Classes of drinking motives among American Indian youth drinkers

Abstract

Research exploring American Indian (AI) youth drinking motives and their relation to negative outcomes is critical due to higher rates of alcohol use and early exposure to intoxication in the population. The purpose of this study is to explore classes of drinking motives as they relate to heavy episodic drinking, perceived discrimination, religious importance, ethnic identity, and ethnic pride. This study is part of an ongoing epidemiologic and etiologic investigation of substance use among AI youth drinkers living on or near reservations (N = 1,934, M_age = 15.31). A Latent Class analysis (LCA) was conducted to discern latent classes of drinking motives. Once latent classes were identified, differences in perceived discrimination, ethnic pride, ethnic identity, spirituality, and heavy episodic drinking were tested. A two-class solution provided the best overall model fit to the data. The higher coping and enhancement motive class was associated with significantly greater heavy episodic drinking, perceived discrimination, and ethnic identity compared to the low motive class. Further, the class structure did not differ between 7^th and 8^th graders and 9^th – 12^th graders. Results indicate that among AI youth, the class with strong motives to drink for coping or enhancement had higher ethnic identity, greater risk of heavy episodic drinking, and greater perceived discrimination compared to the class with low motives. Future research should examine additional factors and stressors that may be associated with these classes of drinking motives and are unique to the AI population.

Introduction

American Indian (AI) youth have significantly higher rates of alcohol use, intoxication, and heavy episodic drinking (women consuming four or more drinks and men consuming five or more drinks in one sitting; Center for Disease Control and Prevention, 2016) compared to the national average (Swaim & Stanley, 2018a). For example, for the years 2016–2017, 11.8% of reservation-based AI 8^th graders reported heavy episodic drinking in the last two weeks compared to 3.4% of 8^th graders nationally. Additionally, AI youth were more than twice as likely to report early initiation of intoxication compared to White youth in the same geographic area (Stanley & Swaim, 2015). These findings are especially troubling because of subsequent increased risk for alcohol-related problems and disorder (Henry et al., 2011, Hingson, Zha, & Weitzman, 2009; National Institute on Alcohol Abuse and Alcoholism [NIAAA], 2015). Given the large persistent disparities in alcohol use among AI youth, we must continue to explore the unique societal and motivational factors associated with these high rates of AI youth alcohol use.

Drinking Motives

Cooper, Russell, Skinner, and Windle (1992) found that people are generally motivated to drink for three reasons: to enhance positive affect, to cope with negative affect, and to be social. These motives are often associated with different alcohol-related outcomes. For example, people who drink to cope are at greater risk of experiencing negative outcomes compared to other use motives (Cadigan, Martens, & Herman, 2015; Cooper, Russell, Skinner, & Windle, 1992; Martens, Cox, Beck, & Heppner, 2003). Various models of drinking motives have been considered including the three-factor model described earlier, four- and five-factor models (Cooper, 1994; Martens, Rocha, Martin, & Serrao, 2008; Grant, Stewart, O’Connor, Blackwell, & Conrod, 2007), and models exploring the validity of motives like conformity (LaBrie, Kenney, Mirza, & Lac, 2011; Müller & Kuntsche, 2011). Differentiation to multiple motives for use does not emerge until earlier adolescence (Anderson, Grunwald, Bekman, Brown, & Grant, 2011; Kuntsche, Knibbe, Gmel, & Engels, 2006), and adolescents generally endorse coping less frequently than other motives (Kuntsche, Knibbe, Gmel, & Engels, 2005). Sex may also influence motives; males endorsed more social, conformity, and enhancement motives than females, while females endorsed a similar amount of coping motives (Cooper, 1994).

Drinking motives have been explored primarily in predominantly White, college student populations. The little research into drinking motives in Indigenous populations suggests that models found in other populations may not be relevant for AI youth. Research on three- and four-factor drinking motives models has demonstrated that Indigenous youth infrequently endorse social motives as a reason for their drinking (Mushquash, Stewart, Comeau, & McGrath, 2008; Mushquash, Stewart, Mushquash, Comeau, & McGrath, 2014) while coping is one of the most frequently endorsed drinking motives (Tingey et al., 2017).

Given this paucity of research, especially as related to previous studies’ limited samples, this study seeks to increase understanding of drinking motives using a population-based sample of AI reservation youth. Its aims are as follows. First, we used a person-centered approach (latent class analysis) to identify classes of AI youth who drink based on their responses to drinking motives survey items. Second, we assessed whether these classes differ across grade. Finally, we examined relations between these classes and four variables that are particularly relevant to this population (religious importance, ethnic identity, ethnic pride, perceived discrimination). These factors were chosen both for their potential relation to drinking and drinking motives, and their relevance to the AI population. We also examined relations to heavy episodic drinking, an outcome that has been found to be positively related to drinking motives (Cooper et al., 1992; Kuntsche et al., 2005).

We begin by briefly reviewing literature on the four variables of interest that have been shown to be related to alcohol use and motives for use, and their relation to AI adolescent substance use and drinking motives.

Perceived Discrimination

Among AI and other North American indigenous adolescents, exposure to discrimination is associated with greater alcohol use and development of an alcohol use disorder (Armenta, Sittner, & Whitbeck, 2016; Cheadle & Whitbeck, 2011; Garrett, Livingston, Livingston, & Komro, 2017, Gilbert & Zemore, 2016). Emotional pain associated with discriminatory acts may be particularly relevant to the coping alcohol use motive. The relation between perceived discrimination and drinking motives has been examined in other groups (Hatzenbuehler, Corbin, & Fromme, 2011; Tse & Wong, 2015), but has not been explored among AI youth.

Religious Importance

In the general population, some form of religious involvement, including religious importance, tends to protect against drug and alcohol use (Ford & Hill, 2012; Wallace, Brown, Bachman, & LaVeist, 2003), with drinking motives mediating the relation between religiosity and alcohol use among college students (Galen & Rogers, 2004). Among AIs, several studies highlight the protective effects of spiritual and religious involvement for substance use among AI adults and youth (Beebe et al., 2008; Kulis, Hodge, Ayers, Brown, & Marsiglia, 2012; Stone, Whitbeck, Chen, Johnson, & Olson, 2006; Yu & Stiffman, 2010). In contrast, Kulis and Tsethlikai (2016) reported no difference in substance use and pro-substance attitudes between high AI spirituality and non-religious classes of middle school students. As with perceived discrimination, much less work has investigated religious importance and drinking motives, with even fewer including AI youth.

Ethnic Identity

Ethnic identity, the degree to which one identifies with a particular ethnicity/culture, does not show a strong relation to substance use, including alcohol, among AI youth (Bates, Beauvais, & Trimble, 1997; Dieterich, Stanley, Swaim, & Beauvais, 2013; Galliher, Jones, & Dahl, 2011; Swaim & Stanley, 2018b; Urbaeva, Booth, & Wei, 2017). However, several studies have found that reasons to drink may be different across countries and cultures (Kuntsche et al., 2015; Mackinnon et al., 2017), suggesting that ethnicity and identity may relate to motives for use. In a study of drinking motives and alcohol consequences among Alaska Native drinkers, researchers found that a component of ethnic identity (measured as affirmation, belonging and commitment) was negatively related to alcohol consequences which was, in turn, related to drinking motives (Skewes & Blume, 2015). Thus, although ethnic identity has not been found to be directly related to alcohol use, it may be related to drinking motives.

Ethnic Pride

Though the concepts of ethnic pride and ethnic identity have been used interchangeably, individuals may identify strongly with their ethnic group yet not be proud of their cultural heritage (Valk & Karu, 2001). While findings related to the protective effects of ethnic identity are mixed, several studies have noted the positive effects of ethnic pride and its role as a protective factor for avoiding substance use (Castro, Stein, & Bentler, 2009; Kulis, Napoli, & Marsiglia, 2002) and other negative effects (Yabiku, Rayle, Okamoto, Marsiglia, & Kulis, 2007). However, as with ethnic identity, no research has examined the relation between pride and drinking motives for AI youth.

Person-Centered Analyses and Drinking Motives

Person-centered analyses such as latent class or latent profile analysis are not often used when examining drinking motives and their relation to other factors. These types of analyses classify individuals based on similar response patterns. Examination of differences between classes can then be assessed on covariates and outcomes (Muthén & Muthén, 1998–2017). Two studies have explored drinking motives using a person-centered approach and have found varying results. Some find that a two-class solution characterized by greater coping or enhancement motives for drinking is the best solution for adolescent alcohol users (Kuntsche, Knibbe, Engels, & Gmel, 2010), and others report a six-class solution including three drinking motives (enhancement, coping, and social) is best for college students (Cadigan et al., 2015). Collectively, the previous studies suggest that class solutions appear to be inconsistent across different populations. Studies examining drinking motives of AI youth using person-centered analyses are limited or absent in the current literature. This study is a beginning step in understanding the underlying structure of drinking motives among reservation-based AI youth.

Method

Sampling Procedure

This study used survey data from 31 schools that participated in an ongoing epidemiologic investigation of substance use among AI youth in grades 7–12 who live on or near reservations. The sampling frame consists of schools that include a grade 7 located in the continental United States (excluding Oklahoma) on or within 25 miles of a reservation or tribal lands with at least 20 students per grade and at least 20% AI students enrolled. Schools were stratified into seven regions (Northeast, Northwest, Southeast, Southwest, Northern Plains, Southern Plains, and Upper Great Lakes) based on cultural and other similarities (Snipp, 2005). From the sampling frame, schools were randomly drawn to reflect the regional distribution of AIs residing in various geographic regions, based on the 2010 Census data. For schools not including high school grades (e.g., middle schools), the high school most likely to be attended by the AI middle school students was determined and added to the drawn sample.

Sample schools were contacted about participating in the survey by mail, email, and phone contact. If they agreed to participate, appropriate tribal and/or school board approvals were obtained. Participating schools received a comprehensive report of the finding from their school and were compensated for the resources used to complete the survey process. Participants were administered the Our Youth, Our Future (OYOF) survey online using Qualtrics software. Parents were provided information on opting their child out of the study and students were allowed to skip questions they did not want to answer. No identifying information was collected from participants, and all procedures were approved by the University Institutional Review Board.

Participants

Participants were 7–12^th grade students from 31 schools during the 2016–2017 academic year. Of the 3,851 total participants that identified as AI in the sample, 1,934 AI youth who reported drinking in their lifetime were included for analyses. The average age of participants in the study was 15.31 years old and 55% of the sample identified as female, 44% identified as male, and 1% chose not to respond. Further, 24% of the sample were in 7–8 grades at the time of completing the survey and 76% were in 9–12 grades. Approximately 69% (n = 1,326) of the sample reported not having consumed five or more drinks in one sitting within the last two weeks.

Measures

The OYOF survey contains a verbatim subset of questions related to substance use asked in the most recently available Monitoring the Future (MTF) survey, including a scale measuring drinking motives. Drinking motives were measured using responses to the question: “What are the most important reasons you drink alcohol?” Fourteen reasons were listed, and students were asked to mark all that apply. Examples of these items are, “to relax or relieve tension”, “to have a good time with my friends”, and “to feel good or get high.” Responses to these motives items were dichotomous (0 = not endorsed, 1 = endorsed).

Additionally, the survey contains questions asking about ethnic identity, religious importance, ethnic pride, and perceived discrimination. Ethnic identity was assessed using the six AI items (Cronbach’s α = 0.91) from the Orthogonal Cultural Identification Scale (Oetting & Beauvais, 1990). Examples of these items are, “How many of these special activities or traditions does your family have that are based on the American-Indian culture” and, “Do you live by or follow the American-Indian way of life” (1 = not important or none/no, 4 = very important or a lot). The mean score across the six items was computed. Ethnic pride was measured with eight items (Cronbach’s α = 0.96) where participants marked how true each statement was for them (1 = not at all true, 4 = very true). Examples of these items are, “I have a lot of pride in my ethnic group” and, “I feel good about my ethnicity.” The mean score of the eight items represented the average ethnic pride for each participant. The relation between ethnic identity and ethnic pride was moderate (Spearman’s r = 0.49). Religious importance was measured with one item, “How important is religion in your life” (1 = not important, 4 = very important). Perceived discrimination was measured using twelve items (Cronbach’s α = 0.94) assessing the frequency of perceived discrimination based on race/ethnicity, adapted from a perceived discrimination scale validated with an Indigenous sample (Whitbeck, Hoyt, McMorris, Chen, & Stubben, 2001). The mean score of the 12 items represented the average perceived discrimination for each participant. Finally, heavy episodic drinking was measured with one item, “During the last two weeks, how many times (if any) did you have five or more drinks in one sitting” (1 = none, 6 = ten or more times).

Analysis Plan

Latent class analysis.

Prior to conducting the latent class analysis (LCA), several items were removed from the analysis due to low endorsement (See Table 1). Additionally, we followed recommendations from DeVellis (2012) and removed an item with a residual negatively correlated with some items and positively with others. This can indicate the item does not fit well with the others LCA (Kreuter, Yan, & Tourangeau, 2008; Yan, Kreuter, & Tourangeau, 2012). The analyses are designed to distinguish homogenous classes from a heterogenous population, and an item that is too commonly endorsed may not discriminate latent classes (Kreuter et al., 2008; Yan et al., 2012). Classes were then derived using a series of LCAs from the ten remaining drinking motive items in the OYOF survey. LCAs were run in MPlus Version 8 using the maximum likelihood with robust standard errors (MLR) estimator (Muthén & Muthén, 1998–2017). To account for the effects of clustering, the model was specified as complex, with school as the clustering variable. Specifying the model as complex allows the model to be estimated using a sandwich estimator which helps reduce measurement bias in the estimate standard errors. Additionally, to account for potential differences across sex, we controlled for sex in all models. While similar studies have not found differences across sex (Coffman, Patrick, Palen, Rhoades, & Ventura, 2007), no studies have explored sex differences in latent class structure among AI youth. One- through five-class models were evaluated for model fit using statistics based on recommendations made by Nylund, Asparouhov, and Muthén (2007) and Tein, Coxe, and Cham (2013), and the criteria recommended by Muthén and Muthén (2000a).

Table 1.

Percent endorsement of drinking motive items by grade group

Item	Grade range
	7th-8th, n = 458	9th-12th, n = 1454
1. To experiment - to see what it’s like*	46.3	45.3
2. To relax or relieve tension	14.6	27.9
3. To feel good or get high	21.2	24.1
4. To have a good time with my friends	23.8	43.7
5. To fit in with a group I like	8.1	5.1
6. To get away from my problems or troubles	29.9	28.8
7. Because of boredom, nothing else to do	15.3	20.5
8. Because of anger or frustration	25.8	20.6
9. To get through the day	9.2	9.9
10. To increase the effects of some other drug(s)	2.6	4.1
11. To decrease (offset) the effects of some other drug(s)	1.7	2.7
12. To get to sleep	7.2	6.8
13. Because it tastes good	10.7	14.8
14. Because I am “hooked” −1 feel I have to drink	1.3	3.1

Note. Items in boldface were removed from analysis due to low endorsement. Items marked with an asterisk were removed due to problematic inter-item correlations.

The four criteria used to determine final model fit were the sample-size adjusted Bayesian Information Criterion (SABIC; Sclove, 1987), entropy, average latent class probabilities, and the Lo-Mendell-Rubin likelihood ratio test of model fit (LMR; Lo, Mendell, & Rubin, 2001). SABIC values closer to zero are best and generally the model with the lowest SABIC is preferred (Tein, Coxe, & Cham, 2013). Entropy values ranging from 0 to 1, with values greater than 0.80 indicate adequate classification quality (Celeux & Soromenho, 1996; Jung & Wickrama, 2008). Average latent class probability values close to 1 on the primary diagonal are preferred. In the off-diagonal, values close to 0 are preferred. A significant LMR p-value indicates that the tested model is a significant improvement over the model with one less class. Goodness of fit indices, parsimony, and substantive interpretability of the model were used to determine the final model.

Once the number of classes was identified, we tested whether the latent classes differed by grade group, with 7^th and 8^th graders together in one group and 9^th through 12^th graders in another. Two models were estimated – one with class probabilities constrained across grade and one that allowed class probabilities to vary across grade. Lower SABIC and significantly different chi-square between the unconstrained and constrained models suggest differences between classes at the structural level.

Finally, the BCH method (Bakk, & Vermunt, 2016) was used to determine whether classes were associated with differences in perceived discrimination, ethnic identification, ethnic pride, religious importance, and heavy episodic drinking. The BCH method outperforms other methods when measuring differences in continuous distal outcomes (Bakk, & Vermunt, 2016; Asparouhov & Muthén, 2015), accounting for the probabilistic nature of class membership and using the Wald test to assess global and pairwise comparisons based on a Chi-square statistic (Clark & Muthén, 2009).

Results

Latent Class Analysis

The two-class model provided the best overall model fit to the data (see Table 2). The two-class model had a lower SABIC than the one-class model, and the highest entropy of all solutions. The LMR tests indicated that the two-class model was a significant improvement over the one-class model, and all subsequent models with greater than two classes did not improve fit. While the five-class model had the lowest SABIC, the average latent class probabilities were much lower and further from a value of 1 than the two-class model, and the LMR test was not significant. Class 1 (n = 423) was characterized by high endorsement of coping and enhancement motive items such as, “To get away from my problems or troubles” and, “To feel good or get high” and moderate endorsement of social motive items (e.g. “To have a good time with my friends”). This class was labeled the Coping and Enhancement class (CE). Class 2 (n = 1496) was characterized by low endorsement of all drinking motives survey items and was labeled the Low Motives class (LM).

Table 2.

Latent Class Analysis model fit statistics for ten-item drinking motives scale

	Class 1	Class 2	Class 3	Class 4	Class 5
SABIC	20443.19	16018.74	15752.94	15638.89	15595.25
Entropy	—	.808	.800	.795	.790
ALCP	—	.89-.96	.84-.94	.75-.93	.75-.91
# of Sample
1	1923	423	331	287	30 (1%)
2		1496	189	114	151
3			1399	1302	284
4				206	1249
5					205
LMR	—	.00	.21	.30	.62

Note. N = 1923. BIC, Bayesian Information Criterion; SABIC, Sample-Size Adjusted Bayesian Information Criterion; ALCP, Average Latent Class Probability; LMR, Lo-Mendell-Rubin likelihood ratio test.

Multi-group Latent Class Analysis

After selection of the number of classes for the overall sample, multi-group latent class analysis was conducted to test if the class structure was significantly different for 7^th and 8^th graders compared to 9^th – 12^th graders. The difference between the SABIC from the constrained model and unconstrained model was small (ΔSABIC = 4.819). A chi-square difference test using the log-likelihood values from each model was not significant (Δχ² = 2.19, Δdf = 1). The item response probabilities for each group and class are shown in Figure 1. The results suggest that the overall two-class model fits equally well for 7^th and 8^th graders and 9^th – 12^th graders.

Figure 1. Drinking motive item response probabilities.

Note. Each radar plot displays the probability of endorsing each drinking motive item for individuals in each latent class and each grade group.

BCH tests of equality, 2-class model

The results of the BCH tests of equality for the two-class model are illustrated in Figure 2, which shows the model-estimated means (±1 SE) for perceived discrimination, heavy episodic drinking, ethnic pride, ethnic identity, and religious importance for the two latent classes. The BCH tests were significant for perceived discrimination (χ²(1) = 22.01, p < .01), AI ethnic identity (χ²(1) = 4.13, p < .05), and heavy episodic drinking (χ²(1) = 47.91, p < .01). Comparisons revealed that the CE class reported higher perceived discrimination (M = 1.96), ethnic identity (M = 2.94), and heavy episodic drinking (M = 2.50) than the LM class (M = 1.70, M = 2.80, M = 1.50). The overall tests were not significant for ethnic pride or religious importance.

Figure 2. Bar plot of model-estimated means for model outcomes.

Note. * p < .05, ** p < .01, *** p < .001

Discussion

The present study identified two latent classes based on ten drinking motive items in a sample of AI youth drinkers. The first class was characterized by higher endorsement of coping and enhancement motive items. The second class was characterized by lower endorsement of any motive items in the survey. The coping and enhancement class was associated with greater perceived discrimination, AI ethnic identity, and heavy episodic drinking compared to the low motive class.

Since drinking has been conceptualized in a developmental framework, it is likely that motives for use and drinking patterns are yet-to-be fully established in younger drinkers (Masten, Faden, Zucker, & Spear, 2009). However, greater drinking at younger ages is associated with a greater likelihood of drinking more later in life and experiencing greater alcohol-related problems (Muthén & Muthén, 2000b). Several studies report that enhancement motives are associated with heavy drinking and coping motives are associated with greater alcohol-related problems (Cooper, 1994; Cooper et al., 1992; Kuntsche et al., 2005). The coping and enhancement class identified in these AI youth suggest that they may be at risk for both high rates of alcohol use in addition to problematic outcomes such as academic low performance (Windle, 1996), and delinquent behaviors (Karwacki & Bradley, 1996). The relation of the coping and enhancement class to heavy episodic drinking portends the future development of these negative alcohol-related outcomes.

Of even more concern is the finding that the same pattern of drinking motives was found between middle school and high school students. This is consistent with past findings in which young reservation AI youth are at much greater risk to initiate alcohol use compared to their white counterparts who reside and attend schools in the same locations (Stanley & Swaim, 2015). Other research has found that the patterning of latent classes of substance use, similar to their older-age high school peers, are established by middle school among AI reservation youth (Stanley & Swaim, 2018). This early initiation, coupled with drinking in order to enhance and cope provides clear warning signs for this group of youth.

Additionally, the findings reported here offer support that the structure of drinking motives for Indigenous youth is different than for a largely White population. Coffman, Patrick, Palen, Rhoades, and Ventura (2007) reported that four classes of drinking motives best fit a national sample of predominantly White adolescent students. However, in the present sample we observed only two classes of drinking motives among AI youth drinkers. Furthermore, the latent classes observed in the present sample appeared to be primarily characterized by high coping and enhancement motives and lower social motives. This finding is consistent with findings from previous studies that report that social motives are less relevant and infrequently endorsed by Indigenous populations (Musquash et al., 2008; Mushquash et al., 2014; Tingey et al., 2017). Overall, these findings suggest that interpretation of and intervention targeting drinking motives should not be treated as universal across ethnicity.

The findings related to perceived discrimination highlight the unique stressors on AI youth and their relation to heavy episodic drinking. AI youth who endorse more motives are more likely to report perceiving more discrimination than the class of low motives endorsement. Further, AI youth who endorse more motives are more likely to report stronger AI ethnic identity. Greater perceived discrimination is associated with more negative mental health outcomes and greater stress (Pascoe & Richman, 2009). And while higher ethnic identity has been found to be protective against, and even a coping strategy for, the negative effects of perceived discrimination (Galliher et al., 2011; Mossakowski, 2003), it may be likely that heavier drinking is also used by AI youth to help cope with stress from perceived discrimination and other negative outcomes. The findings of this study, coupled with research on perceived discrimination and ethnic identity, may help explain why items associated with coping motives appear to be more frequently endorsed in this sample of AI youth.

Limitations and Future Directions

While this study presents findings that contribute meaningfully to the field of alcohol use motives and AI youth alcohol use, it has several limitations. First, the data are cross-sectional and conclusions about the stability and direction of the relations and classes across time cannot be made. It may be possible that the class of AI youth that is more motivated to drink does so because they perceive more discrimination. Alternatively, they may perceive more discrimination because they drink more than the low motive youth. Similarly, conclusions about whether these classes represent developmental stages or if they are stable over time cannot be made. Longitudinal research would address this limitation and explore the stability of classes and directions of the relations more appropriately.

Second, the drinking motives scale is not yet fully validated. This scale was adapted from the Monitoring the Future scale distributed nationwide to United States adolescents. There are several validated drinking motive measures that may assess motives differently (Drinking Motives Questionnaire Revised; Cooper, 1994; Drinking Motives Questionnaire Revised Short Form; Kuntsche & Kuntsche, 2009). Third, the scale measures motivations using a mark all that apply format. While this method is easy to use, it does not collect information on the degree to which these motives are important for AI youth drinking. The latter two limitations could be addressed by using either of the empirically validated motive measures listed above in future research.

Fourth, the methods in which several of the variables were measured likely limited the scope of the conclusions that could be drawn. Religious importance was measured by a single item in the present study. While it is important to note that a single item measure for religious importance may miss other facets related to religion, single item measures have been found to be reliable, specifically with religiosity (Abdel-Khalek & Lester, 2007; Youngblut & Casper, 1993; Zimmerman et al., 2006). In the present study, heavy episodic drinking was assessed by the single item, “During the last two weeks, how many times (if any) did you have five or more drinks in one sitting” which may limit the findings several ways. First, the question does not accurately measure heavy episodic drinking as defined by the Center for Disease Control and Prevention (2016) as at least four drinks in one sitting for women and at least five drinks in one sitting for men. The present item may not be capturing all female participants that engage in heavy episodic drinking by limiting the response to at least five drinks in one sitting. Also, using this single item to measure alcohol use does not account for other aspects of use such as typical use, alcohol use disorder symptoms, and alcohol-related consequences. While the findings suggest that a high coping and enhancement class is related to greater frequency of heavy episodic drinking, that relation may be driven by enhancement due to its stronger association with heavy episodic drinking (Cooper, 1994; Cooper et al., 1992).

Additional stressors and their relation to AI classes of drinking motives should be explored in the future. The additional stress of intergenerational trauma and current public policies in the United States contributes to higher rates substance misuse and other mental health issues in the AI population (American Psychological Association, 2018). Longitudinal research examining the relation between classes of drinking motives and outcomes such as perceived discrimination, access to healthcare and insurance, socioeconomic status, and other stressors may address whether motives (particularly coping motives) remain stable over time and how they are impacted by these additional stressors over time. Additionally, future research should explore community-based interventions that target classes characterized by higher coping and enhancement drinking motives in order to reduce heavy episodic drinking and attenuate additional alcohol-related negative consequences.

Conclusions

This study examined classes of responses to alcohol use motive items among American Indian (AI) youth using person-centered analyses. Further, this study used a national dataset of AI youth and examined the relation to and differences between classes on several outcomes. While controlling for sex, the results indicated two classes of AI youth drinkers. The findings are consistent across grade. The higher enhancement and coping motives class is associated with greater heavy episodic drinking, perceived discrimination, and ethnic identity compared to the low motives class. These findings identify classes that are associated with particular negative outcomes and may inform future interventions. Further research on additional negative outcomes associated with these classes and on motive-centered interventions is necessary.