Brief online interventions targeting risk and protective factors for increased and problematic alcohol use among American college students studying abroad

Abstract

Research documents increased and problematic alcohol use during study abroad experiences for college students yet no research documents effective preventive programs with these students. The present randomized controlled trial was designed to prevent increased and problematic alcohol use abroad by correcting misperceptions of peer drinking norms abroad and by promoting positive and healthy adjustment into the host culture (i.e., sojourner adjustment) through brief online personalized feedback interventions. A sample of 343 study abroad college students was randomly assigned to one of four conditions including a personalized normative feedback intervention (PNF), a sojourner adjustment feedback intervention (SAF), a combined PNF + SAF intervention, and an assessment-only control condition. Generalized estimated equation analyses accounting for baseline drinking and consequences revealed an intervention effect for PNF that was mitigated by baseline drinking level, such that PNF was best for those with lighter baseline drinking, but heavier baseline drinkers receiving PNF alone or PNF + SAF drank comparatively similar or more heavily abroad to those in the control condition. However, PNF + SAF condition participants with greater baseline levels of consequences reported comparatively less consequences abroad than their control participants. Thus, PNF alone may be helpful for lighter drinkers at predeparture and the addition of SAF to PNF may help prevent consequences abroad for those reporting more consequences prior to departure abroad. This research represents an important first step in designing and implementing efficacious interventions with at-risk study abroad college students, for which no current empirically-based programs exist.

Study Abroad Alcohol Risk

Approximately 10% of all American undergraduate college students study abroad, with 304,467 students completing study abroad experiences in over 75 foreign countries in the 2013–2014 academic year (Institute of International Education [IIE], 2015). This number has tripled in the past decade and federal initiatives have been implemented to increase participation rates to about 1 out of every 5 students within the next 10 years and to create more opportunities for lower income and underrepresented groups (NAFSA: Association of International Educators, 2009; IIE, 2015; Murray Brux & Fry, 2009). The personal, cultural, and academic benefits of study abroad are many, including increased global perspectives, respect for other cultures, preparation for international careers, and academic success (Dwyer, 2004; Kauffmann & Kuh, 1984; Kitsantas, 2004; Kitsantas & Meyers, 2001; Lewin, 2009; Sutton & Rubin, 2004; 2010).

Despite the benefits of study in foreign countries, American study abroad students are at-risk for increased and problematic drinking. Students more than double their weekly alcohol use while abroad and the heaviest drinkers abroad return home drinking at higher levels than before they left (Pedersen, Larimer, & Lee, 2010b). In addition to increased and heavy drinking, a substantial portion of these American students face multiple negative alcohol-related consequences while abroad (Hummer, Pedersen, Mirza, & LaBrie, 2010; Pedersen, Neighbors, Lee, & Larimer, 2012b). For example, within only a one-month time frame, more than one-third of male and female students reported drinking on nights they had not planned to drink, taking risks or doing impulsive things when drinking they later regretted, noticing changes in their tolerance level, engaging in regretted sex, and drinking to the point of blacking out (Hummer et al., 2010). Other unfortunate risks reported by at least 1 in 10 students include drinking to the point of passing out, missing classes, feeling guilty or bad about themselves due to their drinking, finding themselves in a dangerous situation they would not have been in if sober, and alcohol-related injuries (Pedersen et al., 2012b). Compounding the negative experiences during the actual trip, students who intend to study abroad in college and those who have previously studied abroad tend to drink more heavily compared to those who do not study abroad (Pedersen, LaBrie, Hummer, Larimer, & Lee, 2010a). Other researchers confirm that students who study abroad in college drink more frequently, drink in greater quantities, and report drunkenness more often than students with no study abroad history and there is potential for these patterns of heavy drinking to influence overall mental health and physical health once students return to campus (Wielkiewicz & Turkowski, 2010). Thus, American study abroad students represent a large and diverse group at-risk for heavy and problematic drinking before, during, and after study abroad experiences.

Student affairs personnel working with study abroad students and university administrators have reported substantial concerns about student drinking while abroad. For example, the Forum on Education Abroad, a collaborative organization of over 800 colleges and universities reported that alcohol misuse by students is the most concerning behavioral problem reported by program staff, with alcohol use and poor judgement cited as the factors most contributing to critical incidents abroad, including robberies, arrests, physical and sexual assaults, and even student deaths (Forum on Education Abroad, 2009; Mello, 2015). In a survey of study abroad directors at 152 institutions, we found that over half believed drinking poses the most serious negative issue for students abroad – placing them at risk for physical and emotional harm and also placing programs at risk for serious liabilities – and 90% of the directors reported it would be important for their office to have access to alcohol risk reduction programs (LaBrie et al., 2013b). These concerns, along with student data, suggest that much more needs to be done to address the harmful role of alcohol in study abroad experiences – both the harms to students and institutions’ liability. Yet despite the clear need for concrete strategies to reduce the risk for detrimental outcomes as a result of student alcohol misuse abroad, there are currently no published empirically-tested prevention programs addressing this issue and no broadly used targeted prevention programs.

Development of a Study Abroad Alcohol Intervention

Given the increasing number of students going abroad each year, a brief, self-delivered, easily accessible, and empirically-validated approach targeted toward the unique abroad environment would help meet the large unmet need at institutions across the U.S. In designing our intervention, we turned to the literature on efficacious interventions with college students and, in particular, to more recent work utilizing “event-specific” interventions that are designed to prevent escalation of heavy and problematic alcohol use during periods of known risk for students (e.g., 21^st birthday celebrations, spring break trips) (Lee, Lewis, & Neighbors, 2009; Neighbors et al., 2007b; 2011; Rutledge, Park, & Sher, 2008). Recent clinical trials of prevention programs tailored toward reducing incidences of problematic drinking during these events have demonstrated efficacy (Lee et al., 2014; Neighbors et al., 2012). An event-specific prevention program makes sense for the study abroad context because these students are an already at-risk subgroup entering into a period of risk where they are known to drink more and experience consequences. Furthermore, trips are preplanned, thus presenting a unique opportunity to prepare students for the upcoming trip during a period of time when they are attentive and excited about their upcoming trip.

We designed a study abroad-specific intervention based on the social norms approach to college drinking and the literature on how acculturation into new environments (even temporarily) can contribute to reduced incidences of problematic drinking. First, a large body of work indicates that young adults are influenced to drink alcohol based on their perceptions of how much and to what extent their peers are drinking in a variety of contexts, including specific events (Borsari & Carey, 2003; Lewis, Neighbors, Lee, & Oster-Aaland, 2008; Neighbors, Lee, Lewis, Fossos, & Larimer, 2007a; Neighbors, Oster-Aaland, Bergstrom, & Lewis, 2006; Pedersen & LaBrie, 2008). Within the study abroad environment, perceptions of general and host country-specific study abroad peer drinking, as well as perceptions of native young adult drinking within the host country, are associated with increased drinking while abroad (Pedersen, Cruz, LaBrie, & Hummer, 2011a; Pedersen, LaBrie, & Hummer, 2009; Pedersen et al., 2010b). Correcting misperceptions of peer drinking norms through personalized normative feedback (PNF) has become one of the prominent strategies for addressing excessive alcohol use among college students (Cronce & Larimer, 2011; Miller et al., 2013; White, 2006) and interventions based solely on PNF have demonstrated small to moderate effect sizes with this population (LaBrie et al., 2013; Lewis, Neighbors, Oster-Aaland, Kirkeby, & Larimer, 2007; Neighbors, Lee, Lewis, Fossos, & Walter, 2009a) even when seeking to prevent escalation of risk during specific events (Lee et al., 2014; Neighbors et al., 2012; Neighbors, Lee, Lewis, Fossos, & Walter, 2009b). Thus, challenging study abroad students’ pre-departure overestimations of salient and proximal reference groups (i.e., study abroad peers; native country-specific adults) by providing them with accurate normative information may help students form more accurate beliefs about drinking within their host country and, in turn, prevent heavy drinking.

Second, theories of acculturation(John W. Berry, 1997; J. W. Berry, 2003) and, more specifically, “sojourner adjustment” (Church, 1982; i.e., the sociocultural and psychological adjustment of relatively short-term visitors to new cultures) posit that immigrants and students who attempt to integrate or assimilate more fully into their new culture are at the least risk for sociocultural and psychological adjustment difficulties (Berry, 1998; Searle & Ward, 1990; Ward & Rana-Deuba, 2000; Ward & Kennedy, 1994). Research and theory also suggest that if young people transition to a temporary novel risky environment (e.g., spring break, Mardi Gras in New Orleans) and they do not feel connected to their environment, they may view their time as a temporary reprieve from real life (i.e., a “backspace”), transgress drinking and sexual norms, act in ways inconsistent with their personalities, and, thus, be more likely to engage in risky behaviors (Jankowiak & White, 1999; Maticka-Tyndale, Herold, & Mewhinney, 1998; Millhausen, Reece, & Perera, 2006; Triandis, 1977). Our research with American study abroad students suggests that those most at-risk for heavy drinking and consequences are those who separate themselves from the host environment (i.e., placing more emphasis on the home/U.S. culture), those who perceive the abroad culture as very different than their home culture, those who spend more time with other Americans while abroad, those who feel out of place being away from home, and those who experience anxiety about interacting in the foreign environment (Hummer et al., 2010; Pedersen et al., 2011a). Other work has confirmed that this negative sojourner adjustment is associated with greater risk for heavy drinking and problems abroad (Mitchell et al., 2016; Wielkiewicz & Turkowski, 2010). Conversely, we have also found that positive sojourner adjustment (i.e., quality/quantity of time with local people, cultural understanding and participation in foreign cultural activities, foreign language development/use, identification as a member of the host culture) protects students from heavy alcohol use and problems.(E. R. Pedersen, Neighbors, et al., 2012). Thus, promoting cultural engagement and helping with adjustment/transition to life in a foreign environment may prevent incidences of problematic drinking abroad. Positive sojourner adjustment can relate to taking advantage of cultural learning experiences by participating in local customs and spending time with local people, rather than focusing on drinking-centered social experiences with other American students. Culturally-engaging activities may also serve as healthy alternatives and provide a means of achieving social and recreational goals without drinking, potentially reducing motivation to drink for both social and coping reasons and, therefore, reducing alcohol use and consequences while abroad. Importantly, preventing heavy drinking patterns from forming abroad may also reduce continued heavy drinking once students return home.

The Present Study

The current study is the first to evaluate an event-specific approach to prevent heavy alcohol use and resulting consequences among study abroad students. Using brief online personalized feedback interventions, we sought to prevent increased and problematic use by correcting misperceptions of study abroad student drinking norms, correcting misperceptions of country-specific native adult drinking norms, and promoting positive adjustment and engagement into the host culture (sojourner adjustment) prior to departure abroad. We hypothesized that those participants who received a PNF intervention to correct descriptive drinking norms for region-specific study abroad peers and country-specific host country nationals would drink less and experience fewer consequences over the course of their trip compared to participants in an assessment-only control condition. Second, we hypothesized that participants receiving personalized Sojourner Adjustment Feedback (SAF) with tips, strategies, and resources to meet cultural engagement goals while abroad would drink less and report fewer consequences than controls. Finally, there is also evidence that limited positive sojourner adjustment combines with greater perceptions of drinking behavior in the study abroad environment to promote even further risk (Pedersen et al., 2011a). Targeting both perceptions and sojourner adjustment during interventions with this group may aid in the prevention effort against heavy drinking and negative consequences. Thus, we hypothesized that participants receiving both PNF and SAF would drink the least and experience the fewest alcohol-related consequences.

Method

Participants

Participants were recruited through the study abroad office at one large university in the northwest United States. Data were collected during three consecutive quarters at the institution during 2010 and 2011. Eligibility requirements were (1) studying abroad for one quarter (between eight and 18 weeks) and (2) studying in one of 30 identified countries. Of the 848 students who signed up for the study, 431 (51%) met criteria and were invited to participate, of which 343 (80%) completed the predeparture survey and were randomly assigned to an intervention condition. Most participants did not meet eligibility criteria due to short-term trips (e.g., one summer month). Participants reported a mean age of 21.14 (SD = 3.11) with 16% reporting first year or sophomore status, 38% reporting junior status, 43% reporting senior status, and 3% reporting graduate student status. Participants were primarily female (78%) and of White race/ethnicity (72%). Other race/ethnicities represented were 14% identifying as Asian/Pacific Islander, 7% as “mixed ethnicity,” 2% as Hispanic/Latino(a), 2% as African-American/Black, and 3% Native American/Alaskan Native or “other ethnicities.” Consistent with the broader study abroad population, 74% of participants studied abroad in Europe (e.g., Ireland, France), 5% studied in Asia (e.g., China, Japan), 5% studied in Oceania (Australia and New Zealand), 8% studied in Latin America (e.g., Mexico, Brazil), and 9% studied in non-traditional study abroad locations encompassing countries in the Middle East, Africa, South Asia, and West Asia (e.g., South Africa, India, Israel). The top three countries represented in the sample were Italy (29% of sample), Spain (13%), and Greece (8%). Participants studied abroad for a mean of 11.15 weeks (SD = 2.50).

Procedures

All procedures were approved by the local Institutional Review Board at the university. All prospective study abroad students received an email from the study abroad office informing them of the opportunity to participate in a research study. This email briefly described the study and incentive structure, ensured participants of the confidentiality of their responses and that personal information would not be shared with the study abroad office, and made clear that the study was conducted by independent researchers. Interested participants clicked a web link to an online sign-up sheet and indicated their contact information, including country of expected study and departure date. Two weeks before their departure abroad, eligible students were emailed a link to an online confidential survey, where they indicated online consent and complete a 20-minute baseline survey. Once completed, participants were electronically randomized to one of four conditions using blocked randomization to promote equal cell sizes based on baseline drinks per week. If randomized to one of the three intervention conditions, participants immediately received feedback based on their responses. Control participants exited the survey. During the first week of their trip, participants in the three intervention conditions were resent their personalized feedback by email. Participants received emails containing links to online follow-up surveys with similar content during their first month abroad (approximately four weeks into the trip) and during their last month abroad (approximately one week prior to their return date to the U.S.). Eighty-seven percent completed the first follow-up survey and 84% completed the last month follow-up survey. Participants received $10 for completion of each survey.

Measures

Items from the predeparture questionnaire were used within the feedback components of the intervention conditions. Relevant information (e.g., intentions to drink while abroad, perceptions of peer and host national behavior within the study abroad environment, sojourner adjustment goals) were preloaded into the feedback presented to intervention condition participants.

Alcohol use, consequences, and intentions

Alcohol use and consequences over the past month were assessed at each time point. Participants completed the Daily Drinking Questionnaire (DDQ; Collins, Parks, & Marlatt, 1985) which allowed for the calculation of typical weekly drinking average drinks per occasion used in analyses and in the PNF condition. At predeparture, participants also completed a modified DDQ for intended weekly drinking while abroad and average drinks consumed while abroad used in the PNF conditions. The Rutgers Alcohol Problem Index (RAPI; White & Labouvie, 1989) assessed alcohol-related problems experienced over the past month (“0 = never” to “4 = 10 or more time”). The RAPI consisted of 23 items and included items related to academic and social problems (e.g., “Not able to do your homework or study for a test,” “Had a fight, argument, or bad feelings with a friend”) as well as items assessing physiological consequences such as increased tolerance, blacking out, and passing out. Internal reliability for the RAPI (α) ranged from 0.85 to 0.90 across assessment points. Prior to all alcohol measures, participants were presented with standard drink definitions (12 oz. beer, 5 oz. wine, 1.5 oz. shot of hard alcohol) as well as region-specific popular drinks (e.g., standard drink sizes for sake and soju if participant was studying in Asia).

Perceived alcohol use

At each assessment period, participants completed a modified Drinking Norms Rating Form (DNRF; Baer, Stacy, & Larimer, 1991) to indicate how many drinks they believed a typical study abroad student from their university living in their host region (Europe, Asia, Oceania, Latin America, and non-traditional study abroad locations) drank on each day of a typical week and how many drinks per occasion they believed these students drank. Participants were also asked to complete a DNRF (yielding perceived drinks per week for host country nationals) and one open-ended question about their perceptions of drinking by native adults of their host country (i.e., estimate the percentage of abstainers in their host country). These items were repeated for the follow-up surveys. Responses to the predeparture perceived norms questions were included in the PNF.

Sojourner Adjustment

The Sojourner Adjustment Measure (SAM; Pedersen et al., 2011b) assessed aspects of the cultural adjustment process among students living temporarily abroad. At predeparture, this 24-item measure was used to assess intended goals relating to the four factors of positive sojourner adjustment: (1) social interaction with host nationals (e.g., “socialize a good deal with local people from my host country;” α = 0.81), (2) cultural understanding and participation (e.g., “enhance my understanding of my host country’s culture;” α = 0.84), (3) language development and use (e.g., “increase my understanding of my host country’s language [or dialect/idioms];” α = 0.81), (4) host culture identification (e.g., “subscribe to the values of my host country;” α = 0.58) and 2 factors of negative sojourner adjustment: (5) social interaction with co-nationals (e.g., “have meaningful social interactions with Americans; α = 0.70”), and (6) homesickness/feeling out of place (e.g., “feel out of place in my host country;” α = 0.75). Each factor contains four items. The SAM has displayed adequate convergent validity with established measures of acculturation (Pedersen et al., 2011b) and theoretically consistent associations with negative consequences and drinking motives while abroad (Pedersen et al., 2012b). At predeparture, participants were asked to consider their upcoming time abroad and indicated the degree with which they agreed they would experience each item (e.g., “During my time abroad, I will spend a good amount of time meeting and conversing with local people”) and these goals were included in the SAF. Participants were asked about actual experience of each of the 24 items “during the past month abroad” on the first and last month surveys (αs ranged from 0.65 to 0.91).

Intervention Conditions

PNF condition

Immediately after completing the predeparture questionnaire, participants in the PNF only condition were presented with on-screen PNF containing information about (1) region-specific study abroad peers and (2) country-specific adults living in the country (i.e., host country nationals). Participants were shown their intended drinks per week and average drinks per occasion while abroad alongside their perceptions of how other study abroad peers drank in their host region. These perceptions were shown alongside the actual drinking norms abroad of university-specific peers collected from our prior work (see Pedersen et al., 2010b). Similar content was presented to challenge perceptions about the drinking behavior of adults living in the host country. Perceptions collected from participants on the DNRF and the abstinence item were presented alongside host country-specific per capita drinking rates and abstinence rates obtained from the World Health Organization Global Status Report (WHO, 2004).

SAF condition

In the SAF, participants were given feedback based on their responses to the 24 SAM items. They were provided with statements such as “You indicated that during your time abroad you will ‘Socialize a good deal with local people from your host country’ and ‘Have meaningful social interactions with local people.’” These individual goals were then followed by tips and strategies for how to meet these sojourner adjustment goals while abroad. Tips and strategies were generated from two focus groups (10–12 students per focus group) with previous study abroad students prior to the development of the SAF. The four positive SAM factors included tips/strategies for approaching these goals, while the two negative SAM factors included tips/strategies for limiting experience with these aspects. We provided links to host country-specific websites for local cultural events, news, and language dictionaries. These links also contained information regarding country-specific cultural activities, practices, and general descriptions of typical adjustment stages while abroad.

Combined PNF and SAF condition

Participants in the combined condition first received PNF followed by SAF. All procedures described in the above conditions were otherwise the same.

Control condition

Participants in the assessment-only control condition received all measures as described with no feedback.

Analytic Plan

Primary analyses were conducted using generalized estimating equations (GEEs; Hardin & Hilbe, 2012). GEE models were developed as an extension of generalized linear models to adjust for correlated outcomes such as in the case of longitudinal data. Unlike generalized linear mixed models (GLMM; cf. Atkins, Baldwin, Zheng, Gallop, & Neighbors, 2013), GEE models do not estimate subject-specific effects (i.e., random effects) and are more appropriate when the focus is on average associations and trends. In addition, GEE are computationally simpler than GLMMs and thus are preferred when random effects are not of substantive interest, which was the case here. GEE have been frequently and similarly used in longitudinal randomized trials evaluating non-normally distributed substance use outcomes (e.g., Bergman et al., 2015; Ondersma et al., 2009; Pettinati et al., 2008; Roux et al., 2008).

Outcomes were drinks per week from the monthly DDQ assessments and a sum score of the RAPI consequences from the monthly RAPI assessments, which were evaluated in separated models and fit using negative binomial distributions with log link. The two dependent variables were correlated at r = 0.46 and thus we included a Bonferroni adjustment of p = .034. Within-person correlations were specified as exchangeable and robust standard errors were estimated. In both models, baseline values of the outcomes were included as covariates, and the time variable references the post-baseline assessment points during the period abroad (first month abroad, last month abroad). The first and last month abroad drinks per week or alcohol-related consequences outcomes did not differ within individuals between the first and last month abroad.

Covariates in each model included age (mean centered), gender, and location of study abroad program (European countries versus other countries) since these have been shown to affect the rates of drinking and consequences among study abroad students in our other work (Pedersen et al., 2009). Also included per the GEE analyses were baseline outcomes (mean centered), time (first month abroad, last month abroad), and three dummy-coded condition variables (PNF, SAF, and PNF+SAF). Coefficients for intervention condition represent differences between the given condition and the control condition. We also examined the interactions between baseline outcomes and condition variables to evaluate whether intervention effects depended on the level of drinking or problems for the participant before the intervention and trip abroad. The interaction analyses were conducted post hoc. Significant interactions were graphed based on approximate 25% quartiles for baseline drinking and on baseline RAPI scores that did not fit specifically into quartiles but that made theoretical sense for this population (i.e., 0 consequences at baseline, 1 to 2, 3 to 4, and 5 to 6). Randomization was successful in that there were no differences at baseline between conditions for demographics or outcome variables. Analyses were conducted in Stata v.13.

Results

Alcohol Use

Table 1 contains the actual means and standard deviations of the sample by condition for baseline, first month abroad, and last month abroad. Findings from the GEE analyses can be found in Table 2. Because of the log link function, the raw coefficient is in the log scale, and thus somewhat similar to logistic regression, coefficients are typically exponentiated (i.e., raised to the base e) and interpreted as rate ratios. For each one point change in the covariate, the predicted mean of the outcome changes by a multiplicative factor of the rate ratio. For example, of the three covariates (age, gender, location), only location of study abroad was predictive of drinking, such that studying in a European country was associated with more drinking over time. The exponentiated coefficient of 1.49 indicates that those who studied abroad in Europe drank 1.49 times more than those who studied elsewhere. The 0.99 exponentiated coefficient for time indicates a decrease of 1% in drinking at the subsequent assessment following the first-month abroad (i.e., at last month abroad). Each unit increase in baseline drinking was associated with a 3% increase in follow-up drinking. The significant main effect for the PNF condition suggested that, controlling for covariates and baseline drinking, PNF participants drank 30% fewer drinks per week than control participants during the follow-up period.

Table 1.

Actual means for drinks per week and RAPI outcomes over time by condition

	Control		PNF		SAF		PNF + SAF
	M	SD	M	SD	M	SD	M	SD
Drinks per week
Predeparture	7.33	8.10	6.88	7.58	6.69	6.66	6.85	7.57
First month abroad	9.71	7.61	9.81	10.33	11.88	9.93	10.42	8.88
Last month abroad	11.12	9.49	8.59	9.03	11.57	9.22	10.82	10.65
RAPI consequences
Predeparture	2.83	3.92	3.19	4.33	3.11	4.10	2.93	4.95
First month abroad	2.44	4.20	2.69	3.71	2.92	4.30	2.32	3.43
Last month abroad	3.32	5.54	2.58	4.37	2.50	3.75	2.50	3.96

Note: PNF = personalized normative feedback; SAF = sojourner adjustment feedback

Table 2.

Main effects and baseline behavior interaction effects controlling for age, gender, and location of program

	Coefficient	Robust SE	Exponentiated Coefficient	Z-score	p-value	95% CI
						Lower	Upper
Drinks Per Week
Constant	2.26	0.09	9.57	24.72	0.000	2.08	2.44
Age	−0.02	0.02	0.98	−1.14	0.254	−0.05	0.01
Gender1	0.16	0.12	1.18	1.34	0.179	−0.07	0.40
Location of program2	0.40	0.13	1.49	3.02	0.003	0.14	0.66
Time	−0.01	0.04	0.99	−0.16	0.870	−0.09	0.07
Baseline DDQ	0.03	0.01	1.03	3.10	0.002	0.01	0.05
PNF3	−0.36	0.13	0.70	−2.64	0.008	−0.62	−0.09
SAF	0.11	0.13	1.12	0.86	0.390	−0.14	0.36
PNF + SAF	0.19	0.18	1.21	1.02	0.305	−0.17	0.54
Baseline DDQ × PNF	0.07	0.01	1.08	5.15	0.000	0.05	0.10
Baseline DDQ × SAF	0.02	0.02	1.02	1.03	0.303	−0.02	0.05
Baseline DDQ × PNF + SAF	−0.06	0.02	0.94	−2.74	0.006	−0.10	−0.02
Alcohol-related Consequences
Constant	0.91	0.17	2.48	5.38	0.000	0.58	1.24
Age	−0.03	0.02	0.97	−1.11	0.266	−0.07	0.02
Gender1	−0.13	0.18	0.88	−0.71	0.480	−0.48	0.23
Location of program2	0.39	0.16	1.47	2.37	0.018	0.07	0.71
Time	0.03	0.10	1.03	0.28	0.777	−0.16	0.22
Baseline RAPI	0.14	0.02	1.15	5.83	0.000	0.09	0.18
PNF	−0.39	0.22	0.68	−1.77	0.077	−0.81	0.04
SAF	−0.26	0.22	0.77	−1.14	0.252	−0.70	0.18
PNF + SAF	0.38	0.29	1.46	1.31	0.190	−0.19	0.95
Baseline RAPI × PNF	0.09	0.04	1.10	2.67	0.008	0.02	0.16
Baseline RAPI × SAF	0.04	0.03	1.04	1.41	0.158	−0.02	0.10
Baseline RAPI × PNF + SAF	−0.12	0.05	0.88	−2.73	0.006	−0.21	−0.04

Note: PNF = personalized normative feedback; SAF = sojourner adjustment feedback; DDQ = Daily Drinking Questionnaire total drinks per week; RAPI = Rutgers Alcohol Problem Index composite score

¹gender coded 1 for males, 0 for females

²location of program coded 1 for European countries, 0 for non-European countries

³PNF coefficient in the model without interactions was −0.30 (SE = .14), p = .030.

Main effect results did not change when we ran the models without the interaction effects included and thus we present the full model only in Table 1. There were significant interaction effects for baseline drinking × PNF and for baseline drinking × the combined PNF + SAF conditions. Graphing these interactions revealed that the intervention worked better for lighter baseline drinking PNF participants (Figure 1). For example, PNF participants reporting drinking 8 or less drinks per week at baseline reported fewer drinks per week abroad compared to control. The protective effect was less evident if lighter baseline drinkers also received SAF, as per the significant interaction between the combined PNF + SAF condition and baseline drinks per week. Indeed, while heavier drinkers in the PNF condition drank at approximately the same level abroad as control, heavier baseline drinkers receiving SAF in addition to PNF drank more than control (e.g., about one drink more per week for those drinking 12 baseline drinks per week). The SAF condition was graphed for reference purposes, since there was no observed SAF × baseline drinking interaction; however, those in the SAF condition drank the most abroad regardless of their baseline drinking (see also means in Table 1).

Figure 1. Baseline Drinks per Week Moderates PNF and PNF + SAF Conditions versus Control.

Note: The dependent variable (abroad drinks per week) functionally represents the average drinks per week across the two abroad time points (first month abroad, last month abroad) within each individual.

Alcohol-related Consequences

Table 1 also contains the actual means and standard deviations by condition and time point for the RAPI consequences sum score. Findings from the GEE analyses can be found in Table 2. Neither gender nor the age covariates associated with more reported consequences over time; however, those who studied in Europe experienced 1.47 times more consequences than those who studied elsewhere. The time effect was significant, such that after controlling for baseline consequences, consequences increased by 3% over time across first month abroad to last month abroad. The effect for RAPI was significant, with those reporting greater consequences at baseline reporting more consequences during follow-up.

Main effect results again did not change when we ran the models without the interaction effects included. We present the full model only in Table 1. There were significant interaction effects for baseline RAPI × PNF and for baseline RAPI × the combined PNF + SAF conditions. Graphing these interactions revealed that while PNF participants of varying baseline RAPI levels reported fewer consequences than control during abroad, the effect was strongest for PNF participants reporting fewer consequences at baseline (see Figure 2). As baseline consequences increased from 0 to 6, the difference between control and PNF participants was less pronounced, to the point where PNF and control participants with baseline RAPI consequences scores of 6 nearly matched in terms of the consequences experienced abroad. However, regardless of whether PNF + SAF participants reported few or many consequences at baseline, they experienced fewer consequences abroad compared to control. Overall, combined condition participants with greater baseline consequences experienced the fewest consequences abroad compared to participants in the other conditions. This demonstrated the added protective effect of SAF when combined with PNF. The SAF condition was also graphed for reference purposes as this interaction was not significant in the model.

Figure 2. Baseline RAPI Moderates PNF and PNF + SAF Conditions Versus Control.

Note: The dependent variable (abroad RAPI score) functionally represents the average RAPI score per month across the two abroad time points (first month abroad, last month abroad) within each individual.

Discussion

The present research study addressed the public health concern of problematic drinking among an at-risk population lacking empirically supported intervention. Utilizing prior research to inform intervention content, these brief Internet-based interventions targeting established risk (misperceived drinking norms; PNF) and protective factors (positive sojourner adjustment; SAF) aimed to prevent increased and problematic drinking for college students while abroad. After controlling for covariates and baseline drinking, we observed main effects for the PNF intervention only, with those participants receiving PNF drinking less over the course of the trip compared to control participants. However, findings suggested that a PNF “one-size-fits-all approach” may not be appropriate for all study abroad students. Indeed, baseline heavier drinkers that received PNF reported greater levels of drinking during abroad than those heavy drinkers not receiving PNF. The addition of SAF to the PNF intervention appeared to exacerbate the effects of increased drinking by the PNF condition, with combined condition participants with heavier levels of predeparture drinking reporting comparatively more drinking than their control counterparts. For alcohol-related consequences, there were no significant main effects of any intervention condition on preventing consequences compared to control. However, PNF participants with fewer baseline consequences reported fewer consequences abroad compared to control. Also, combined PNF + SAF condition participants with varying levels of predeparture consequences reported comparatively less consequences abroad than their control counterparts, with differences between this intervention condition and control most pronounced for those with the greatest number of predeparture consequences. Taken together, PNF may be helpful for lighter drinkers and those with few consequences at predeparture, and the addition of SAF to PNF may help prevent consequences abroad for those reporting more consequences at baseline. In other words, SAF appears important to present in addition to PNF to those reporting more consequences at predeparture, while PNF alone may be sufficient for lighter predeparture drinkers.

It is noted that we did not observe intervention effects for the SAF condition alone, suggesting that there is something important about the combination of SAF and PNF that is beneficial for students. That is, simply receiving encouragement and information about engaging the culture abroad is not sufficient to prevent heavy drinking and consequences. Indeed, those who received SAF alone reported the most drinking abroad. In our prior work, we found that students who reported little assimilation into the host culture (i.e., limited sojourner adjustment) while they were abroad who also believed that host country people drank heavily were the heaviest drinkers abroad (Pedersen et al., 2011a). Thus, the combined intervention of PNF and SAF may work via the mechanism of targeting important risk (correcting high perceived norms) and protective (promoting sojourner adjustment) factors of study abroad drinking. During the PNF presentation, students learn that drinking is less than they perceived it to be, but without promotion of active sojourner adjustment, they may never engage the culture enough while abroad to observe that the norms we presented them were actually accurate. Thus, those heavy drinkers receiving PNF alone may discount the norms at predeparture and spend their time abroad with other Americans in bars and clubs never learning that drinking is quite moderate in their host country. On the contrary, those primed to look for moderate norms by the PNF who also review SAF, may have been influenced by the SAF intervention to engage their environment (e.g., talking with local people, exploring cultural activities) and been in drinking situations to observe the moderate drinking behavior of local people (e.g., drinking wine with dinner). As students engage their environment more, they may drink more moderately and limit their experience of alcohol-related consequences to fit in with the observed cultural norms in the abroad environment.

All participants experienced increases in their drinking behavior from baseline to abroad. Yet while the figure for drinks per week indicates that heavier predeparture drinkers receiving SAF alone or SAF combined with PNF drank more abroad than control, differences were perhaps not clinically meaningful. For example, there is a difference of only one greater drink per week for combined PNF + SAF compared to control for the heaviest baseline drinkers. Drinking within a culture that promotes moderate drinking (e.g., drinking with meals in European countries) would likely explain increases in total drinking behavior. Still, participants in all conditions increased their drinking abroad, with even lighter drinkers reporting almost double their drinking abroad compared to predeparture levels. Even in the absence of the heavy consequences we assessed, these students may be placing themselves at increased health risks. Targeted efforts for the heaviest drinkers are still needed.

The most striking findings from this study were for consequences and consequences may be the more important outcome to target in prevention efforts. Control participants with both many and few consequences at predeparture experienced more consequences abroad than participants in any of the intervention conditions and there were significantly greater consequences for control participants compared to PNF and PNF + SAF participants with multiple predeparture consequences. Consequences that seem benign on campus may be exacerbated abroad or develop into long-term problems due to limited access to resources and familiar coping strategies (e.g., being far from friends/family; being unfamiliar with local healthcare locations or law enforcement policies). The lack of practice drinking in this unique context could also create a wider range of abroad-specific consequences than those that are present on campus (e.g., offending host families, losing passports, disrupted travel plans). Moreover, even though we found superior effects for the combined PNF +SAF condition, these participants still reported a number of consequences abroad, which could have the potential to ruin study abroad experiences, have lasting psychological and academic effects, and cause liabilities for home institutions. This suggests that while our approach was promising, more efforts are needed to better prevent the escalation of both drinking and consequences abroad for students.

Nevertheless, this study demonstrates the promise of providing students with a theoretically-informed program, as the current status of effective programs for many institutions is very limited. Student affairs personnel working with study abroad students cite drinking abroad as a major unaddressed problem and recognize the need for targeted empirically-supported efforts with these students (Forum on Education Abroad, 2009; LaBrie et al., 2013; Mello, 2015). Despite a need to address drinking while studying abroad, there have been no prior evaluations of empirically-supported interventions tailored toward preventing problematic drinking during this period of known high-risk, which from a public health perspective makes the contribution of the present study pronounced.

Limitations

Limitations related to the sample utilized and the design of the intervention were evident. While the recruitment and data collection techniques yielded similar demographics to typical study abroad students at the university and adequate retention rates, these findings may not be generalizable to students in other regions of the U.S. or to students from smaller home institutions. For example, our sample contained more women and ethnic minority students than might be seen in the study abroad population as a whole (IIE, 2015). Data collection was also based on self-report, which has potential bias. However, research suggests confidential surveys enhance reliability and validity of self-report (Babor & Higgins, 2000; Babor, Stephens, & Marlatt, 1987; Chermack, Singer, & Beresford, 1998, Darke, 1998) and response rates are higher for web than mailed surveys (McCabe, Couper, Cranford, & Boyd, 2006). Participants may even be more comfortable reporting on illegal or socially undesirable behaviors using online and computer-based methods (Pedersen, Grow, Duncan, Neighbors, & Larimer, 2012a; Turner et al., 1998). In addition, we recruited a large selection of study abroad students but only invited participants who were studying abroad for approximately two to three months. It became clear after study design was finalized that nearly 60% of the participants who did not meet our screening criteria failed to be invited because they were studying abroad for approximately one month. These brief trips abroad are becoming increasing popular (IIE, 2015) and finding ways to engage the culture even within a brief time period can possibly help students increase connection to the environment and prevent negative drinking incidents. It will also be important to examine the drinking behavior of these short-term students once they come back to determine if immediate intervention is needed to prevent continued heavy drinking at home. Lastly, personalized feedback (especially the SAF) was text heavy and required active motivation to engage the content provided in the links. While the content for the PNF was developed based on theory and empirical research, the content also was informed by the available data we had for study abroad peers (from our prior work) and from host national referents (from the WHO’s Global Status Report in 2004).

Conclusion and Future Directions

This research represents a first step in a novel area with an at-risk young adult populations that currently lack empirically supported interventions targeting risky drinking behaviors. The effects observed for interventions in this study suggest that future work with larger and more diverse samples is warranted. Researchers can examine similar intervention designs with students from different regions of the U.S. or can determine if in-person delivery of intervention content in group formats is equally or more efficacious compared to the online delivery of content. As research with study abroad students is nascent, it is important to expand understanding of the mechanisms behind increased and problematic drinking abroad so we can continue to design efficacious interventions. As most universities and colleges do not have empirically-supported predeparture prevention programs for these students but find alcohol use abroad concerning, it will be especially important to inform study abroad office personnel and university administration of the research findings resulting from these studies so policies and programs can be implemented prior to departure to aid in the prevention effort

In addition, the ideas of promoting positive sojourner adjustment into novel foreign environments and correcting foreign environment perceived drinking norms are innovative and may be applicable for adaption of use with the over 6 million young adults currently living abroad temporarily (Association of Americans Resident Overseas, 2013; Department of Defense, 2013); with diverse groups such as travelers, expatriates, foreign aid workers, and military personnel in non-combat areas also at-risk for problematic drinking or other risky behaviors such as unprotected sex and cigarette/drug use abroad (Bray et al., 2007; Cardozo & Salama, 2002; Hawkes et al., 1994; Moore, Beeker, Harrison, Eng, & Doll, 1995). While notable differences exist between American college students studying abroad and these other groups, developmental and contextual similarities are evident, such as susceptibility to peer influence, increased independence and unstructured time, developmental processes related to identity exploration and risk of engaging in risky behaviors, and difficulties with transitions to life away from familiar family, friends, and environments (Baer, 2002; Brown et al., 2008; Erikson, 1968; Masten, Faden, Zucker, & Spear, 2009; National Institute of Alcohol Abuse and Alcoholism, 2006). Examination of the risk and protective factors specific to other groups transitioning to life abroad can help inform future prevention of alcohol and other risk behaviors. Predeparture interventions with these groups in accessible and relevant formats that utilize group-specific risk and protective factors may help prevent increased and problematic behavior during transitions to new environments.