Socio-Cultural Subgroups of Latina/o Immigrants: A Latent Profile Analysis

Abstract

Socio-cultural factors, such as familismo, social support, machismo, and multigroup ethnic identity, are strongly associated with Latina/o immigrants’ alcohol misuse and depressive symptoms. However, research has rarely explored whether unobserved groups of Latina/o immigrants with similar socio-cultural factors exist. Latent Profile analysis can illuminate which subgroups to target, or which socio-cultural factors need to be supported, to have an impact on the prevention and treatment of alcohol use and/or depression in the Latina/o immigrant population. Cross sectional data from on ongoing longitudinal investigation was utilized (N= 518 Latina/o adults living in Miami-Dade County, Florida and have immigrated to the US within one year prior to assessment). Latent Profile Analyses (LPA) were conducted utilizing validated measures of familismo, social support, neighborhood collective efficacy, ethnic identity, machismo, caballerismo, and marianismo. The LPA revealed three, significantly different profiles: (1) low socio-cultural protection (n=155, 29.98%), (2) high socio-cultural protection (n=21, 4.06%), and (3) high socio-, low-cultural protection (n=341, 65.96%). Profile membership was associated significantly with immigrant documentation status, education level, and past family history of substance use. Results indicate that Latina/o immigrants in the low socio-cultural protection group had significantly higher alcohol use compared to high socio-, low cultural protection group. No significant differences were found for depressive symptoms. We discuss implications of our findings and encourage researchers to continue to unpack the complexities associated with socio-cultural factors and Latina/o mental and behavioral health. Specifically, research should focus on socio-cultural factors can provide protection from negative health outcomes and increase resiliency among this population.

Alcohol misuse and depression tend to co-occur and are strongly associated with one another (Lai, Cleary, Sitharthan, & Hunt, 2015). Alcohol use severity and depressive symptoms exacerbate each other’s symptomatology and can lead to increased adverse health consequences at the individual and community level (Sullivan, Fiellin, & O’Connor, 2005). Differences in alcohol use severity and depressive symptoms have been demonstrated among racial/ethnic groups (Evans, Grella, Washington, & Upchurch, 2017). For instance, recent evidence suggests Latino/a’s are more likely to be at risk for depression than their non-Latina/o White counterparts (Rodriquez et al., 2018). Findings indicate that Latina/o immigrants are susceptible to experiencing stress associated with adverse circumstances stemming from the migration process. The adverse circumstances include family separation, loss of social supports and social status, legal status, discrimination, economic instability, and language barriers which in turn increases their risk for depression (Ding & Haragraves, 2009; Garcini et al., 2016; Gutierrez-Vazquez, Flippen & Prado, 2018; Portes & Rumbaut, 1996). This association is further exacerbated by engaging in health compromising behaviors including alcohol misuse (De La Rosa et al., 2012). Previous studies have reported that Latina/o’s (compared to non-Latina/o ethnic groups) are more likely to engage in heavy drinking (Field et al., 2015) and experience alcohol-related disparities, including higher rates of chronic liver disease and a greater likelihood to receive a citation for drinking and driving (National Institute on Alcohol Abuse and Alcoholism, 2019).

To adequately address and prevent depression and alcohol misuse in the Latina/o community, an ecological approach that comprehensively addresses multiple levels of influence including individual, interpersonal, community, and cultural factors, is warranted (Sudhinaraset, Wigglesworth, & Takeuchi, 2016).

Individual Factors

Demographic differences such as age and gender have been shown to predict greater substance use and depressive symptoms among Latina/os: higher prevalence of alcohol misuse among males, more depression among females, and younger age being associated with both greater substance use (De La Rosa, Dillon, Sastre, & Babino, 2013) and depressive symptoms (Alarcón et al., 2016). Furthermore, alcohol use and depression prevalence differs by country of origin among Latina/os (Alegría et al., 2008). For example, Latinas from South and Central America have been shown to have higher rates of binge drinking compared to their Mexican and Cuban counterparts, while Mexican men have been shown to have higher rates of alcohol use disorder compared to Central American and Cuban men (Chartier & Caetano, 2011; Castaneda et al., 2019). Other important sociodemographic factors that can influence both substance use and depression include education, marital status, and undocumented immigration status (Ramos-Sánchez, 2009).

Interpersonal Factors

The presence of interpersonal factors—such as familism (“familismo”) and social support have been found to be particularly relevant protective factors against depressive symptoms and alcohol misuse among Latina/o immigrants (Cano, Sánchez, et al., 2017; Perreira et al., 2019). Familismo encompasses Latina/o families’ emphasis on ‘desirability to maintain close relationships’, obligation to take care of one another, and ‘reliance on communal interpersonal reflection to define self’ (Knight et al., 2010). Literature suggests that the protective effects of familismo may function to buffer against substance use and adverse mental health outcomes by facilitating communication among family members and increasing the perception of social and emotional support (Calzada, Tamis-LeMonda, & Yoshikawa, 2013; Campos & Kim, 2017; Cervantes & Kuotantos, 2020; Valdivieso-Mora et al., 2016). Maintaining close relations with nuclear and extended family members throughout the lifespan is a hallmark of Latina/o culture. As such, familismo is a particularly salient protective factor against alcohol misuse and depressive symptoms among Latina/o immigrants (Cano et al., 2018; Dillon, De La Rosa, Sastre, & Ibañez, 2013), and has even been found to moderate intergenerational acculturation conflict (Pina-Watson, Gonzalez, & Manzo, 2019).

Social support refers to the degree to which members of an individual’s social network serve particular functions (e.g., provide information and guidance, affection, material aid, and empathetic understanding (Cohen, 2004; Sherbourne & Stewart, 1991). Prominent theories of health behavior propose that higher levels of social support may help an individual manage psychosocial stress more effectively, in turn, reducing the probability of using maladaptive coping strategies, such as substance use (Cohen, 2004). Additionally, investment in social relations facilitates the flow of information which can ease the immigration process and reduce stress (Portes & Rumbaut, 2006). Social support has been found to be an important mitigating factor in buffering immigration stress among Latino immigrants (Concha, Sanchez, de la Rosa, & Villar, 2013; Salgado, Castañeda, Talavera, & Lindsay, 2012) and a protective factor against alcohol use and adverse mental health outcomes in this population (Sanchez et al., 2016).

Community Factors

There is a growing body of evidence linking neighborhood level factors to health outcomes including substance use and mental health outcomes (Diez-Roux & Mair, 2010; Vaeth, Caetano, & Mills, 2016). The community has characteristics that can influence substance use and mental health outcomes including neighborhood collective efficacy. Neighborhood collective efficacy can be defined as the social linkages that exist among community members (Sampson, Raudenbush, & Earls, 1997). It is conceptualized as consisting of combination of informal social control (i.e., neighbors’ willingness to intervene in an effort to maintain social order in the neighborhood) with social cohesion and trust (i.e., neighbors’ capacity to collectively bond for common rules and values) (Sampson et al., 1997). Residing in neighborhoods with high neighborhood collective efficacy can foster “loose” interpersonal connections that potentially lead to positive structural benefits, including better access to social and health services that can mitigate health outcomes including substance use and depression among Latina/os (Vaeth, Caetano, & Mills, 2016; Vega, Ang, Rodriguez, & Finch, 2011). Conversely, neighborhoods with poor physical infrastructure, low social cohesion, and inadequate health care resources have been found to place residents at risk for alcohol misuse, adverse mental health outcomes, as well as other health risk behaviors (Latkin & Curry, 2003).

Cultural Factors

Finally, cultural factors including cultural norms and identity are constructs that should be accounted for when attempting to understand and address mental and behavioral health outcomes among Latina/os. Traditional gender roles in Latina/o culture including machismo and marianismo have been found to substantially impact health behaviors in this population. Machismo is a traditional Latino male gender norm that encompasses hyper masculinity, aggression, and dominance (Arciniega, Chicago, Tovar-Blank, & Tracey, 2008) and has been linked to depression and binge drinking among men (S. J. Schwartz et al., 2015; Perrotte, Baumann, & Knight, 2018; Perrotte & Zamboanga, 2019).

Scholars have argued that viewing male gender norms from the unidimensional aspects of machismo in the context described above is narrow and provides negative perspective of a complex gender norm (Arciniega, Anderson, Tovar-Blank, & Tracey, 2008; Falicov, 2010; Torres, Solberg, & Carlstrom, 2002). Researchers have posited that Latino masculinity is more accurately represented in a multi-faceted manner. For instance, another dimension of machismo is caballerismo, which encompasses attributes such as chivalry, respect, honor, and being the family provider (Arciniega et al., 2008). Previous studies have found caballerismo to be associated with health-promoting outcomes including a greater sense of well-being and increased problem-solving (Arciniega et al., 2008), which have been found to be protective against problematic drinking (Corbin, Farmer, & Nolen-Hoeksema, 2013). Notably, recent evidence suggests that the strong link between hypermasculinity and drinking is indeed multifaceted and requires further inquiry to disentangle its effects on alcohol (Zamboanga, Audley, Iwamoto, Martin, & Tomaso CC, 2017).

Marianismo, which encompasses traditional female gender norms, emphasizes the role of women as family- and home-centered, while encouraging passivity, self-sacrifice, and chastity (Castillo, Perez, Castillo, & Ghosheh, 2010). Previous studies have suggested that marianismo among Latina women is associated with lower levels of emotional well-being and increased negative emotions, particularly with regard to higher depression symptoms (Piña-Watson, Castillo, Ojeda, & Rodriguez, 2013). Notably, adherence to traditional Latina gender norms may have health compromising effects as well. Aspects of marianismo, such as respecting and honoring women, and the belief that women are the source of strength for their families, have demonstrated to be protective factors against substance use among Latinas (Sanchez, Vanderwater, & Hamilton, 2017).

It is important to note that the influence of gender norms on Latina/os is not gender specific. Based on prior research higher levels of marianismo in males has been associated with higher level of depression while adherence to machismo among females has been linked to negative mental health and behavioral health outcomes including higher levels of anxiety and unsafe sex practices (Nuñez et al., 2016; Cianelli, Ferrer, & McElmury, 2008).

Ethnic identity refers to one’s sense of belonging to and feelings of connectedness to a group of people. It encompasses an understanding of the meaning of membership with, and positive attitudes towards, an ethnic group as well as familiarity with the history and culture of that group and involvement in its practices (Phinney and Ong, 2007). Ethnic identity is thus conceptualized as a multi-dimensional, dynamic construct that develops over time via exploration and commitment to one’s ethnic group (Phinney and Ong, 2007). A meta-analysis of 184 studies found positive associations between ethnic identity and well-being, with stronger associations for adolescents and emerging adults (Smith and Silva, 2011). Specifically, ethnic identity has been linked with greater positive psychological well-being and lower depressive symptoms and alcohol use among Latina/os and other immigrant groups in the U.S. (Cobb et al., 2018; Perreira et al., 2019; Thibeault. Stein, & Nelson-Gray et al., 2018).

Objectives

While it is well understood that the aforementioned socio-cultural factors are associated with and affect Latina/o immigrant mental and behavioral health (Alegría, Álvarez, & DiMarzio, 2017), a person-centered approach to understanding the effects of socio-cultural factors in Latina/os is not as frequently explored. A person-centered approach, Latent Class or Latent Profile analysis, classifies similar individuals from a population into subgroups based on complex patterns of variables (Collins & Lanza, 2010; Howard & Hoffman, 2018). Compared to a variable-centered approach that detect correlations in a single set of parameters for a sample, a person-centered approach allows researchers to be more specific in identifying who needs to be targeted within the sample (Howard & Hoffman, 2018). Lanza and Rhoades (2013) further described how a person-centered approach overcomes methodological challenges related to traditional subgroup analysis (i.e., moderation), such as ‘high Type I error rate, low statistical power, and limitations in examining higher-order interactions’ (Lanza & Rhoades, 2013). Several studies exist which have looked at different socio-cultural factors in Latina/o adolescents (Prado et al., 2013), emerging adults (Vaughan, Wong, & Middendorf, 2014), or a specific subpopulation (e.g., Mexican-Americans) (Safa et al., 2019). However, these studies have not focused on individual, interpersonal, community, and cultural variables to classify subgroups (e.g., only focusing on traditional gender roles;Vaughan et al., 2014). They also have not looked at these variables in a diverse group of Latina/o immigrant adults.

Thus, the objectives of this study were to: 1) identify subgroups of Latina/o adult immigrants based on socio-cultural factors utilizing Latent Class analysis, 2) if subgroups exist, test for significant differences between the subgroups, 3) determine demographic predictors of subgroup membership, and 4) assess whether subgroup membership can predict alcohol use severity and depressive symptoms. Ultimately, our findings may inform the prevention and treatment of alcohol misuse and/or depression in the Latina/o immigrant population. Specifically, a person-centered approach may illuminate which subgroups to target, or which socio-cultural factors need to be supported, to have an impact on mental and/or behavioral health.

Methods

This study is a secondary data analysis of cross-sectional data (N=518) from an on-going longitudinal investigation examining pre- to post-immigration alcohol use trajectories among adult Latina/o immigrants during their first decade in the US. The main objective of the original study is to examine pre- to post-immigration alcohol use behaviors and alcohol use trajectories of early adult recent Latina/o immigrants. The study was appr17 latoved by the Institutional Review Board (IRB) of a large public university in South Florida.

Participants

Baseline inclusion criteria was: (1) being an adult (18 years or older), (2) having immigrated to the US from a Latin American country within one year prior to baseline assessment, (3) self-identifying as Latina/o, (4) living in Miami-Dade County, Florida, (5) and having the intention to stay in the U.S. Notably, the present study was conducted with the 4^th wave of the on-going study when participants had been in the U.S. approximately 10 years.

Data Collection

Respondent-driven sampling was the primary recruitment strategy. Specifically, we asked each participant (the seed) to refer three individuals in his/her social network who met eligibility criteria. Seeds were recruited via fliers and in-person throughout Miami-Dade County across neighborhoods and businesses with substantial Latina/o recent immigrant populations as well as community-based agencies, local Latina/o festivals, health fairs, soccer fields and other parks. This procedure was followed for a maximum of three legs per seed. Efforts were made to recruit participants across a broad range of zip codes in Miami-Dade in an effort to increase the representativeness of the sample to the general geographic region.

Trained bilingual research staff obtained consent and administered the computer assisted personal interviews in a safe location agreed upon by both the interviewer and participant. Approximately, 80% of the interviews were conducted in the participant’s home while the remaining were completed in other locations such as a participant’s place of employment, restaurant or coffee shop.

Each interview was approximately one hour in duration. For their participation, participants were compensated $50. Seeds received $20 for each successfully recruited participant (with a possible maximum of 3 participants). All questionnaires were administered in Spanish. Further methodological details of this investigation are detailed elsewhere (De La Rosa et al., 2012).

Measures

Individual.

Age, gender, country of origin, education, documentation status, marital status, and family history of alcohol use were all included as covariates. Age was measured as a continuous variable, while gender was assessed as a dichotomous variable (1=male and 0=female). Country of origin was categorized into three groups (1=Cuban, 2=South American, 3=Central American). Educational status was recoded as a dichotomous variable (0=high school or less and 1= some college training or more). Immigration status was recoded as a dichotomous variable (0= temporary or undocumented and 1= permanent or resident status). Marital status was recoded as a dichotomous variable (0= single, 1= married/living with partner). Family history of substance use was measured (0= no and 1= yes) if the participants reported that biological or non-biological mother or father had been an alcoholic of problem drinker at any time in their life.

Interpersonal.

Familismo and Social Support encompassed interpersonal factors. Familismo was assessed by the Family Support Subscale of the Mexican American Cultural Values Scale (MACVS) (Knight et al., 2010). This subscale contains 6 items on a 5-point Likert scale ranging from 1=not at all to 5=completely. Example items include: (1) ‘It’s always important to be a united family,’ and (2) ‘Family provides a sense of security because they will always be there for you.’ A total scale score was calculated by obtaining a mean value with higher values indicating placing greater value on family support. In the present study, the scale demonstrated good internal consistency (a =.88). Social support was assessed by the Medical Outcomes Study (MOS) Social Support Survey (Sherbourne & Stewart, 1991). The instrument contains 19 items set on a five-point Likert-type scale ranging from 1 = none of the time to 5 = all of the time, with higher scores indicating more social support. Example items include: (1) ‘Someone you can count on to listen to you when you need to talk,’ and (2) ‘Someone to turn to for suggestions on how to deal with a personal problem.’ The total score of the four subscales of the MOS was utilized and demonstrated high internal consistency (a =.97).

Community.

Neighborhood Collective Efficacy was assessed by the Neighborhood Collective Efficacy Scale (Sampson, Raudenbush, & Earls, 1997), which consists of two 5-item subscales: the Social Cohesion and Informal Social Control. Responses are on a 5-point Likert Scale ranging from 1= strongly disagree to 5=strongly agree. Example items include: (1) ‘If there was a fight in front of your house and someone was being beaten or threatened, how likely is it that your neighbors would break it up?’ and (2) ‘If a group of neighborhood children were skipping school and hanging out on a street corner, how likely is it that your neighbors would do something about it?’ The scale indicated high internal consistency (a =.85).

Cultural.

Ethnic identity was assessed using the Multi-Group Ethnic Identity Measure (Roberts et al., 1999). The scale has 12 items on a 5-point Likert scale ranging from 0= strongly disagree to 4= strongly agyee. Example items include: (1)’I have spent time trying to find out more about my ethnic group, such as its history, traditions, and customs,’ and (2) ‘I have a lot of pride in my ethnic group.’ A total scale score was calculated by obtaining a mean value with higher values indicating greater ethnic identity. The scale exhibited high internal consistency (a =.90). Traditional male gender norms were measured for both male and female participants using the Machismo and Caballerismo Scale, a widely recognized and validated 20-item bi-dimensional scale that measures machismo and caballerismo on two separate subscales (Arciniega, Anderson, Tovar-Blank, & Tracey, 2008). Items were measured using a 7-point Likert scale ranging from 1 =not at all to 7=very much so. Example items for the machismo subscale include ‘it’s important not to be the weakest man in a group’ and ‘real men never let their guard down.’ Items on the caballerismo subscale include ‘men should be affectionate to their children’ and ‘family is more important than the individual.’ Subscale scores were measured by calculating mean values, with higher values indicating greater adherence to traditional gender norms. The machismo subscale demonstrated good internal consistency (a =.83), while the caballerismo scale indicated marginal internal consistency (a =.63).

Traditional female gender norms were assessed for both females and male participants via the Marianismo Beliefs Scale (Castillo, Perez, Castillo, & Ghosheh, 2010). The instrument contains 24 items set on a four-point Likert-type scale ranging from 1 = strongly disagree to 4 = strongly agree, with higher scores indicating greater adherence to mariansimo. The measure contains 5 subscales: Family Pillar, Virtuous and Chaste, Subordinate to Others, Self-silencing to Maintain Harmony, and Spiritual Pillar. Example items include: (1) ‘A Latina is considered the main source of strength of her family,’ and (2) ‘A Latina should not speak out against men.’ A total scale score was calculated using mean values and revealed good internal consistency (a = 81).

Depressive symptoms.

Depressive symptoms were measured with the Center for Epidemiologic Studies Depression (CES-D) Scale (Andresen, Malmgren, Carter, & Patrick, 1994). The CES-D is a 10-item scale measured on a 4-point Likert scale ranging from 0=Rare or none of the time less than 1 day to 3=Most or all of the time/5-7 days. Participants were asked to rate how often they experienced specific feelings over the past week. Example items include: (1) ‘I felt lonely,’ and (2) ‘I felt depressed.’ A sum scale score was calculated with higher scores indicating greater levels of depressive symptoms. The CES-D indicated marginal internal consistency (a =.68).

Alcohol use severity.

Alcohol use severity was measured with the Alcohol Use Disorder Identification Test (AUDIT), which has been validated in Spanish (Babor, Biddle-Higgins, Saunders, & Monteiro, 2001). The AUDIT consists of 10-items with varied response choices on a Likert-scale. Summed scores range from 0 to 40 with higher scores indicating higher alcohol use severity. An example item is ‘Has a relative or friend or a doctor or another health worker been concerned about your drinking or suggested you cut down?’. AUDIT revealed good internal consistency (α =.77).

Analysis

To fulfill the first objective, we utilized Latent Profile Analysis (LPA). LPA is a mixture modeling approach which identifies latent, or unknown, subgroups in the study sample (Lanza & Rhoades, 2013; McLachlan & Peel, 2000; Titterington, Smith, & Makov, 1985). The subgroups are defined by individual characteristics (Collins & Lanza, 2010; Lanza & Rhoades, 2013) which in this study are familismo, social support, neighborhood, and culture. We sequentially ran models with two, three, four, and five subgroups and employed several criteria to choose the optimal number of latent subgroups. Firstly, a model with a smaller value for Bayesian Information Criterion (BIC) and Sample Size Adjusted BIC (SSABIC), compared to the previous model, was considered preferable. Secondly, a model with a statistically significant (p < .05) Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) signified that the model was not statistically different from the previous model. Thirdly, entropy, which is a measure of the degree of profile separation, closer to 1.0 indicated greater separation of subgroups (Nagin, 2005).

To fulfill the second objective, we utilized mean plots and overall Wald Tests for Mean Equality. If the overall Wald test was significant (p <.05), post-hoc comparisons were run to understand which specific subgroup comparisons were statistically different (Wickrama, Lee, O’Neal, & Lorenz, 2016). To fulfill the third objective, predictors were added to the optimal profile solution utilizing an automatic, three-step approach (Muthén & Asparouhov, 2015). In this approach, the statistical software estimates the effects of the predictors while taking into account uncertainty of profile membership. The predictors in our model included: age, gender, education, documentation status, marital status, and family history of alcohol use. Logistic regression coefficient estimates and standard errors of the predictors on the profile membership were transformed into odds ratios with 95% confidence intervals. Confidence intervals without which did not include the number one demonstrated significance between predictor and subgroup. To fulfill the fourth objective, to determine the effect of subgroups on alcohol and depression, we utilized a three-step procedure (DU3STEP in MPlus) because it is the most recommended for continuous outcomes (Muthén & Asparouhov, 2015).

Results

Descriptives

All variables were normally distributed, as demonstrated by skewness and kurtosis. The present study was conducted with a sample of (N = 518) Latina/o immigrants (Table 1 demonstrates the participant characteristics in the overall sample). The sample was fairly representative of the Miami-Dade County Latino community, which is 53.7% Cuban, 17.8% South American, 13.6 % Central American, and 3.5% Other Caribbean (Statistical Atlas, 2018). Participants in the present sample represented 17 Latin American countries. The distributions by country/region of origin were as followed: 41.8% Cuban, 28.2% South American (i.e., 17.5% Colombia, 2.8% Peru, 2.6% Venezuela, 1.9% Argentina, etc.) and 30% Central American (12.4% Honduras, 8.7% Nicaragua, 3.4% Guatemala, 1.9% El Salvador, etc.). US Census data has indicated substantial national shifts in immigration patterns with steep increases in immigrants from Central and South America arriving in the U.S. over the past decade (Noe-Bustamante, 2019). For instance, rates of immigrants from South American countries such as Venezuela have increased up to 76% to 421,000, while immigrants from Central American countries such as Guatemala have increased by 37% to 1.4 million (Noe-Bustamante L, 2019). This likely explains the over-representation of South and Central Americans in this immigrant sample.

Table 1.

Participant Characteristics.

Participant Characteristics	Data (N= 518)
Predictors
Age (M, SD, Range)	35.23, 5.08, 24-51
Gender (% Female)	45.9
Country of Origin (%)
Cuban	41.8
South American	28.2
Central American	30.0
Documentation Status (%)
Temporary/Undocumented	21.1
Permanent/Citizen	78.9
Greater than High School Education (%)	48.3
Married (%)	68.3
Family History of alcohol abuse (%)	9.6
Interpersonal Variables
Familismo (M,SD, Range)	4.59(0.45), 1-5
Social Support (M,SD, Range)	4.09 (0.66), 1-5
Community Variable
Neighborhood Collective Efficacy (M, SD, Range)	3.30 (0.47), 1-5
Cultural Variables
Multi-group Ethnic Identity (M, SD)	2.39(0.65), 0-4
Machismo (M, SD)	2.38(1.30), 1-7
Caballerismo (M, SD)	6.53(0.63), 1-7
Marianismo (M, SD)	2.53(0.32), 1-4
Outcomes
Depression (M,SD)	14.85(4.25), 10-40
AUDIT (M, SD)	3.99(3.76), 0-40

Latent Profile Analysis

Table 2 demonstrates the LPA. In summary, a 3-profile solution was determined to be optimal. None of the solutions provided a significant LMR-LRT; however, other criteria were utilized. The 3-profile solution had a smaller BIC and SSABIC compared to the 2-profile solution, and the highest entropy compared to the 2, 4, and 5-profile solution, respectively. Figure 1 demonstrates the mean plot of the 3-profile solution. Table 3 demonstrates the overall Wald Tests for Mean Equality, and Wald Tests post-hoc comparisons. In summary, across the three subgroups, only familismo and machismo were significantly different, and neighborhood collective efficacy and marianismo were not significantly different. Social support, multi-group identity, and cabellerismo varied in their significance in subgroup comparisons. The profiles were named Low Socio-Cultural Protection (Profile 1), High Socio-Cultural Protection (Profile 2), and High Socio-, Low Cultural Protection (Profile 3).

Table 2.

Optimal Profile Solution.

Fit statistics	2 Profiles	3 Profiles	4 Profiles	5 Profiles
BIC	6223.794	6110.787	6040.867	5968.924
SSABIC	6153.961	6015.561	5920.248	5822.911
Entropy	0.927	0.935	0.866	0.870
Adj. LMR-	0.0001	0.0350	0.0000	0.0253
LRT (p-value Group size (n, %)
Profile 1	161, 31.14%	155, 29.98%	153, 29.59%	37, 7.16%
Profile 2	356, 68.86%	21, 4.06%	261, 50.48%	117, 22.63%
Profile 3		341, 65.96%	23, 4.45%	21, 4.06%
Profile 4			80, 15.47%	261, 50.48%
Profile 5				81, 15.67%

Note.BIC= Bayesian Information Criteria, SSABIC= Sample Size Adjusted Bayesian Information Criteria, Adj. LMR-LRT= Adjusted Lo-Mendell-Rubin Likelihood Ratio Test.

Figure 1. Mean Plot of Optimal Profile Solution.

Note. 0= Familismo, 1= Social Support, 2=Neighborhood Efficacy, 3= Multi-Group Ethnic identity, 4= Machismo, 5= Caballerismo, 6= Marianismo.

Table 3.

Wald Tests for Mean Equality.

	Profile 1: Low Socio-Cultural Protection M (SE)	Profile 2: High Socio-Cultural Protection M (SE)	Profile 3: High Socio-, Low Cultural Protection M (SE)	Wald-Test for Mean Equality
Familismo	3.988(.03)	4.559(.14)	4.858(.01)	1269.329*a,b,c
Social Support	3.738(.06)s	4.26(.11)	4.23(.03)	51.100*a,b
Neighborhood	3.254(.03)	3.254(.07)	3.28(.03)	3.005
Multi-group Identity	2.196(.05)	2.623(.17)	2.464(.04)	20.592*a,b
Machismo	2.656(.13)	1.495(.09)	2.326(.07)	80.764*a,b,c
Caballerismo	6.579(.04)	4.529(.25)	6.648(.03)	94.881*a,c
Marianismo	2.501(.032)	2.551(.08)	2.537(.02)	1.049

Note.

^* = Overall significant wald-test, p <.05.

^a= Profile 1 vs. 2, p <.05.

^b= Profile 1 vs. 3, p <.05.

^c = Profile 2 vs. 3, p <.05.

Predictors of Profile Membership

Table 4 demonstrates the adjusted odds ratios representing predictors of profile membership. In summary, age, gender, and marital status were not significant predictors of profile membership. Participants were less likely to be low socio-cultural protection (profile 1), compared to the high socio, low cultural protection group (profile 3), if they were documented and educated. Participants were more likely to be in in the low socio-cultural protection group (profile 1), compared to high socio, low cultural protection group (profile 3), if they had family history of substance use. Participants were more likely to be in high socio-cultural group protection (profile 2), compared to high socio, low cultural group protection (profile 3), if they had a family history of substance use.

Table 4.

Adjusted Odds Ratios of Predictors of Profile Membership.

	Age	Gender	Marital Status	Documentation Status	Education	Family History
Low Socio-Cultural vs. High Socio-Cultural	1.062	0.661	2.286	1.043	0.873	1.420
Low Socio-Cultural vs. High Socio-, Low Cultural	0.973	1.190	1.289	0.580*	0.591*	2.354*
High Socio-Cultural vs. High Socio-, Low Cultural	1.034	0.786	2.948	0.605	0.516	3.343*

Note.

^*=Significant Odds Ratio

Alcohol Use Severity and Depressive Symptoms Outcomes

There was a significant overall difference in alcohol use severity among the three subgroups (χ²= 8.6287, p=.013). Specifically, the low socio-cultural protection group had a significantly higher (p=.005) AUDIT Score (M=4.813, SE=.346), compared to the high socio, low cultural protection group (M=3.674, SE=.199). There was a marginally significant difference (p=.054) between the low socio-cultural protection groups and the high socio-cultural protection group (M=3.244, SE=.199). There was not a significant difference in depression score among the three subgroups (χ²= 1.028, p=.598).

Discussion

Given the disproportionate rates of misuse and depression in the Latina/o population, it is necessary to understand how we can predict these outcomes and describe profiles of people who are affected by them. In summary, we found three subgroups, which we named low sociocultural protection, high sociocultural protection, and high socio-, low cultural protection. Documentation status, education level, and family history of alcohol use were significant predictors of profile membership. These socio-cultural subgroups predicted alcohol use severity differences but did not predict depressive symptoms. Below we discuss the applicability of our findings to the literature and future intervention development.

In comparison to the other profiles, Latina/os with low sociocultural protection were those with the lowest familismo, social support, and ethnic identity, and in turn, the people who reported higher alcohol use compared to those with higher levels of familismo, social support, and neighborhood collective efficacy. Theoretically, this finding aligns with the literature: familismo has demonstrated to buffer against alcohol misuse and Latina/o immigrants (Dillon et al., 2013); thus, low levels of familismo may indeed be a risk factor. Higher levels of social support has also been shown to be protective against alcohol use (Cano et al., 2018); therefore, lower levels may confer more risk. Furthermore, social support has been shown to be an important mediator between family cohesion and alcohol use (Cano et al., 2018), and, thus, could be a modifiable target for future interventions. Interestingly, neighborhood collective efficacy was not statistically different across the subgroups. This could reflect the unique cultural context of Miami as a well-established Latina/o immigrant receiving community, which will be further elaborated upon later the discussion.

Additionally, we explored cultural factors including gender norms (machismo, caballerismo, marianismo) and ethnic identity. Individuals with low sociocultural protection were characterized by high levels of machismo and reported to have higher alcohol use severity. These findings are in line with previous research that have demonstrated machismo to be a risk factor for alcohol misuse as it tends to uphold positive attitudes towards alcohol use, socialize drinking behaviors and may promote reckless drinking as a way to showing one’s “manhood” (Lee et al., 2019; Perrotte & Zamboanga, 2019). Furthermore, we found that membership in the low sociocultural protective group was characterized by lower levels of ethnic identity compared to the other profiles. Previous evidence suggests ethnic identity has protective effects against perceived stress among diverse immigrants (Espinosa et al., 2018). Among Latina/o adolescents, strong ethnic identity has been inversely associated with alcohol misuse (Banks, Winningham, Wu, & Zapolski, 2019). Based on a theory of cultural identity proposed by Unger (2011), ethnic identity and other shared cultural characteristics prevent health-risk behavior via cultural values characterized by norms and cognitions that are not supportive of risk behavior. Applied to alcohol, Unger’s (2011) model posits that cultural values aligned with strong ethnic identity prevent alcohol use by shaping protective attitudes against alcohol use. While much of the work on ethnic identity among Latina/o’s has focused on youth (Meca et al., 2020; Schwartz et al., 2014), the present study supports the notion of the negative association between ethnic identity and substance use across the lifespan.

Moreover, the traditional Latina/o gender role, marianismo, was not found to be statistically significant across subgroups for both females and males. Among Latina/o immigrants in the United States, traditional Latina/o female gender roles that promote collectivist values which restrain the adoption of drinking norms are less likely to be reported (Lee et al., 2019). This may be due to changes in the social context during the acculturative process to the U.S that may interact with the need to adhere to traditional gender roles. Exposure to a new social environment and economic opportunities may result in Latina/o immigrants acculturating to the more permissive alcohol use norms across genders associated with mainstream American culture. Indeed, contextual factors such as immigration-related stressors may play a more influential role in perceived opportunities to engage in alcohol drinking behaviors for both males and females (Perrotte & Zamboanga, 2019). As a result, both male and females may have been more likely to score higher on the machismo scales that coincide with traditional roles that hold positive attitudes toward alcohol drinking behaviors. Unexpectedly, gender was not found to be a significant predictor of group membership in the present study. The null findings for gender are in contrast to existing evidence indicating the influence of gender norms in alcohol use behaviors in this population (Cano et al., 2017).

Other predictors of subgroup membership were noted. The people in the low sociocultural protection subgroup were more likely to be undocumented, have a lower education level, and greater family history of substance use, compared to people in the high socio/low cultural protection group. The low sociocultural protection group was also the subgroup which had significantly higher alcohol use severity. Undocumented immigrants and those with lower levels of education face high levels of acculturation-related stress, which have been found to be associated with alcohol use among Latina/o immigrants (Blackson, De La Rosa, Sanchez, & Li, 2015). Undocumented immigrants are also more likely to have less access to interpersonal support, and experience discrimination, stigmatization, marginalization, fear of deportation, and live in unsafe neighborhood, all of which can lead to adverse mental health outcomes including depression and anxiety as well as substance use disorder (Garcini et al., 2016; Garcini et al., 2017; Sanchez, Dillon & De La Rosa, 2015). This study solidifies this variable-centered evidence through a person-centered approach.

Our null findings regarding depressive symptoms may be related to previous research that have found negligible associations between high levels of machismo and depression (Nunez et al., 2016), but significant relations between machismo and alcohol misuse (Perrotte et al., 2018). Similarly, low levels of familismo has been linked to alcohol misuse, but only small effects have been found between familismo and depression (Valdivieso-Mora, Peet, Garnier-Villarreal, Salazar-Villanea, & Johnson, 2016). Previous evidence suggests that depression differs by Latina/o subgroups with Mexican Americans and Puerto Ricans more likely to have major depressive disorder (Jetelina, Reingle Gonzalez, Vaeth, Mills, & Caetano, 2016). These Latina/o subgroups were not represented in our sample, which was primarily Cuban, South American, and Central American, which may explain we did not have findings for depression. Another potential explanation is the low level of internal consistency of the CES-D in our sample. Notably, further psychometric examination of the item level statistics on the CES-D scale indicated that the two reverse-scored items were problematic in the present sample and drove down the internal consistency of the scale to marginal levels. While removing the items from the summary score would have improved the internal consistency, a decision was made to keep the scale in its current form given its status as a widely used, accepted, and validated scale.

It is also important to note the community context in which this study took place. As of 2019, nearly 70% of Miami Dade County’s 2.7 million residents self-identify as Latina/os and 53% are foreign-bom making the County one of the largest metropolitan area in the US with a Latina/o’s foreign-born majority (US Census Bureau, 2020). With Latina/os immigrants from a large variety of Caribbean, and South and Central American countries, today Miami Dade County is home to the most diverse Latino population of any County in the US. Notably, Latina/os in Miami tend to enjoy more political and economic advantages compared to other cities (S. J. Schwartz, Unger, Zamboanga, & Szapocznik, 2010; Stepick & Stepick, 2002). Well-established immigrant receiving communities like Miami Dade County, with dense ethnic enclaves can provide increased availability to culturally and linguistically appropriate support, including tangible and intangible support systems. These strong ethnic enclaves have the potential to shield its residents from exposure to discrimination, and facilitate the maintenance of heritage culture identity, beliefs and values including the preservation of family cohesion (Seth J. Schwartz et al., 2014; S. J. Schwartz et al., 2010). These well-established immigrant receiving communities can reduce exposure to immigration related stressors and its related deleterious impacts including alcohol use and depressive symptoms.

We anticipate that knowledge gained from the present study can be used to develop effective programs that incorporate sociocultural factors found to be associated with adverse outcomes such as substance use among Latina/o immigrants. These interventions could include strategies that capitalize on the collectivist nature of the Latina/o culture through developing interventions that maintain or increase familismo or decrease machismo. Programs targeting Latina/o immigrant populations that account for risk and protective sociocultural factors operating at the varying systemic levels (i.e., individual, interpersonal, community) could also be informed. Interventions that foster proactive connections between Latina/o immigrants and other important systems such as community-based organizations as a means of increasing social support.

Limitations and Future Directions

Results from this study should be interpreted in light of several limitations. First, although respondent-driven sampling is a preferred method to recruit “hard to reach” populations such as recent immigrants it does not ensure a representative sample. Additionally, information regarding the chain-referral method utilized as part of the sampling strategy was not available. As such, we were not able to control for this in the analyses. Participants in this study were also from Miami, Florida, where Latina/os are the majority population (69% of the population according to the Census) (US Census Bureau, 2017) so our results may not be generalizable to other Latina/os in the United States where they may be the minority. As such, it is important to conduct similar studies in new immigrant receiving communities without well-established immigrant networks and resources, as they may provide distinctly different results compared to those found in the present study. Third, we looked at Latina/os as a homogenous group regardless of Cuban, South American, or Central American origin. Previous studies have found that alcohol use and depression differ by country of origin (Jetelina et al., 2016). Future studies should consider analyzing Latinos based on country of origin as these groups could substantially differ (Chartier et al., 2015). This study also focused solely on immigrants, but it would be interesting to replicate these analyses comparing US born vs. foreign born Latina/o participants, as nativity status has been found to predict alcohol misuse (Salas-Wright et al., 2018). Additionally, this data was cross sectional, thus it would be interesting to apply longitudinal finite mixture modeling to see if socio-cultural subgroups can change over time.

New Contribution to the Literature

Our study identified subgroups of Latina/o adult immigrants based on socio-cultural factors. Membership in these subgroups could predict alcohol use severity; thus, implications of our findings could inform current and future interventions for alcohol misuse in this population. We encourage researchers to utilize person-centered approaches to continue to unpack the complexities associated with socio-cultural factors affecting mental and behavioral health in Latina/o immigrants. Future research should further investigate the mechanisms for these sociocultural factors particularly the importance of age and gender. Future interventions should be developed with cultural values and traditional gender roles in mind. Specifically, research should focus on socio-cultural factors can provide protection from negative health outcomes and increase resiliency among this population.