The Effects of Positive and Negative Consequences of Drug Misuse on Television on Behavior

The Effects of Positive and Negative Consequences of Drug Misuse on Television on Behavior Alexandra T. Cass

Master’s Thesis

Graduate School of Communication Master’s Program Communication Science

11099615

Supervisor: Bas van den Putte 29 June 2018

Abstract

The objective of this current study was to investigate the effects of the portrayal of positive and negative consequences of drug misuse on a television show on viewer attitudes towards drug use and intention to use drugs. This study also looked at the impact of narrative engagement and connectedness as moderator variables. Participants completed an online-embedded survey that included a five-minute video clip that portrayed either positive or negative consequences of drug misuse. The research findings demonstrated that positive and negative consequences of drug misuse on a television show effect viewer attitudes towards drug use in a story-consistent way. Positive and negative consequences of drug misuse on a television show did not significantly impact viewer intention to use drugs. In addition, narrative engagement and connectedness were not found to be significant moderators of the positive and negative consequences on attitudes and intention. This current research showed that an existing television show and other destructive health behaviors can be used to test and replicate the effects of this study.

Introduction

From adolescence to young adulthood, people are heavily influenced by the behavior of those around them (Knoll, Magis-Weinberg, Speekenbrink, & Blakemore, 2015; Matos, Tome, Simoes, Diniz, & Camacho, 2012). During these times, people feel the need to conform to and imitate the behaviors of their peers to better fit in (Knoll et al., 2011; Harakeh & Vollebergh, 2012). It has also been shown that television has influential power on the behavior of people of all ages (Moyer-Guse & Nabi, 2010). Media is able to be persuasive due to the fact that people construct their reality based on what they see and hear (Singhal, Cody, Rogers, Sabido, & Bandura, 2010). People can identify and connect with characters as if they are real peers, as well as get completely engulfed into the fictional world of the show, making it possible to be

influenced by television as if it was real life (Fiske, 1992).

Existing literature has examined the extent to which people are influenced by the behavior of others on television. Research in the field of entertainment-education has looked at the power of narrative persuasion on attitudes and behavior. A meta-analysis conducted on the effectiveness of entertainment education in communicating health information looked at messages containing alcohol use, safe sex, and many other behaviors often portrayed on

television (Shen & Han, 2014). The results of the studies reviewed in the meta-analysis prove a significant effect on knowledge, attitudes, intention and behaviors (Shen & Han, 2014). Thus, those explicit behaviors on television are predictive of one's own knowledge, attitudes, and intention in regard to those behaviors.

Drug use is a common destructive behavior witnessed and partaken by people of all ages. A recent survey showed that 81% of U.S. teens, aged 13 to 18, witnessed drug use and 43% used drugs (Castillo, 2012). More specifically, the use of prescription drugs used for unintended

purposes is on the rise, due to the increase in the amount of prescriptions (NIDA, 2018). Amphetamine, a drug prescribed to treat ADHD and narcolepsy and commonly referred to as Adderall and/or Speed, is often misused for the purpose of increasing energy and staying awake. Prescription drugs are being misused second most often after marijuana (SAMHSA 2014). It was found that in 2017 6% of all U.S. 18 year olds misused amphetamines (NIDA, 2018). In

addition, the misuse of prescription drugs is also rising in older adults (ages 50+), thus making all age groups relevant to study. The effect of drug misuse on television on behavior has not been extensively studied. However, a recent study found that watching reality television that contained drug misuse and identifying with the people on the reality show increased viewer illegal drug use and prescription drug misuse (Fogel & Shlivko, 2015). In connection, another study looked at how portrayal of positive and negative consequences of alcohol drinking on television affected attitudes towards drinking (Russell, Russell, Grube, & Mcquarrie, 2017). The results showed that portraying positive consequences, i.e., characters were rewarded or not punished for their

behavior, had a positive significant effect on attitudes but negative consequences, i.e., characters were punished for their behavior, did not have an effect (Russell et al., 2017). However, the authors created a television series themselves, thus compromising the authenticity of the episode. They also only tested the effect on drinking, leaving other behaviors a future area of study. My study examines the effect of drug use portrayed on television that either result in positive or negative consequences on attitudes and behavior and poses the research question (RQ1): Does the portrayal of drug misuse displayed with positive consequences on television influence the attitudes and intentions to perform the behavior more favorably than drug misuse that is portrayed with negative consequences?

Given that the Russell et al. study (2017) only found significant results for the positive consequence condition and not the negative, it is relevant to replicate the findings with drug misuse as well as look at how moderator variables may place a part in increasing the effect of negative consequences. Counter-arguing, or arguing against what one is told or shown, is often seen as a problem in entertainment-education as viewers argue against the negative consequences of a behavior and thus talk themselves out of the fact that there could be negative consequences (Slater & Rouner, 2002).It has been shown that narrative engagement, which is how much a viewer is transported into the world of the television show and identifies with the characters, reduces counter-arguing and thus decreases the amount of resistance towards persuasion (Busselle & Bilandzic, 2009). This reduction in counter-arguing may make viewers accept that negative consequences can happen to them instead of arguing that it won’t happen to them. Thus, narrative engagement may offer a solution to this problem. This poses research question two (RQ2): Does the level of narrative engagement affect the strength of positive and negative consequences on attitude and intention?

Because a real television show is being used in this study, it is appropriate to look at the main effect on viewers who have seen the show before. Connectedness is described by Russell, Norman, and Heckler (2004) as “the level of intensity of the relationship(s) that a viewer develops with the characters and contextual settings of a program in the parasocial television environment”. For a viewer to have connectedness, they have to have watched the television show previously. Russell, Norman, and Heckler designed a connectedness scale that describes how viewers identify with, socially compare themselves to, and get inspiration from characters. This poses research question three (RQ3): Does the level of connectedness towards the television

show and its characters affect the strength of the of positive and negative consequences on attitude and intention?

Theoretical Framework Entertainment-education & Narrative Persuasion

Entertainment-education involves the use of prosocial, educational messages embedded in a narrative storyline in entertainment media, either intentionally added to persuade or educate; or added for another purpose that still has a persuasive effect, such as keeping the plot

interesting, which in return influences viewer attitudes towards the behavior and intentions (Moyer-Guse, 2008). Over time, storylines in popular media have included alcohol, drugs, smoking, disease, safe-sex, and many other controversial topics. Entertainment-education has researched the influence of these controversial topics (Shen & Han, 2014). More specifically, in similar studies to mine, it has been shown that those who view television shows with the

portrayal of negative consequences (i.e., unprotected sex leading to an unwanted pregnant) and with the portrayal of positive consequences (i.e., a character partakes in breast cancer detection and screenings and catches breast cancer early on) are influenced and produce story-consistent health attitudes, intentions, and behavior (Collins, Elliott, Berry, Kanouse & Hunter, 2003; Hether, Hang, Beck, Murphy, & Valente, 2008). This demonstrates that both positive

consequences of a portrayed behavior and negative consequences of a portrayed behavior can impact attitudes and intentions. In terms of relevancy to my study, drug misuse has not been as extensively researched in fictional narrative storylines. Drug misuse refers to using prescription and/or legal drugs not for their intended use. In this context, positive and negative consequences

refer to the either positive or negative results of misusing prescription drugs in the narrative television story.

Social Cognitive Theory and Modeling

There are multiple theories that surround entertainment-education and help explain the influence of narrative messages on behavior. Social cognitive theory (SCT) uses the term

modeling to describe how people copy, or model, behaviors of those they observe. The adoption of observed behavior becomes easier through entertainment mediums as they directly inform, enable, motivate and guide viewers through narrative storylines (Bandura, 2001). Media is able to be persuasive due to people constructing their reality based on what they see and hear (Singhal et. al., 2010). Bandura (2001) explains that modeling has the ability to motivate, or persuade, one to perform a modeled behavior. On television, among other mediums, viewers can live

vicariously through characters. If a character is rewarded for a behavior or is punished for a behavior, the viewer would expect the same outcomes, or consequences, if they were to perform that same behavior. A specific type of modeling, called social modeling, explains how viewers model the emotional experiences of a character and/or storyline, such as fear and gratification (Singhal et. al., 2010). Social modeling helps create outcome expectances that match those of the experience of a character(s). Television also persuades viewers to discuss the behaviors in a storyline with friends and family, connecting the social circle of the viewer to the behaviors, which in return creates an even stronger impact of media on behavior.

For example, a television show created in Mexico that followed population-representative illiterate characters become literate (a positive consequence) increased the rate of enrollment in a Mexico literacy program by 800% (Singhal et. al., 2010). Another study showed that a soap opera containing an AIDS prevention storyline (a positive consequence) reduced the amount of

unsafe sexual practices (Rogers et al., 1999). In addition, television shows with unprotected sex resulting in unwanted pregnancies (a negative consequence) also resulted in story-consistent attitudes and behavior (Collins et. al., 2003). In the more recent study by Russell et. al. (2017), it was shown that a television show portraying positive consequences to alcohol consumption results in more favorable attitudes towards alcohol. With this research and findings two main effect hypotheses were formed:

H1: Participants exposed to drug misuse portrayed with positive consequences will be more likely to produce more favorable attitudes towards drug misuse, as compared to those exposed to drug misuse portrayed with negative consequences.

H2: Participants exposed to drug misuse portrayed with positive consequences will be more likely to produce behavioral intentions to partake in drug misuse, as compared to those exposed to drug misuse portrayed with negative consequences.

The Extended Elaboration Likelihood Model

The extended elaboration likelihood model (E-ELM) explains that people often resist persuasive messages in the form of prosocial, educating storylines found in

entertainment-education scenarios (Slater & Rouner, 2002). Viewers often form counter-arguments towards the message that is received. For example, they take the message that was given to them and create an argument to convince themselves that the negative outcome of the portrayed situation would never happen to them because of specific reasons. A viewer may watch a drug overdose on television and convince themselves that would never happen to them because they are educated about what they take and know their limits. The E-ELM explains that through viewer

the story, the more engaged a viewer becomes, and the less likely they will resist the persuasive message, reducing counter-arguing. The E-ELM also explains the role of identification in mediating the amount of narrative engagement which ultimately reduces counter-arguing (Slater & Rouner, 2002). Identification involves the bonding of a viewer with a character over personal similarity or as someone they have a real relationship with (Cohen 2001). The more a viewer feels that they identify with a character and feels as if they are friends in real life, the more likely the viewer is to want to model their behavior, increasing the likelihood that a fictional storyline that begins to seem real, thus allowing a prosocial, educational message to become more persuasive.

Narrative Engagement

In this study, narrative engagement is analyzed for potential moderation effects, as engagement is used to explain the process of the fictional world becoming closer to the viewer than the real world. Prior research has shown that the more viewers are immersed into the story, the more their attitudes and behavior will become story consistent (Green & Brock, 2000). According to Buselle and Bilandzic (2009), engagement with the narrative is combination of the following subcategories: narrative understanding, attentional focus, narrative presence, and emotional engagement. Narrative understanding includes and measures narrative realism, cognitive perspective taking and ease of cognitive access. If understanding of the storyline, characters, and thread of the story is unclear, there is low narrative understanding and vice versa. For example, if a viewer watches a television show with drug misuse in which the storyline switches between multiple places and plot points to the point it becomes confusing, the viewer may not understand the narrative and thus will not become engaged. Attentional focus includes and measures distraction levels and mind wandering. If a viewer is watching a clip of a character

misusing drugs but is thinking about what they will have for dinner, their focus will be not on the narrative and thus they are not engaged in the narrative. Narrative presence measures how

present one is in the story and the created world as opposed to the outside real world. If a viewer is watching a clip of a party in which drugs are used with severe consequences, but they still hear and process his or her friends talking about the difficulty of the last exam, they may still be present in the outside world and thus not fully engaged in the narrative. Emotional engagement focuses on and measures empathy and sympathy with the story and characters. If a viewer watches a character die of a drug overdose and they do not feel sad or sorry for them and their family, the viewer may not be fully emotionally immersed into the story and the character’s life and is thus not fully engaged. Combined high levels of these factors define narrative

engagement.

Along with high narrative engagement comes less resistance to a persuasive message from a reduction in counter-arguing (Busselle & Bilandzic, 2009). This is what separates those with higher and lower levels of narrative engagement. Those with lower levels are less likely to be influenced by a negative consequence storyline, or a storyline in which a character is punished for a behavior, because of preexisting views and experiences that will cause counter-arguments. In this preliminary study of Busselle and Bilandzic (2009), narrative engagement correlated with story related attitudes and did not correlate with story unrelated attitude; such that, if a viewer watched a clip about drug misuse and was asked to rank attitudes towards premarital sex, the attitudes towards premarital sex were not related to the level of narrative engagement. This shows that significant levels of narrative engagement in a story result in story-consistent attitudes. This was further proven in other studies, such as a recent study by van Leeuwen, van den Putte, Jan Renes, and Leeuwis (2018) that found that those with high narrative engagement

shown a shortened episode of a drama series with negative consequences of binge drinking had lower intention to binge drink. With this research and findings, two moderation hypotheses were formed:

H3a: The effect of exposure of the positive versus negative consequence clip on attitude towards drug use is moderated by narrative engagement, that is, participants who have higher levels of narrative engagement will have a more favorable attitude towards drug use if shown the positive clip or have a less favorable attitude towards drug use if shown the negative clip than those who have lower levels of narrative engagement.

H3b: The effect of exposure of the positive versus negative consequence clip on intention to use drugs is moderated by narrative engagement, that is, participants who have higher levels of narrative engagement will have higher intentions to use drugs if shown the positive clip or have lower intentions to use drugs if shown the negative clip than those who have lower levels of narrative engagement.

Connectedness

Because this study is using an existing television show, it is relevant and important to look at how a viewer’s existing relationship with a television show impacts story-consistent attitude and behavior results. Connectedness, a term created by Russell et. al. (2004), explains the extent viewers are connected to the characters and the storyline. In addition to being engaged, connectedness adds elements only a viewer who has seen the show before could have. An intense relationship with characters and the story cannot be formed instantly and must be formed over time, which is possible with an already existing television show (Russell et. al., 2004). A connectedness scale was created to measure how viewers identify with, socially compare

themselves to, and get inspiration from characters in which they are familiar with. In my study, connectedness can be measured because an existing television show was used. The

connectedness scale was initially developed to see how viewer connectedness impacted

consumer behavior and market behavior with regards to product placement in television shows. Other studies have shown that as a viewer’s relationship with a television show grows stronger, the television show becomes more influential to the viewer, thus showing that highly connected viewers may be more influenced by a television show than lowly connected viewers (Russell & Puto, 1999). This makes it important to look at how a connected relationship, or high

connectedness, affects the influence educational, persuasive messages in a preexisting television show. Thus, the following moderation hypotheses were formulated:

H4a: The effect of exposure of the positive versus negative consequence clip on attitude towards drug use is moderated by connectedness, that is, participants who have higher levels of

connectedness will have a more favorable attitude towards drug use if shown the positive clip or have a less favorable attitude towards drug use if shown the negative clip than those who have lower levels of connectedness.

H4b: The effect of exposure of the positive versus negative consequence clip on intention to use drugs is moderated by connectedness, that is, participants who have higher levels of

connectedness will have higher intentions to use drugs if shown the positive clip or have lower intentions to use drugs if shown the negative clip than those who have lower levels of

Method Design

This study was conducted using an online survey-embedded experiment. Subjects were randomly assigned to two conditions (television clip with positive consequences; television clip with negative consequences) within a 2x2 between-subjects, posttest only, factorial design. The independent variable, drug misuse, consists of two levels (positive consequences and negative consequences). Participants shown the clip with negative consequence were presented with the clip containing drug misuse that resulted in negative consequences while those who were shown the clip with positive consequences were presented with a clip containing drug misuse that resulted in positive consequences. In addition to the independent variable, two moderator variables will be studied: narrative engagement (low/high) and connectedness (low/high).

Participants

An online survey was distributed via social media platforms and sharing among acquaintances. A convenient, self-selected sample of participants (N = 222) responded to the survey online regarding drug misuse in television shows. Those who did not consent to the survey, indicated they were under the age of 16, indicated they were uncomfortable watching a video clip in English with no subtitles, or those who answered a content question about the clips incorrectly were excluded from the survey. After these exclusions, the sample yielded 183

participants (N = 183). Of these participants, the reported ages ranged from 16 to 85 years old (M = 32.39, SD = 12.83). 87 participants indicated themselves as female (47.30%) and 97 indicated themselves as male (52.70%). The participants indicated their country of origin in 23 different countries (N = 183), with the majority coming from the United States (41.53%) and the second

most from India (37.16%). A randomization check was conducted to see if the participants (N = 183) were randomly and equally assigned to the two experimental conditions. It can be assumed that participants were randomly and equally assigned, as there were 93 participants in the first condition (N = 93) and 90 participants in the second condition (N = 90). A Levene’s F test for equality of variances was performed to test this. This test was found to be not significant for age of participants (F = 0.63, p = .429) and sex of participants (F = 1.20, p = .275), so equal

variances across conditions can be assumed.

Stimuli

The two conditions of the independent variable (positive consequences and negative consequences) were manipulated by taking an already existing television show, Pretty Little Liars, and clipping scenes from a singular storyline across two episodes. Both clips contained the same characters and the same drug misuse. The drug misuse storyline contained the misuse of amphetamines by a character named Spencer. In the positive consequences clip, Spencer is seen tired and under stress and notices that a boy that she knows, Andrew, somehow does it all (gets good grades, plays sports, is involved in school activities, etc.) without feeling or looking tired. She knows he takes something and convinces him to give her some of his prescription pills. After taking the pills, she is able to stay up all night to complete a task. In the negative

consequences clip, Spencer’s best friends are made aware of Spencer’s amphetamine problem, that she has been falling asleep at school, and learn that she has had the problem in the past. They stage an intervention where they confront Spencer, and Spencer is clearly distraught and high. Her best friends are clearly upset and disappointed in her and Spencer becomes

Procedure

The online survey was distributed through social media platforms and other mediums seem fit to gather the most respondents. Because the survey was distributed online, it was a self-administered survey and responses were voluntary. Upon opening the survey, participants were provided with an informed consent form. Along with a brief overview, participants were asked if they understood that their participation was voluntary and that all responses were anonymous, with which they had to indicate ‘I agree’ or ‘I do not agree’ to participate in the study. Those who indicated ‘I do not agree’ were excluded from the study. After the informed consent form, two filler questions were asked with the purpose of excluding those not able to partake in the survey. Any participants who indicated they were under the age of 16 were excluded from the survey due to their minor status. Any participants who indicated that they were not confident with watching a video clip in English with no subtitles were excluded because the stimulus television clip was in English. Three sociodemographic questions were asked regarding age, sex, and country of origin. Next, the survey randomly presented each participant with one of the two stimulus video clips. The participants could proceed to the next page only after five minutes, as each of the video clips were around five minutes long. Participants were then asked a series of questions regarding their attitudes, intentions, and narrative engagement. Next, participants were asked if they had ever seen the show, Pretty Little Liars, prior to the study. If ‘yes’ was

indicated, participants were asked a series of questions about connectedness. If ‘no’ was indicated, participants skipped forward to the manipulation check question. To make sure participants understood the video clipped, they were asked a content question (“what is an amphetamine?”) with three multiple choice options. Those who did not answer correctly were excluded (N = 1). At the end of the survey, participants were debriefed and informed of the

purpose of the study and were asked to optionally enter their email address to be entered in a raffle to win a free movie ticket.

Measures

Attitude toward drug use. This variable was measured with five items derived from the attitude scale of Russell et. al. (2017) on a five-point Likert scale from 1 to 5. The items included ranking drug use from (1) bad to (5) good, (1) unappealing to (5) appealing, (1) uncool to (5) cool, (1) boring to (5) fun, and (1) unsafe to (5) safe. The scale had a Cronbach’s ⍺ of 0.92 (M = 2.58, SD = 1.18).

Intention to use drugs. This variable was measured with three items derived from the behavioral intention scale of adolescents’ intentions to use marijuana by Skenderian et. al. (2008) on a five-point Likert scale. The items included the questions “How likely is it that you will”… “Use drugs, even once or twice, in the next year?”, “Use drugs every month for the next year?”, and “Do enough drugs to get high in the next year?”. The scale ranged from (1) highly unlikely to (5) highly likely. The scale had a Cronbach’s ⍺ of 0.96 (M = 2.34, SD = 1.29).

Narrative Engagement. This variable was measured with 12 items on a five-point Likert scale derived from the narrative engagement scale of Busselle and Bilandzic (2009). The items included statements such as “My understanding of the characters in the clip is unclear” and “At times during the clip, the story world was closer to me than the real world” in which participants were asked to rank from (1) extremely disagree to (5) extremely agree. The scale had a

Cronbach’s ⍺ of 0.80 (M = 3.02, SD = 0.68). The scale was recoded to represent participants who were highly engaged; or had a mean score of three and over.

Connectedness. This variable was measured with 7 items on a five-point Likert scale derived from the connectedness scale created by Russell, Norman, and Heckler (2004). The items included statements such as “I would love to meet the characters of Pretty Little Liars” and “Watching Pretty Little Liars is an escape for me” in which participants were asked to rank from (1) extremely disagree to (5) extremely agree. Only participants who had seen Pretty Little Liars (N = 82) were presented the statements. The scale had a Cronbach’s ⍺ of 0.95 (M = 3.47, SD = 0.91). The scale was recoded to represent participants who were highly connected; or had a mean score of three and over.

Results

A manipulation check was conducted to ensure the two manipulated conditions were perceived as either positive or negative depending on if they viewed the positive consequence clip or the negative consequence clip. Participants were asked to rank on a five-point Likert scale ranging from (1) extremely negative to (5) extremely positive the extent in which they felt the clip was negative or positive. An independent t-test was used to test if the mean score for the negative consequence condition would be significantly lower than the mean score for the positive consequence condition. It was found that those in the positive consequence condition had a significantly higher mean score (M = 3.49, SD = 0.64) than those in the negative

consequence condition (M = 1.64, SD = 0.64) and thus clips were perceived as expected and future analysis was continued t (181) = 19.59, p < .001, d = 3.27, 95% CI [1.66; 2.04].

Hypothesis one (H1) predicted that participants exposed to the clip with positive consequences would have more positive attitudes towards drugs and drug misuse. With equal variances assumed (F = 1.46, p =.228) a one-way ANOVA was conducted to see if those shown the positive clip had significantly different attitudes than those who were shown the negative clip. Age of participants and sex of participants were included as covariates, as they were highly correlated with attitudes. Through the analysis, the overall model was found to be significant F(3, 182) = 6.93, p < .001, η2 = .10. It was shown that those shown the positive consequence clip (M = 2.74, SD = 1.12) did differ significantly from those shown the negative consequence clip (M = 2.41, SD =1.22), showing that attitudes towards drugs did differ between the two groups F(1,182) = 3.85, p = .051, η2 = .02 (Table 1). With this slight difference between groups, H1 was supported.

Hypothesis two (H2) predicted that participants exposed to the clip with positive consequences would have higher intentions to misuse drugs than those exposed to the clip with negative consequences. With equal variances assumed (F = .04, p =.852), a one-way ANOVA was used to see if those exposed to the positive clip had higher intentions to misuse drugs compared to those exposed to the negative clip. Again, age and sex of the participants were included as covariates, as they were highly correlated with intention. The overall model was found to be significant F(3,182) = 4.96, p = .002, η2 = .08. It was found that participants shown the positive consequence clip (M = 2.38, SD = 1.28) did not differ significantly than those shown the negative consequence clip (M = 2.29, SD = 1.31), thus concluding that intention to misuse drugs did not differ between the two groups F(1,182) = 0.31, p = .579, η2 = .00 (Table 2). H2 was thus not supported and rejected.

Hypothesis 3 (H3a and H3b) predicted that high levels of narrative engagement would moderate the relationship between the positive and negative consequence clips and attitude and intentions, respectively. For both analyses, age and sex were included as covariates A two-way ANOVA was used to see if participants with high levels of narrative engagement exposed to the positive consequence clip would have more favorable attitudes towards drugs than those with

low levels of narrative engagement. Equal variances could not be assumed with a Levene’s F test (F = 5.88, p = .001), thus a Welch test was conducted. The Welch test was found to be

significant and further analysis was conducted (F(2,112.18) = 19.83, p < .001). The overall model for H3a was found to be significant F(5,182) = 13.13, p < .001, η2 = 0.26. There was found to be no significant moderation effect as high levels of narrative engagement did not strengthen the effect of the conditions (positive consequence clip and negative consequence clip) on attitudes towards drugs F(1,182) = 1.30, p = .256, η2 = 0.01. Thus, H3a was rejected.

H3b produced similar results. A two-way ANOVA was used to see if participants with high levels of narrative engagement exposed to the positive consequence clip would have higher intentions to use drugs than those with low levels of narrative engagement. Equal variances were assumed (F = 2.03, p = .112). The overall model for H3b was found to be significant F(5,182) = 8.96, p < .001, η2 = 0.20. There was also found to be no significant moderation effect as high levels of narrative engagement did not strengthen the effect of the conditions (positive

consequence clip and negative consequence clip) on intention to use drugs F(1,182) = 0.89, p = .347, η2 = 0.01. Thus, H3b was rejected.

Hypotheses 4 (4a and 4b) predicted that high levels of connectedness would moderate the relationship between the positive and negative consequence clips and attitude and intention, respectively. For both analyses, age and sex were included as covariates. A two-way ANOVA was used to see if participants who had seen the show Pretty Little Liars before (N = 82) with high levels of connectedness exposed to the positive consequence clip would have more favorable attitudes towards drugs than those with low levels of connectedness. Equal variances could not be assumed with a Levene’s F test (F = 6.46, p = .001), thus a Welch test was

(F(2,50.67) = 18.77, p < .001). The model for H4a was found to be significant F(5,81) = 5.59, p <.001, η2= 0.27. There was found to be no significant moderation effect as high levels of

connectedness did not strengthen the effect of the conditions (positive consequence clips and negative consequence clip) on attitudes towards drugs F(1,81) = 0.00, p = .951, η2 = 0.00. Thus, H4a was rejected.

H4b predicted that participants with high levels of connectedness exposed to the positive consequence clips would have higher intentions to use drugs than those with low levels of connectedness. Equal variances could not be assumed with a Levene’s F test (F = 5.00, p = .003), thus a Welch test was conducted. The Welch test was found to be significant and further analysis was conducted (F(2,47.84) = 18.66, p < .001). The model for H4b was found to be significant F(5, 81) = 4.44, p = .001, η2 = 0.23. There was found to be no significant moderation effect as high levels of connectedness did not strengthen the effect of the conditions (positive consequence clip and negative consequence clip) on intention to use drugs F(1, 81) = 0.53, p = .469, η2 = 0.01. Thus, H4b was rejected.

Discussion

The aim of the current study was to test the effect of positive and negative consequences of drug misuse on televisions shows on the attitudes towards drug misuse and intentions to use drugs of viewers. The effects of narrative engagement and connectedness as possible moderators were also researched. Firstly, it was predicted that viewers exposed to a video clip with positive consequences of drug misuse would have higher positive attitudes towards drugs. Furthermore, it was predicted that viewers exposed to a video clip with positive consequences of drug misuse would have higher intention to use drugs compared to those who viewed the negative

consequence clip. This was supported by previous research, such as the Russell et. al. (2017) study which found that viewers who were shown a clip with high alcohol drinking with positive consequences had more positive attitudes towards drinking. Similarly, Collins et. al. (2009) found that showing unprotected sex with negative consequences resulted in more negative attitudes towards unprotected sex and lower intentions to have unprotected sex.

Results showed that H1 was supported, showing that attitudes of participants were story-consistent and showed a slight significance difference from each other. However, H2 was not supported, contrary to existing literature that found that the SCT predicts behavioral intention in entertainment-education (Smith, Downs, & Witte, 2007). The social cognitive theory explains how and why viewers adopt story-consistent attitudes and behaviors, such as positive or negative attitudes towards misusing drugs or high or low intentions to misuse drugs. Thus, it was expected that those who viewed the negative consequence clip would produce story- consistent negative attitudes and intentions and those who viewed the positive consequence clips would produce story-consistent positive attitudes and intentions.

Several limitations exist in this study, and may explain the resulting unsupported H2 and only slightly significant H1. This was due to the limited time-frame of the research and the need for many participants to partake without much incentive. A longer video clip may have led to less completion of the survey. Similar studies that produced significant results used longer clips and in some cases full length episodes and series (Russell et. al., 2017; Singhal et. al., 2010; van Leeuwen et. al., 2018). Another factor may have been that the positive clip was not viewed as positive as the negative clip was viewed negative. Even though those who viewed the positive consequence clip ranked the clip significantly more positive than those who viewed the negative consequence clip, the mean rank was still only slightly above neutral. However, viewers who

watched the negative consequence clip ranked, on average, between the clip being extremely negative and negative.

This current study also predicted that viewers with high levels of narrative engagement would be more affected by the video clips than those who were not highly engaged. This was supported by the E-ELM and narrative engagement theory that explains that high levels of engagement result in less counter-arguing towards the educational and/or persuasive message and thus less resistance to the message which produces story-consistent attitudes and behavior (Busselle & Bilandzic, 2009). These theories were supported with the findings of other studies that showed that high levels of narrative engagement produced story-consistent attitudes and behavioral intentions (van Leeuwen et. al., 2018). Even with this support, H3a and H3b were found to be not significant. Again, this may be due to the short video length, allowing only a short time for viewers to become engaged and due to the fact that the positive consequence clip was not viewed as being overtly positive. However, narrative engagement was found to be a conditional main effect and should be studied more closely one standard deviation above and below the mean score of narrative engagement.

Lastly, my study predicted that viewers with high levels of connectedness, or viewers who had strong relationships with the television show, would be more affected by the positive or negative consequence clips than those viewers who were not highly connected. Connectedness theory demonstrates that those with strong relationships with a television show can be persuaded to partake in specific market and consumer behavior (Russell et. al., 2004). Previous studies also support this influential power by showing that as relationships with television shows grow stronger they become more influential and thus persuasive, which may produce story-consistent attitudes and behaviors (Russell & Puto, 1999). Even with the support, H4a and H4b were not

found to be significant. Since Pretty Little Liars is a real television show, it is possible that those who were highly connected were familiar with the presented storyline. In the real show, the drug misuse storyline results in almost all negative consequences, which may have made the positive consequence clip not effective to highly connected viewers who knew the real outcome.

However, connectedness was found to be a conditional main effect and should be studied more closely one standard deviation above and below the mean score of connectedness.

Limitations and Future Research

The current study had a wide range of demographics and did not target a specific group, although a number of teen and young adults were included. Due to the limited time and resources available to this current study, the ages of participants ranged from 16 to 85 and only 44

participants (N = 183) were between the ages of 16 and 24. Given that adolescents, teenagers, and young adults are the most vulnerable group when it comes to being influenced by what they hear and see, the lack of this age group in this current study was a substantial limitation (Knoll et. al., 2015; Matos et. al., 2012). In addition, it may have been helpful to run a pre-test that checked the extent to which the clips were found to be either positive or negative. This may have provided insight into the positive clip not being viewed as positive as it could have been, and thus changes could have been made to create a bigger contrast between the negative consequence clip and the positive consequence clip.

Future research should focus more specifically on the adolescent to young adult age group to see if this would significantly affect the results. It is also relevant to look at the effects of sex of participants due to the high correlation with the attitude and intention scales. It may be beneficial to use a longer video clip to ensure enough time for the growth narrative engagement. Given the significant conditional main effects of connectedness and narrative engagement, future

research should use highly engaged and highly connected and lowly engaged and lowly connected viewers as manipulated independent variables to investigate how narrative

engagement and connectedness directly impacts story-consistent attitudes and behaviors. The findings suggest practical implications that researchers can replicate this current study with other destructive health behaviors and existing television shows.

Conclusion

My study demonstrates through significant findings that people who view positive or negative consequences of drug misuse produce story-consistent attitudes. This is an important finding that shows the need for replication with other lesser-studied destructive health behaviors. This current study also showed that using an existing television show can be effective and this should also be replicated with other behaviors. Although there was no significant effect on intention, and although the moderator variables did not strengthen the effects of the conditions on attitude and intention, as long as television remains a popular form of media consumption, it is important to keep researching the effects and keep looking for new causes of behavior change and moderators.

References

Bandura, A. (2010). Social Cognitive Theory for Personal and Social Change by Enabling Media. In A. Singhal, M. J. Cody, E. M. Rogers, & M. Sabido (Authors), Entertainment-education and social change: History, research, and practice (pp. 75-96). New York: Routledge.

Busselle, R., & Bilandzic, H. (2009). Measuring Narrative Engagement. Media Psychology, 12(4), 321-347. doi:10.1080/15213260903287259

Castillo, M. (2012, April 03). Survey reveals shocking levels of teen drinking, drug abuse. Retrieved from https://www.cbsnews.com/news/survey-reveals-shocking-levels-of-teen-drinking-drug-abuse/

Cohen, J. (2001). Defining Identification: A Theoretical Look at the Identification of Audiences With Media Characters. Mass Communication and Society, 4(3), 245-264.

doi:10.1207/s15327825mcs0403_01

Collins, R. L., Elliott, M. N., Berry, S. H., Kanouse, D. E., & Hunter, S. B. (2003). Entertainment Television as a Healthy Sex Educator: The Impact of Condom-Efficacy Information in an Episode of Friends. Pediatrics, 112(5), 1115-1121. doi:10.1542/peds.112.5.1115

Fogel, J., & Shlivko, A. (2015). Reality Television Programs Are Associated With Illegal Drug Use and Prescription Drug Misuse Among College Students. Substance Use & Misuse, 51(1), 62-72. doi:10.3109/10826084.2015.1082593

Green, M. C., & Brock, T. C. (2000). The role of transportation in the persuasiveness of public narratives. Journal of Personality and Social Psychology, 79(5), 701-721. doi:10.1037//0022-3514.79.5.701

Harakeh, Z., & Vollebergh, W. A. (2012). The impact of active and passive peer influence on young adult smoking: An experimental study. Drug and Alcohol Dependence, 121(3), 220-223. doi:10.1016/j.drugalcdep.2011.08.029

Hether, H. J., Huang, G. C., Beck, V., Murphy, S. T., & Valente, T. W. (2008). Entertainment-Education in a Media-Saturated Environment: Examining the Impact of Single and Multiple Exposures to Breast Cancer Storylines on Two Popular Medical Dramas. Journal of Health Communication, 13(8), 808-823. doi:10.1080/10810730802487471

Knoll, L. J., Magis-Weinberg, L., Speekenbrink, M., & Blakemore, S. (2015). Social Influence on Risk Perception During Adolescence. Psychological Science, 26(5), 583-592.

doi:10.1177/0956797615569578

Leeuwen, L. V., Putte, B. V., Renes, R. J., & Leeuwis, C. (2017). Do Narrative Engagement and Recipients’ Thoughts Explain the Impact of an Entertainment-Education Narrative on Discouraging Binge Drinking? Media Psychology, 20(2), 194-220.

doi:10.1080/15213269.2016.1142379

Matos, M., Tome, G., Simoes, C., Diniz, J. A., & Camacho, I. (2012). How Can Peer Group Influence the Behavior of Adolescents: Explanatory Model. Global Journal of Health Science, 4(2). doi:10.5539/gjhs.v4n2p26

Moyer-Guse, E., & Nabi, R. L. (2010). Explaining the Effects of Narrative in an Entertainment Television Program: Overcoming Resistance to Persuasion. Human Communication Research, 36(1), 26-52. doi:10.1111/j.1468-2958.2009.01367.x

Moyer-Guse, E. (2008). Toward a Theory of Entertainment Persuasion: Explaining the Persuasive Effects of Entertainment-Education Messages. Communication Theory, 18(3), 407-425. doi:10.1111/j.1468-2885.2008.00328.x

National Institute on Drug Abuse. (2017, December). Monitoring the Future Survey: High School and Youth Trends. Retrieved from https://www.drugabuse.gov/publications/drugfacts/monitoring-future-survey-high-school-youth-trends

Rogers, E. M., Vaughan, P. W., Swalehe, R. M., Rao, N., Svenkerud, P., & Sood, S. (1999). Effects of an Entertainment-education Radio Soap Opera on Family Planning Behavior in

Russell, C. A., Norman, A. T., & Heckler, S. E. (2004). The Consumption of Television Programming: Development and Validation of the Connectedness Scale. Journal of Consumer Research, 31(1), 150-161. doi:10.1086/383431

Russell, C.A. & Puto, C.P. Marketing Letters (1999) 10: 393. https://doi.org/10.1023/A:1008170406363

Russell, C. A., Russell, D. W., Grube, J. W., & Mcquarrie, E. (2017). Alcohol Storylines in Television Episodes: The Preventive Effect of Countering Epilogues. Journal of Health

Communication, 22(8), 657-665. doi:10.1080/10810730.2017.1341564

SAMHSA. (2014, August 08). Retrieved June 28, 2018, from https://www.samhsa.gov/prescription-drug-misuse-abuse/specific-populations

Shen, F., & Han, J. (. (2014). Effectiveness of entertainment education in communicating health information: A systematic review. Asian Journal of Communication, 24(6), 605-616. doi:10.1080/01292986.2014.927895

Shulman, L. (2009). Does Watching Sex on Television Predict Teen Pregnancy? Findings From a National Longitudinal Survey of Youth. Yearbook of Obstetrics, Gynecology and Women's Health, 2009, 154-155. doi:10.1016/s1090-798x(09)79044-7

Skenderian, J. J., Siegel, J. T., Crano, W. D., Alvaro, E. E., & Lac, A. (2008). Expectancy change and adolescents' intentions to use marijuana. Psychology of Addictive Behaviors, 22(4), 563-569. doi:10.1037/a0013020

Slater, M. D., & Rouner, D. (2002). Entertainment?Education and Elaboration Likelihood:

Understanding the Processing of Narrative Persuasion. Communication Theory, 12(2), 173-191. doi:10.1111/j.1468-2885.2002.tb00265.x

Smith, R. A., Downs, E., & Witte, K. (2007). Drama Theory and Entertainment Education: Exploring the Effects of a Radio Drama on Behavioral Intentions to Limit HIV Transmission in

Ethiopia. Communication Monographs, 74(2), 133-153. doi:10.1080/03637750701393048 What is the scope of prescription drug misuse? (2018). Retrieved from

https://www.drugabuse.gov/publications/research-reports/misuse-prescription-drugs/what-scope-prescription-drug-misuse