The initial phase of this investigation sought to qualitatively explore the contextual factors infuencing one's responsible drinking practices. Due to the dearth of systematic, published investigations into responsible drinking, this phase encompassed a series of less structured focus group sessions. Less structured groups are an ideal choice when researchers do not have prior knowledge/insight into the topic they are investigating (Morgan, 1998
). An “emergent design“ approach also guided the data collection process (Lincoln & Guba, 1985
Participants. Participants were recruited from several health promotion core-content courses offered at a large, four-year public university in Texas. The first author visited randomly chosen class sessions to provide information regarding the purpose and overall objectives of the study. Once informed about the objectives, students could indicate their interest in participating in the study, by providing (on an index card) their name, e-mail address, and the most convenient day(s) and time(s) for meeting with a focus group. The final sample size comprised 13 individuals, two men and 11 women. The majority of the sample was Caucasian (n = 11), with nominal representation of African-Americans (n = 1; female), and Hispanics (n = 1; male).
Data Collection. Prior to beginning each focus group session, participants reviewed and signed an IRB-approved informed consent form. As outlined in the consent document, each session was audio-taped. In total, four focus group sessions and three individual interviews (necessary in order to accommodate participants' scheduling conflicts) were conducted. The first author facilitated each focus group session and conducted each interview. During focus group or interview sessions, participants were asked to discuss the contextual factors infuencing their responsible drinking behaviors. Specifically, participants were asked a series of questions including, but not limited to, “Can you think of any barriers and/or obstacles that would prevent you or your peers from drinking re-sponsibly?”, “What types of situations would impact your ability to drink responsibly?”, “Can you think of a situation in which it would be impossible for you or someone else to drink responsibly?”, and “Can you provide some examples of how you could drink in a more responsible manner?” During discussions, participants were probed to provide Specific norms, attitudes, and/or beliefs associated with their contributions. All identifying characteristics or personal descriptions (i.e., name, age, etc.) were removed from the typed transcripts as well as from any presented or published accounts of the sessions, to ensure confidentiality.
. At the conclusion of each session, typed transcripts were developed from the audio-recordings. Using the constant comparison model (Lincoln & Guba, 1985
) to sort and classify recurrent or significant themes within participants' responses, we identifed (highlighted the text) and extricated (placed the text into an unconnected document) each distinct idea/thought from all transcripts. Each idea/thought unit was then grouped with similar ones, and each group of similar ideas formed a category and received a label. Each labeled category became an overarching theme. This particular method of thematic analysis has also been referred to as a general inductive approach (Thomas, 2006
(1) FINDINGS: Contextual Factors Infuencing Responsible Drinking. Overall, the emerging contextual factors infuencing the practice of responsible drinking were either intrapersonal (i.e., internal) or environmental (i.e., external) in nature. Grounded in the Ecological Perspective (McLeroy, Bibeau, Steckler, & Glanz, 1988), intrapersonal factors comprise a person's cognitive, affective, or behavioral traits/characteristics that influence behavior, such as knowledge, attitudes, and beliefs. The intrapersonal factors emerging from the focus groups and interview sessions, deemed to influence one's decision to drink responsibly, centered on “personal responsibilities” and “emotional status” at the time of drinking. Environmental factors infuencing responsible drinking, on the other hand, comprised interactions between family, friends, and peers as well as social networks and/or norms associated with group membership. Specific environmental factors emerging from the focus groups and interview sessions included subthemes of “surrounding environment,” “monetary considerations,” “drinking games,” and “other people.”
(2) Intrapersonal Factors: Personal Responsibilities. Participants believed their school-related obligations represented a salient factor impacting their responsible drinking practices. For example, the following days' school schedule emerged as a factor affecting decision-making: “Do I have class at eight in the morning? If so, maybe I shouldn't stay out until four O'clock.” Another participant echoed the impact of school-related factors, saying:
You may go somewhere [social environment, bar, etc.] knowing you have a test the next day and say “I am just going to go for a little bit. I will just have one or two drinks.” But then if you have one or two drinks, you might come home and be tired, not study for that test like you were supposed to, and do poorly.
In addition to school-related factors, scenarios involving personal accountability and responsibility also emerged. One such example was the health of an unborn baby, or “not drinking during pregnancy.” This participant elaborated on the manner in which her motherly duties influenced her drinking behaviors, stating:
I know that if I go somewhere and I drink and then I have to come home and take care of my kids I know I cannot do it well if I have had drinks, or maybe the next morning I will not be able to be as attentive as I would have been if I didn't have drinks.
Thus, personal responsibilities, whether school or family-related, emerged as an intrapersonal factor infuencing one's decision to drink responsibly.
(1) Emotional Status. An individual's emotional state/status also emerged as a factor potentially impacting responsible drinking behavior. This subtheme revolved around feelings, such as anxiety, stress, or depression. For example, one participant stated, “sometimes you are feeling depressed or stressed out and you want to get away from that, and that causes people to drink more than they usually would.” Participants cited Specific, hypothetical, instances that could impact one's decision to drink responsibly. Some of these instances inhibiting ability and/or decision to drink in a responsible manner, ranged from work-related circumstances [“a bad day at work”] to relationship problems [“boyfriend or girlfriend broke up with them”].
(2) Environmental Factors: Monetary Considerations. Participants felt carrying an excess of cash while in a social setting, such as a bar, could influence drinking behavior. Specifically, excess money allowed additional drinks, should he/she feel the urge, to be purchased at whim. As a preventative measure, and strategy to limit consumption, participants allocated a set amount of money to an evening's-worth of alcohol. Participants felt this strategy prevented them from giving in to temptation and assisted in limiting alcohol consumption. In other words, persons used money as a meter for determining how much had been (and could be) consumed; once the money was gone, drinking stopped. For instance, one participant revealed “I will only bring enough money for a certain amount of drinks. If you cannot afford to have more, you aren't tempted to drink more because you don't have the money for it.” Another participant echoed this sentiment, stating “I make sure not to bring a bunch of money to the bar that I could spend on drinks.” In addition to applying this concept in a social setting, others implemented this strategy when purchasing alcohol for consumption at home: “I only like to buy a certain amount because I will know exactly how much I am going to drink. That makes me feel very comfortable.” When probed to develop this point, the participant offered the following analogy: “It is like cooking extra; if you don't cook extra food then you will not eat more.” Similarly to cash, the utilization of credit cards emerged as an influence on responsible drinking. Specifically, the “open tabs” that credit cards allow was characterized as a detriment to responsible drinking behavior. By keeping an “open tab,” participants felt it was easy to lose track of both the amount spent on alcohol, as well as the number of drinks consumed. For these reasons, one participant identifed credit card use at a bar as “terrible.” This individual elaborated, stating “my friends get open tabs with their credit cards and by the end of the night they are like ‘wow, I did not realize this!”’
(3) Surrounding Environment. Another external factor potentially impacting responsible drinking included the surrounding environment. This notion applied not only to the location where drinking occurs (physical environment), but also to the individuals present at the time (social environment). With regard to the physical environment, participants felt one needs to “be in a safe setting.” Participants depicted a safe environment as, “go to someone's house to drink all night and stay the entire night there.” In addition to considering where one will drink, participants insisted the individuals present should also be considered. As one person stated, “If you are going to drink alcohol you should feel comfortable and trust people you are around.” Thus, participants felt it was important to consider not only where one will be drinking, but also who is present and how well they're known. One participant asserted,
You need to make sure that you are around people you are familiar with. You have to trust the people you are around. If you are just in a bar getting drunk, you do not know the people you come into contact with.
Elaborating on the importance of one's drinking environment, and those in it, multiple female participants highlighted the dangers of leaving a drink unattended: “Watch your drink. Do not set them down because people could always put something in them.” Multiple female participants were wary of other individuals and fearful of being dosed with a “date rape drug.” These admissions clearly point to a gender-related concern, as none of the males in the current sample discussed or alluded to consequences associated with leaving a drink unattended.
(4) Drinking Games. The current sample also identifed participation in drinking games as precluding drinking in a responsible manner. Borsari (2004) asserts, “Drinking games often stimulate a competitive environment, replete with winners, losers and spectators” (p. 37). One participant echoed this sentiment, saying participation in a drinking game was “the same as if you were in sports or an athletic competition.” The participant elaborated, “this person says that they can do this many shots, well then I have to one up that. I cannot let that person out show me.”
(5) Other People. Lastly, other persons were also cited as impacting one's decision to drink responsibly. One form of influence alluded to in our sample was peer pressure. Specifically, participants provided insight into two distinct types of peer pressure infuencing responsible drinking: indirect and direct. Borsari and Carey (2001) characterize these separate, Specific types of peer pressure in a review outlining the peer in-fuences on college alcohol use, stating: “Direct (or active) peer influences explicitly focus on getting a person to drink, and can range from polite gestures (e.g., offering to get a peer a drink, buying a round) to overt commands or encouragement to drink (e.g., forcing others to drink during drinking games)” (p. 393). Indirect peer pressure, however, includes providing “information about what behaviors are accepted and admired, what is considered appropriate in a given social context, and therefore what behaviors are likely to lead to social acceptance and reinforcement” (p. 393).
Instances exemplifying direct peer pressure among our sample were primarily related to hazing. However, the apparent pressure applied on participants was not restricted to brute force. For instance, the simple gesture of offering a drink could have profound influence: “I think a single girl just going to the bar with her friends to have fun and plans on not drinking, then a cute guy buys her a drink, I think that she would probably take it.” Thus, whether being propositioned or forced to consume alcohol, one's ability to drink responsibly was cited as being hindered by direct peer pressure. Examples of indirect peer pressure primarily stemmed from social norms and group membership. One participant explained, “if all your friends' activities revolve around drinking, then it becomes part of hanging out. You don't want to miss out on time with your friends so you drink too.” Said differently, “If everyone is at the party and everyone is drinking then you might feel left out.”
A subtheme associated with the influence others exert on responsible drinking behavior emerged in the concept of a “designated caretaker.” Participants in our study defined this drinking “buddy-system,” as having another individual, whom the drinker trusts, make decisions for the drinker while he/she is intoxicated. This practice was identifed as a method not only to ensure the drinker's safety, but also to accomplish responsible drinking. As one participant explained, a caretaker is “there to watch out for you and make sure that you are going to have a safe way to get home and that no one can take advantage of you.” At its core, this theme centered on the concept of having a known, trusted friend who could “make wise choices about any situation that may arise that night.” A participant-provided analogy equated a designated caretaker to “a sort of mother hen.”
At the conclusion of Phase One, the aforementioned themes were used as the foundation for the development of two quantitative scales. In other words, based upon the participant-provided contributions, a set of items was developed to address the contextual themes emerging from Phase One. Once developed, these scales allowed for the second phase of this study to be completed. Specifically, we (a) assessed the prevalence of the contextual factors among a sample of college students attending a large, Texas, public institution, and (b) determined whether the prevalence of these factors varied as a function of both sex and alcohol consumption.
Sample. A randomly selected sample of currently enrolled students was solicited via e-mail to participate. Of the 5,000 invitations sent, 4,985 were received. Of those successfully contacted, a total of 729 (15% response rate) students returned usable surveys. On average, respondents were 22 years of age (SD = 5.49). Females (55%) represented a slight majority in this sample. Ethnically, the majority of the sample was Caucasian (76%); while the remainder were Hispanics (10%), Asian or Pacifc Islander (7%), African-American (2%), Eastern Indian (1%), American Indian or Alaskan Native (0.5%), Middle Eastern (.5%), or Other (2.3%). That said, it is important to note that the demographic (gender and ethnic) distribution of this sample is comparable to the population from which it was drawn. While primarily Caucasian (73%), the university's three largest ethnic groups are also Hispanics (11%), Asians (4%), and African-Americans (3%). Females represent 47% of the institution's student body and were, therefore, slightly over-represented in our sample. Considering the low proportion of individuals in ethnic backgrounds other than Caucasian and Hispanic, for data analysis purposes all non-Caucasian and non-Hispanic ethnicities were collapsed into an “Other” category in order to ensure enough participants were included in each ethnic group.
Measures. To assess the prevalence of contextual factors infuencing college students' responsible drinking practices, two distinct scales were employed. All included scale items were developed based upon the themes emerging from the qualitative data from Phase One. For example, on the responsible drinking motivations scale (21 items), respondents were asked to indicate if work- or school-related obligations, monetary concerns, or the actions of others would motivate drinking responsibly. Additionally, as part of the responsible drinking barriers scale (16 items), respondents indicated if being challenged to a drinking contest, peer pressure (both direct and indirect), or performing badly on school assignments would be a barrier to drinking responsibly. For each item in the scales utilized, respondents were asked to indicate whether the given conditions/situations served as a potential motivator/barrier to drinking responsibly (1) never, (2) seldom, (3) some of the time, (4) most of the time, or (5) always. See and for Specific wording of each item and its associated response scale.
Responsible drinking motives across genders
Responsible drinking barriers across genders
Before analyzing the data produced from these scales, the internal consistency (i.e., reliability) of the motivation and barriers scales employed in this investigation were evaluated. With this study's sample, the motivation scale exhibited a Cronbach coeffcient alpha of 0.87 and the barriers scale, a 0.91. Scales exhibiting coeffcient alphas within this range have been deemed “very good” (DeVellis, 2003
). Both scales also exhibited strong score validity. Specifically, a principal component factor analysis (PCA) revealed the 21 items of the motivation scale accounted for 57.7% of the total variance, while the barrier scale's 16 items accounted for 62.2% of the variance.
Data Collection. In order to obtain a representative sample, the investigators obtained a list of all currently enrolled undergraduate and graduate students attending the institution. This list contained students' full names and their corresponding e-mail addresses. Using the list as a sampling frame, a random sample was invited to participate in the study via e-mail. Contacted individuals had approximately seven days to complete the online survey before the link would become inactive. Reminders were sent to those who had not yet accessed or completed the survey on days three and five. In order to access the survey, individuals clicked on a hyperlink embedded within the invitation and/or reminder e-mail. Persons clinking on the weblink were subsequently directed to an Information Sheet outlining informed consent matters. A “Go to Survey” link was located at the bottom of the information sheet. By clicking on the “Go to Survey” link, participants were confrming their understanding and voluntary acceptance of the study procedures.
. Prior to data analysis, participant non-response was examined. After taking into account non-response due to embedded skip patterns (e.g., abstained from alcohol consumption), all motivation scale items exhibited less than 1% missing data and all barriers scale items exhibited no more than 5%. Thus, the quantity of missing data for the measures was quite low. With regard to the pattern of missing data, nonresponse on both the motivation and barrier scales were examined. Independent sample t
-tests confrm that respondents with fully completed motivation scales did not significantly differ from those with missing responses with regard to gender [t
(689) = −0.530, p
< .107] or binge drinking status [t
(687) = 0.302, p
< .510]. Similarly, those who fully completed all the barrier scale items did not significantly differ from respondents with missing responses based on gender [t
(689) = −0.193, p
< .650] or binge drinking status [t
(687) = −0.426, p
< .447]. Due to the lack of large amounts of missing data, as well as the manner in which data were missing, it could not be justifed to employ complex analytical technique(s) to calculate imputable scores. Moreover, it was determined that data were missing at random (MAR). Researchers categorize data MAR as “ignorable” (Buhi, Goodson, & Neliands, 2008
). Consequently, incomplete surveys were retained for data analysis and missing data were deleted (listwise) from the analysis.
Utilizing the Predictive Analytics SoftWare (PASW) (version 18.0), basic descriptive statistics (Mean ± SD) were generated for each item on the two included scales. Sex- and consumption-based differences for each scale were calculated using one-way ANOVA. Cohen's d effect sizes were calculated for all statistically significant ANOVA fndings in order to further understand the strength of the observed relationship. Lastly, we employed logistic regression to assess the strength of the association among one's responsible drinking motivations, barriers, and reported drinking behavior. Age, sex, and race served as covariates for the logistic regression analysis.
Results. In establishing the extent to which a factor served to facilitate or impede responsible drinking practices, there was a clear distinction between male and female responses across the motivations and barriers scales. Females were more likely to indicate that an item on the motivations scale facilitated responsible drinking, while less likely to indicate that an item on the barriers scale impeded responsible drinking. In other words, males identifed more factors as obstructing responsible drinking practices, while at the same time identifying fewer factors aiding drinking responsibly.
Women were significantly more likely to identify the following factors as a motivator to responsible drinking: When I drink responsibly, one of my motivations is… “because of my religious convictions” [F(1,513) = 4.546, p < .033]; “because I do not want to do anything out of my character I may later regret” [F(1,512) = 25.109, p < .0001]; “because I do not want someone to take advantage of me” [F1(1,512) = 170.784, p < .0001]; “because I want to have control over my actions” [F(1,513) = 18.643, p < .0001]; “because I do not want to get nauseous or vomit” [F(1,512) = 10.217, p < .001]; “because I want to be aware of and understand what is going on around me” [F(1,511) = 12.529, p < .0001]; “because I want to remember what happens” [F(1,513) = 19.489, p < .0001]; “because I want to keep my blood alcohol concentration (BAC) under 0.08%” [F(1,514) = 4.912, p < .027]; and “because I am with people I do not know very well or in a new environment” [F(1,512) = 11.444, p < .001]. See for mean male and female responses to the motivations scale items as well as gender-based differences.
Regarding factors serving as barriers to responsible drinking, women were significantly less likely to identify the following factors as barriers to responsible drinking: The next time I drink alcohol, I would not be able to drink responsibly if… “an attractive person wanted to buy me a drink(s)” [F
(1,491) = 11.393, p
< .001], “I was playing a drinking game” [F
(1,494) = 4.411, p
< .036], “I had someone challenge me to a drinking contest” [F
(1,491) = 10.726, p
< .001], and “I felt pressured by friends to drink” [F
(1,491) = 3.752, p
< .05]. See for mean male and female responses to the barriers scale as well as gender-based differences. Effect sizes (Cohen's d
) are reported in conjunction with all significant group differences presented in both and . To interpret this index, Cohen (1988)
supplies the following categorizations: small (.20), moderate (.50), and large (.80). While these designations have been criticized, these categories do provide insight into the degree to which the null hypothesis is false,
ANOVAs were also conducted to determine whether factors facilitating or impeding responsible drinking varied as a function of one's alcohol consumption (i.e., binge drinking status). Respondents were asked to respond to the following question “Think back over the last two weeks. How many times, if any, have you had fve or more alcohol drinks at a sitting?” Participant responses ranged from none (65.2%), to one time (13.2%), two times (8.0%), three times (5.5%), four times (3.3%), fve times (1.6%), six times (1.9%), seven times (0.3%), eight times (0.1%), and nine or more times (0.9%). In order to group respondents based upon binge drinking status, responses were dummy coded as “0” for no times within the past two weeks, and “l” for one or more times within the past two weeks. After grouping respondents based upon their binge drinking behaviors, the majority of the items for both the motivations and barriers scales now violated Levene's statistic (an underlying assumption of homogeneity of variance). Consequently, we employed Brown–Forsythe's robust test of equality of means to account for the statistically significant Levene's statistic. Thus, the p-values outlined in both and are based upon the Brown–Forsythe test.
Responsible drinking motives as a function of binge drinking
Responsible drinking barriers as a function of binge drinkings
Those who had consumed fve or more drinks in one sitting at least once within the past two weeks also exhibited significantly different motivators to responsible drinking. Specifically, binge drinkers nearly universally reported each of the specifed factors would be less of a motive for drinking responsibly than nonbinge drinkers. In other words, nonbinge drinkers acknowledged more motives to drink responsibly than binge drinkers. Of the 21 items included in the scale, there was a statistically significant group difference for 16 of them. Effect size relationships ranged from 0.21 (small—When I drink respon-sibly one of my motive is because I have to look out for one of my friends) to 0.83 (large—When I drink responsibly one of my motive is because I do not want to get drunk). See for mean responses of binge drinkers and nonbinge drinkers and effect sizes.
While binge drinkers reported nearly all the motives scale factors as less likely to facilitate responsible drinking, they also identifed the vast majority of items on the barriers scale as factors inhibiting responsible drinking. Thus, binge drinkers identifed more barriers to drinking responsibly when compared to their nonbinge drinking peers. Binge drinkers were significantly different from nonbinge drinkers in 13 of the 16 items on the barriers scale. Specifically, binge drinkers identifed all but one factor on the barriers scale as an impediment to responsible drinking, compared to their nonbinge drinking peers. Effect size relationships ranged from 0.28 (small—I would not be able to drink responsibly if I felt depressed or stressed out) to 0.75 (large—I would not be able to drink responsibly if I was playing a drinking game). See for mean responses of binge drinkers and nonbinge drinkers and effect sizes.
In order to assess the strength of the association among one's responsible drinking motivations, barriers, and reported drinking behavior, we conducted a multivariate logistic regression analysis. Specifically, we sought to determine whether one's composite score on the motivations and barriers scales (independent variables) would be associated with individuals' binge drinking behavior within the past two weeks (dependent variable), after controlling for sex, age, and ethnicity. Prior to the analysis, multicollinearity was assessed to determine if any of the independent variables and covariates were highly correlated. Since the highest Pearson correlation between variables was 0.199, it was concluded that collinearity would not influence the logistic regression analysis. Overall, the full model was statistically significant (χ2 with 6 df = 99.422, p < .0001), indicating the model was able to distinguish between those who had, and had not, binged within the past two weeks. Further supporting the model's goodness-of-fit and overall reliability, the Hosmer and Lemeshow test was not significant (χ2 with 8 df = 10.803, p < .213). In all, the model as a whole explained between 20.8% (Cox & Snell R2) and 27.9% (Nagelkerke R2) of the variability associated with one's binge drinking status, and correctly classifed 71.2% of cases, an increase of 14% over the original model. As shown in , even when controlling for the covariates of age, sex, and ethnicity, the motivations scale (OR = .396, Wald = 22.044, df = 1, p < .0001) and barriers scale (OR = 2.548, Wald = 39.382, df = 1, p < .0001) exhibited strong associations with binge drinking status. Participants more motivated to drink responsibly were less likely to have engaged in binge drinking within the past two weeks (B = −0.926), while participants who perceived more barriers to drinking responsibly were more likely to have engaged in binging (B = 0.935). In addition to being Hispanic (OR = 3.423, Wald = 5.201, df = 1, p < .023), age was also significantly related to binge drinking status (OR = 0.944, Wald = 4.840, df = 1, p < .028): younger participants (B = −0.058) were more likely to have binged in the weeks prior to the survey.
Logistic regression predicting binge drinking status
In order to further assess the independence of the motivations and barriers scales, with regard to their association with binge drinking, a separate model was tested. Specifically, we conducted a subsequent multivariate logistic regression analysis that excluded the barriers scale, to determine the amount of shared variance across the two scales and their association to binge drinking status. Without the barrier scale included, the motivations scale (B = −1.041) association to binge drinking remained statistically significant (OR = 0.353, Wald = 35.022, p < .0001); however, changes in the odds ratio were minimal. Thus, while both the motivations and barriers scales are clearly associated with one's binge drinking, these two factors account for different aspects of this behavior.