Drinking at Work: The Portrayal of Alcohol in Workplace-Related TV Dramas

This study analyzed the portrayal of alcohol in the most popular workplace-related TV dramas. Based on social cognitive theory, we coded the extent and context of all alcoholic and nonalcoholic beverage appearances as well as character-beverage interaction. Motivations to drink, topic, portrayed outcomes related to alcohol, and the characteristics of models associated with alcohol were also analyzed. More than 90% of the episodes contained depictions or references to alcohol. Although less alcohol was portrayed at the workplace compared to other locations, still nearly one third of all consumed beverages at work were alcoholic. Regardless of the location, character–beverage interaction was more likely for alcoholic than nonalcoholic beverages. The main motivation to drink alcohol was social ease, and only a few outcomes of alcohol consumption were presented. Overall, findings suggest that television dramas convey the image that alcohol at the workplace is ubiquitous, socially accepted, and without many serious consequences. Implications for mass communication and society are discussed.


INTRODUCTION
We see two of Manhattan's top lawyers, sitting in their office, contemplating work problems: "I need to hear some music. And drink some Scotch. How often do you really do one without the other?" (Sparling & Cropper, 2014). Dr. Bernard Prince (Nurse Jackie) is lying in a hospital chair commenting on his hangover and regretting those two bottles of wine before his shift (Mensch & Bernstein, 2015). Frank Underwood, vice president in House of Cards, and Charlie Skinner, CEO of a big news corporation in The Newsroom, frequently greet their guests with a whiskey or scotch, and of course every work success is celebrated with the best champagne. Those few examples taken out of popular American workplace-related TV dramas illustrate different ways that alcohol can be embedded in workplace situations on television.
Based on the mechanism of observational learning (Bandura, 1969), research suggests that alcohol on entertainment television acts as an "alcohol educator" (Wallack, Breed, & Cruz, 1987, p. 34). That is, alcohol depicted at the workplace can be a potentially consequential source of information for the audience regarding socially acceptable patterns of drinking and drinking situations, common motivations, or consequences of consumption. Yet the societal costs of alcohol at work are undeniable (Anderson, 2012). Despite the relevance of this topic, only a few studies so far have looked at the presence of alcohol in a televised workplace environment (Diener, 2007;Russell, Russell, & Grube, 2009;Van Hoof, De Jong, Fennis, & Gosselt, 2009).
In fact, there are two pressing research gaps: First, the few existing studies have treated the workplace as only one location among many others. However, workplace dramas are more likely to depict the workplace in great detail compared to other TV programs in which workplace scenes occur occasionally. Rephrased, if we want to find out how alcohol is depicted at the workplace in television entertainment, we need to analyze formats that deal with the workplace, that is, workplace dramas. Second, and more importantly, no study so far has analyzed the specific ways with which alcohol is depicted in a workplace environment. The present study attempts to fill these research gaps by, for the first time, investigating the presence of alcohol portrayals in workplace-related TV dramas. Based on social cognitive theory (Bandura, 1969), we content analyzed the interaction between characters and alcohol, the situations in which alcohol is shown, the displayed motivations and outcomes of alcohol consumption, and the characteristics of the models depicting alcohol-related behavior. These variables are key to understanding the potential effects of alcohol depictions in workplace dramas on viewers.

Factors Driving Retention
Extent of Alcohol Portrayal. A necessity in observational learning is the retention of the observed specific behavior. As Bandura (2002) stated, "People cannot be much influenced by observed events if they do not remember them" (p. 272). The most important factor increasing retention of the observed behavior is the extent of exposure (see Matthes & Naderer, 2016 for research on product placements). Therefore, it is crucial to investigate the amount of alcohol portrayals in televised workplace environments. A broad body of literature reports a strong presence of alcohol in the mass media. Mathios, Avery, Bisogni, and Shanahan (1998) stated that alcohol "is by far the most frequently shown food/beverage item" on prime-time television (p. 308). Pendleton, Smith, and Roberts (1991), for instance, found that 89% of all fictional programs in British television contained alcohol (see also Ryan & Hoerrner, 2004;Stern, 2005). In U.S. prime-time television series, Wallack et al. (1987) reported 78% of episodes showing at least one alcoholic drink. In television series across genres,  found that 88% of all episodes contain alcohol, and it is "present in at least one episode of every program coded" (p. 118). Furnham, Ingle, Gunter, and McClelland (1997) observed that 86.7% of all soap opera episodes depict alcohol, and Van Hoof et al. (2009) as well as Coyne and Ahmed (2009) found that even more than 90% of soap opera episodes feature alcohol.
Regarding alcohol at the workplace, however, there is only some preliminary evidence based on depicted workplace environments shown in general television series. For instance, Diener (2007) observed a 5% to 7% rise in soap opera alcohol portrayals at the workplace from 1986 to 2001, Van Hoof et al. (2009) found, in the same genre, that 12% of all alcohol depictions occurred in work-related situations, and  reported an average of three seconds of alcohol depiction at work per single episode of a youth-oriented drama series. Still, it is questionable if these findings can be generalized to series that show the workplace in greater detail. A TV series that mainly occurs at the workplace may totally differ in its depiction of alcohol compared to a series that merely features some scenes related to workplace topics. First, the quantity of alcohol depiction is arguably different. In nonworkplace series, the workplace is shown less frequently, and viewers thus receive less information about how alcohol is used at the workplace. Second, and because of that, workplace dramas are likely to show a larger variety of workplace situations, such as work frustration, success, interpersonal conflict, or after-work socials. If we want to understand how alcohol is depicted at the workplace, we thus need to analyze workplace series. In fact, no previous study available to date has done that. Therefore, and as one aspect driving retention suggested by social cognitive theory, we state the first research question: RQ1: To what extent is alcohol portrayed in workplace-related TV dramas? PORTRAYAL OF ALCOHOL Interaction. Another key aspect found to increase retention of product depictions in entertainment content is the interaction between depicted character and product (Naderer, Matthes, & Zeller, 2017). Kamleitner and Jyote (2013) stated that placements with character-product interaction "are better recalled and lead to more favorable attitudes, higher purchase intentions" (p. 643). Several prior works have looked at variables comparable to character-beverage interaction. Breed and De Foe (1984) defined a "drinking act" as ingestion or "being prepared to consume" (p. 260). Diener (2007) focused on the differentiation between no interaction and interaction, which was defined as "consumed/held" (p. 46), stating that more than 50% of alcoholic beverages were consumed or held in the 2001 sample. Van Den Bulck, Simons, and Van Gorp (2008) looked at passive (i.e., holding) and active (i.e., consumption, p. 935) interaction. Coyne and Ahmed (2009) recorded only cases in which interaction took place, differentiating between ingestion and preparation. They found 2.03 acts of holding a drink and 2.83 acts of ingestion per episode. Although experimental research suggests that the main difference is interaction versus no interaction, it is common in content analytical studies to also record explicit consumption. Therefore, our coding scheme will take all three aspects into account. We thus ask in RQ2: RQ2: What is the extent of alcohol interaction within workplace-related TV dramas?

Situation Characteristics
At the center of observational learning stands a behavior in a specific situation. Alcohol-related behavior is special in the sense that specific behaviors are more socially acceptable or more problematic than others. There are several important situational characteristics. First, we need to look at the location in which alcohol is shown, that is, if it is shown in a workplace environment or outside of work, as alcohol at work is one of the "generally acknowledged forms of alcohol misuse" (Van Hoof et al., 2009, p. 423). Several studies have recorded the location (e.g., Diener, 2007;, and others have mangled location with motivation (Van Hoof et al., 2009). This is the first study to differentiate between location, motivation, and topic of a situation. Related to this is the basic question of whether drinking is related to a work topic or a personal topic. A situation might take place at a bar but show the characters coping with work-related issues (i.e., workrelated topic). In this case, the viewer could still learn that alcohol is a way to deal with work problems, even if the situation does not take place at work.
Also, time of day matters because drinking in the morning is arguably socially less tolerated and more problematic than drinking in the evening. In the Alcohol Use Disorders Identification Test, morning drinking presents an item in the dependency section (Babor, Higgins-Biddle, Saunders, & Monteiro, 2001). Van Hoof et al. (2009) showed that 20% of alcohol consumption in a soap opera sample happened before 5 p.m. They stated that "adolescents may get the impression that alcohol consumption in the morning and early afternoon is a normal thing to do" (p. 428). Consequently, we analyzed the time of day. As to the number of characters, research suggests that heavily drinking alone can have more severe consequences than heavily drinking in a group (Christiansen, Vik, & Jarchow, 2002; see Van den Bulck, 2008). We thus need to analyze the number of characters (e.g., alone, talking to a friend, crowded party scene) in more detail. Taken together, in terms of observational learning, it is crucial to examine the characteristics (i.e., location, topic, time of day, number of characters) of the situations in which a model performs alcohol-related behaviors. This has not been done for work-related dramas. RQ3: What are the characteristics (i.e., location, topic, time of day, number of characters present) of the situations in which alcohol-related behaviors are shown?

Vicarious Incentives
Social cognitive theory suggests that the outcome of a certain behavior shapes the viewer's learning process of it. "The observed detriments and benefits experienced by others influence the performance of modeled patterns in much the same way as do directly experienced consequences" (Bandura, 2002, p. 129).
In the case of televised alcohol-related behavior, two types of vicarious incentives matter: depicted motivations and depicted outcomes.
Motivations. Studies suggest that exposure to alcohol-related scenes can shape audiences' expectations toward alcohol (Osberg et al., 2012). That is, viewers learn what kind of motivation can lead to alcohol consumption and whether such needs are successfully met by drinking. Carey and Correia (1997) distinguished two types of motives: escapist, personal, or coping motives and social, convivial, or celebrative motives. Those are directly related to different outcome expectancies. "The process may be consistent with either positive reinforcement (enhancing pleasure in social situations) or negative reinforcement (avoidance of social censure or ridicule)" (Carey & Correia, 1997, p. 104).
Aside from one study (Coyne & Ahmed, 2009), previous research has overlooked motivations in building outcome expectancies (e.g., Diener, 2007; or blurred the lines between motivations, situations, and consequences. In Van Hoof et al.'s (2009) study, for instance, 38% of the scenes were coded as social settings, 11% as culinary situations, 24% as to suppress problems, 12% before or during work, and 14% as other. It seems questionable, however, that those categories can be coded as mutually exclusive. That is, actors can consume alcohol in a social setting to suppress problems during work, or they can drink in a social setting to have fun. Also, Van Den Bulck et al. (2008) conducted an inductive qualitative analysis of the frames presenting alcohol in the TV series The O.C. "The top three reasons for drinking alcohol were a party (43.4%), a habit (22%), and stress relief (14.3%)" (p. 936). Again, the lines between setting and motivational variables are blurred in this study. Coyne and Ahmed's (2009) article reported the categories of coping and social facilitation in soap operas. They found that coping appeared in 7.4% of all scenes and that social facilitation was present in an overwhelming majority of 92.5%. Yet this could strongly differ in a TV drama focused on a work environment because there are arguably different motives for drinking at work compared to outside work. This leads to the next research question. RQ4: Which motivations to consume alcohol are depicted in workplace-related drama?
Outcomes. Social cognitive theory suggests two possible outcomes: observed benefits and observed costs. Alcohol consumption is special in this regard though, because Bahk (1997) showed that the mere lack of negative consequences with a portrayal of alcohol already has a positive influence on the attitudes toward alcohol. In other words, no explicitly depicted cost of alcohol consumption may be interpreted as a benefit. Research on alcohol depiction in non-work-related dramas consistently states that outcomes of alcohol consumption are rarely shown (Coyne & Ahmed, 2009;Furnham et al., 1997;Van Den Bulck et al., 2008;Van Hoof et al., 2009). Regarding the valence of the outcomes, "only 1.82% of all alcohol acts resulted in negative behavior by the character" (Coyne & Ahmed, 2009, p. 352). Likewise, Van Den Bulck et al. (2008) stated that "in 94.4% of the cases no consequences of drinking were shown; 3.7% of the alcohol acts were followed by a negative consequence; and the remaining 1.4% of the alcohol acts had a positive effect" (p. 936). Again, this might strongly differ in a work-related drama, as alcohol at work may cause a plethora of different problems. Therefore we ask the following: RQ5: How are the outcomes of alcohol consumption depicted in workplace-related TV dramas?

Model Characteristics
To elicit observational learning, the specific characteristics of the model matter (see, e.g., Matthes & Naderer, 2015). Bandura (2002) distinguished three factors: similarity between viewer and model, the prestige, and the attractiveness of a model.
Based on the assumption that models similar to the viewer are more likely to be trusted and copied, several studies have analyzed the gender of models (Bandura, 2002;Meijnders et al., 2009). For instance, analyzing a teen drama, Van Den Bulck et al. (2008) recorded the gender of the characters and found that "of all visual references to alcohol, 55% involved female characters" (p. 936; see also Coyne & Ahmed, 2009). To our knowledge, McIntosh, Bazzini, Smith, and Mills's (1999) study is the only one incorporating socioeconomic status (SES) and attractiveness. Analyzing models displaying alcohol-related behaviors in movies, they stated that alcohol drinkers were portrayed as having a higher SES than nondrinkers, being more attractive and romantically active. Regarding gender, however, they did not find any differences in the depiction of alcohol use. No studies so far have looked at model characteristics for work-related dramas. We thus coded gender, SES, and attractiveness in order to answer RQ6: RQ6: What is the gender, the SES, and the attractiveness of the models displaying alcohol-related behaviors?

METHOD Sample
We chose to focus on dramas only, because workplace-related dramas are more popular and have a higher international distribution than workplace-related comedy series. We drew a sample of the most popular (at the time of sampling) workplacerelated TV dramas from the CBS portal TV.com. We chose all workplace-related dramas falling under the top 150 shows, which resulted in a sample of eight workplace TV dramas: Grey's Anatomy; The Good Wife; Suits; House, M.D.; House of Cards; Nurse Jackie; Private Practice; and The Newsroom. We excluded crime dramas because their narrative usually does not unfold in one specific work environment. This sample choice encompasses both cable and broadcast programs. To avoid distortions caused by social or regulative changes and increase comparability between the series, we selected only episodes aired in the years 2010-2015. For each TV series, we randomly selected 10 episodes. This resulted in a final sample of 80 episodes, equaling around 53 hours of programming. This sample is much larger than those of previous studies. Furnham et al. (1997), for instance, analyzed five episodes from six series. Pendleton et al. (1991) included 24 hours of programming, whereas Van Hoof et al. (2009) looked at approximately 15 hours in total.

Measures
The main unit of analysis was any visual or verbal appearance of beverages or beverage brands, both alcoholic and nonalcoholic (n = 1,356), as well as depicted alcohol outcomes when no beverage was present (n = 24; e.g., a hangover after consumption). We call this the case level (Furnham et al., 1997). To put the presence of alcohol into perspective, we coded nonalcoholic beverages as a benchmark. This is important for the interpretation of alcohol appearances (e.g., Van Den Bulck et al., 2008). For instance, Harvey Specter (Suits) and his colleague drink liquor in the morning on the way to a meeting (Cowan & McDonough, 2013). However, considering the numerous times he has coffee in the morning, it becomes clear that alcohol in the morning is not his usual drinking pattern. Furthermore, if there were no differences between alcoholic and nonalcoholic beverages, we may simply observe a general pattern of beverage depiction, which is not specific to alcohol. As in previous studies (e.g., Van Den Bulck et al., 2008), the substance was defined as beer, wine/champagne, spirits/cocktails, and nonalcoholic beverage. In addition, we added hidden alcohol (i.e., wrapped in a paper bag) and, in case of a verbal reference to alcohol and alcohol-related behavior in general, alcohol overall. As mentioned, references to alcohol outcomes without alcohol being present in the scene were coded as outcome. This was the case, for example, when Dr. Bailey (Grey's Anatomy) enters the hospital and is asked if she is drunk; she answers, "Tipsy, a little tipsy" (Clack & Bailey, 2010). For all observations, we coded the modality (i.e., if the placement was visual, audio, audiovisual; see Van Den Bulck et al., 2008) and regarding visual placements, we coded prominence as well. A case was perceived as prominent if it was at least 30% of the screen size (e.g., a rack of alcohol behind a barkeeper) or at the same screen level as the characters. All other placements were considered background. In case of a prominent visual placement or an audio placement, we assessed the character-beverage interaction, that is, if a beverage was actually consumed, interacted with (i.e., held or poured), or not interacted with (i.e., standing on a table next to the character).
Location was coded for all observations to separate work from outside-ofwork contexts (i.e., home, work, bar, club, restaurant and café, store, outdoors, or other). For prominent visual placements, we assessed time of day in blocks as morning (6 a.m.-11: 59 a.m.), afternoon (12 p.m.-5:59 p.m.), evening and night (6 p.m.-5:59 a.m.), and unknown time. The time was coded based on explicit references (e.g., character greeting each other with good morning, characters referencing lunch), information out of the time line of the plot (e.g., characters getting out of bed, characters meeting after lunch), and the light of day. In the case of a prominent alcohol appearance, an audio alcohol placement, or an outcome scene, the number of characters present was counted up to five, then it was coded as a crowd scene. The topics were analyzed, differentiating between work topics and personal topics. If the topic was unclear, it was recorded as such. For example, a work topic would be McKenzie and Sloan (The Newsroom), sitting in a bar, having a drink, and discussing the program content of the next newscast (Sorkin & Glatter, 2012). Arguably, for nonprominent placements, the topic is uninformative (i.e., a glass of wine in the background, independent of the main character or situation).
If there was a character-beverage interaction with alcohol, we coded drinking motivations. Following Coyne and Ahmed (2009), social ease would encompass social, convivial, or celebrative motives of drinking, for example, celebration of work success, a party, or just relaxing in a bar together. Problem coping was coded, for example, when a character like Dr. Jake Reilly (Private Practice) sits in his office, having a drink while being upset about a work problem (Lim & Purple, 2012). Personal ease would be drinking alone to relax.
We reported alcohol outcomes by their time frame (short, medium, long) and valence (positive, negative). Short-term outcomes were defined as happening directly after the alcohol consumption, for instance, feeling dizzy. Medium term would describe all outcomes on the next day, for instance, a hangover or explicit reference to "a great night." Long term would be an outcome affecting the character more than 1 day after the alcohol consumption (e.g., getting married while drunk or having a car accident). In addition, we coded all specific types of referenced outcomes as an open variable.
To characterize the models related to beverage references, we coded the gender, SES, and attractiveness of recurring characters. SES took into consideration the occupation of the character and additional information on wealth. For instance, the lawyers in the series Suits or doctors in Grey's Anatomy were all considered of high SES. Donna, the secretary in Suits, or Jackie (Nurse Jackie), a nurse, were given an average score. A low score was given to Rachel (House of Cards), a prostitute, or Alvi (House, M.D.), a thief. Attractiveness took into consideration physical attributes (average, above-average, and below-averagelooking characters; see McIntosh et al., 1999).
We analyzed the data descriptively using chi-square tests. On the case level, each single case was taken into consideration in the analysis as one unit. This level was used to make statements regarding the type of alcohol that appeared or was referenced to, character-beverage interactions, motivation of interaction, topics, and outcomes. Regarding the portrayal of alcohol, a scene was qualified as such if one or more cases had the same location, characters present, and point in time (N = 774). For example, if numerous people were each holding a glass of an alcoholic beverage, this would only count as a single alcohol-related scene. We used this level of analysis to determine the number of scenes that are alcohol related and their setting.

PORTRAYAL OF ALCOHOL
We tested the reliability of the codebook in several steps. First, all variables were thoroughly defined and discussed, then two coders independently coded 80 minutes of randomly chosen parts of nonsampled episodes. We tested the coding using Krippendorff's alpha and discussed each case in order to fine-tune the coding scheme. In a second round, two coders independently analyzed 160 minutes of randomly chosen parts of nonsampled episodes. This led to satisfying Krippendorff's alpha scores of all variables: substance (α = 1.00), modality (α = 1.00), prominence (α = 0.72), location (α = 0.93), daytime (α = 0.71), interaction (α = 0.73), number of characters (α = 0.82), motivations (α = 0.93), and topic (α = 0.76), as well as gender (α = 1.00), SES (α = 0.89), and attractiveness (α = 0.73) of the model (Krippendorff & Craggs, 2016). Because there was no variation for the variable called outcome, Krippendorff's alpha could not be calculated, but the agreement between coders was at 100%. After ensuring the reliability, the sampled episodes were coded by one coder.

Extent of Alcohol Portrayal (RQ1)
To answer our RQ1, we first analyzed the extent of alcohol portrayal. In 92.5% (n = 74) of all analyzed episodes, alcohol is depicted or commented upon. House, M.D. had the lowest rate, with seven of 10 coded episodes containing alcohol depictions or references. Nurse Jackie, The Newsroom, The Good Wife, and House of Cards visually or verbally contained alcohol in each coded episode. There was no significant difference in the extent of alcohol portrayal between broadcast and cable content. On a scene level, this results in 774 scenes, out of which 48.4% (n = 375) were alcohol related. Moreover, 401 scenes took place at work, out of which 32.2% were alcohol related, whereas 373 scenes took place at another location, out of which significantly more-namely, 66.0%-were alcohol related, χ 2 (1, N = 774) = 88.30, p < .001. On the case level, we coded alcohol 645 times and nonalcohol beverages 711 times, which amounts to 46.7% versus 51.5%. An outcome without alcohol reference or appearance was reported in 1.7% (n = 24) of cases. The difference between work and nonwork mirrors the situation on a scene level, with significantly less alcohol appearances or references in the workplace (30.5% vs. 63.4%), χ 2 (1, N = 1,356) = 146.29, p < .001. For the exact results, see Table 1. Looking at the type of alcohol that was shown or referenced in detail, the substance was clearly identifiable in 578 cases. Of interest, strong liquor, was nearly as often depicted as weak liquor (i.e., beer and wine) with 52.8% of beer and wine (n = 305) and 47.2% (n = 273) of spirits. This was the case outside of work as well as at the workplace.

Character-Beverage Interaction (RQ2)
Regarding character-beverage interaction (RQ2) with alcohol at the workplace, in 75.5% (n = 71) of the cases in which alcohol appeared at work, the characters interacted with the alcohol, for instance, they held the glass or poured it. For alcohol shown outside of work, characters interacted with alcohol in 76.2% (n = 253) of all cases. This is not a significant difference. More important, overall, characters interacted with alcohol 76.1% (n = 324) of the time, whereas there was an interaction with nonalcoholic beverages only 44.5% (n = 272) of the time, a significant difference, χ 2 (1, N = 1,037) = 102.15, p < .001. This is the case for all locations (see Table 2). The same pattern appears for consumption, with 46.9% (n = 200) of alcohol shown being consumed against only 24.1% (n = 147) of nonalcoholic beverages, χ 2 (1, N = 1,037) = 59.06, p < .001. This is again the case at work and out of work. Furthermore, it has to be noted that out of all beverages actually consumed at work, 32.2% were alcoholic. This is significantly less than outside of work, where 70.7% of all consumed beverages were alcoholic χ 2 (1, N = 347) = 47.37, p < .001.  Note: Interaction includes holding the beverage as well as consuming it.

Situation Characteristics (RQ3)
Next we analyzed location, topic, time of day, and number of characters present (RQ3). If a scene was alcohol related, it took place at work 34.4% (n = 129) of the time, closely followed by home 24.3% (n = 91) of the time and bars 17.6% (n = 66) of the time. Of interest, only 8.5% (n = 32) of alcohol-related scenes happened in restaurants. In 51.9% (n = 291) of alcohol-related cases, the topic was an aspect of personal life, whereas in 43.7% (n = 245), work was discussed. In 4.5% (n = 25) of cases, the topic was unclear (those were taken out of the subsequent analysis). The topics were related to the location in which the alcohol appearance, reference, or outcome took place. For example at work, 58.5% (n = 100) of alcohol-related cases were work related and 41.5% had a personal topic, whereas outside of work, 60.3% (n = 220) were personal and 39.7% (n = 145) were work related, which is significant, χ 2 (1, N = 536) = 16.50, p < .001. It is interesting to note that 145 alcohol-related cases outside of work had a work topic. For instance 52.5% (n = 63) of cases in a bar had a work topic.
Regarding the time of day, 91.3% (n = 190) of beverages shown in the morning were nonalcoholic, whereas in the evening and night, 77.6% (n = 357) of the depicted beverages were alcoholic. Concerning the number of characters present in alcohol-related scenes, in the majority of scenes (57.1%, n = 196), two people are interacting; 12.0% (n = 41) were crowd scenes with more than five characters present, 24.5% (n = 84) showed two to five characters, and 6.4% (n = 22) were scenes with only one character. At work, significantly more alcohol-related scenes showed two characters (51.4% vs. 67.8%), χ 2 (1, N = 343) = 9.78, p < .05; there was no significant different in location for the other categories.
Overall, if a character is shown alone with alcohol, in 69.2% (n = 18) of these cases a spirit was depicted, whereas in scenes with five or more people, spirits were seen only in 24.4% (n = 21) of the cases. It is exactly the other way around for weak alcohol: The more people, the more beer and wine. If one person is shown, beer and wine were shown in 30.8% (n = 8), and five or more people were shown with beer and wine in 75.6% (n = 65). This is a significant difference, χ 2 (3, N = 477) = 26.64, p < . 001. This difference is significant at work, χ 2 (3, N = 121) = 18.79, p < . 001, and out of work, χ 2 (3, N = 356) = 14.75, p < . 05.

Motivations (RQ4)
To answer RQ4, we analyzed the depicted motivations. It can be stated that the most common drinking motive was social ease, with 61.4% (n = 205) of the alcohol interactions driven by it. The most typical situation in this context was a small group of main characters enjoying a drink together. This reoccurred with pairings such as Alicia Florrick and her colleague Kalinda Sharma (The Good Wife) or Harvey Specter and his colleague Mike Ross (Suits). Problem coping was recorded in 26.3% (n = 88) of the cases and personal ease in only 2.4% (n = 8). It must be mentioned that we coded no motive in 8.7% (n = 29) of the cases, because context was unclear. Looking at the motivations to drink depending on location, only social ease and problem coping can be considered, due to the extremely rare appearance of personal ease. There was no significant difference between locations.
Furthermore, in the context of a problem-coping situation, the characters chose strong liquor in 70.5% (n = 62) of the time; beer and wine totaled only 29.5% (n = 26). However, beer and wine were the choice in 65.2% (n = 131) of all alcohol interactions motivated by social ease, with strong liquors in only 34.8% (n = 70), χ 2 (1, N = 289) = 31.31, p < .001. There was no significant difference between locations in that matter.

Outcomes (RQ5)
Out of all prominent and verbal alcohol-related cases, 82.2% (n = 467) did not show any outcomes (RQ5). With regard to the workplace, out of all alcoholrelated cases, 26.3% (n = 46) were related to outcomes, whereas only 14.0% (n = 55) of alcohol-related cases in other locations were related to outcomes, χ 2 (1, N = 568) = 12.51, p < .001. Looking at all the outcomes depicted or mentioned in the series, it becomes clear that negative outcomes by far outweigh the positives ones-74.3% (n = 75) against 25.7% (n = 26), respectively. The most referenced short-term negative outcomes were drunken conflict (n = 5), vomiting (n = 4), and drunken depression (n = 4). Drunken conflict ranged from a drunken relationship fight between Don and Maggie in The Newsroom to a victim of sexual harassment testifying in court on her attacker being drunk (The Good Wife). Medium-term negative outcomes were hangovers, being drunk at work the next day, and comments on drunken dialing (n = 2). Longterm negative alcohol outcomes were mostly alcoholism (n = 17), followed by problems at work due to alcohol-related behavior (n = 14). Positive short-term outcomes of alcohol were references to happiness and fun thanks to alcohol (n = 15) and sexual courage due to alcohol (n = 4). In those cases, alcohol plays a major role in plot advancement, for example, when Dr. Christina Yang finally admits her romantic attraction to Dr. Owen Hunt while being drunk (Grey's Anatomy). Positive medium-term outcome was coded in only one case, in which drunken sex was commented upon the next day. Positive long-term outcomes were also coded only once, where the glamour of alcoholism in Hollywood was referenced. Outcomes are significantly more often negative at work 87.0% (n = 40) than outside of work 63.6% (n = 35), χ 2 (1, N = 101) = 7.13, p < .05 (see Table 3). However, it must be stated that the number of cases were small, especially for positive outcomes at work (n = 6). Furthermore, short-term PORTRAYAL OF ALCOHOL outcomes were the most prevalent, with 58.4% (n = 59) of all outcomes shown. When it comes to negative outcomes, there was a significant difference regarding the location. At work, the long-term ones were depicted or referred to in 60.0% (n = 24), whereas outside-of-work short-and medium-term outcomes were more common with 71.4% (n = 25), χ 2 (1, N = 75) = 7.44, p < .05 (see Table 3).

Model Characteristics (RQ6)
A total of 68.4% (n = 944) of all cases could be matched to a recurring character (RQ6). Out of those, 52.3% were male characters associated with a beverage (47.7% female). Men were connected to alcohol references in 46.0% (nonalcoholic beverages = 54.0%), whereas women were connected to alcohol in 49.3% (nonalcoholic = 50.7%). This is not a significant difference. There was also no significant difference if only visual depictions are taken into account, and there were no significant difference depending on location. If men referenced alcohol, it was significantly more often hard liquor (52.6%, n = 102); by contrast, woman choose weak alcoholic beverages more often (58.6%, n = 102), χ 2 (1, N = 368) = 4.62, p < .05. The majority (78.7%) of the recurring characters had a high socioeconomic background, because the analyzed workplace-related TV dramas focus on lawyers, doctors, politicians, and journalists. A fourth (19.7%) was categorized as average and only 1.6% as having a low socioeconomic background. Regarding beverage references, there was no difference between characters of different socioeconomic backgrounds. Also, there was no difference in the choice of hard versus weak alcohol. The majority of beverages (60.6%) were referenced by averagelooking characters (34.2% above average; 4.8% below average). Four cases were not rated because the characters were kids. There was no significant difference in beverage choice regarding attractiveness.

DISCUSSION
This study is the first to investigate alcohol portrayals in the genre of workplacerelated TV dramas in general and in the workplace environment in particular.
Results indicate a problematic portrayal of alcohol in workplace-related TV dramas and in the workplace context in which alcohol is embedded. First, alcohol was present in 92.5% of all episodes. However, contrary to Furnham et al. (1997), who looked at British soap operas and found "over twice as many alcohol as non-alcohol drinking scenes and visual references to alcohol" (p. 523), this study reports 48.4% of beverage scenes as being alcohol related.
Considering that these are drama series centered on a workplace environment, this is still a surprisingly high number. Regarding actual consumption, the numbers become even more remarkable, as 32.2% of beverages consumed at work are alcoholic, whereas 70.7% of beverages consumed outside of work are alcoholic. According to the World Health Organization (2011), the average amount of pure alcohol consumed per capita worldwide is 6.13 liters per year, whereas the yearly per capita amount of soda consumption comes to 43.15 liters (Basu, McKee, Galea, & Stuckler, 2013), not to mention the most frequently consumed beverages, such as water, tea, and coffee. Clearly, alcohol is highly overrepresented in TV shows. Nearly one in three beverages at work is alcoholic, and more than two thirds of beverages outside of work are alcoholic. For instance, Christina Yang, a main character in Grey's Anatomy, is connected to alcohol when shown at home, in a bar, or in a restaurant, never to a nonalcoholic beverage. To add to that, we find that around half of alcohol shown or referenced at work is even strong liquor (i.e., spirits). With social cognitive theory in mind, which postulates that viewers can vicariously learn behaviors and attitudes through observing models on television (Bandura, 2002), it is rather alarming how television conveys an overstated presence and acceptance of alcoholespecially spirits-at the workplace. Another worrying finding is the fact that characters interact with alcohol significantly more often, and they consume it significantly more often, as compared to nonalcoholic drinks. This finding is particularly important because Kamleitner and Jyote (2013) observed that placements with character-product interaction "are better recalled and lead to more favorable attitudes, higher purchase intentions" (p. 643). Therefore, it can be assumed that even if there are more nonalcoholic placements, the viewer may be influenced more strongly PORTRAYAL OF ALCOHOL by the alcoholic ones when there is a character-placement interaction. Following social cognitive theory (Bandura, 2002), it can also be assumed that a learning effect is stronger if the actual use of the product is depicted, compared to depictions with no direct connection to the character.
When it comes to learning from TV series, the setting of alcohol-related scenes is also important to look at. As expected, alcohol is mainly shown evenings and nights, but still, 7.2% (n = 15) of beverages shown in the morning are alcoholic. Furthermore, the fewer people are present, the stronger the alcohol of choice gets, even though this finding is based on a small number of cases. Further research employing larger samples should look deeper into this.
Regarding motivations, findings show that spirits are more likely to be chosen in cases of problem coping compared to soft alcohol. In cases of social ease, by contrast, beer and wine are the likely drink of choice, less so strong alcohol. Furthermore, results suggest that problem coping is much more likely to be depicted or referenced to in workplace-related dramas than in other genres. For example, Coyne and Ahmed (2009) recorded problem coping in only 7.4% of soap operas, whereas in this study it was the motivation in 26.3% of cases. This is an alarming result, because research suggests that negative motives especially predict alcohol-related problems (Carey & Correia, 1997) and that "increased alcohol consumption is reliably related to the move from social to escape drinking" (Farber, Khavari, & Douglass, 1980, p. 781). In terms of social learning, it seems problematic that escaping work problems is depicted as motive of alcohol consumption because, as mentioned before, viewers learn from benefits and punishments vicariously experienced by television characters (Bandura, 2002).
In line with prior research, outcomes were depicted in only a very few cases drawing a picture of high-success role models consuming alcohol without any substantial negative effects. For example, main characters such as Alicia Florrick (The Good Wife) and Harvey Specter (Suits) are shown explicitly consuming alcohol on several occasions but never display any sign of inebriation. Also, keeping in mind that short-and medium-term outcomes like conflicts, absences, sleeping on the job, or decreased work performance are the main cost drivers related to drinking at work (Ames et al., 1997;Martin et al., 1994), work-related TV dramas give a skewed portrayal. Of interest, when outcomes were shown, they were significantly more often present in a work environment than outside of work. Regarding the portrayal of negative outcomes, long-term outcomes were significantly present more often at work than outside of work. The reason for this finding might be that alcohol addiction or accidents under alcohol influence are indeed tackled in medical or legal series. Even main characters in Grey's Anatomy, such as Dr. Webber, or in Private Practice, such as Dr. Shepherd, struggle with addiction problems at the workplace.
Last but not least, findings with respect to model characteristics suggest that men are more likely associated with hard liquor compared to women. However, gender, attractiveness, and SES did not explain whether characters were associated with alcohol or not. Put bluntly, everybody was shown drinking alcohol, independent of individual characteristics. Translated to observational learning, this suggests that male or female viewers, those with low or high SES, and those low or high in attractiveness find alcohol-consuming role models in workplace dramas. Yet the majority of depicted models were high in SES. This may imply that alcohol and economic success are associated.
As one limitation of this study, we focused on workplace-related dramas as the most important workplace series. Of course, it would be interesting to look at subgenres, such as medical dramas or comedy shows. Our goal was, however, to gain insights into the portrayal of alcohol at the workplace and especially the context analyzing the most popular series in one genre as a starting point. Of course, one could argue that only specific workplace settings (i.e., medical, law, political, or journalistic) are depicted in those series. Still, the situations depicted in those series, such as coping with work problems, celebrating a work success, or handling interpersonal relationships with colleagues, are common to most jobs. Therefore, if a viewer observes a doctor drinking scotch because he failed a patient, or a lawyer drinking champagne because he won a case, the viewer might perceive it as socially acceptable to drink alcohol in such workplace situations. It would also be interesting to look at changes in alcohol depiction in entertainment series over time and across different cultures. In addition, the combination of drugs and alcohol in entertainment media deserves more attention. In our study, we looked at situations in which beverages were shown. We did not code the characteristics of situations in which no beverages were shown. Thus, we cannot interpret our findings relative to all depicted scenes. This, however, would not make much sense. It makes only sense to interpret the findings relative to scenes in which a beverage can be shown in principle. For instance, a scene showing a heart surgeon in the operating room or a man in a shower is unlikely to depict alcohol. Coding nonalcoholic beverages as benchmark allows us to compare drinking situations that are similar (e.g., Does the character choose a soda or a beer to refresh?). This gives a more accurate picture of the importance of alcohol in a show than looking at scenes without any beverages.
To conclude, our findings show that workplace series highly overrepresent alcohol consumption at work, especially strong alcohol. Apart from that, alcohol is more likely involved in character-beverage interaction and consumption than nonalcohol. The shows convey the questionable image that alcohol, and even strong alcohol, can be consumed at work with very few outcomes. To give the audience the impression that alcohol consumption at work is socially accepted and without outcome may have harmful effects on society. Our findings have thus important practical implications. In short, policymakers, drama, and program directors should aim to reduce the problematic depiction of alcohol in these shows.
Likewise, audience members, educators, schools, and parents need to be made aware of the problematic depiction of alcohol on such popular programs. This may help to raise the awareness and potentially counter this important phenomenon.