Mexican-origin youths’ language brokering for fathers and mothers: Daily experiences and youths’ diurnal cortisol slopes

Abstract

This study examined the associations of language brokering stress intensity and exposure with Mexican-origin youths’ cortisol responses when brokering for fathers and mothers, and the moderating role of youths’ brokering efficacy in these relations. Participants were 289 adolescents (M_age = 17.38, SD = .94, 52% girls) in immigrant families. When brokering for mothers, stress exposure was related to flatter (less healthy) same-day diurnal slopes in youth. When brokering for fathers, daily brokering efficacy buffered the detrimental link between stress intensity and youths’ same-day cortisol slopes. When brokering for fathers/mothers, stress intensity and exposure were related to flatter (less healthy) next-day diurnal slopes. Although daily brokering stress can relate to youth physiologic functioning, feeling efficacious about brokering may buffer the negative ramifications of stress.

Children in immigrant families often translate for their English-limited family members in daily life (Chao, 2006), such as at the grocery store or when watching a show for entertainment (Roche et al., 2015). This translating activity, known as language brokering, is essential for immigrant families’ survival and adaptation in the United States (Morales & Hanson, 2005). Even though some prior research has focused on language brokering as a stressful experience, because it can be a large or even overwhelming responsibility (e.g., Morales & Hanson, 2005), the physiological effects of language brokering stress are relatively unknown. In fact, according to the cultural and stress biology framework, language brokering is a cultural practice that may influence ethnic minority adolescents’ physiological stress responses (Haft et al., 2021). Emerging work suggests that brokering experiences are nuanced—specifically, that individuals’ efficacy with brokering may negate or buffer any tension or difficulties around translation (Kam & Lazarevic, 2014; Kim, Zhang, et al., 2018). Moreover, although language brokering is a day-to-day activity, prior studies have examined language brokering experiences at only one time point, which means they have failed to capture daily fluctuations in language brokering experiences (Guan et al., 2020; Kam, 2011). By considering daily fluctuations in brokering with greater methodological rigor, we gain a better understanding of the microprocesses through which language brokering affects youths’ physiological stress responses in naturalistic settings. Thus, the current study adopted a 4-day diary study to investigate adolescents’ physiological stress response to daily language brokering stress and feelings of efficacy when brokering for fathers and mothers.

We focus on cortisol, which is the end hormonal product of the hypothalamic–pituitary–adrenal (HPA) axis, one of the body’s major stress response systems (Starr et al., 2019). Cortisol can be measured easily through saliva. Cortisol levels follow a strong diurnal pattern that is reactive to daily stressors (Pruessner et al., 1997). These characteristics of cortisol make it an ideal indicator for the physiological stress response. Specifically, we operationalized physiological stress response as the rate of decline in cortisol level from waking to bedtime (i.e., diurnal cortisol slope). According to the appraisal theory (Scherer et al., 2001), physiological responses to stressful events depend on individuals’ cognitive appraisal of those events. Therefore, we explore the potential buffering role of language brokering efficacy in the link between daily language brokering stress and diurnal cortisol slope, in order to capture a nuanced picture of daily language brokering experiences among adolescents in immigrant families.

The cultural and stress biology framework

According to the cultural and stress biology framework, an individual’s cultural context can both elicit stress and function as a resource (Haft et al., 2021). Cultural practices, such as language brokering, can be part of an individual’s cultural context, and may influence minority children’s development through the activation of their physiological stress systems. Specifically, language brokering stress may activate the HPA axis, one of the major physiological stress systems in the human body. The HPA axis functions through interactions among the hypothalamus, pituitary gland, and adrenal cortex (Herman et al., 2016). Specifically, when external stressors are experienced, the HPA axis is activated and releases cortisol to prepare the body and brain to cope with stress. In addition to the acute stress response, the HPA axis follows a typical basal activation pattern exhibited by the diurnal cortisol rhythm: cortisol levels start out relatively high at waking, increase around 50%–60% at 30 to 40 min after waking, and then consistently decline throughout the day (Pruessner et al., 1997). Theory and empirical work both suggest that chronic or persistent activation of the HPA axis may result in changes in the diurnal cortisol pattern (Gunnar & Quevedo, 2007). That is, changes can be seen in the diurnal slope. Individuals experiencing higher levels of stressors exhibit flatter diurnal slopes (lower levels of cortisol in the morning and higher levels in the evening), which in turn relate to a range of mental and physical health problems (Starr et al., 2019).

Second, the cultural and stress biology framework (Haft et al., 2021) suggests that while brokering stress can evoke physiological stress responses directly, there are individual resources related to cultural practices that can function as protective factors to attenuate the adverse effect of brokering stress on stress responses. According to the cognitive appraisal theory, individuals react differently to the same events as their evaluations of the events vary (Scherer et al., 2001). Based on the cultural and stress biology framework and cognitive appraisal theory, positive evaluations of language brokering, such as high brokering efficacy, may be a resource that protects adolescents from the detrimental effect of brokering stress on their physiological stress response systems.

Language brokering stress

Children in Mexican-origin immigrant families often master both Spanish and English, and thus are qualified, and often required, to be a language broker for their parents. As language brokers, children translate a broad array of items, from entertainment to official documents necessary to promote the family’s survival and adaptation (Chao, 2006). Child language brokers may find the translation process stressful, as they are responsible for the accuracy of the translation (Morales & Hanson, 2005).

There are two separate lines of research on language brokering. The first line of research has relied on qualitative approaches to understand the subjective feelings and experiences of children who are language brokers (e.g., Corona et al., 2012; Morales et al., 2012). This line of research has revealed that language brokering often relates to experiencing stress, which may be due to the complex and “adult-like” translating situations facing child language brokers (Roche et al., 2015). Also, this line of research shows that language brokering is an everyday activity influenced by contextual factors, such as language brokering purpose and content (Corona et al., 2012). As a bridge between their parents and their host communities, language brokering adolescents are responsible for translating items of differing levels of complexity, from daily conversations to official immigration documents (Morales et al., 2012). Given that the difficulty and importance of brokering may vary across different items translated, the stressful feeling of language brokering may also vary across different language brokering situations. Descriptive information about daily language brokering activities and experiences for different items and situations is therefore needed.

The second line of research utilizes quantitative methods to investigate the influence of language brokering on children’s development, particularly their physical and mental health (e.g., Guan et al., 2020; Kam, 2011; Kim et al., 2017). Considering the prevalence and potential detrimental effects of language brokering, this line of research has focused on the consequences of language brokering stress (Kam & Lazarevic, 2014; Shen et al., 2019). Given the difficulty of assessing daily language brokering stress in different situations, the bulk of this research assesses only general language brokering stress. However, Roche et al. (2015) point out that brokering is a situation-specific activity with varying stress levels across different situations. Therefore, we examine stress intensity and stress exposure as distinct concepts to yield a more nuanced picture of language brokering stress across various brokering situations at the daily level. The concept of stress intensity comes from literature focused on racism-related stressors; specifically, the Index of Race-Related Stress refers to stress intensity, or how upset individuals are when encountering discriminatory events (Utsey & Ponterotto, 1996). We adapt the construct to language brokering stress intensity, defined as the highest level of stress across different language brokering situations within a day. Adopting the concept of stress exposure as the total number of stressful activities (Koffer et al., 2016), we conceptualize language brokering stress exposure as the extent to which language brokering stress is experienced across different language brokering situations (e.g., doctor’s office, grocery store).

Even though nearly 68% of Latinx children in the U.S. live with two parents (U.S. Census Bureau, 2019), language brokering research has largely focused on experiences with parents generally, without distinguishing between fathers and mothers (e.g., Guan et al., 2020; Kam & Lazarevic, 2014). However, there is a growing recognition of gender differences in brokering activities, suggesting it may be important to consider the gender of not only the child broker (Buriel et al., 2006), but also the parent for whom the child is translating (Hua & Costigan, 2017). In fact, brokering experiences may differ according to whether children are translating for fathers or mothers.

Paquette (2004) posited that fathers and mothers play different roles in child development because of particular interaction patterns with children. The mother–child relationship is most strongly linked to attachment, whereas the father–child relationship aims to develop child competence (Paquette, 2004). In Mexican-origin families more specifically, children tend to have closer relationships and more open conversations with mothers, whereas fathers tend to be reliable supporters, financially and instrumentally (Crockett et al., 2007). Thus, findings pertaining to the influence of mother–child interactions may not be directly transferable to father–child interactions, and vice versa. Even though mother–child interactions are more frequent, father–child interactions, such as brokering for fathers, may have unique influences on children’s responses (Flouri, 2003).

There are a few studies that have investigated children’s language brokering activities and experiences with fathers and mothers separately in naturalistic settings. For instance, Hua and Costigan (2017) assessed language brokering overall frequency and feelings for fathers and mothers separately among children in Chinese immigrant families in Canada. They found that adolescents translated more frequently for mothers than for fathers, and compared to language brokering for mothers, language brokering for fathers was characterized by more challenges and fewer positive feelings. In Latinx families, even though mothers are more likely to be the primary caretakers, with a profound influence on child development (Updegraff et al., 2014), Latino fathers are more likely to be involved in parenting compared to fathers from other racial/ethnic minority groups, which makes the paternal role potentially more prominent for Latinx adolescents (Cabrera & Bradley, 2012). The current study investigates adolescents’ language brokering experiences with fathers and mothers separately, in order to examine the similarities and differences between language brokering experiences among Mexican-origin youth.

Language brokering stress and physiological response

Language brokering stress is linked to multiple indicators of well-being (Kam, 2011; Kam & Lazarevic, 2014; Kim, Hou, et al., 2018; Morales & Hanson, 2005; Shen et al., 2019). For instance, Mexican-origin adolescents reported more depressive symptoms, higher anxiety, and more delinquent behaviors when experiencing higher language brokering stress while translating for fathers and mothers (Shen et al., 2019). Also, among Mexican-origin adolescents, negative language brokering feelings have been found to relate to a higher probability of consuming alcohol and engaging in risky behaviors (Kam, 2011). Despite the consistent results showing detrimental effects of language brokering stress on mental health and risky behaviors, the physiological mechanisms, such as diurnal cortisol, that underlie language brokering stress are relatively understudied.

Day-to-day fluctuations in cortisol patterns are related to daily stressors, especially for individuals in their late adolescence, a period when physiological responses are particularly sensitive to stressors (Romeo, 2013). For instance, among adolescents in immigrant families, daily discrimination has been found to relate to greater diurnal cortisol overall output (Zeiders et al., 2012) and flatter diurnal cortisol slope (Huynh et al., 2016). Hence, language brokering, a stressful daily activity for many adolescents in immigrant families, may also result in variations of diurnal cortisol slopes. Specifically, higher language brokering stress may be related to flatter cortisol slopes. Although few studies have examined the relation between language brokering stress and diurnal cortisol slopes, one prior study provides preliminary evidence of the relation by revealing that language brokering experiences can be manifested physiologically in cortisol levels (Kim, Zhang, et al., 2018).

Rather than examining a single indicator of stress (Fabes & Eisenberg, 1997), we examine the effects of both stress intensity and stress exposure on diurnal cortisol patterns. Stress intensity (i.e., stress level of the most salient stressor within a specific period) may produce an immediate HPA axis activation response, and this may last across the day (Hastings et al., 2011). Previous studies have shown that high stress intensity is related to prolonged responses and slower post-stress recovery of the HPA axis (Herman et al., 2016; Márquez et al., 2002). Therefore, higher language brokering stress intensity may also be a direct and effective predictor of diurnal cortisol rhythm, specifically flatter diurnal cortisol slope. However, it remains unclear whether the influence of brokering stress intensity is restricted to the same-day diurnal cortisol slope or whether it can extend to the following day. On the other hand, stress exposure (i.e., the number of stressors in daily life) has also been found to predict higher psychological distress and more self-regulatory difficulty (Evans & English, 2002). Given that language brokering stress may be spread widely across multiple aspects of life, more language brokering stress exposure may lead to prolonged activation of adolescents’ physiological stress response system, as represented by flatter diurnal cortisol slopes. Since repeated daily stress is related to maladaptive physiological responses (Epel et al., 2018), the adverse influence may accumulate across different brokering situations and persist to the following day. High-stress intensity may not necessarily correspond to high-stress exposure, as adolescents who translate less often may feel brokering stress infrequently (i.e., low-stress exposure) but could experience a high level of stress related to the few items they do translate (i.e., high-stress intensity). To yield a more nuanced understanding of the physiological responses to brokering stress, the current study adopts the daily diary method to examine how brokering stress intensity and exposure influence same-day and next-day diurnal cortisol slopes.

Language brokering stress and efficacy

Language brokering efficacy, which refers to individuals’ confidence in their translating ability (Kam & Lazarevic, 2014), is one of the key evaluative dimensions of language brokering. Enhanced efficacy in brokering is associated with positive child outcomes, such as fewer depressive symptoms, lower parent–child alienation, and stronger academic competence (Chen et al., 2020; Shen et al., 2018). In a longitudinal study of Latino youth, Kam and Lazarevic (2014) found that brokering efficacy buffered the association between brokering stress and youths’ alcohol consumption; higher brokering stress related to greater alcohol consumption 6 months later only among youth with low (vs. high) brokering efficacy.

Self-efficacy may also buffer youths’ physiological responses to language brokering. Individuals with high self-efficacy are better at coping with stressors or obstacles (Bandura, 1977), suggesting that individuals with a higher sense of brokering efficacy may show lower cortisol output when facing stressors, as one major function of cortisol is to increase glucose in the blood to prepare the body to cope with stress (Hellhammer et al., 2009). In other words, feeling efficacious when translating for parents may mitigate the physiological response to language brokering stress. Kim, Zhang, et al. (2018) conducted laboratory experiments and found that adolescents who reported feeling low (vs. high) efficacy about language brokering for their mothers demonstrated cortisol patterns characterized by greater baseline stress before the language brokering task and a slower recovery of cortisol output after the tasks. Although this research provides preliminary evidence for the buffering effect of efficacy in the relation between language brokering stress and acute cortisol reactivity, it is relatively unknown whether language brokering efficacy could attenuate the effect of language brokering stress on diurnal cortisol rhythms in a naturalistic setting, and whether the buffering effects of brokering efficacy can extend to subsequent days. Additionally, different language brokering contexts may confer various demands for adolescents and thus result in varied brokering experiences (Roche et al., 2015). As sense of efficacy is situation-specific and context-dependent (Bandura, 1982), adolescents may perceive various levels of efficacy when translating different content in various situations. For example, adolescents may be more confident when translating the names of products in a grocery store, but less confident about translating a complex financial agreement detailing annual percentage rates and terms of a loan. Given that prior studies have only tested the general sense of language brokering efficacy, without considering the varying nature of sense of efficacy (Shen et al., 2018), the current study utilized the daily diary approach to investigate whether daily brokering efficacy would influence the relation between brokering stress and diurnal cortisol slopes.

The current study

Guided by the cultural and stress biology framework (Haft et al., 2021) and appraisal theory (Scherer et al., 2001), we aimed to fill the research gaps concerning how daily brokering experiences are associated with physiological stress responses. Specifically, we conducted a relatively exploratory study to investigate the influence of brokering experiences with fathers and mothers on adolescents’ diurnal cortisol rhythm through a 4-day daily diary study. First, to understand the extent to which brokering stress influenced adolescents’ diurnal cortisol rhythm, we examined the effect of brokering stress exposure and intensity, separately for fathers and mothers, on adolescents’ same-day and next-day cortisol slopes. Second, the current study examined the moderating role of language brokering efficacy in the relation between language brokering stress and adolescents’ diurnal cortisol slopes. We hypothesized that higher levels of language brokering stress exposure and stress intensity would be related to flatter same-day cortisol diurnal slopes, and the relation would be weakened by high levels of language brokering efficacy, whether adolescents translated for their mother or father. However, we made no specific hypotheses about whether the effect of language brokering stress would persist to predict next-day diurnal cortisol slopes, due to lack of compelling research evidence.

METHOD

Participants

The current study utilized data from a larger longitudinal study of Mexican-origin immigrant families with language-brokering adolescents, started in 2012 (Kim, Hou, et al., 2018). In the three-wave larger study, potential participants were originally recruited through school presentations, public records, and community recruitment in and around a metropolitan city in Central Texas, and families that were Mexican-origin and had a child functioning as a language broker qualified for the study and were recruited. Participants from the 342 families at Wave 3 of the larger study were asked to participate in the daily diary project from 2017 to 2020, and 289 Mexican-origin adolescents chose to participate in the 4-day diary project. Participating adolescents (52% girls) were, on average, 17.38 years old (SD = .94) and a majority were born in the United States (76%). All fathers and 99% of mothers were born in Mexico. A majority of families (72% of fathers; 74% of mothers) reported an education level of less than high school. Most of the fathers were manual laborers (e.g., construction worker, janitor, mover, etc.). A majority of parents (52% of fathers; 61% of mothers) reported family incomes in the range of $10,000 to $30,000.

Procedures

Parental consent and adolescent assent were obtained at an acquaintance meeting if the family decided to participate in the project. Participants were asked to complete a survey regarding their daily experience of language brokering and daily activities on a tablet before going to bed for four nights. Survey items were presented in both English and Spanish simultaneously. Spanish survey items were translated by bilingual Spanish speakers and then back translated to English to validate the accuracy of translation. In addition to the daily survey, adolescents provided salivary samples at waking, 30 min post-waking, and bedtime on each of the days (12 samples in total). Adolescents were compensated up to $20 for participating in the daily diary study. Among the 289 participants, 256 (88.5%) provided survey data for all four study days, and 245 (84.8%) provided cortisol data for all four study days.

Measures

Salivary cortisol

Salivary samples were collected using SaliCaps (IBL International GMBH) and were frozen at −32°C before being sent to Biochemisches Laboratory, Trier, Germany, to be assayed for cortisol. Cortisol concentration was determined by using a time-resolved fluorescence immunoassay (Dressendörfer et al., 1992). Intra-assay coefficients of variation (CVs) were between 4.0% and 6.7%, and inter-assay CVs ranged from 7.1% to 9.0%. In this study, we focused on the diurnal cortisol slope, which was calculated by subtracting each day’s waking sample from the bedtime sample and dividing by the time discrepancy between the wake-up time and bedtime of each day. Diurnal cortisol slope ranged from −3.18 to 1.06 nm/h with a mean of −0.48 nm/h for the current sample. The intraclass correlation coefficient (ICC) of diurnal cortisol slope is .263.

Stress intensity/exposure of language brokering for fathers/mothers

The measurement of daily language brokering stress was modified from the yearly assessment of language brokering stress (Shen et al., 2019). Adolescents reported whether they had engaged in any of eight translating activities for their father and/or mother across 4 days (see details of items in Table 1). For items they did translate, a follow-up question was asked to rate how stressful each specific activity was on a five-point scale (ranging from 0 = not stressful to 4 = very stressful). A sample item is: “How stressed did you feel when you translated the phone calls from English to Spanish for your father today?” We calculated two different scores (i.e., intensity and exposure) of language brokering stress, separately for translating for father versus mother. We measured stress intensity by using each participant’s maximum value across the language brokering stress items reported for translating for father or mother on each day. The ICC of brokering stress intensity for fathers is .602 and for mothers is .412. We assessed stress exposure by taking a count of the number of translation activities that each participant reported as stressful (i.e., brokering items rated as 1 through 4 were counted, and items rated as 0 were not counted) on each day, for each parent. The ICC of brokering stress exposure for fathers is .730 and for mothers is .723.

TABLE 1.

Descriptive information of language brokering activity and stress for fathers and mothers across four days (N = 289)

	Day 1	Day 2	Day 3	Day 4	Total
Number (%) of adolescents engaged in language brokering activities for fathers on a particular day
Phone calls	19 (6.57%)	14 (4.84%)	19 (6.57%)	15 (5.19%)	46 (15.92%)
Explaining the use of products	19 (6.57%)	21 (7.27%)	18 (6.23%)	16 (5.54%)	50 (17.30%)
Entertainment (television shows, movies, music)	18 (6.23%)	13 (4.50%)	22 (7.61%)	13 (4.50%)	46 (15.92%)
Use of technology (computer, Internet, smartphone, tablet)	28 (9.69%)	26 (9.00%)	22 (7.61%)	23 (7.96%)	66 (22.84%)
Receipts, prices, or other things at a store	16 (5.54%)	18 (6.23%)	11 (3.81%)	13 (4.50%)	43 (14.88%)
Explaining work notices, manuals, or instructions for father's job	29 (10.03%)	12 (4.15%)	14 (4.84%)	14 (4.84%)	46 (15.92%)
Filling out applications (for job, housing, insurance)	14 (4.84%)	10 (3.46%)	9 (3.11%)	6 (2.08%)	29 (10.03%)
Filling out government documents or other legal documents	13 (4.50%)	8 (2.77%)	11 (3.81%)	9 (3.11%)	31 (10.73%)
Total	53 (18.34%)	48 (16.61%)	47 (16.26%)	46 (15.92%)	120 (41.52%)
Number (%) of adolescents engaged in language brokering activities for mothers on a particular day
Phone calls	42 (14.53%)	36 (12.46%)	41 (14.19%)	33 (11.42%)	100 (34.60%)
Explaining the use of products	62 (21.45%)	39 (13.49%)	40 (13.84%)	42 (14.53%)	120 (41.52%)
Entertainment (television shows, movies, music)	52 (17.99%)	45 (15.57%)	42 (14.53%)	38 (13.15%)	114 (39.45%)
Use of technology (computer, Internet, smartphone, tablet)	72 (24.91%)	50 (17.30%)	43 (14.88%)	39 (13.49%)	126 (43.60%)
Receipts, prices, or other things at a store	47 (16.26%)	38 (13.15%)	28 (9.69%)	28 (9.69%)	97 (33.56%)
Explaining work notices, manuals, or instructions for mother's job	49 (16.96%)	25 (8.65%)	25 (8.65%)	27 (9.34%)	83 (28.72%)
Filling out applications (for job, housing, insurance)	27 (9.34%)	16 (5.54%)	17 (5.88%)	16 (5.54%)	49 (16.96%)
Filling out government documents or other legal documents	27 (9.34%)	20 (6.92%)	15 (5.19%)	16 (5.54%)	47 (16.26%)
Total	125 (43.25%)	102 (35.29%)	89 (30.80%)	89 (30.80%)	206 (71.28%)

Language brokering efficacy for fathers/mothers

Adolescents’ perception of efficacy in translating for each parent was measured using a three-item scale (Kim et al., 2017). Items are, “I am good at translating for my father/mother today,” “I was skilled at translating for my father/mother today,” and “I translated correctly for my father/mother today.” Adolescents self-reported on a scale ranging from 1 (strongly disagree) to 5 (strongly agree), with higher scores indicating a higher sense of efficacy. The scale demonstrated adequate reliability across each day (for fathers: α_efficacy = .96, .94, .95, and .95 for day 1, day 2, day 3 and day 4, respectively; for mothers: α_efficacy = .93, .94, .95, and .95 for day 1, day 2, day 3 and day 4, respectively). The ICC of brokering efficacy for fathers is .603 and for mothers is .464.

Individual and behavioral covariates

Adolescents self-reported their age, gender, and nativity. Adolescents also reported their daily consumption of alcohol/caffeine/nicotine, medication use, wake time, bedtime, hours of exercise, and whether they had oral disease/injury. Day of the study was also included in the analysis as a covariate.

Analytic strategy

In the current study, 289 adolescents participated in the 4-day diary study, and thus reported on their language brokering activities. Descriptive statistics in SPSS were used to understand frequencies of translation for fathers and mothers, respectively, across adolescents. A paired t-test examined differences in translating frequency for fathers and mothers. An independent t-test examined adolescent gender differences in brokering frequency, stress intensity, and stress exposure when translating for father and mother, separately. For analyses linking language brokering activities to diurnal cortisol, we included only adolescents who provided at least one cortisol sample and completed at least one translation activity for their father (n = 108) or mother (n = 193) across the four survey days.

Given the nested structure of the data (i.e., days nested within individuals), multilevel regression analyses linking language brokering stress to diurnal cortisol slopes were conducted in Mplus 8.3 (Muthén & Muthén, 2017). Full information maximum likelihood was used to handle missing data in covariates (Muthén & Muthén, 2017). Language brokering stress intensity and exposure were tested in separate models to investigate their effects on same-day and next-day diurnal cortisol slopes. In terms of the same-day effects regarding brokering stress intensity for fathers, we tested the main effects of brokering stress intensity and efficacy on the same-day diurnal cortisol slope in Model 1 and then added the interaction term between stress intensity and efficacy in Model 2. For the next-day association analysis, all predictors were lagged by 1 day to assess their effects on the next-day diurnal cortisol slopes. The same set of analyses was conducted for brokering for fathers stress exposure. Then, we repeated the same analyses for models with mothers. All time-varying predictors (e.g., brokering for mothers stress intensity/exposure and brokering for mothers efficacy) and behavioral covariates (i.e., alcohol, medication, nicotine and caffeine consumption, exercise hours, wake time, bedtime, and oral disease or injury) were centered at each participant’s personal mean, as recommended, to remove between-person variability and yield a within-person regression coefficient estimation (Curran et al., 2012). Unlogged diurnal cortisol slope served as the dependent variable, as diurnal cortisol slope was neither highly skewed (−1.19) nor kurtotic (4.78). Two sets of sensitivity analyses were conducted to assess the validity of the results (see Supporting Information for details).

RESULTS

Descriptive statistics

Table 1 displays descriptive information on the frequency of language brokering activities engaged in for fathers and mothers across 4 days. Among eight language brokering activities, translating the use of technology for their father or mother was the most common brokering activity. On each survey day, whereas only 15% to 18% of adolescents translated for their fathers, 30% to 43% of adolescents translated for their mothers. Across the 4 days, 40% of adolescents had translated for their fathers and 71% of adolescents had translated for their mothers. The paired t-test showed that, overall, adolescents translated more frequently for mothers than for fathers (t = 8.340, p < .001). The independent t-test showed that whereas there was no significant adolescent gender difference in translating frequency for fathers (t = 0.678, p = .498), girls (vs. boys) reported a higher frequency of translating for mothers (t = 2.317, p = .021). There was also no difference by adolescent gender in brokering stress intensity or exposure when translating for either fathers or mothers. Table S1 displays means and standard deviations of language brokering experiences variables, separately for fathers and mothers. Table S2 shows the count and percentage of different levels of brokering stress intensity on each study day, separately for mothers and fathers.

Language brokering stress and diurnal cortisol slopes

The effects of language brokering experiences on diurnal slopes were tested in multilevel models. In Model 1, the main effects of language brokering stress and efficacy were examined. The within-person interaction term between language brokering stress and efficacy was added in Model 2 (see both Model 1 and Model 2 in Supporting Information). Tables 2 and 3 display the results of Model 2 for the same-day and next-day effects of stress intensity and exposure.

TABLE 2.

Multi-level moderation model examining within-person effects of language brokering for fathers and mothers efficacy and stress intensity on the same-day and next-day diurnal slope (nm/L)

	Outcomes: same-day diurnal slope				Outcomes: next-day diurnal slope
	Brokering for fathers		Brokering for mothers		Brokering for fathers		Brokering for mothers
	b(SE)	P	b(SE)	P	b(SE)	P	b(SE)	P
Intercept	−0.399 (0.134)	.003	−0.336(0.07)	<.001	−0.691 (0.199)	.001	−0.368 (0.103)	<.001
WP LB stress intensity	−0.087 (0.068)	.202	0.039 (0.043)	.369	0.192 (0.046)	<.001	0.059 (0.032)	.064
WP LB efficacy	−0.135 (0.09)	.134	0.049 (0.058)	.401	−0.05 (0.099)	.616	0.094 (0.053)	.076
WP LB stress intensity × WP LB efficacy	−0.557 (0.185)	.003	−0.139 (0.086)	.103	0.135 (0.368)	.713	−0.098 (0.071)	.166
Control variables
WP diurnal slope					−0.267 (0.132)	.044	−0.564 (0.114)	<.001
Day of study	−0.083 (0.032)	.010	−0.061 (0.020)	.002	0.029 (0.048)	.545	−0.061 (0.030)	.040
WP exercise hours	0.034 (0.048)	.482	0.026 (0.029)	.375	0.042 (0.052)	.420	0.001 (0.032)	.973
WP caffeine consumption	0.083 (0.088)	.343	−0.004 (0.054)	.947	−0.044 (0.134)	.744	−0.012 (0.076)	.877
WP alcohol consumption	−0.643 (0.710)	.365	0.194 (0.209)	.352	3.730 (1.757)	.034	−0.115 (0.236)	.625
WP nicotine consumption	0.271 (0.152)	.074	0.014 (0.160)	.929	−1.002 (2.031)	.622	0.212 (0.109)	.051
WP medication consumption	0.188 (0.133)	.159	−0.090 (0.171)	.600	0.654 (0.248)	.008	−0.083 (0.145)	.569
WP oral disease or injury	0.476 (0.208)	.022	−0.057 (0.387)	.883	1.681 (0.603)	.005	−0.154 (0.515)	.766
WP bedtime	0.06 (0.048)	.211	0.054 (0.031)	.084	0.049 (0.028)	.085	0.061 (0.035)	.085
WP wake time	−0.103 (0.038)	.008	−0.034 (0.026)	.193	0.022 (0.047)	.639	−0.035 (0.026)	.165
BP age	0.045 (0.045)	.323	0.031 (0.027)	.244	0.104 (0.049)	.035	0.021 (0.028)	.457
BP nativity	0.074 (0.096)	.442	−0.044 (0.055)	.421	0.056 (0.089)	.528	0.024 (0.065)	.713
BP gender	0.148 (0.072)	.039	0.063 (0.054)	.242	0.175 (0.082)	.033	0.084 (0.049)	.085

Note: Observations are only counted on days when adolescents translated for a specific parent (father or mother). The same-day model for brokering for fathers: N = 108 adolescents; number of observations = 173; for mothers: N = 193 adolescents; number of observations = 374. The next-day model for brokering for fathers: N = 97 adolescents; number of observations = 130; for mothers: N = 182 adolescents; number of observations = 292. Gender: 1 = boy, 0 = girl; Nativity: 1 = U.S.-born, 0 = Mexico-born. Significant effects from language brokering experiences (i.e., brokering stress and/or efficacy) on diurnal cortisol are highlighted in bold.

Abbreviations: BP, between-person; LB, language brokering; WP, within-person.

TABLE 3.

Multi-level moderation model examining within-person effects of language brokering for fathers and mothers efficacy and stress exposure on the same-day and next-day diurnal slope (nm/L)

	Outcomes: same-day diurnal slope				Outcomes: next-day diurnal slope
	Brokering for fathers		Brokering for mothers		Brokering for fathers		Brokering for mothers
	b(SE)	P	b(SE)	P	b(SE)	P	b(SE)	P
Intercept	−0.399 (0.123)	.001	−0.329 (0.07)	<.001	−0.678 (0.173)	<.001	−0.374 (0.104)	<.001
WP LB stress exposure	−0.077 (0.094)	.411	0.068 (0.034)	.044	0.14 (0.044)	0.002	0.063 (0.026)	.016
WP LB efficacy	−0.129 (0.09)	.149	0.041 (0.054)	.440	−0.075 (0.089)	0.403	0.08 (0.049)	.106
WP LB stress exposure × WP LB efficacy	−0.404 (0.275)	.142	−0.179 (0.159)	.261	0.092 (0.18)	0.61	−0.071 (0.106)	.499
Control variables
WP diurnal slope					−0.268 (0.131)	0.04	−0.577 (0.111)	<0.001
Day of study	−0.082 (0.031)	.007	−0.060 (0.019)	.002	0.022 (0.042)	0.605	−0.064 (0.030)	.029
WP exercise hours	0.034 (0.048)	.476	0.034 (0.029)	.244	0.015 (0.051)	0.771	0.005 (0.033)	.878
WP caffeine consumption	0.077 (0.088)	.386	−0.009 (0.052)	.855	−0.062 (0.126)	0.622	−0.001 (0.072)	.993
WP alcohol consumption	−0.643(0.708)	.364	0.195 (0.215)	.366	3.607 (1.523)	0.018	−0.166 (0.234)	.478
WP nicotine consumption	0.290 (0.154)	.059	−0.016(0.175)	.928	−0.866 (1.709)	0.612	0.229 (0.110)	.037
WP medication consumption	0.341 (0.179)	.056	−0.095 (0.168)	.571	0.570 (0.214)	0.008	−0.077 (0.143)	.591
WP oral disease or injury	0.456 (0.197)	.020	−0.061 (0.392)	.877	1.697 (0.525)	0.001	−0.157 (0.509)	.758
WP bedtime	0.049(0.049)	.318	0.059 (0.031)	.059	0.061 (0.034)	0.074	0.064 (0.035)	.072
WP wake time	−0.106(0.038)	.005	−0.027 (0.026)	.289	0.011 (0.046)	0.812	−0.031 (0.025)	.219
BP age	0.047 (0.045)	.297	0.001 (0.027)	.226	0.109 (0.050)	0.028	0.022 (0.028)	.429
BP nativity	0.077 (0.092)	.403	−0.051 (0.055)	.348	0.065 (0.084)	0.44	0.018 (0.065)	.782
BP gender	0.141 (0.068)	.039	0.068 (0.053)	.198	0.179 (0.081)	0.027	0.090 (0.049)	.065

Note: Observations are only counted on days when adolescents translated for a specific parent (father or mother). The same-day model for brokering for fathers: N = 108 adolescents; number of observations = 173; for mothers: N = 193 adolescents; number of observations = 374. The next-day model for brokering for fathers: N = 97 adolescents; number of observations = 130; for mothers: N = 182 adolescents; number of observations = 292. Gender: 1 = boy, 0 = girl; Nativity: 1 = U.S.-born, 0 = Mexico-born. Significant effects from language brokering experiences (i.e., brokering stress and/or efficacy) on diurnal cortisol are highlighted in bold.

Abbreviations: BP, between-person; LB, language brokering; WP, within-person.

Same-day effects of stress intensity

Table 2 displays significant interactions between language brokering stress intensity and brokering efficacy for fathers (b = −0.557, SE = 0.185, p = .003), indicating that the same-day effect of language brokering stress intensity for fathers on adolescents’ diurnal slope varied by different levels of language brokering efficacy after controlling for age, gender, nativity, and time-varying covariates. When adolescents translated for fathers, simple slope analyses showed that, on days when adolescents experienced low language brokering efficacy, stress intensity was not associated with adolescents’ diurnal slopes (b = 0.070, SE = 0.054, p = .195; Figure 1); on days that adolescents experienced high language brokering efficacy, higher stress intensity was associated with steeper diurnal slope (b = −0.244, SE = 0.109, p = .025; Figure 1).

FIGURE 1.

Within-person level interaction effect of language brokering for fathers efficacy and stress intensity of language brokering for fathers on same-day cortisol slope (nm/L)

Next-day effects of stress intensity

Language brokering stress intensity when translating for fathers (b = 0.193, SE = 0.045, p < .001; Table S5) or mothers (b = 0.062, SE = 0.030, p = .042; Table S6) predicted adolescents’ next-day diurnal slope, after accounting for the prior day’s diurnal slope, age, gender, nativity, and time-varying covariates. This suggests that, on days when adolescents experienced higher stress intensity when translating for fathers or mothers, they exhibited flatter diurnal slopes the following day. There were no significant interactions between stress intensity when translating for fathers or mothers and brokering efficacy on next-day diurnal slope.

Same-day effects of stress exposure

There was a main effect of language brokering stress exposure when translating for mothers on the same-day diurnal slope (b = 0.072, SE = 0.036, p = .044; Table S8). On days when adolescents reported higher stress exposure when translating for mothers, they exhibited a flatter diurnal slope. There was no association between language brokering stress exposure when translating for fathers and the same-day diurnal slope. There were no significant interactions between stress exposure when translating for fathers or mothers and brokering efficacy on same-day diurnal slope (Table 3).

Next-day effects of stress exposure

Language brokering stress exposure when translating for both fathers (b = 0.141, SE = 0.043, p = .001; Table S9) and mothers (b = 0.063, SE = 0.026, p = .016; Table S10) did predict next-day diurnal slopes, after controlling for the prior day’s diurnal slope, age, gender, nativity, and time-varying covariates. This result suggests that on the days when adolescents reported more diverse, stressful language brokering activities for either parent, they demonstrated flatter diurnal slopes the next day. Language brokering stress exposure when translating for fathers or mothers did not interact with efficacy to predict next-day diurnal slopes (Table 3).

DISCUSSION

Language brokering, as a day-to-day activity in the lives of many Mexican-origin adolescents, is often essential for Mexican-origin immigrant families to survive in the United States (Morales & Hanson, 2005). However, most prior research has examined language brokering experiences at a single time point only (Guan et al., 2020; Kam, 2011), ignoring the day-to-day fluctuations of such experiences. Furthermore, language brokering has been considered a stressor due to the challenging and demanding nature of the experience (Kam & Lazarevic, 2014). Like many other stressors, language brokering can activate physiological stress responses, which subsequently contribute to compromised health outcomes (Geronimus et al., 2006). However, the link between daily language brokering stress and stress-response systems has remained unexplored. Moreover, most extant research has focused on adolescents’ experiences of brokering for mothers, with very limited attention paid to brokering for fathers. Guided by the cultural and stress biology framework (Haft et al., 2021) and appraisal theory (Scherer et al., 2001), the current study addressed these limitations by using a daily diary approach to examine the relations between brokering stress, brokering efficacy, and adolescent brokers’ diurnal cortisol rhythms, separately for translating for fathers and mothers. Overall, our findings reveal that the influence of brokering stress can persist from the same-day to the next-day diurnal cortisol slope, and that brokering efficacy can buffer the negative influence of the highest level of brokering stress (i.e., stress intensity) on diurnal cortisol slope. However, brokering efficacy does not seem to moderate the effect of being exposed to many types of brokering stress in different contexts (i.e., brokering exposure) on the same-day diurnal cortisol slope.

We first examined whether brokering stress intensity and exposure related to same-day and next-day diurnal cortisol slopes, and whether brokering efficacy would moderate such links. For brokering stress intensity, we did not find a main effect on the same-day diurnal cortisol slope when brokering for fathers or mothers. However, language brokering stress intensity for fathers did interact with brokering efficacy to predict diurnal slopes. Specifically, on days when adolescents had high brokering efficacy, they also had steeper diurnal cortisol slopes (a healthier pattern) when their highest brokering stress of that day (i.e., stress intensity) was greater than their average level. This finding supports our hypothesis of the moderating role of brokering efficacy in the relation between brokering stress and brokers’ physiological reactivity, and is also in line with the tenet of cognitive appraisal theory (Scherer et al., 2001) that the positive evaluation of brokering experiences (i.e., brokering efficacy) is beneficial. Furthermore, this finding is consistent with the cultural and stress biology framework (Haft et al., 2021) and prior research suggesting that the confidence Mexican-origin brokers derived from their brokering experience is a cultural asset (Borrero, 2015; Chen et al., 2020), which could work together with brokering stress to influence adolescent physiological stress responses. Our study extends this literature by suggesting that brokering efficacy has the potential to buffer the negative effects of brokering stress. This result highlights the importance of taking a strength-based approach to identify factors that may help Mexican-origin adolescents adjust to living in the United States.

We also observed a lagged effect of brokering intensity when adolescents translated for fathers on next-day diurnal cortisol slope. Specifically, we found that, on days when brokers reported greater levels of their highest brokering stress when translating for fathers, they had flatter diurnal slopes (a less healthy pattern) the next day. This finding extends prior research documenting the detrimental effects of brokering stress on adolescent brokers’ socioemotional well-being (Martinez et al., 2009) by highlighting that high levels of brokering stress could also get “under the skin” to dysregulate adolescents’ stress response system. This finding is consistent with previous research, which suggests that experiencing uncontrollable and socially evaluative stressors can activate individuals’ HPA axis responses (Dickerson & Kemeny, 2004). The significant next-day effect may be due to the fact that ruminations on stressful and negative experiences are likely to last across days to impede adolescents’ physiological stress response (Harrell et al., 2011).

In terms of brokering stress exposure, we found that translating for both fathers and mothers influenced the adolescent broker’s diurnal cortisol slope. Specifically, on days when brokers were exposed to more types of brokering stress in different contexts (i.e., stress exposure) when translating for mothers, they had a flatter diurnal slope, and the negative effect of brokering stress exposure carried over to the next day. We also observed the harmful effect of brokering exposure when adolescents translated for fathers; however, we only observed a next-day effect. That is, when brokers reported more types of brokering stress across diverse contexts when translating for fathers, they had flatter diurnal slopes the next day. Our results suggest that the accumulation of these stressful brokering moments (i.e., stress exposure) were likely to take a toll on adolescent brokers’ stress physiology. This finding aligns with prior research that has found deleterious effects of chronic stress exposure on HPA axis functioning (Starr et al., 2019).

Interestingly, brokering efficacy was not observed to buffer the negative effect of stress exposure on cortisol slope, the way it did for brokering stress intensity. Even though the different results for brokering stress intensity and exposure might be partly due to statistical power and measurement reliability issues, it is also possible that the intensity of a stressful moment may evoke the moderating role of brokering efficacy in ways stress exposure does not. Future studies are needed to replicate our finding before making any definitive conclusions. Prior language brokering literature has categorized brokering experiences into high-stakes, low-stakes, and everyday duties (Anguiano, 2018). Our results highlight that it is also important to examine brokering stress appraisal, and suggest that brokering intensity and exposure are likely to operate in different ways to influence brokers’ well-being. We encourage future studies to continue examining stress appraisals of brokering experiences within different brokering contexts.

It is also worth noting that the finding for the modulating effect of brokering efficacy for the highest brokering stress across the 4 days of our study is shown only for brokering for fathers but not for mothers. The different roles of brokering efficacy for fathers and mothers may be explained by parenting roles in Mexican culture. In Mexican-origin families, mothers often serve as the primary caretakers and provide daily caregiving to children, whereas fathers often serve as the authority figures and provide financial support to the family (Updegraff et al., 2014). Adolescent brokers are likely to spend more time with mothers, and to form stronger bonds with them. Therefore, it is possible that stress experienced when translating for mothers may be resolved more easily, without the engagement of cognitive appraisal. If so, perhaps brokering efficacy was not evoked in the brokering stress response, the way it was when adolescents translated for fathers. However, these findings should be interpreted with caution. Due to sample size differences, we were unable to examine brokering stress when translating for father and mother in the same model. Future studies should continue to consider adolescents’ experiences of brokering for fathers as well as mothers, ideally with more fathers participating, in order to reach more definitive conclusions about the influence of brokering for mothers and fathers on adolescents’ development.

To understand brokering activities at the daily level, we also explored the prevalence of different language brokering activities when adolescents translated for fathers and mothers, separately. Our study focused on two important brokering contexts: community and home (Roche et al., 2015). We found that translating the use of technology (e.g., computer, Internet, smartphone, tablet) was the most common brokering activity when brokering for both fathers and mothers, followed by translating how to use a certain product. The least frequent activities were translating application materials and government documents. The current frequency distribution was different from what was found previously. For example, Weisskirch and Alva (2002) found that, for early adolescents, translating notes or letters from school was the most popular activity, followed by translating household-related and medical bills, credit card statements, and bank statements. This difference suggests a potential shift in the landscape of translating activities for Mexican-origin adolescents, but it may also have to do with age differences between the samples or the fact that school items requiring brokering were not measured in the current study. The current finding of a high frequency of translating the use of technology was consistent with recent research showing that child brokers often translate digital media and technology for their parents, and that such experiences served as a key way for parents to adopt and use new technology (Correa et al., 2015). With the increasing demand for and reliance on technology, future studies should continue to include translating new technologies as a part of their language brokering measure.

In addition, we found that adolescents translate more frequently for mothers than for fathers, which is consistent with prior studies (Chao, 2006). This is probably because adolescent brokers tend to spend more time with their mothers on a daily basis, and thus are likely to perform brokering tasks more frequently for them. Along with prior research showing that brokering is a gendered activity, and that girls are more likely to serve as language brokers than boys (Buriel et al., 2006), our results suggest a more nuanced role of gender. Given the finding that brokering for mothers takes place more frequently, there may be a potential interaction between the genders of the parent and child on brokering activity. However, despite evidence of adolescent gender differences in brokering frequency, our results showed that brokering stress intensity and exposure did not differ by gender, suggesting that the stress engendered by brokering activities is the same across boys and girls.

Contributions, implications, and limitations

There are several strengths of the current study. First, it is the first study, to our knowledge, to document the relations between a culturally relevant stressor—brokering stress—and Mexican-origin adolescents’ physiological response in a naturalistic daily setting. The current study provides an important first step in understanding the physiological mechanism that may underlie adolescents’ stressful brokering experiences. Future empirical work should continue this line of research by examining how both cortisol responses and efficacy may continue to play out across time in health outcomes. Second, the current study employed a novel daily diary design to understand brokers’ translating activities. This approach is advantageous in capturing the daily fluctuations of brokering activities. It can also reduce participants’ recall bias and result in more accurate reports. Third, the current study differentiated adolescent brokering experiences when translating for fathers versus mothers and found that fathers play an important role in adolescent brokers’ day-to-day brokering experiences.

Fourth, the current study took a strength-based approach by identifying the protective role of brokering efficacy in buffering the detrimental effects of brokering stress. This finding extends the cultural and stress biology framework (Haft et al., 2021) by identifying the distinct role of a specific cultural resource, namely brokering efficacy, in modulating the influence of different aspects of cultural stress on physiological stress responses. Intervention efforts that focus on the strengths and benefits of retaining Spanish within immigrant families may aid in building stronger brokering efficacy among youth. Furthermore, parents could be encouraged to provide praise and positive feedback to their children for performing translation duties. Fifth, our study took a nuanced approach to operationalizing brokering stress by differentiating brokering stress exposure from brokering stress intensity. Our results suggest that these two different aspects of stress operate differently in how they influence adolescents’ outcomes, and it is important for future studies to consider both aspects to gain a comprehensive understanding of adolescents’ brokering experiences.

At the same time, there are important limitations of the current study to consider. First, despite the novel daily diary design, the current study recorded participants’ daily brokering for only four consecutive days, and thus was not able to capture great within-person variability of the study variables. The short time period also limits our understanding of language brokering across more extended periods in participants’ day-to-day lives. Future studies should adopt a daily diary design with a longer timeframe. Second, as the current sample is Mexican-origin adolescents from low-income immigrant families in central Texas, caution is warranted when generalizing the results to adolescents from other ethnic and socioeconomic backgrounds and regions. Third, the current study only listed brokering activities in eight common contexts, but brokering can happen in various other situations (e.g., school, doctor’s visit, phone call). For example, Roche et al. (2015) suggest that brokering at school is common for adolescents, which is not assessed in the current study. Future studies could examine brokering stress in additional brokering contexts to better understand how the nature of language brokering affects youths’ biological stress processes. Fourth, as adolescents translate more frequently for mothers than fathers, the analyses were conducted with different sample sizes in different models for mothers and fathers, and thus, mother-father differences were not directly tested in the current study. Future studies could reveal further differences between brokering for mothers versus fathers by including brokering for mothers and fathers in the same model using a larger sample size. In a similar vein, the sample size for brokering for fathers is small, and we encourage future studies with larger sample sizes. Fifth, the current study only examined adolescents’ brokering experiences when translating for fathers and mothers. However, studies have shown that adolescents also broker for other relatives, or even strangers (Weisskirch & Alva, 2002). It is important for future studies to investigate how the stress associated with brokering for other people influences adolescents’ well-being (Chao, 2006). Sixth, language brokering for fathers may be more stressful not because of the brokering itself but possibly due to the quality of the father-adolescent relationship. Future studies could explore the relations between brokering for father experiences and adolescent stress responses by considering the moderating role of the father–adolescent relationship, or by controlling for the father–adolescent relationship as it relates to stress physiology.

CONCLUSION

Using a novel daily diary design, the current study took an important first step in understanding how language brokering stress links to adolescents’ physiological stress response. The current study also identified the buffering role of brokering efficacy, which can attenuate the effects of brokering stress on youths’ physiological stress response. The current findings are important, given the growth of the Mexican-origin population in the United States and the prevalence of brokering among Mexican-origin adolescents.

Abbreviations:

CV

coefficients of variation

HPA

hypothalamic–pituitary–adrenal

ICC

intraclass correlation coefficient