GENDER DIFFERENCES IN COLLEGE STUDENTS' ALCOHOL REPRESENTATIONS IN FACEBOOK STATUS UPDATES.
Gender and college student alcohol use
Gender differences are consistently present in studies of alcohol-related perceptions and behaviors. Males are more likely than females to describe intoxication as acceptable and suggest that heavy drinkers do not suffer health problems (Svenson, Jarvis, & Campbell, 1994). Males also consume more alcohol per week, and are more likely to binge drink and receive a 12-month diagnosis of dependence compared to females (Bewick et al., 2008; Knight et al., 2002; Wechsler, Dowdall, Davenport, & Castillo, 1995). In contrast, females tend to perceive greater risk associated with alcohol use (Spigner, Hawkins, & Loren, 1993).
These gender differences in behavior are also reflected in language used to describe alcohol (Levitt, Schlauch, Bartholow, & Sher, 2013). A previous study used a survey-based vignette about a college student consuming alcohol to examine characterizations of peers' levels of inebriation. Male college students were more likely to use alcohol-related terms connoting intoxication, like "wasted". However, females tended to use language connoting more general alcohol use, such as "tipsy". This study concluded that gendered use of intoxication terms could contribute to increased alcohol use among males by suggesting that drinking to excess is normative. While the study examined verbal communication, it is possible that these gender norms could be conveyed through other means such as social media.
Social media and college student alcohol use
Social media sites such as Facebook are extremely popular and influential venues, particularly for conveying norms regarding alcohol use. As many as 88% of young adults report ownership of a social media account (Duggan & Brenner, 2013). Further, up to 60% of college students reference alcohol on Facebook by the end of their first year (Moreno et al., 2014). The Facebook Influence Model supports the influence of these alcohol references on their viewers' behavior (Moreno, Kota, Schoohs, & Whitehill, 2013). Personally referencing alcohol on Facebook may also influence one's own behavior, through the incorporation of the online self-portrayal into one's offline identity (D'Angelo, Kerr, & Moreno, 2014). These findings highlight the potential of social media in communicating norms regarding alcohol use.
It is possible that social media sites such as Facebook represent communication venues which convey and reinforce gender norms around alcohol use and binge drinking. However, little is known about gender differences in communication about alcohol on social media sites. Specifically, there is a gap in understanding whether male and female students use alcohol use terms, like "tipsy", versus intoxication terms, like "drunk", at different rates. Therefore, this study had two purposes. First, we sought to compare the proportion of male and female college students using general alcohol use versus intoxication terms on Facebook. Second, we aimed to understand other ways college students use language to reference alcohol on Facebook, and whether gender is associated with differences across these dimensions.
Data for this study was collected between fall of 2011 and spring of 2013. The setting included two large state universities, one in the Midwest and another in the Northwest, as well as the social media website Facebook. com. The relevant university institutional review boards approved this project.
Incoming freshmen students at either of the two study universities were identified through registrar lists, and approximately 600 were randomly selected to receive recruitment contacts. Over a four-week period, potential participants received up to four rounds of phone calls, emails and Facebook messages from the study team. Eligible participants were 18 years or older, English speaking, and owners of a Facebook profile. As baseline measures were intended to assess precollege experiences, students were excluded if they were already attending university in an early-enrollment program.
Phone interviews. Upon giving consent, participants were asked to accept a friend request from a researcher's Facebook profile, designed for the study, and to maintain open security settings with this profile. Investigators explained that participants' profiles would be evaluated toward gaining a sense of health messages viewed on the site. At this time, participants also completed phone interviews which assessed demographic variables including age, gender, and ethnicity/race.
Content Analysis. Throughout participants' first two academic years in college, this content analysis study evaluated displayed Facebook alcohol references present in status updates. Starting in the fall of the first academic year, each participant's Facebook profile underwent a monthly evaluation of the Timeline, such that all status updates created during each academic year were reviewed.
Obtaining Status Updates. While alcohol references may appear in many sections of Facebook, and in many forms such as memes and gifs, this study aimed to understand males' and females' own choices in language used in alcohol references. Thus, investigators recorded all status updates which included alcohol references, defining a status update as any original content by the profile owner which constituted the first post in a thread on the Timeline. We included user generated, text-based status updates only; we excluded shared status updates, photos, memes, gifs, song lyrics, and links to external content.
Extensive training, lasting a minimum of 3 months, preceded the evaluation of status updates for the presence of alcohol references. Each investigator evaluated a random subset of 20% of profiles to test interrater reliability. Fleiss' Kappa was used to evaluate overall agreement. For the presence or absence of alcohol references in each profile Fleiss' Kappa was 0.82, and regarding the number of alcohol references it was 0.74. These tests suggested high interrater reliability.
Once trained, investigators used an existing codebook to evaluate the presence of alcohol references in status updates (Moreno, Egan, & Brockman, 2011). This codebook was developed in accordance with the Theory of Reasoned Action constructs: attitude, intention and behavior (Montano, 2002). Facebook content was considered as an alcohol reference if it represented a positive attitude, intention or behavior regarding alcohol.
Status update codebook development. Status updates containing alcohol references underwent further evaluation. Three investigators developed a codebook in order to meet the study's two purposes. Toward the first study purpose, to understand gender differences in representations of general alcohol-related versus intoxicated states on Facebook, we used a deductive approach. Prior to analysis, we developed the code Level of Inebriation Term to capture references to states of inebriation, such as "tipsy" or "drunk", such as were studied by Levitt et al. (2013). Toward the second study purpose of understanding gender differences in the likelihood of using other manners to reference alcohol, we used an inductive, open coding technique. This process began with an initial review of eligible posts which led to the development of categories informed by the data, discussed below. Subsequently, researchers applied these categories to the data independently and refined categories as they identified discrepancies in their evaluations. After the codebook was finalized, interrater agreement between two researchers was high: 83.3%.
Variables. Our codebook included one deductive and six inductive categories which were not mutually exclusive; any status update could meet the criteria for multiple categories. The deductive category was Level of Inebriation Term, which was defined as a reference to any inebriation state such as "tipsy" or "drunk". Status updates in this group were subcategorized as either an Intoxication (e.g., "wasted") or a General Alcohol Use Term (a state of inebriation that does not necessarily imply intoxication, e.g., "buzzed").
The first inductive category was Type of Beverage, which applied to status updates mentioning specific alcoholic beverages such as beer or wine. Because this category emerged as an important one in our data, we conducted further analyses on status updates which referenced it. These status updates were subcategorized to capture specific alcoholic beverages mentioned including: Beer, Mixed Drinks and Cocktails, Vodka, Wine, Whiskey, Tequila, Champagne, and Other alcohol beverages.
Our codebook included five other inductive categories. A Mode of Consumption category referred to a physical means by which to consume alcohol such as a shot or a beer bong. An Event/Location category captured mentions of places, events, or parties at which drinking may occur. An Action Verb category referred to status updates using verbs such as "chug" to refer to the act of consuming alcohol. A Drinking Game category applied to mentions of games intended to encourage alcohol consumption, such as beer pong. Finally, an Alcohol as Concept category was defined as a comment on the concept of alcohol use in general, such as "I need some alcohol right now." See Table 2 for a list of example references from each deductive and inductive category.
Demographic variables as well as deductive and inductive codebook variables were evaluated with descriptive statistics. Investigators used Chi Squared tests to detect differences in presence and manner of alcohol references across gender, using the Bonferroni correction due to multiple comparisons.
A total of 338 students enrolled in this study, yielding a response rate of 54.6%; the retention rate after two years was 97.3%. Among these participants, 56.1% were female, 74.8% were Caucasian, and 58.8% were from the Midwestern University. More information on participants' demographic background is available in Table 1.
Prevalence of status updates with alcohol references
A total of 37.3% of participants shared an original, text-based status update including alcohol, and 58.2% of these posts were generated by female participants. Over this two-year study, participants generated 170 status updates which included alcohol references and fit our inclusion criteria of participant-generated language.
Prevalence of deductive and inductive categories
Among status updates with alcohol references, 87.1% fit into a single category, and 1.8% did not fit into any categories. 22.3% fit into multiple categories. The most common category referenced was Type of Beverage (34.1%). One participant posted, "If anyone else dumps their samples of free wine, I'll drink from the waste basin." Participants commonly referenced Level of Inebriation, which was present in 25.3% of eligible status updates. For example, one participant posted, "I'm very drunk. Don't judge me. #goodgrammarwhiledrunk." Participants also referenced Event/Location (18.2%), Action Verb (12.4%), Drinking Game (5.3%), and Alcohol as Concept (7.6%), and Mode of Consumption (7.0%). See table two for a full list of category frequencies and prevalence.
Gender differences in deductive and inductive categories
We found significant gender differences in three deductive and inductive categories. First, females were 2.5 times more likely to reference the deductive category Level of Inebriation compared to males (p=0.003). Second, females were more likely to reference an Action Verb (p=0.01). Third, males were more likely to reference Alcohol as Concept (p=0.01). We did not observe gender differences across Type of Beverage, Mode of Consumption, Event/Location, and Drinking Game. See table two for a list of gender differences across inductive and deductive categories.
Prevalence and gender differences in Level of Inebriation Term Subcategories
The majority of Level of Inebriation Term references (92.9%) were subcategorized as Intoxication Terms while few counted as General Alcohol Use Terms (7.1%). An example status update that fit into the Intoxication Terms category was: "My dorm window was just bombarded by snowballs #Fridaynight #Drunkpeople." A status update meeting the criteria for the General Alcohol Use Terms category was: "Tipsy Tuesday." Specific Level of Inebriation Terms included: drunk, blacked out, buzzed, crunk, not drunk, fucked up, get weird, go hard, hangover, messed up, passed out and sobering up. Of these terms, "drunk" was the most commonly referenced for both males (71%) and females (64.3%). No gender differences were observed in referencing Specific Level of Inebriation Terms (p=0.209). The frequency of references to these Specific Level of Inebriation Terms is represented in the word cloud in Figure 1.
Prevalence and gender differences in Type of Beverage Subcategories
The most commonly referenced Type of Beverage Subcategory was beer (30.5%), followed by mixed drinks and cocktails (22.0%), followed by vodka (11.9%). One participant's status update read: "Oh my! This beer is so cheap!" Participants also referenced wine (13.6%), whiskey (8.5%), tequila (7.8%), champagne (1.7%), and Other alcohol beverages (5.1%). For example, one participant shared: "Boys bring flowers, men bring wine." Male participants were more likely to reference whiskey than females (p=0.007). However, we did not observe gender differences in the likelihood of referencing any hard liquor (tequila, vodka, and whiskey combined) (p=0.133).
This content analysis study sought to understand gender differences in alcohol-related Facebook status updates posted by college students. Most commonly status updates with alcohol references included a Level of Intoxication Term or a Type of Beverage. Importantly, women were more likely than men to reference a Level of Intoxication Term. On the other hand, men were more likely to refer to Alcohol as Concept, describing their relationship to the substance, without commenting specifically on their state of inebriation. Over a third of students referenced alcohol in status updates that they generated.
Our main finding was that Level of Inebriation Terms were present in almost a quarter of alcohol-related status updates observed in this study, and nearly all of these terms connoted intoxication rather than general alcohol use. One possible explanation of this result is that, in general, only certain Level of Inebriation Terms are perceived as potentially interesting enough to warrant a Facebook post. A student may not anticipate status updates about being tipsy to generate a positive audience response, such as via Likes, but may suppose that a post about drunkenness would. In this sense, the norms around alcohol-related communication online may differ from those offline, with drunk-related posts favored. As a consequence, Facebook may over-represent the prevalence and frequency of binge drinking among college alcohol users.
A second important finding was that women were more likely than men to reference Level of Inebriation Terms on Facebook. Given that nearly all Level of Inebriation Terms referenced being drunk, this result is different compared to past studies on gender differences in alcohol use. Previous work suggests men have riskier alcohol-related perceptions and behaviors compared to women (Bewick et al., 2008; Knight et al., 2002; Spigner et al., 1993; Svenson et al., 1994; Wechsler et al., 1995). Further, online survey studies have supported the notion that men are more likely to use a term suggesting an intoxicated state than women (Levitt et al., 2013; Levitt, Sher, & Bartholow, 2009). A number of possibilities may explain this contrast. Previous studies used a survey with researcher-established terms indicating level of inebriation, whereas our study examined participant-generated terms. It is possible that a student's self-reported likelihood of using a given term differs from one's terms used in real life settings such as Facebook. However, another explanation may be that the online context of a social networking site plays a role in the decision to use a term indicating state of inebriation. While females may perceive social rewards associated with posting about drinking to excess, males may believe these rewards are primarily associated with offline conversations. Alternatively, it is possible that our study reflects shifting paradigms in the frequency of female college students drinking to excess.
This study's findings regarding a tendency to use terms related to drunkenness online are important to consider in light of the way college students interpret Facebook alcohol references. According to one study (Moreno, Briner, Williams, Walker, & Christakis, 2009), although adolescents acknowledge that alcohol references on Facebook may represent efforts to appear cool, these references are still taken at face value as indications of actual behavior. Therefore, based on our findings, a new college student inquiring which alcoholic beverage choice might assist with fitting in may consider beer. If this student were curious about what state of intoxication to reach, they might consider binge drinking to be normative, especially for women.
Our study's third finding was that over a third of college students referenced alcohol in user-generated status updates. This estimate is lower than previous work has suggested, with one study estimating that over 60% of college students reference alcohol on Face-book (Moreno, D'Angelo, Kacvinsky, Zhang, & Eickhoff, 2014). However, previous work has included content analysis of the entire Facebook profile, including the About, Likes, Timeline, and Photos sections, whereas the current study focused on a subset of content from the Timeline. It is possible that content from these sections, for example, tagged photos, accounts for this difference between our findings and past work. That said, it is worth noting that user-generated status updates account for the Facebook alcohol references of a sizable minority of college students. Given the prevalence of alcohol-related status updates among college students, examination of the manner of representing alcohol in these posts is warranted.
Our study's limitations warrant consideration. First, this study examined the Facebook profiles of students from two large state universities, and the extent to which its results can be generalized to other universities is unclear. However, the demographic makeup of study participants is consistent with that of the involved universities. Second, this study examined only user-generated status updates on Facebook, and this is not the only format in which an alcohol reference may be displayed. However, our results showed that as many as a third of college students created status updates referencing alcohol in two years. Third, the present study exclusively reviewed posts from Facebook, and it is not clear how these findings translate to other popular social media sites such as Snapchat. Future research should explore gender differences in alcohol references posted on other social media sites.
Despite these limitations, our study has important implications regarding how college alcohol-related norms may be propagated on Facebook. User-generated alcohol references in status updates, which often use terms connoting drunkenness, may lead to normalization of binge drinking. Further, women's increased likelihood of using these terms suggests differences in how women express alcohol use in person compared to on social media. These findings underscore the importance of media literacy efforts around social media use. Particularly, these efforts should encourage college students to identify and challenge their assumptions regarding the commonness of drinking to intoxication.
This work was supported by the National Institutes of Health [Grant R01DA031580-03].
Bewick, B. M., Mulhern, B., Barkham, M., Trusler, K, Hill, A. J., & Stiles, W. B. (2008). Changes in undergraduate student alcohol consumption as they progress through university. BMC Public Health, 8, 8. doi:16310.1186/1471-2458-8-163
D'Angelo, J., Kerr, B., & Moreno, M. A. (2014). Face-book Displays as Predictors of Binge Drinking: From the Virtual to the Visceral. Bulletin of Science, Technology, and Society. 34(5-6), 159-169. doi: 10.1177/0270467615584044
Duggan, M., & Brenner, J. (2013). Demographics of Social Media Users-2012. Washington, D.C.: Pew Research Center.
Hingson, R. W., Zha, W., & Weitzman, E. R (2009). Magnitude of and trends in alcohol-related mortality and morbidity among U.S. college students ages 18-24, 1998-2005. Journal of Studies on Alcohol and Drugs. Supplement. (16), 12-20.
Hingson, R., Heeren, T., Winter, M., & Wechsler, H. (2005). Magnitude of alcohol-related mortality and morbidity among U.S. college students ages 18-24: changes from 1998 to 2001. Annual Review of Public Health. 26. 259-279. doi: 10.1146/annurev.publhealth.26.021304.144652
Knight, J. R., Wechsler, H., Kuo, M. C., Seibring, M., Weitzman, E. R., & Schuckit, M. A. (2002). Alcohol abuse and dependence among US college students. Journal of Studies on Alcohol. 65(3). 263-270.
Levitt, A., Schlauch, R. C. Bartholow, B. D., & Sher, K. J. (2013). Gender differences in natural language factors of subjective intoxication in college students: an experimental vignette study. Alcoholism: Clinical and Experimental Research. 37(12), 2145-2151. doi:10.1111/acer.l2200
Levitt, A., Sher, K. J., & Bartholow, B. D. (2009). The language of intoxication: preliminary investigations Alcoholism: Clinical and Experimental Research. 55(3). 448-454. doi: 10.1111/j.1530-0277.2008.00855.X
Montano, D. E. K. D. (2002). The theory' of reasoned action and the theory of planned behavior. In K. R. B. Glanz & F. Lewis (Eds.). Health Behavior and Health Education: Theory. Research and Practice (pp. 67-98). Philadelphia. PA: John Wiley and Sons.
Moreno, M. A., Briner, L. R., Williams, A., Walker, L., & Christakis, D. A. (2009). Real use or "real cool": adolescents speak out about displayed alcohol references on social networking websites. Journal of Adolescent Health, 45(4). 420-422. doi: 10.1016/j.jadohealth.2009.04.015
Moreno, M. A., D'Angelo, J., Kacvinsky, L. E., Zhang, C., & Eickhoff, J. (2014). Emergence and predictors of alcohol reference displays on Facebook during the first year of college. Computers in Human Behavior. 50(2014), 87-94.
Moreno, M. A., Egan, K. G., & Brockman, L. (2011). Development of a Researcher Codebook for Use in Evaluating Social Networking Site Profiles. Journal of Adolescent Health. 49(1), 29-35. doi: 10.1016/j.jadohealth.2011.04.015
Moreno, M. A., Kota, R., Schoohs, S., & Whitehill, J. M. (2013). The Facebook Influence Model: A Concept Mapping Approach. Cyberpsychology. Behavior and Social Networking, 16(1), 504-511. doi:10.1089/cyber.2013.0025
National College Health Association. (2017). National College Health Assessment: Reference Group Executive Summary, fall 2016 Retrieved from Imp: www.acha-ncha.org/reports_ACHA-NCHAIIc.html
Spigner, C., Hawkins, W., & Loren, W. (1993). Gender differences in perceptionof risk associated with alcohol and drug-use among college students. Women & Health. 20(1), 87-97. doi:10.1300/J013v20n01_06
Svenson, L. W., Jarvis, G. K., & Campbell, R. L. (1994). Gender and Age-Differences in the drinking behaviors of University-Students. Psychological Reports. 75(1), 395-402.
Wechsler, H., Dowdall, G. W., Davenport. A., & Castillo, S. (1995). Correlates of college student binge drinking. American Journal of Public Health, 85(1), 921-926. doi: 10.2105/ajph.85.7.921
BRADLEY KERR, MS
Center for Child Health, Behavior and Development, Seattle Children's Research Institute Department of Pediatrics, University of Wisconsin-Madison
ANNIE YI, BS
Center for Child Health, Behavior and Development, Seattle Children's Research Institute Ellen S. Floyd College of Medicine, Washington State University
MEGAN MORENO, MD, MSED, MPH
Center for Child Health, Behavior and Development, Seattle Children's Research Institute Department of Pediatrics, University of Wisconsin-Madison
Table 1 Participant Demographics Demographic Variable Number (%) Gender Female 189(56.1) Male 148(43.9) University Midwestern 198(58.8) Northwestern 139(41.3) Ethnicity Caucasian/White 252(74.8) Asian 39(11.6) More than One 21(6.2) Hispanic 13(3.9) African American/Black 5(1.5) East Indian 3(0.9) Native American/Alaskan 2(0.6) Other 2(0.6 Table 2 Status Updates in Each Deductive and Inductive Category across Gender Reference Type Total, n Females, n Males, n (%) p value (*) (%) (%) Level of Inebriation 43 (25.3) 31 (31.3) 12 (16.9) 0.003 Term Type of Beverage 58 (34.1) 32 (32.3) 26 (36.6) 0.44 Mode of Consumption 12 (7.0) 9 (9.0) 3 (4.2) 0.07 Event/Location 31 (18.2) 17 (17.2) 14 (19.7) 0.45 Action Verb 21 (12.4) 16 (16.2) 5 (7.0) 0.01 Drinking Game 9 (5.3) 3 (3.0) 6 (8.5) 0.04 Alcohol as Concept 13 (7.6) 4 (4.0) 9 (12.7) 0.01 Reference Type Example Level of Inebriation I felt drunk off my a** and Term could barely stay awake after one sip Type of Beverage Yay, tequila! Mode of Consumption Clearly a good Friday night if I came home with a flamingo bong Event/Location My first time in Canada! Drinking legally is the best! Action Verb I'll drink beers for everyone! Drinking Game I play flip cup with wap because it starts the party Alcohol as Concept Never thought I'd say this, but I could really use some alcohol right now (*) p value obtained by chi square test with Bonferroni correction for multiple comparisons, significant values bold and italics
|Printer friendly Cite/link Email Feedback|
|Author:||Kerr, Bradley; Yi, Annie; Moreno, Megan|
|Publication:||College Student Journal|
|Date:||Dec 22, 2018|
|Previous Article:||COLLEGE STUDENTS' DISORDERED EATING, SUBSTANCE USE, AND BODY SATISFACTION.|
|Next Article:||DESIGNING MENTORSHIP: EXPLORING THE CHALLENGES AND BENEFITS OF UNDERGRADUATE RESEARCH.|