpretations (e.g., “Why is the boy angry?”). This category serves as a broad measure of cognitive distancing rather than an exclusive measure of emotional or moral socialization, as it includes both socioemotional inferences and general narrative predictions. For this study, the frequency of maternal questions at each of the four levels was calculated and used in the data analysis.

The two trained research assistants, one fluent in Spanish and English and another fluent in Cantonese and English, independently coded 20% of the translated transcripts to ensure reliability across both cultural samples. Interrater reliability, calculated using Cohen's kappa, was 0.770 for the Mexican American subsample, 0.706 for the Chinese American subsample, and 0.726 for both groups combined. After establishing reliability on the English transcriptions, each coder independently coded the transcriptions in the original language for their respective cultural group. The two coders met regularly to review the English translation, resolve discrepancies, and prevent coding drift between the two languages.

Child narrative production

Child narrative production was measured using the same verbatim transcriptions from the mother-child shared book-reading task described above. Previous studies with bilingual populations have used the mean length of utterance (MLU), the total number of utterances, and the total number of word tokens (e.g., Gorman et al., 2016; Rezzonico et al., 2016). These indices have been validated as significant indicators of language development in both Spanish-speaking English learners (Castilla-Earls et al., 2022; Fiestas and Peña, 2004; Gorman et al., 2016) and Cantonese-speaking children (Rezzonico et al., 2016; Wong et al., 2004). The total number of utterances and total tokens are used to represent children's overall language productivity during narrative tasks (Fiestas and Peña, 2004; Wood et al., 2018), while MLU provides an estimate of children's ability to produce expanded linguistic structures (Rezzonico et al., 2016; Losi et al., 2022).

Child shyness

Child shyness was assessed using the 32-item very short form of the Child Behavior Questionnaire (CBQ-VSF; Rothbart et al., 2001). The translated version has been validated for use across diverse contexts in non-Western populations, including in China (Zentner, 2020) and Spanish-speaking communities (De La Osa et al., 2014), and has been used in cross-cultural comparison studies (Chen et al., 2014). Child shyness was derived from the eight-item shyness subscale, which measures a child's reaction to novel or challenging social situations (e.g., “acts shy around new people”). In this study, mothers were asked to indicate how well each item described their child's behavior over the previous 6 months on a seven-point Likert scale from 1 (Extremely Untrue) to 7 (Extremely True). Two items, “seems to be at ease with almost any person” and “is comfortable asking other children to play,” were reverse-coded. The mean for the eight items was then calculated to create a composite child shyness score. Internal reliability of the shyness subscale, measured by Cronbach's alpha, was calculated separately for each cultural group. Reliability for the Mexican American (α = 0.79) and Chinese American (α = 0.72) samples was satisfactory.

Data analysis

Data were first screened for outliers, normality, linearity, and homoscedasticity. Skewness and kurtosis indices were examined. Descriptive statistics were calculated for all demographic and primary variables. Since all variables were non-normally distributed, Mann-Whitney U tests were conducted to examine differences between the Mexican American and Chinese American groups. Zero-order Kendall's correlations were also computed to examine the associations among variables. Fisher's transformations were calculated to compare the strength of the correlation coefficients between the two cultural groups.

A two-step analytic approach was used to examine the relationships between maternal elicitation and child shyness. First, to determine whether the overarching linguistic environment differed across cultures and according to child shyness, a two-way Multivariate Analysis of Variance (MANOVA) was conducted to test the main effects and interaction of cultural group and child shyness on the four types of maternal cognitive demand elicitation. Second, to isolate the specific cognitive demand elicitation uniquely associated with these environments, partial Kendall's correlations were conducted. Child age and per capita income were included as covariates for both cultural groups. Finally, a sensitivity power analysis was conducted to evaluate the statistical power of the partial correlations. All statistical analyses were conducted using RStudio Version 2026.01.0+392 (Posit Team, 2026).

The descriptive statistics for the demographic variables, key predictors, and outcomes, stratified by cultural group, are presented in Table 1. Skewness and kurtosis indices indicated that several maternal elicitation categories were significantly non-normally distributed, but child shyness fell within acceptable ranges. Specifically, among the Mexican American group, yes/no, behavioral, and inferential questions showed high positive skewness and kurtosis. Linearity was also confirmed by inspecting scatterplots. In terms of maternal elicitation during shared book reading, Chinese American mothers asked more yes/no questions (U = 1,242.5, p < 0.001), referential questions (U = 614.5, p < 0.001), behavioral questions (U = 1,928, p = 0.003), and inferential questions (U = 1,460.5, p < 0.001) than Mexican American mothers. Additionally, Quade's non-parametric ANCOVA was conducted to account for the significant differences observed in maternal education. The results showed that the group differences in all four types of maternal elicitation remained statistically significant even after controlling for maternal education: yes/no questions (F = 38.44, p < 0.001), referential questions (F = 112.92, p < 0.001), behavioral questions (F = 9.46, p = 0.003), and inferential questions (F = 24.12, p < 0.001). In terms of child narrative production, Chinese American children produced significantly more total utterances (Mdn = 20.5; IQR = 9.75–32.2) than Mexican American children (Mdn = 12; IQR = 3–20.2; U = 1,705.5, p < 0.001). No group differences were found for MLU, total tokens, or child shyness. Additionally, gender differences were not detected in any of the primary study variables. Hence, child gender was not included as a covariate in subsequent analyses.

The zero-order correlations presented in Table 2 were conducted separately for the Mexican American and Chinese American subsamples to examine differences in the associations among demographic variables, maternal elicitation, child narrative production, and child shyness. The results revealed distinct patterns across the two groups. In the Chinese American sample, maternal yes/no questions were positively correlated with child shyness (τ = 0.20, p = 0.011). Additionally, maternal behavioral questions were negatively correlated with child age (τ = −0.17, p = 0.031). Total number of utterances was negatively correlated with child shyness (τ = −0.16, p = 0.031).

In contrast, the Mexican American sample showed a different pattern of correlations. Child shyness was correlated with demographic factors rather than maternal elicitation. Specifically, child shyness was negatively correlated with per capita income (τ = −0.25, p = 0.009) and child age (τ = −0.22, p = 0.015). Child age was negatively correlated with the frequency of maternal yes/no questions (τ = −0.32, p < 0.001). Maternal education was positively correlated with the child's total tokens (τ = 0.21, p = 0.023). Despite these distinct group patterns regarding shyness and demographic variables, the correlations between the higher cognitive demand questions and child total utterances were positive and significant in both groups. This shared pattern for Mexican American and Chinese American groups was observed for referential (τ = 0.38 and τ = 0.29, respectively, p < 0.001), behavioral (τ = 0.29 and τ = 0.24, respectively, p = 0.003), and inferential questions (τ = 0.37, p < 0.001 and τ = 0.21, p = 0.006, respectively).

Fisher's transformations were conducted to test whether the magnitude of the highlighted correlations above differed significantly between the two cultural groups. The results showed that the positive correlation between maternal yes/no questions and child total utterances was significantly stronger in the Mexican American sample than the Chinese American sample (z = 2.11, p = 0.035). Furthermore, no significant differences were detected between groups for the negative correlation between child age and maternal yes/no questions (z = −1.66, p = 0.097) and the positive correlation between maternal education and child total tokens (z = 1.67, p = 0.095). No differences were observed in the correlations between per capita income and total number of utterances, and between child shyness and total number of utterances.

A two-way MANOVA was conducted to test whether the relationship between child shyness and maternal elicitation differed by cultural group. The dependent variables were the four categories of maternal cognitive demand elicitation, and the independent variables were cultural group, child shyness, and their interaction. The results showed significant multivariate main effects of cultural group (Pillai's Trace = 0.437, p < 0.001) and child shyness (Pillai's Trace = 0.102, p = 0.004). However, the multivariate interaction between cultural group and shyness was not significant [Pillai's Trace = 0.021, F (4, 141) = 0.74, p = 0.564]. Although this omnibus multivariate interaction was not significant, the MANOVA model does not accommodate demographic covariates identified in our preliminary correlations. Therefore, we proceeded with the planned partial correlations to test our a priori hypotheses, applying appropriate contextual covariates and addressing heteroscedasticity.

For the Mexican American group, partial Kendall's correlation results showed that maternal inferential questions were negatively associated with child shyness (partial τ = −0.19, p = 0.046), after controlling for child age and per capita income. No other cognitive demand questions yielded significant partial correlations. For the Chinese American group, partial correlations showed that maternal yes/no questions were positively associated with child shyness (partial τ = 0.16, p = 0.041), after controlling for the same demographic covariates. All other cognitive demand questions were non-significant.

Sensitivity power analyses were conducted to determine the minimum detectable effect size for the partial correlation analyses. The results showed that the Mexican American sample was adequately powered to detect a partial correlation effect size of r = 0.34 or larger, while the Chinese American sample was adequately powered to detect a partial correlation effect size of r = 0.30 or larger. These results indicated that the sample sizes were sufficient to detect moderate to large effects of maternal elicitation and child shyness after accounting for demographic covariates. The current findings should be interpreted with caution as the observed partial correlations were of a lower magnitude than the minimum detectable thresholds.

Findings from this study offer one of the most nuanced examinations to date of how cultural practices, maternal elicitation styles, and child temperament intersect within low-income, home-language-dominant Mexican American and Chinese American immigrant families. Findings also highlight the importance of considering families' repertoires of practice and the ways in which maternal scaffolding may function differently across cultural ecologies.

To illustrate these patterns in greater depth, we begin by examining cultural variations in maternal elicitation styles and children's narrative production during shared book reading. The study results supported the hypothesis and indicated significant differences between Mexican American and Chinese American dyads. Chinese American mothers in the present study utilized significantly more questions across all levels of cognitive demand, yes/no, referential, behavioral, and inferential, compared to Mexican American mothers. This finding aligns with previous literature suggesting that Chinese American parents often adopt a more didactic and instructional approach during shared book reading (Huang et al., 2026; Luo et al., 2014). For these first-generation immigrant mothers who remain dominant in their home language, the high volume of maternal elicitation likely reflects the Chinese socialization goal of 做人 (zuò rén), which emphasizes the cultivation of moral character, linguistic precision, and academic readiness through parent-led instruction (Chen, 2019; Wang, 2001). The findings from the present study support the contemporary views that Chinese maternal elicitation, particularly within low-income immigrant families maintaining traditional heritage practices, represents a high level of engagement geared toward factual and instructional knowledge transmission (Huang et al., 2026), in addition to being labeled as less dialogic in earlier studies (Luo et al., 2014).

Mexican American mothers in the present study, on the other hand, asked fewer questions overall. This does not necessarily indicate lower engagement but rather a different narrative style. Mexican American mothers, who are dominant in their home language, often prioritize narrative coherence and embellished elaborations (Escobar et al., 2017; Schick et al., 2017). Therefore, these mothers may be providing more space for co-constructing a storyline than testing for specific vocabulary or actions seen in the book. This practice aligns with the values of bien educado, where families prioritize social harmony, respect, and holistic storytelling over direct instruction (Valdés, 1996).

Although there are differences in the frequency count of maternal elicitation, the proportional distribution of these elicitation patterns was consistent across both cultural groups. Both Mexican American and Chinese American mothers asked referential questions most frequently, followed by yes/no, behavioral, and lastly inferential questions. This suggests that, even though elicitation frequency varies by culture, the underlying structure of scaffolding and elicitation during shared book reading may be a shared feature across cultures. This finding is consistent with a universal pedagogical progression from literal identification of story characters and objects toward inferential discussion (Sigel, 2002; Zucker et al., 2010).

Furthermore, Chinese American children produced significantly more total utterances than their Mexican American counterparts. This directly supports the hypothesis and suggests that the high elicitation frequency among Chinese American mothers is closely related to more verbal utterances from the child. However, it is important to note that no significant group differences were observed in total tokens or MLU. This pattern also suggests that the differences observed in child narrative production are mainly interactional rather than structural. The higher utterance count in Chinese American children likely reflects the higher frequency of turn-taking storytelling style by their mothers, consistent with the didactic instructional goal of 做人 (zuò rén). In this context, children may provide more frequent but shorter responses to maintain the pace of maternal elicitation (Huang et al., 2026; Luo et al., 2014).

The zero-order correlations revealed that the associations between higher cognitive demand elicitation, specifically referential, behavioral, and inferential questions, and child total utterances were positive and largely similar across both the Mexican American and Chinese American samples. This indicates a universal mechanism in which higher levels of cognitive demand questions are related to child narrative production regardless of cultural background. However, the comparison of correlations showed that the positive association between maternal yes/no questions and child total utterances was significantly stronger in the Mexican American sample. This suggests that while the function of cognitive demanding questions is shared across groups, these closed-ended questions may function as an effective prompt to encourage Mexican American children's verbal participation, which aligns with previous findings that Mexican American mothers actively match and co-construct narratives based on their children's engagement (Luo and Tamis-LeMonda, 2017). On the other hand, since yes/no questions were not significantly correlated with Chinese American children's utterances, their high utterance rate may be driven by the overall volume of maternal elicitation across all cognitive demand levels, which reflects a cultural prioritization of instruction flow (Nan and Tian, 2025). The comparable total tokens and MLU across groups suggest that Mexican American children may make more descriptive contributions when speaking, even though they contribute fewer turns than Chinese American children in the shared reading task.

The second aim of the study was to examine associations between maternal elicitation and child shyness within each cultural group. Notably, we did not find group differences in children's shyness, a pattern that diverges from prior research documenting cultural variations in childhood shyness among Chinese American and Mexican or Hispanic American children (e.g., Balkaya et al., 2018; Gudiño and Lau, 2010). Several important distinctions may account for this inconsistency. Child shyness in our study was assessed via parent report only, without supplementary teacher report or observer ratings of shyness. However, this approach is largely consistent with the cited works in Balkaya et al. (2018) and Gudiño and Lau (2010). Therefore, the divergence in our findings is more likely driven by demographic and sociocultural differences. For instance, Balkaya et al. (2018) examined Chinese American families in which all mothers had at least a college education, and Gudiño and Lau (2010) included Hispanic American parents who were both first-generation immigrants and U.S. born. In contrast, our sample consisted exclusively of low-income, first-generation immigrant parents. These demographic and sociocultural differences shape children's everyday participation in cultural routines and expectations, which in turn influence how behaviors such as shyness are expressed and interpreted. As (Gutiérrez and Rogoff 2003) emphasize, it is essential to distinguish an individual's temperament traits from their repertoire of practice; otherwise, researchers risk mislabeling children's histories of participation in particular sociocultural activities as temperamental characteristics. From this perspective, the absence of group differences in our study may reflect the shared sociocultural and socioeconomic conditions of our sample.

The subsequent aims of this study examined the unique associations between maternal elicitation, child narrative production, and child shyness. The first step was to conduct a MANOVA to confirm that the overarching profiles of maternal elicitation differed significantly by cultural group and child shyness. The significant main effect of shyness demonstrates that it is associated with a distinct overall profile of maternal elicitation styles across both cultural contexts. This finding extends prior literature reporting no parental language adjustments in response to child shyness (e.g., Laible, 2004; Reynolds and Evans, 2009) and confirms that shy children in these immigrant samples navigate different communicative environments than their non-shy peers. However, the multivariate interaction between cultural groups and child shyness was not significant. This suggests that the effect of shyness on these elicitation profiles may share structural similarities across the two cultures.

The next step was to unpack these broad findings while avoiding the imposition of a directional framework on cross-sectional data, so we used a variable-centered approach based on non-parametric partial correlations. This approach allowed us to identify the specific maternal elicitation types that uniquely characterized the communicative environments of shy children while accounting for relevant demographic covariates.

The partial correlation analyses revealed diverging patterns within each cultural group. For Mexican American dyads, maternal high-demand inferential elicitation, such as questions involving character motives, emotions, or predictions, was negatively associated with child shyness, after controlling for demographic variables. From the perspective of bien educado, Mexican American mothers often prioritize narrative flow and emotional warmth (Escobar et al., 2017; Schick et al., 2017). This negative association suggests that shy children are embedded in environments where mothers may use fewer high-demand inferential questions to maintain a supportive and stress-free interaction. Pushing a shy child with too many complex questions might cause them to become uncomfortable and disengage from the activity, as shared book reading can act as a social-evaluative task (Schick et al., 2017). Hence, adapting the interaction with shy children by using fewer high-demand inferential questions may allow these mothers to co-construct the narrative while respecting the child's temperament.

In contrast, the positive partial correlation observed between low-demand yes/no questions and child shyness among Chinese American dyads likely reflects the Chinese cultural strategy of didactic instruction and linguistic precision (Huang et al., 2026; Wang, 2001). If a child exhibits temperamental shyness, Chinese American mothers may utilize simpler, closed-ended questions to maintain social harmony and ensure their shy child can participate easily without the stress of open-ended prompts (Chen, 2019; Liu et al., 2020). This aligns with the concept of regulated shyness (Xu et al., 2007; Xu and Zhang, 2025), where a mother adapts her linguistic demands to match her child's comfort level. Although the shared book reading task in the present study did not require child speech, the high overall frequency of elicitation by Chinese American mothers may lead to more intense interactions. They may have used more yes/no questions to ensure their shy children could participate successfully to fulfill the instructional goals of a shared reading session (Luo and Tamis-LeMonda, 2017).

Limitations and future directions

While the present study provides important insights into the cultural and temperamental dynamics of shared book reading in low-income immigrant families, several limitations should be considered when interpreting the findings. First, this study did not directly measure cultural values such as 做人 (zuò rén) or bien educado, as there are no widely validated measures specifically designed to quantify these multifaceted cultural values. Our coding for high-demand inferential questions captures utterances related to internal states, social knowledge, and cognitive leaps, such as plot prediction and making inferences related to real life. Thus, our coding scheme cannot definitively disentangle whether a high frequency of inferential questions was driven by emotional socialization, behavioral lessons, or general narrative complexity. Future research should focus on developing and validating culturally specific measures that capture the socialization goals and practices of immigrant families and incorporate a coding scheme that assesses both the structural cognitive demand and the specific thematic content of the conversations, rather than relying on broad measures such as immigration status and language dominance. Furthermore, future studies should directly measure acculturation levels to better isolate the unique contributions of heritage culture from differential acculturation or maternal scaffolding.

Second, the current study utilized a parent-reported measure of child shyness. These parent-reported measures of child temperamental traits may be subject to perception bias. Although the present study included an objective observation of parent-child interaction during a shared reading task, future research should incorporate measurements from other informants, such as teachers, and objective observations of social wariness with unfamiliar researchers or peers to further validate the relationship between temperament and narrative production.

Finally, the study's sample size likely constrained the statistical power to identify small effects. Future studies with larger samples would allow for more complex modeling to explore potential mediators between maternal narrative style and child outcomes. Furthermore, future research should adopt a longitudinal design to examine the inherently bidirectional parent-child interactions, specifically whether maternal scaffolding and elicitation shape or adapt to the child's temperament over time.

Theoretical and practical implications

The findings of the present study offer several key implications for understanding the relationships between culture, temperament, shared reading, and language development in low-income immigrant families. First, this study provides robust empirical support for Sigel's (2002) Psychological Distancing Model and Gutiérrez and Rogoff's (2003) repertoires of practice by bridging these two frameworks to demonstrate that the hierarchical structure of cognitive demand remains consistent across distinct cultural groups, particularly Mexican American and Chinese American dyads, even when the raw frequencies vary. These shared reading interactions are embedded within culturally patterned repertoires of practice (Gutiérrez and Rogoff, 2003). Second, this study extends the literature on temperamental shyness from a cultural perspective. The findings highlight that maternal elicitation or scaffolding is a transactional process that accommodates the child's comfort level in social settings, aligned with specific cultural socialization goals such as 做人 (zuò rén) and bien educado. This suggests that maternal behavior can be understood by integrating the structure of cognitive demand (Sigel, 2002) with the child's temperamental needs and their distinct cultural repertoires of practice (Gutiérrez and Rogoff, 2003).

The findings also challenge the deficit-oriented assumptions about the home literacy environments of the low-income immigrant families. As the positive association between higher cognitive demanding questions (referential, behavioral, and inferential) and child utterances is shared across both cultural groups, educators should recognize that mothers of both groups are effectively scaffolding their children's speech when using these complex prompts. Furthermore, educators should recognize that Mexican American mothers' lower overall use of questions and their tendency to use fewer inferential questions when interacting with a shy child may reflect a culturally grounded narrative style that accommodates a child's temperament by maintaining a supportive and low-stress communicative environment (Schick et al., 2017). Additionally, although the high frequency of elicitation from Chinese American mothers may encourage more child verbal turns overall, the reliance on closed-ended yes/no questions may be explained not as a drive of child speech but as their cultural adaptation to maintain social harmony for shy children (Chen, 2019).

Finally, the present study underscores the importance of dyadic assessment, as a familiar caregiver can act as a buffer for social wariness (Coplan and Rubin, 2010). This is particularly crucial for assessing shy bilingual children's performance, as a home-based assessment is likely to provide a more accurate representation of their linguistic skills than a school- or lab-based assessment conducted with unfamiliar research staff (Crozier and Perkins, 2002; Spere and Evans, 2009). Researchers should incorporate observing parent-child interactions to gain a more holistic view of a child's true linguistic potential.

A key strength of this study is its explicitly asset-based and culturally grounded approach to understanding parent-child processes. By combining fine-grained linguistic analyses of parent-child discourse and narratives during shared book reading with parent reports of children's temperament, particularly shyness, a construct interpreted differently across highly individualistic and collectivistic cultural contexts, the study illuminates developmental pathways in ways that honor families' cultural knowledge and practices. The participating parents were low-income, first-generation, home-language-dominant immigrants from Chinese or Mexican backgrounds, many of whom likely maintain heritage-based values that shape their socialization goals. For Chinese American parents, these goals often align with the indigenous concept of 做人 (zuò rén), or becoming a proper person who embodies moral character, humility, and respect for others, especially elders. For Mexican American parents, similar aspirations are reflected in the indigenous notion of bien educado, which emphasizes being well-mannered, courteous, and respectful toward others and authority figures. Taken together, these cultural frameworks highlight that both Chinese and Hispanic heritage communities hold parallel expectations and repertoires of practice (Gutiérrez and Rogoff, 2003) for raising children to become morally grounded, respectful, and socially attuned individuals. By situating parenting and child temperament within these culturally meaningful constructs, the study advances a strengths-based understanding of immigrant families and the cultural wisdom embedded in their home learning environments and home literacy practices (see Liew et al., 2020; Gonzalez et al., 2023).

Furthermore, study findings also highlight that parental narratives and parent-child conversations during home literacy activities vary across immigrant families. How parents prompt, guide, and scaffold children's meaning-making during shared book reading reflects both the foundational framework of cognitive demand (Sigel, 2002) and their heritage cultural repertoire of practice and socialization goals (Gutiérrez and Rogoff, 2003). For Chinese American and Mexican American parents, these practices stem from commitments to moral character, humility, courtesy, respect, and social responsibility. Such goals shape parents' narratives and questions as well as their expectations for children's participation, emotional expression, and conversational style. While the positive association between higher levels of cognitive demand questioning and child verbal output indicates a universal mechanism, the variability in how mothers use specific prompts underscores that there is no uniform pattern of parent-child discourse. Importantly, the present study suggests that mothers adapt their conversations to accommodate their child's temperamental shyness. Their strategies reflect culturally embedded forms of scaffolding through Mexican American mothers using fewer high-demand inferential questions to maintain a supportive narrative flow and Chinese American mothers using more low-demand yes/no questions to encourage shy children's verbal participation. Thus, understanding narrative and conversational practices requires attention to the cultural frameworks that guide parents' intentions and the competencies they aim to cultivate. By highlighting these nuances, our study affirms the importance of interpreting parental behaviors within their cultural ecologies and repertoire of practice rather than through a one-size-fits-all lens.