Media Language Preferences and Mental Illness Stigma among Latinx Adolescents

Abstract

Purpose:

Media—a powerful influence on mental illness stigma—varies by language and culture. Nevertheless recent meta-analyses have demonstrated scant attention to Spanish language media as well as historically low Latinx participation in mental illness anti-stigma intervention. To better inform how to improve equity in mental health service utilization, this study assessed how language preferences in mass media influence stigma among Latinx adolescents, compared to family language and social preferences.

Methods:

Sixth-graders self-identifying as Latinx self-completed assessments of mental illness knowledge/positive attitudes and desired separation from peers and adolescent vignette characters experiencing mental illness (N=179; Texas, U.S., 2011–2012). Participants also responded to measures of language preferences (any Spanish versus only English) for consuming media (film/television, music/radio) and speaking with family (parents/grandparents), and social preferences for parties or social gatherings (Latinx versus Anglo persons). Linear regression models adjusting for student and household factors examined the associations between media and family language and social preferences on mental illness stigma.

Results:

Latinx adolescents preferring any Spanish versus English only media reported less mental illness knowledge/positive attitudes and greater social separation from peers and vignette characters with a mental illness, net of all covariates. Family language and social preferences were not associated with any mental illness stigma outcomes.

Conclusions:

Spanish media preference is associated with greater stigma suggesting more stigmatization may exist in Spanish- versus English-media. Attention to mental illness stigma in Spanish media may reduce disparities by media language preference among Latinx adolescents. These findings have implications for populations with other non-English media preferences.

INTRODUCTION

Adolescents identifying as Latinx—a gender-neutral, nonbinary alternative to Latina/o to describe a person of Latin American origin or descent¹—represent a large, growing segment of the United States (U.S.) population. Depression and anxiety are consistently more prevalent among Latinx compared to non-Latinx (NL) white adolescents in nationally representative studies and public health surveillance.^2–6 Studies spanning decades have also reported that Latina girls have higher rates of suicidality than their peers.⁷ Despite this mental health burden, mental health services are underutilized among Latinx adolescents compared to their NL-white counterparts. Moreover, Latinx adolescents are less likely to receive services even when experiencing severe impairment and accounting for health insurance coverage and economic factors.^2,3,8–11 Left untreated, mental illness can lead to further problems for youth including social, behavioral, and academic problems, increased severity and chronicity, comorbidity with other mental illnesses, and chronic physical illness.^12–15 We propose that mental illness stigma contributes to this set of circumstances, leading us to investigate sources of stigma among Latinx adolescents.

Mental Illness Stigma

Mental illness stigma can block the use of mental health services, especially for U.S. Latinx populations.¹⁶ Latinx adults hold greater stigma compared to NL-groups despite roughly equal mental health literacy.^17–19 Stigma among Latinx adults also influences adherence to antidepressant medications as the use of such medications is often viewed as a sign of weakness and considered addictive.¹⁹ Stigmatizing views regarding medications also contribute to family dysfunction among family-oriented Latinx adults when family members discourage the use of treatment.¹⁹ Stigma has been found to be greater among foreign-born and less acculturated Latinx adults.^20,21

Among Latinx adolescents, most of the existing literature has applied qualitative methodology in clinical samples. In general, Latinx compared to NL-white adolescents in treatment for mental health issues receive more stigmatizing attitudes and behaviors from their family and peers.^22–25 Among Mexican-American youth, those with stronger versus lower orientation to Mexican culture (e.g., language use/preference, ethnic identity/interaction) reported fewer negative attitudes towards mental health help-seeking; no association was found between cultural values of familism, folk illness, religiosity, or machismo on help-seeking.²⁶

To promote mental health service use and improve life circumstances for people with mental illness, public anti-stigma efforts can be effective.^27,28 However, these efforts have included little Latinx representation despite the aforementioned disparities. From a meta-analysis of 79 anti-stigma intervention studies that included a pool of 38,364 participants world-wide,²⁷ only 5.6% Latinx compared to 61.1% European/European-American participants were represented across these studies. Less than 1% of the studies included children under 12 years old. Language preferences of the participants and language proficiency of the interventions were not reported. Collectively the studies demonstrate a lack of representation of young Latinx individuals in anti-stigma clinical trials and unknown potential of anti-stigma efforts in reaching all Latinx populations. Failure to examine younger Latinx populations limits our understanding of how policies can close mental health treatment gaps for medically under-served Latinx youth.

Latinx Heterogeneity

Prior stigma research has mostly considered Latinx populations homogenously by lumping multiple groups into one ethnic category (Hispanic/Latino), as some attributes are often shared including familiarity with Spanish language, cultural values, and not belonging to the socio-historical dominant group.²⁹ However, the U.S. Latinx population is heterogeneous across many factors including ancestral heritage and language proficiencies (e.g., Indigenous, Spanish, or English tongues) where Latinx populations experience structural (i.e., healthcare costs, workforce/supply shortages) and cultural barriers (i.e., language of services, familismo) to mental health services differently.^30,31 Conceptualizing Latinx populations heterogeneously in stigma research in particular can help yield more specific knowledge about unique mechanisms through which some Latinx populations develop greater stigma, a disadvantage related to receiving mental health treatment, compared to other Latinx and NL-white populations. With more specific knowledge in hand, interventions will be better positioned to tailor their approach to cultural preferences to reduce service use disparities.

Prior research among U.S. Latinx adult participants (ages 18+) has examined how language preferences within families and for media consumption (aggregated as one variable) and the degree of Latinx social homophily/preference (e.g., in social gatherings/parties) are related to mental health services use.³¹ Adjusting for individual and household factors, the study found that greater Latinx social homophily/preference and Spanish language preference, as well as greater importance of Latinx identity, predicted lower service use for mood and anxiety disorders.³¹ As social and language preferences predict mental health treatment utilization among Latinx adults, and because stigma is related to mental health service use, measures of social and language preferences can allow for a test of these preferences on the development of stigma as a critical component of mental health services.

Media Influence

Mass media (e.g., film/television, music/radio) is a powerful influence on health knowledge, norms, and behavior, particularly among young people who may lack fully developed critical thinking skills to evaluate the motivation behind, and possible negative consequences of, media messages.³² In fact, mass media plays an important role in shaping mental illness stigma among children and adolescents. Pejorative portrayals of mental illness in media fosters the development of negative stereotypes among youth, which leads to avoidant behaviors towards people with mental illness.^33–35 Extant research in this area predominantly assessed English language media^35,36 and prompted mental illness anti-stigma efforts to advocate for change in mass media with some success.^27,35,36

However, heterogeneity within Latinx populations has not been fully addressed by these efforts. In particular, variation by media language preference is often overlooked, likely masking existing disparities for those who may not benefit from interventions targeting English language media. In two recent systematic reviews examining the impact of media portrayal of mental illness on stigmatization and discrimination and the impact of anti-stigma interventions among media professionals across a combined 39 total studies,^35,36 none of the included studies came from countries where Spanish is the official language (20 total countries worldwide). Further, no study from the U.S.^37–44 where large Latinx populations reside assessed or included Spanish language media or its consumers, specifically. Therefore, the generalizability of this research to media in other languages, particularly consumers of Spanish media, is unknown.

To put this issue to scale, Spanish is a leading “power” language worldwide among native speakers and speakers in general along with Mandarin and English; in the U.S., Spanish is the second most spoken language behind English with more than 37 million speakers ages five and older.^45,46 Related to mass media, Spanish language television networks available in the U.S. (e.g., Univision, Telemundo) have set record-high ratings in 2019 among young adults, beating out major English-language networks (e.g., ABC, NBC).^47,48

To the extent that younger consumers of Spanish media read newspapers, one study for informing state-wide prevention strategies in California, U.S.²⁸ found substantial differences in mental illness-related content across the largest Spanish- and English-language California newspapers.⁴⁹ Negative portrayals of people with mental illness occurred 70% of the time in Spanish newspapers compared to 37% in English newspapers.⁴⁹ Just 14% of stories in Spanish acknowledged treatment compared to 51% in English, and only 1% provided help-seeking resources in Spanish newspapers versus 55% in English newspapers.⁴⁹ A newspaper content analysis from Spain also found high levels of stigmatization in Spanish press.⁵⁰ It is important to consider, however, that newspapers from Spain are culturally and ethnically different from Spanish newspapers found in the U.S., and in the era of digital and social media, children and adolescent consumers of Spanish language media may rarely read newspapers.^50,51 Nevertheless, this pair of studies demonstrates that as a result of these differences across Spanish and English mass media, young Latinx consumers of Spanish media may have stigma reinforced. In contrast, youth who use English language media, which has been the focus of anti-stigma work, may encounter less pejorative content and more help-seeking resources.

Study Purpose

The current study builds on the aforementioned knowledge base by providing quantitative assessments of the contribution of media language, family language, and social preferences on shaping stigma in a socioeconomically and culturally diverse sample of Latinx adolescents. For two reasons, we hypothesize finding increased stigma among Latinx adolescents preferring Spanish media as an indicator of the type of mass media that they would encounter. The first reason is that the paucity of research in this area reflects a lack of scrutiny of Spanish media by mental health researchers and advocates. Being less subject to scrutiny may have allowed Spanish media to default to views and language that have historically stigmatized mental illness, which young people are susceptible to absorbing. The second reason concerns the developmental context of adolescence. Young adolescents begin discussing mental health issues at roughly the same time they begin focusing much more on youth culture that they consume through television/movies and music/radio. At the same time, they also start placing lower value on spending time with their family and their family’s social ties. Therefore, outside of the influence of Spanish media, use of Spanish language in the family or preferring Latinx social ties would not be associated with increased stigma in early adolescence. Instead, we hypothesize that external influences that are not controlled by the youth themselves but are part of the media context that they encounter results in increased stigma for some Latinx groups. Specifically, we propose that exposure to Spanish media that has not been as thoroughly vetted for stigmatizing messages will be associated with greater stigma in the youths exposed to it. These hypotheses are assessed across a range of stigma measures including responses towards peers with a mental illness and adolescent vignette characters who are described as experiencing a mental illness.

METHODS

Data on Latinx youth, language and social preferences, and mental illness stigma came from the baseline assessment of the Texas Stigma Study (2011–2015; Texas, U.S.), a longitudinal evaluation of an anti-stigma intervention among a large multi-ethnic sixth-grade sample. The selection of participants, design, and procedures of the intervention are described in detail elsewhere.⁵² Briefly, 14 schools in an urban area in Texas, U.S. agreed to participate in an anti-stigma intervention study where each school’s sixth-grade class was randomized to receive a mental illness anti-stigma intervention. Both parents/guardians and students gave active assent and consent for participation after receiving information about the study. Students were not included in the study without signed forms. Students and parents/guardians received a modest monetary incentive ($10 U.S. dollars) for returning signed forms and completing the surveys. Assessments were administered on laptop computers and offered in English or Spanish; all students chose the English survey.

Study Sample

A total of 1260 sixth-graders received a study packet and invitation to participate. Of those who responded (n=882), 751 (85%) agreed to participate in the study and completed the survey at baseline. The study sample is representative of publicly available classroom enrollment data across age, race/ethnicity, gender, and socioeconomic status; therefore, there is little evidence of non-response bias at baseline. As the current study focuses on the Latinx sample (n=328; 44.4% of all adolescents at baseline), participants with missing race/ethnicity and those identifying as a member of a race/ethnic group other than Latinx (n=423) were excluded. In 2011 when this sample was drawn, Texas public schools reported that Latinx students represented the largest enrolled group (51%).⁵³ As such, acculturation measures were administered in the same academic year in a sub-sample of Latinx adolescents who agreed to the 6-month follow-up assessment (n=204, or 62% of the 328 Latinx students at baseline); 179 Latinx adolescents completed all acculturation measures (88% of those agreeing to 6-month follow-up) representing the total analytic sample for the current analysis. Completers versus non-completers of acculturation measures did not significantly differ on gender, household income, familiarity with mental illness, or stigma outcomes.

Among the 179 Latinx adolescents in the analytic sample, mean age at baseline was 11.5 years; more than half were girls (Table 1). The majority identified as Mexican (93%) while the remainder identified as Puerto Rican (2%) or another heritage (5%). About two-thirds of the sample came from lower socioeconomic households, while 35% came from a household with income of $30,000 or more ($49,392, 2011 Texas household median income).⁵⁴ About 32% had a parent/guardian educational attainment of some college or more.

Table 1.

Characteristics of the total Latinx adolescent sample and across language and social preferences; Texas Stigma Study, 2011–2012 (N=179)

	Full Sample		Only English Media Preference		Any Spanish Media Preference
	N	%	N	%	N	%
Adolescent Characteristics
Age (M±SD)	11.5±0.6		11.6±0.6		11.5±0.6
Female	106	59.6	40	53.3	66	64.1
Household income ≥$30,000	57	35.4	31	43.7	26	28.9
Parent/guardian education ≥some college	53	32.1	28	38.4	25	27.2
Familiarity with mental illness (M±SD)	2.1±1.7		2.1±1.7		2.1±1.7
Latinx Ethnic Heritage
Mexican/Mexican-American	163	93.1	66	90.4	97	95.1
Puerto Rican	3	1.7	1	1.4	2	2.0
Other Latinx group	9	5.1	6	8.2	3	2.9
Media Language Preference
Only English	75	41.9		n/a		n/a
Any Spanish	104	58.1		n/a		n/a
Family Language Preference*
Only English	22	12.3	19	25.3	3	2.9
Any Spanish	157	87.7	56	74.7	101	97.1
Social preference
Only/mostly Anglos	75	41.9	37	49.3	38	36.5
About equal	55	30.7	21	28.0	34	32.7
Only/mostly Latinx	49	27.4	17	22.7	32	30.8

NOTE:

^*p<.05 for differences between media language preference groups

Measures

Using Link and Phelan’s conceptualization of stigma,⁵⁵ measures of mental illness stigma assessed the stereotyping component—undesirable characteristics and negative stereotypes applied to labeled persons—and the separation component—the sense of separation between groups (“us” versus “them”). Measures were adapted from existing instruments for measuring stigma among children and adolescents.⁵⁶ All stigma composite scales were pilot-tested with an ethnically diverse group of youths in the target age range and developed using exploratory factor analyses, as previously reported.⁵²

First, assessments of mental health knowledge and positive attitudes examined the pervasiveness of mental illness stereotypes (e.g., mental illness as a weakness) as cognitive knowledge structures in the general public (21-items; α=0.78).⁵⁷ Items included ‘People with mental illnesses tend to be violent and dangerous’ (1=Strongly Agree to 5=Strongly Disagree). Higher mean scores suggest better knowledge/positive attitudes (i.e., lower mental health stereotyping). Second, the separation component of stigma was assessed using the social distance scale (6-items; α=0.89)—the degree to which adolescents are unwilling to interact with a peer with mental illness in different contexts (e.g., work on a class project with a peer with a mental illness; 1=Definitely No to 4=Definitely Yes).⁵⁷ Higher mean scores suggest greater social distance.

Similar attitudes and behaviors towards two adolescent vignette characters described as meeting Diagnostic and Statistical Manual of Mental Disorders, IV⁵⁸ criteria for bipolar disorder and social anxiety disorder were also assessed.⁵² After reading each vignette, participants responded to whether they believed the character was a bad person (1=Not At All Likely to 4=Very Likely). Vignette-based social distance to each character was assessed using four items similar to those reported above (1=Definitely No to 4=Definitely Yes; α=0.92). Higher scores suggest greater social distance. Analyses were first run to assess the “bad person” and “social distance” outcomes for each character separately. No differences in the effect of media language, family language, and social preferences on the stigma outcomes were found between the two characters (i.e., bipolar versus social anxiety disorder). Therefore, responses for both characters were combined into a single “bad person” and “social distance” outcome to improve model parsimony and fit.

The Short Acculturation Scale for Hispanics⁵⁹ measured language and social preferences. Item responses about language preferences used a 5-point Likert scale ranging from ‘Only Spanish’, ‘More Spanish than English’, ‘Both Equally’, ‘More English than Spanish’, to ‘Only English.’ Similarly, responses to social preference items ranged from ‘All Latinx’, ‘More Latinx than Anglos’, ‘About Half and Half’, ‘More Anglos than Latinx’, to ‘All Anglos.’ Principal axis factor analysis suggested the data best fit on three factors: 1) media language preference; 2) family language preference; and 3) social preference. Items with high loadings (>0.40) on the same factor were summed to create composite scales after examining the content validity of the clustered items. Media language preference (3-items, scores 3–15; α=0.83) measured the preferred language for watching television/film and listening to radio/music where higher scores indicated greater Spanish language preference. Family language preference (3-items, scores 3–15; α=0.75) concerned the language used with parents/grandparents and at home; higher scores also indicated greater Spanish language preference. Finally, social preference (3-items, scores 3–15; α=0.68) assessed social embeddedness across Latinx and NL-white social contexts with higher scores indicating greater Latinx homophily. Due to non-normal distribution of the language scales (<5% reported a score of 10 or higher on the media language preference scale), the total score was categorized dichotomously as no differences in results were found between higher ordered tertiles and when using a continuous scale. Media language preference scores between 3–5 on the continuous scale were categorized as ‘0=Only English’ while scores >5 on the continuous scale were categorized as ‘1=Any (Some/Mostly/Only) Spanish’; likewise, family language preference scores were categorized. Similarly, the social preferences scale was not normally distributed and categorized as tertiles (0=Only/Mostly Anglos, 1=About equal, 2=Only/Mostly Latinx).

Analyses controlled for self-reported gender (female—referent) and parent/guardian reported household income (0≥$30,000—referent, 1<$30,000). Familiarity with mental illness was also controlled for using an adapted scale⁶⁰ where adolescents were asked about personal interaction with individuals who have a mental illness in different contexts: seeing a mentally ill person on television (least contact); having a friend/relative to living with someone who has a mental illness; and whether the adolescent has a mental illness (most contact). Combined item responses created a ranked familiarity score (0–6; α=0.57) indicating the most intimate level of contact reported.

Data Analysis

Differences by language and social preferences in stigma outcomes measured at baseline in sixth-grade prior to intervention were examined using a series of linear regression models. First, bivariate associations were examined between the media language, family language, and social preference variables and the stigma outcome variables. Covariates were then entered into the models individually with the language and social preference variables to examine if regression coefficients for the variables changed with the addition of each covariate. We tested for potential interactions between media language preference and the covariates in the fully adjusted models and none were statistically significant. Multi-collinearity was not observed between language and social preference variables; all correlations between variables were <0.4. All fully adjusted models include the language and social preference variables plus control for gender, household income, and familiarity with mental illness. P-values <0.05 were considered statistically significant. Stata SE 14.2 was used for all analyses.

RESULTS

The Latinx adolescent sample were diverse with respect to their combinations of preferences for media language, family language, and social ties (Table 1). While 17% (n=31) reported Spanish media/family language preference and also Latinx social ties (i.e., very strong Spanish/Latinx preferences), only 8% (n=14) reported English only media/family language preferences and NL-white social preferences (i.e., very strong English/Anglo preferences); the remaining 75% reported a mix of preferences indicating great biculturalism in the sample.

In terms of media language preferences, 22% reported English only preferences for all three media items, while no adolescent reported Spanish only preferences for all three media items on the 5-point Likert scale. Thus, Spanish media preferences did not reflect extremely strong preferences for media exclusively in Spanish but rather students who consumed “some” or “mostly” Spanish media. Latinx adolescents preferring media in English only versus any Spanish were comparable across sociodemographic factors, familiarity with mental illness, and social preferences (Table 1). The two groups significantly differed by family language preference only (p<0.05).

Media Language Preference and Stigma

Concerning stigma, results found significant variation by media language preferences among Latinx adolescents both before and after controlling for covariates (Table 2). Latinx adolescents with any level of Spanish media preference reported significantly less mental illness knowledge/positive attitudes at baseline compared to those preferring English only media (Table 2). This pattern was found in bivariate analyses (not shown) and after adjusting for family language and social preferences (p<0.01; Model 1). This difference between Spanish versus English media preferences remained statistically significant net of gender, household income, and prior familiarity with mental illness (p<0.01; Model 2).

Table 2.

Effects of language preferences in media (television/film, radio/music) and with family (parents/grandparents), and social preferences (Latinx versus Anglo) on beliefs and behaviors about mental illness among Latinx sixth-graders; Texas Stigma Study, 2011–2012 (N=179)

	Knowledge and positive attitudes (1–5)				Social distance to mental illness (1–4)
	Model 1		Model 2		Model 1		Model 2
	B	95% CI	B	95% CI	B	95% CI	B	95% CI
Media language preference
Only English	ref.		ref.		ref.		ref.
Any Spanish	-0.15**	−0.26, −0.04	−0.17**	−0.28, −0.07	0.30*	0.05, 0.55	0.32*	0.07, 0.56
Family language preference
Only English	ref.		ref.		ref.		ref.
Any Spanish	0.01	−0.16, 0.17	0.04	−0.13, 0.20	<0.01	−0.37, 0.38	−0.05	−0.42, 0.33
Social preference
Only/mostly Anglo	ref.		ref.		ref.		ref.
About equal	0.09	−0.03, 0.21	0.08	−0.05, 0.20	−0.13	−0.41, 0.14	−0.15	−0.43, 0.13
Only/mostly Latinx	−0.05	−0.17, 0.08	<0.01	−0.13, 0.13	−0.02	−0.31, 0.26	−0.14	−0.44, 0.15
Adolescent Characteristics
Female			−0.07	−0.17, 0.04			0.02	−0.21, 0.26
Household income <$30,000			0.05	−0.06, 0.17			−0.20	−0.46, 0.06
Familiarity with mental illness			0.06***	0.03, 0.09			−0.15***	−0.22, −0.08
Constant	3.59***	3.44, 3.73	3.46***	3.27, 3.66	2.07***	1.74, 2.41	2.56***	2.13, 2.99

NOTE: Model 1 includes language and social preference variables. Model 2 includes Model 1 variables plus controls for gender, household income, and familiarity with mental illness. All scales for outcome variables are scored such that a higher score indicates more of the named construct.

^*p<.05,

^**p<.01,

^***p<.001

With respect to desired social distance, or separation from peers with mental illness, those with any Spanish media preference reported significantly greater social distance compared to those with English only preference before and after adjusting for family language and social preferences (p<0.05; Model 1). Latinx adolescents with any Spanish versus English only media preference reported significantly greater social distance even after controlling for gender, household income, and familiarity with mental illness (p<0.01; Model 2).

These results concerning the stigma components of stereotyping and separation were replicated in the vignette-based outcomes (Table 3). Before and after controlling for family language and social preferences, adolescents with any Spanish media preference more often believed that the characters experiencing mental illness were bad people compared to those with English only media preference (p<0.05; Model 1), as well as net of gender, household income, and familiarity with mental illness (p<0.01; Model 2). Finally, Latinx adolescents with any Spanish versus English only media preference also desired greater social distance from the characters experiencing mental illness before and after controlling for all covariates (p<0.05; Models 1–2).

Table 3.

Effects of language preferences in media (television/film, radio/music) and with family (parents/grandparents), and social preferences (Latinx versus Anglo) on vignette-related beliefs and behaviors about mental illness among Latinx sixth-graders; Texas Stigma Study, 2011–2012 (N=179)

	Vignette is a bad person (1–4)				Social distance to vignette (1–4)
	Model 1		Model 2		Model 1		Model 2
	B	95% CI	B	95% CI	B	95% CI	B	95% CI
Media language preference
Only English	ref.		ref.		ref.		ref.
Any Spanish	0.28**	0.07, 0.49	0.27*	0.05, 0.49	0.24*	0.01, 0.47	0.26*	0.01, 0.51
Family language preference
Only English	ref.		ref.		ref.		ref.
Any Spanish	0.12	−0.20, 0.44	0.04	−0.29, 0.38	0.07	−0.28, 0.43	0.06	−0.32, 0.43
Social preference
Only/mostly Anglo	ref.		ref.		ref.		ref.
About equal	−0.09	−0.32, 0.15	−0.06	−0.31, 0.19	−0.17	−0.43, 0.09	−0.19	−0.46, 0.09
Only/mostly Latinx	−0.21	−0.46, 0.03	−0.24	−0.50, 0.03	0.07	−0.20, 0.34	0.03	−0.26, 0.33
Adolescent Characteristics
Female			−0.08	−0.29, 0.14			−0.05	−0.29, 0.19
Household income <$30,000			−0.10	−0.33, 0.13			−0.08	−0.33, 0.18
Familiarity with mental illness			−0.02	−0.08, 0.04			−0.08*	−0.15, −0.01
Constant	1.54***	1.25, 1.82	1.71***	1.32, 2.10	1.99***	1.68, 2.31	2.28***	1.85, 2.71

NOTE: Model 1 includes language and social preference variables. Model 2 includes Model 1 variables plus controls for gender, household income, and familiarity with mental illness. All scales for outcome variables are scored such that a higher score indicates more of the named construct.

^*p<.05,

^**p<.01,

^***p<.001

Family Language and Social Preferences and Stigma

There is no support in our data of family language and social preferences being associated with any of the stigma outcomes at baseline either before or after controlling for covariates (Table 2–3). Nevertheless, we observe a marginally significant trend with respect to social preferences, after controlling for media and family language preferences, and all covariates. Latinx homophily/social preference compared to those with a social preference for NL-white persons was marginally associated with a lower belief that the vignette characters were bad people, adjusting for controls (p<0.10; Table 3).

Sensitivity Analyses

We conducted sensitivity analyses where we examined whether strength of ethnic identity (measured at the 24-month follow-up; n=137) as a binary measure (1=‘Very close Latinx identity’; 0=All other closeness levels’) explained away, or moderated, the associations between media language preference and the stigma outcomes (results available upon request). Results found that Latinx closeness was not associated with any stigma outcome, and the effects of media language preferences on stigma remained unchanged. Concerning moderation, one statistically significant interaction (p=0.02) was found between closeness and media language preference on mental illness knowledge/positive attitudes. Among those with very close Latinx identity, no statistically significant difference was observed in knowledge/positive attitudes by media language preference; in contrast, among those without a very close Latinx identity, any Spanish versus English only media preference was associated with significantly lower knowledge/positive attitudes.

DISCUSSION

Important to mental health research and policy, particularly mental health service utilization and help-seeking among Latinx adolescents, our study provides new evidence on the influence of media language preferences on mental illness stigma. Together the evidence supports our hypothesis that Spanish media preference is associated with greater mental illness stigma among U.S. Latinx adolescents rather than family language or social preferences, at least in this period of development of early adolescence. These patterns are a compelling call for future research to examine mental illness-related messaging in Spanish mass media in order to further elucidate if these patterns are driven by more existing stigmatizing messages in Spanish mass media as compared to English media in the U.S. Specifically, Latinx adolescents preferring Spanish media report less mental illness knowledge/positive attitudes and more social separation towards people with mental illness in all bivariate and multivariate analyses. No effects on stigma were found for family language or social preferences in any analyses. These patterns were observed across the different facets of stigma (stereotyping, social separation) that are thought to influence mental health help-seeking as well as comprehensive measures that elicit responses about both peers and vignette characters with mental illness. Our results suggest that Latinx adolescents with Spanish media preference are at risk of not receiving anti-stigma messaging or information about help-seeking through their preferred media, a salient point of intervention for combating mental illness stigma; instead, they may be receiving negative messaging.

To our knowledge, Spanish language media outlets other than newspapers in California and Spain^49,50 have not been examined with respect to stigma proliferation. In contrast, previous research on how media influences stigma in younger populations has primarily focused on English language media and has been instrumental for informing anti-stigma strategies (Figure 1A), concentrating anti-stigma efforts mostly in English language mass media.^33,35,36,61 Although stigma remains a complex, persistent social problem, small gains in how mental illness is portrayed in English mass media have changed cultural norms towards better public mental health. If Spanish language mass media has been historically excluded from these efforts, thereby introducing a disparity, this exclusion would help explain the patterns that we find in our data among Latinx adolescents (Figure 1B). It is therefore plausible that more pejorative representations of mental illness exist in Spanish versus English media for young Latinx populations with fewer help-seeking resources provided.⁴⁹ However, no study to date has examined this question for Latinx adolescents, or has directly intervened with or involved the Spanish/Latinx media industry.

Figure 1.

Summary and diagram of media-related factors impacting mental illness stigma among Latinx populations

Several limitations of this study warrant discussion. First, although language and social preferences are dynamic over time, this cross-sectional study measured these preferences at one time in sixth-grade. An assessment of language and social preferences among Latinx older adolescents and adults may further our understanding of these relationships across the life course and cohorts of Latinx populations as dynamic processes.⁶² Such an assessment can help determine if stigma that is developed from media influence during adolescence diffuses in networks through family language and social preferences over time. Second, the study’s local context did not permit assessment of sensitive immigrant-related factors (e.g., birthplace, duration living in the U.S.); thus, we cannot assess how these factors relate to media language preferences (e.g., U.S.-born versus foreign-born). Nevertheless, language and social preference measures offered a practical, intentional alternative for estimating meaningful Latinx heterogeneity as well as factors that may influence stigma. Third, media language preference was based on self-report rather than a direct observation of consumption. Assessments of media consumption including specific mediums (i.e., basic cable, Netflix, movies), content (i.e., cartoon, telenovela, mature audience), and hours consuming (i.e., dose-response) could improve measurement for identifying the scope of the disparity. Improving measures may also promote collaboration with media/communication experts. Finally, as the sample was predominantly Mexican-heritage, results may not generalize to Latinx adolescents of other heritage. Future research should examine other Latinx subgroups, because, for example, some Spanish networks cater to Latinx groups broadly (e.g., Univision, Telemundo), while other networks are region-specific (e.g., CentroamericaTV, Caracol) though subscription-based. Despite these limitations, this study provided a comprehensive battery of stigma measures that assessed the kinds of knowledge, attitudes, and behaviors that are thought to facilitate or impede mental health help-seeking in a culturally diverse sample of Latinx adolescents.

It is unlikely that anti-stigma efforts have fully penetrated Spanish language media and its consumers, possibly leading to high stigma and low help-seeking in the young and diverse, growing Latinx population. Our study is the first to our knowledge to identify a relationship between media language preference and mental illness stigma, demonstrating disadvantages among young Latinx consumers of Spanish media. This finding also has implications for consumers of mass media in other non-English languages. In light of this novel finding and because data on Spanish media and stigma is extremely little for adult consumers of Spanish media and virtually nonexistent for child and adolescent consumers of Spanish media, we highlight four key areas for future research to invest its efforts: (1) the presence of mental illness stigma and help-seeking resources in Spanish media, and comparatively to English media; (2) the inclusion of Latinx participants and language preference measures in stigma research, particularly interventions/trials; (3) the effectiveness of anti-stigma interventions for populations with varying language and cultural preferences; and (4) the adaptation of evidence-based anti-stigma interventions to language and cultural preferences. The inclusion of Latinx culture and linguistics in the development, evaluation, and dissemination of anti-stigma strategies, to our knowledge, is also unknown. Attentiveness to these gaps in stigma and mental health services research is critical and will help facilitate greater inclusion of Latinx populations, thereby achieving a more complete public mental health competency.