Message Production

WhiteandMalkowski.pdf

Home >Social Science homework help >Message Production

Communicative Challenges of Bystander Intervention: Impact of Goals and Message Design Logic on Strategies College Students Use

to Intervene in Drinking Situations

Cindy H. White and Jennifer Malkowski Department of Communication

University of Colorado

A key concern on college campuses is how to help students avoid negative outcomes related to alcohol use. One way to address this is to encourage students to intervene when they see others engaged in high-risk behavior. This approach, referred to as bystander intervention, requires that individuals seek to influence others; however, research on bystander intervention has not explored how students communicate when they intervene. Drawing from a multiple goals perspective, we examined individuals’ goals during intervention, the types of messages individuals use to intervene, and the relationship of design logic to sense of efficacy and mes- sages used. Participants (n = 212) identified their goals and indicated what they would say to intervene in two college-drinking scenarios. We found that differences in goals were related to the types of messages used, with more specific goals leading to messages better aligned with achieving those goals. Additionally, more sophisticated design logic was related to a higher sense of efficacy and was reflected in the types of messages individuals used to intervene. The study views bystander intervention as an influence process and highlights the communicative challenges that can impact how people intervene.

When college students take responsibility for keeping each other safe, there can be positive effects on campus climate, including a reduction in problem behaviors such as exces- sive drinking and sexual assault (Banyard, Moynihan, & Plante, 2007; Oster-Aaland, Lewis, Neighbors, Vangsness, & Larimer, 2009). This approach, often referred to as bystander intervention, emphasizes the importance of getting other students (bystanders) to recognize high-risk behavior and to offer help when it is needed. Researchers who have studied problems like sexual assault and excessive drink- ing in college communities have noted the importance of bystander behavior in managing high-risk situations (Banyard et al., 2007; Banyard, Plante, & Moynihan, 2004) and have demonstrated that intervention can reduce risk. Although prior research has considered when students will intervene (Thomas & Seibold, 1995), it has not considered how they do so. Individuals seeking to intervene typically

Correspondence should be addressed to Cindy H. White, Department of Communication, University of Colorado, Boulder, 270 UCB, Boulder, CO 80309. E-mail: [email protected]

want to influence another, but they also want to avoid things like embarrassing themselves or offending the target of their influence attempt. This project expands our understanding of bystander intervention by taking a communicative approach to bystander intervention processes. We draw from commu- nication research on goals (Dillard, 1990), influence pro- cesses (Wilson, 2002), and message design logic (Caughlin et al., 2008; O’Keefe, 1988) to examine individuals’ goals during intervention, the types of messages individuals use to intervene, and the way that differences in individuals’ views of communication are related to self-reported intervention efficacy and strategies.

Bystander intervention approaches seek to establish a sense of responsibility among students for others’ safety. Banyard et al. (2004) have demonstrated that training about the risks of sexual assault can increase students’ willingness to intervene in situations that might lead to sexual assault. Additionally, research on the protective behaviors enacted by students to avoid problems with alcohol reveals that friends and other students are often called upon to help keep one another safe (Howard, Griffin, Boekeloo, Lake, & Bellows,

94 WHITE AND MALKOWSKI

2007). Howard et al. (2007) identified a number of protective strategies that students use to reduce risks around alcohol use, including behaviors such as giving others non-alcoholic drinks early in the night or ensuring that friends get home safely. Similarly, Menegatos, Savage, and Roberto (2011) highlighted what they call interpersonal protective behav- iors: that is, behaviors individuals enact to protect other students, especially friends, from negative consequences of alcohol consumption. They found that individuals were likely to engage in interpersonal protective behaviors when a drunken friend was likely to engage in a “hookup” (sex- ual encounter)—particularly if the friend was female. Thus, bystander intervention can help to reduce risky behavior on college campuses.

However, to fully understand bystander intervention pro- cesses, we need to consider the communication aspects of intervention. Individuals who decide to intervene may want to help, but they likely weigh their goals for helping against identity concerns such as being embarrassed if their help is not needed. Additionally, individuals who decide to help must figure out what to say to effectively achieve their goals. To explore communication in intervention, we used two common college drinking situations where another student was clearly intoxicated and could be at risk. In one case, the student could be at risk for an unwanted sexual encounter; in the other, the student was being pressured by others to con- tinue drinking despite obvious intoxication. We used these scenarios to elicit information about the goals students have for intervening and what they would say to intervene.

A COMMUNICATIVE VIEW OF BYSTANDER INTERVENTION

In order to understand what individuals are doing when they seek to intervene, we adopt a multiple-goals perspective on intervention interaction. The notion that communicators pur- sue multiple goals within any interaction is central to this approach (Berger, 2005). Another fundamental assumption of this approach is that the messages individuals produce are tied to the goals they possess in the interaction (Caughlin, 2010). Research on multiple goals has made a distinc- tion between primary and secondary goals (Dillard, 1990). Primary goals frame interaction by defining the focus of what is happening. Secondary goals, which often include identity or relational concerns, exert important influence on message production by shaping what individuals say or how they understand their role in the interaction. In much of the research, “the desire to modify the target’s behavior” has been understood to be the primary goal across influ- ence situations (Wilson, 2002, p. 137). However, individuals who appear to be engaged in similar interactions may be pursuing different primary goals. For example, Sabee and Wilson (2005) found that differences in self-identified pri- mary goals impacted politeness as well as face-threatening

actions students used when discussing disappointing grades with instructors. Additionally, individuals may vary in the importance they assign to different secondary goals. Trost and Yoshimura (2006) considered how secondary goals influence adolescents’ anticipated ability to resist an offer of alcohol from a friend or romantic partner. They found that strong secondary goals related to self-control (not harm- ing self; acting on own beliefs) made adolescents better able to resist offers of alcohol, but secondary goals related to interaction management (maintaining smooth interaction with others) made it harder for them to resist. In short, com- munication in influence situations is shaped by individual understandings of both primary and secondary goals.

Since individuals can vary in the primary goal they see as relevant to a situation (Sabee & Wilson, 2005) and in the importance they place on different secondary goals (Dillard, 1990; Trost & Yoshimura, 2006), one way to better under- stand bystander intervention is to consider the goals individ- uals are pursuing when they seek to intervene. However, in bystander situations the target of influence and the benefits of intervening are not always clear. For instance, imagine a situation where a bystander notices another student who has clearly had too much to drink but is being pressured by a group of peers to drink more. A bystander who inter- venes could be seen as trying to influence the situation. But in this situation the bystander might do so because he is concerned the peer will suffer alcohol poisoning or he might do so because he is concerned that the peer will get sick and create a big, unsightly mess. As this example demonstrates, bystander intervention represents an influence situation where different approach motivations, and there- fore different primary goals, may be operating. Sabee and Wilson (2005) noted that by examining self-reported primary goals, we can learn about how individuals frame interaction using different “approach motivations” (p. 187). Therefore, it seems useful to identify the primary goals students articu- late for intervening. Thus, we asked: (RQ 1) What primary goals do students report in intervention situations?

In addition to knowing the goals students possess when they intervene, we need to know how students go about accomplishing those goals. Prior research on bystander inter- vention has typically focused on whether individuals are willing to intervene rather than how they do so. Although pri- mary goals frame the interaction, the way individuals pursue a goal can vary, and differences in the strategies individu- als use can significantly impact whether a message is seen as appropriate or effective (Caughlin et al., 2008; Wilson & Kunkel, 2000). In the bystander context, we use the term strategy to refer to any verbal and/or nonverbal attempt(s) to influence a situation. Using the example already given, a student could intervene with the drunk peer using a number of different influence strategies: He could tell the individual not to have another drink; he could strike up a conversa- tion and then offer a nonalcoholic beverage instead; or he could complain of hunger and ask the intoxicated peer to

BYSTANDERS AND COMMUNICATION 95

join him in a search for food. Each of these strategies seeks to influence another but does so in a different way and may be guided by different goals. To further explore this issue, we asked students to indicate what they would say to intervene in two college drinking scenarios and then examined those responses to identify different communicative intervention strategies. Thus, our second research question asked: (RQ 2) What types of interaction strategies are evident in the mes- sages students use when they intervene? This provides the opportunity to explicate different ways students seek to influ- ence situations. We anticipated that the strategies individuals use to intervene might vary based on the primary goal they report pursuing, and so we also asked: (RQ 3) Are the types of strategies used to intervene related to primary goals?

Although a number of factors can influence willingness to intervene (Levine & Crowther, 2008), one individual difference—gender—has been shown to play an impor- tant role in how individuals understand their obligation to intervene. Men are more likely to intervene in emergen- cies when doing so highlights their physical strength or competence (Becker & Eagly, 2004); women, on the other hand, are more likely to help friends and family, or some- one who has a personal or emotional problem (Cross & Madson, 1997). Menegatos et al. (2011) found that in a sit- uation where a friend was headed toward a drunken sexual hookup, individuals were much more likely to engage in protective behavior if the friend was female. Additionally, research on compliance-gaining has shown small but signifi- cant differences in the types of strategies men and women use in influence situations (Wilson, 2002). We therefore deemed it useful to explore gender differences in interven- tion. Thus, we asked: (RQ 4) Are there differences between men and women in the goals reported during intervention? We also asked: (RQ 5) Are there differences between men and women in the types of strategies used to intervene?

MESSAGE DESIGN LOGICS AND INTERVENTION

The notion that individuals have multiple goals during inter- action raises questions about how individuals decide which goals to pursue and how (or whether) individuals seek to reconcile competing goals. Message design logic provides one framework for considering situations where multiple, competing goals may shape interaction (O’Keefe, 1988). According to O’Keefe, message design logic (MDL) reflects a set of beliefs about communication—an implicit theory of communication that guides how individuals approach interaction (Edwards & Shepard, 2007). O’Keefe (1988) maintains that these differences in beliefs about communi- cation are evident in messages individuals construct. Her work identifies three different design logics that are evident in messages individuals produce in multiple goal situations: expressive, conventional, and rhetorical.

The first and least sophisticated logic articulated by O’Keefe (1988) is termed expressive and considers com- munication an opportunity to express thoughts and feel- ings. O’Keefe’s next logic, termed conventional, understands “communication [as] a game played cooperatively, accord- ing to socially conventional rules and procedures” (p. 86). For conventional messages, emphasis tends to be on apply- ing rules and explaining what will happen if these rules are violated. Finally, the most sophisticated of O’Keefe’s logic triad, rhetorical, reflects an understanding of com- munication as creating an opportunity to negotiate social selves. A rhetorical communicator tends to see possibilities for managing interaction and invokes communication to cre- ate a suitable context whereby intended outcomes can be achieved.

The growing body of MDL research provides a frame- work for understanding how differences in beliefs about communication impact message production as well as the effectiveness of different messages in multiple goal situ- ations. For example, Lambert and Gillespie (1994) exam- ined pharmacy students’ messages to patients about com- pliance with hypertension medication routines; they found that patients rated rhetorical messages as most likely to be effective in achieving compliance. Caughlin and col- leagues (2008) utilized the MDL framework to examine design logic of comforting responses in a hypothetical sce- nario situation that involved HIV-positive status disclosure between siblings. They found that message design could be assessed in comforting messages and that HIV aver- sion was related to the sophistication of messages pro- duced. They also found that expressive messages were perceived as least effective while rhetorical messages were perceived as most effective. Edwards and Shepard (2007) demonstrated that social experiences like civic participa- tion, altruism, and trust shape the beliefs individuals develop about communication, and Edwards and Graham (2009) investigated family definitions (identity-based, role-based, or shared reality-based) and MDL to reveal that beliefs about communication relate to how individuals see the relational world. Taken together, past research on MDL indicates that differences in beliefs about communication are likely to influence how individuals view complex influ- ence situations and the strategies they use to influence others.

Early work on MDL found differences between men and women with regard to MDL sophistication. Specifically, O’Keefe (1988) found that women produced more rhetorical messages than men, and Caughlin et al. (2008) found that in a comforting situation men were more likely to produce expressive messages, women were more likely to produce conventional messages, but no gender difference emerged for messages displaying rhetorical logic. Because gender differ- ences are relevant to many bystander intervention situations (Menegatos et al., 2011) and because it seemed possible that design logic might be related to gender, we asked:

96 WHITE AND MALKOWSKI

(RQ6) Are there differences between men and women in message design logic displayed?

In addition to gender differences, beliefs about com- munication could influence how people view their ability to influence others or change a situation. These types of differences could be related to perceptions of bystander efficacy. Bystander efficacy measures how confident indi- viduals are that they could do specific things to intervene (Banyard, Plante, & Moynihan, 2005). Banyard and col- leagues have demonstrated that bystanders’ decisions to help are influenced by their sense of efficacy. Reasoning from the MDL framework, it seems possible that individ- uals with more sophisticated design logic would be more confident in the belief that communication can change the situation. However, it also seems possible that individuals with more sophisticated understandings of communication might hold less confidence in their ability because they more fully consider the complexities of the situation. Because it was unclear how MDL would influence bystander efficacy, we asked: (RQ 7) Do individuals with different MDL report differences in bystander efficacy?

Intervention is frequently regulative in that an individual may want to change the behavior of another. For instance, if a student thinks it is not safe for an intoxicated peer to leave a party and walk home alone, the student could try to con- vince the peer to stay at the party until others are ready to leave. The assumption that beliefs about communication (as indicated by MDL) could influence how individuals assess bystander situations where intervention is needed makes sense. However, whether differences in MDL are related to the strategies individuals use to intervene is an issue that has not been investigated. Given that virtually no research on intervention messages exists, we asked: (RQ 8) Are differ- ences in MDL reflected in the types of intervention strategies individuals use?

METHOD

Participants

Participants in this study were first-year college students at a large university in the southwestern United States. First- year students were recruited for three reasons. First, the university where this study was conducted has been working to improve students’ willingness to intervene when they see others at risk. Training related to intervention occurs at this university during new-student summer orientation, so it seemed useful to examine the goals and anticipated responses of first-year students. Second, first-year college students may be confronted with new forms of risk-taking behavior and understanding their anticipated responses to such situations seemed useful. Third, Banyard et al. (2007) have argued that a community focus on high-risk behavior

can lead to changes in the types of behavior that are deemed acceptable by students and can make students more willing to intervene. Thus, it made sense to focus on incoming students to examine how early views of intervention could be addressed to increase a sense of student responsibility. A random sample of 600 entering first-year students was generated, and students were solicited to participate by e-mail. In total, 212 participants completed the survey for a response rate of 35%.

Procedure and Measures

Students received an e-mail inviting them to complete a survey about their experiences with communication and problem solving in social settings. Students were offered a $10.00 gift card to the university bookstore for completing the survey. Students accessed the survey via a link in the e-mail. After reviewing informed consent information, par- ticipants agreed to complete the survey and were directed to survey instructions. The survey took approximately 30 min- utes to complete.

Intervention scenarios. Two intervention scenarios (presented in the appendix) were presented to students to generate data about student goals and intervention messages. The first situation, which we refer to as the guy–girl hookup scenario, involved two students who appeared to be headed toward a sexual encounter. The female student was described as “pretty out of it (drunk)” and “reluctant” to leave with the male. The second scenario described a situation where stu- dents were playing a drinking game. A male student, who appeared to be very drunk, tells others he is “done” and opts out of the game. Others in the game pressure him to keep playing. We refer to this situation as the drinking game sce- nario. We selected these two scenarios for several reasons. First, the campus community health group, which works with students on issues related to alcohol risks, identified these two scenarios as relevant and realistic for students. Second, prior research on alcohol risk indicates that group attitudes toward drinking impact student behavior, and it is important to help students recognize the influence of others on their drinking choices (Johnston & White, 2003). Finally, research by Menegatos, Lederman, and Hess (2010) indi- cates students recognize the hookup scenario as an appro- priate context for interpersonal protective behaviors—which are similar to intervention behaviors.

Since our purpose in this study was to identify what students would say when they intervene, the instructions after each scenario stated: “You decide to say something.” Students provided their response in a text box. (For the guy–girl hookup scenario, students were then asked to indicate whether they would speak to the “guy” or the “girl.” This was done because we were not sure it would be obvious, in this scenario, who the comments were directed

BYSTANDERS AND COMMUNICATION 97

to.) After indicating what they would say, participants were asked to describe all the goals they were trying to accomplish by saying this. Participant goals were gathered via a listing box that provided lines on which to record each goal. Finally, participants were asked, “Of the goals you listed, which do you see as the most important to accomplish in this situation?” There are instances when primary goals can define an interaction without necessarily being labeled the most important goal (see, e.g., Wilson, 2002, p. 176). However, primary goals are often regarded to be the most important goals; therefore, for coding purposes, we elected to treat them as such.

Goals coding. The focus of this analysis was the pri- mary goal identified by participants. Since many of these responses included single words (like “safety”) or short phrases, an undergraduate coder generated a comprehen- sive list of the wording used by participants in responses. Working together, the first author and the coder then con- densed this list and created categories and definitions of goals; this was done separately for each scenario. The condensed lists of goal categories—including “other” and responses that were deemed “uncodable”—produced 11 cat- egories for scenario one and 10 categories for scenario two. However, this more exhaustive scheme of goal codes pro- duced categories with a small number of responses. In order to facilitate analyses, it was determined that the goal cate- gories needed to be further condensed to create overarching goals. The first author and one undergraduate coder worked to identify categories by discussing similarities and differ- ences between terms used by participants. This resulted in four goal categories for the guy–girl hookup scenario and three goal categories for the drinking game scenario (see Table 1). The first author and an undergraduate coder then coded 30% of the data, using these goal categories. For the guy–girl hookup scenario, coder agreement was acceptable,

TABLE 1 Primary Goals

Frequency Percent

Guy-Girl Hookup Scenario Goals Safety 64 32.8% Separation 48 24.6% Prevent isolation 23 11.8% Determine girl’s intention 20 10.3% Other 25 12.8% Uncodable 15 7.7%

Drinking Game Scenario Goals Safety 48 24.6% Stop peer from drinking 69 35.4% Reduce peer pressure 41 21% Other 20 10.3% Uncodable 17 8.7%

Note. Due to rounding, percentages may not total 100%.

Cohen’s kappa = .84. For the drinking game scenario, coders agreement was good, Cohen’s kappa = .92.

Intervention strategy coding. This coding was designed to categorize the intervention message strategies participants provided in their open-ended responses to each scenario. The intervention responses participants provided were relatively short, so we did not unitize the data. We uti- lized inductive coding to create the coding scheme. The first author and one undergraduate coder each reviewed the data and created lists of types of responses represented in the data. Our goal was to develop an exhaustive list of the types of responses. Next, these responses were compared to identify overlap between categories. The guy–girl hookup scenario resulted in five categories of responses, as well as an “other” category for responses that were reasonable but did not fit any of the definitions and an “uncodable” category for responses that did not make sense or were too cryptic. The drinking game scenario yielded five categories of responses plus “other” and “uncodable” categories (see Table 2).

Two undergraduate coders were then trained and asked to apply the coding scheme to the responses. Both coders examined 30% of the data. Cohen’s kappa was used to assess reliability. For both scenarios, reliability was adequate. For scenario one, tests of coder reliability yielded a Cohen’s kappa of .81; for scenario two, agreement was acceptable with a Cohen’s kappa of .84. Discrepancies between coders were resolved through discussion. After reliability was established, one of the coders coded the remaining messages.

Message design logic coding. To assess message design logic, we used O’Keefe’s (1988) Ron Task prompt. In the task, students are asked to imagine that they are part of a group working on a class project. One member (Ron) is described as not having completed necessary work for the group. Participants are asked what they would say to Ron when he calls at the last minute to indicate that he has not completed his portion of the report, which is due very soon. This task was used to assess message design logic because it has been verified in prior research. O’Keefe (1988) has noted that differences in design logic are not evident in all influence situations, and it was not clear that intervention scenarios would satisfy the context requirements to generate clear evi- dence of differences in design logic. Thus, MDL is used here as an indicator of beliefs about communication. Coding for the MDL responses was completed by the authors. Using a coding guide developed from prior MDL work, the authors met to discuss the categories and examine a sample of responses from the data set. Once the authors felt comfort- able they could distinguish each category, both coders rated 30% of the responses to assess reliability. Coder reliabil- ity was acceptable, Cohen’s kappa = .76. Disagreements were reviewed together and resolved through discussion. The remaining responses were coded by the second author.

98 WHITE AND MALKOWSKI

TABLE 2 Intervention Strategies

Frequency Percent

Guy-Girl Hookup Scenario Inquire 53 26.6%

Inquiry of girl (37) Examples: “Are you okay?” “Are you sure you want to be doing this?”

Inquiry + alternative (16) Examples: “Are you sure you are okay to go back with that guy? If not, you can come hang out with

me.” “Hey, are you okay? I think you left your drink—let’s go get it over there.” Insert 89 44.7%

Engage girl (66) Examples: “Hey, how’s it going? Where did you get those (mention clothing item)?” “Oh, there you are.

Your boyfriend’s looking for you.” Disengage the guy (23)

Examples: My friends and I would pretend to know the guy, swarming him with high fives and say “Hey man, I haven’t seen you in years. Do you remember me?”; “Hey, my name’s Lea and I think my friend knows you from COMM class.”

Warn guy 42 21.1% Examples: “Leave her alone man. She’s obviously out of it.” “Yo dude, relax—she’s saying no.”

Other 6 3% Uncodable 9 4.5%

Drinking Game Scenario Substitute 52 26.1%

Speaker offer to play (16) Examples: “I’ll drink it instead!” “Come on guys, I’ll take his place. I’ve been wanting to play all night.”

Encourage others to play (36) Examples: “Let him do what he wants. Someone else play for him.” “Why don’t one of you guys take

his drink.” Warning the group 112 56.3%

Statement regarding limits (79) Examples: “Hey guys, knock it off. He’s had enough.” “No, I think he’s had enough tonight. Don’t

pressure him to do something he doesn’t want to do.” Consequences (33)

Examples: “Dude, this guy doesn’t seem okay. Let’s let him sit out for a bit—I don’t like seeing people puke.” “Do you really want to deal with him throwing up?”

Escape 14 7% Examples: “No, he’s done. Here, I’m done too. I’ll get us some food or water or something.” “C’mon

don’t make him drink anymore. (To guy) Do you want me to show you that party trick in the kitchen?” Other 15 7.5% Uncodable 6 3%

Note. Due to rounding, percentages may not total 100%.

Bystander efficacy. Bystander efficacy was assessed with 12 items taken from a scale developed by Banyard et al. (2005); three items from the Banyard et al. scale were not relevant to the scenarios used in this study and thus were not included. Each item asked how confident an individual is that he or she could do or say something to assist in the situation described. For instance, one item asked participants to rate their confidence from not at all confident to completely con- fident on the following: “Check on someone who seems to be passed out near my room or apartment.” In prior work, Banyard et al. have used a scale of 0–100 to solicit ratings of confidence. However, we used a scale of 1 to 10 because we anticipated that such a wide rating range was not neces- sary to assess perceptions. This adapted scale showed good reliability, α = .90.

RESULTS

Primary Goals in Intervention Situations (RQ 1)

Descriptions of each goal and percentage of respondents identifying it as their primary goal are presented in Table 1. For the guy–girl hookup scenario, the most frequent primary goal was “safety” (32.8% of participants), and responses coded in this category included the specific word “safety”; some respondents indicated only the word “safety” while others said things like “keep her safe” or “make sure every- one is safe.” The second most frequently reported goal was “separation” (24.6%); this goal included references to get- ting the guy to leave the girl alone or breaking up their conversation. The third most frequent goal was “prevent iso- lation” (11.8%), which referred to the introduction of a third

BYSTANDERS AND COMMUNICATION 99

party intended to ensure that the girl did not end up alone with the guy in a remote location. Stated participant goals such as “keep them from being alone together” or finding her friends to bring them over were categorized under the “prevent isolation” category. The fourth goal was “determine the girl’s intention” (10.3%). In these situations, participants said they wanted to find out what the girl wanted or to get more information from her.

For the drinking game scenario, the most frequently iden- tified goal was “stop the guy from drinking” (35.4%). The next most commonly identified goal was “safety” (24.6%). The third category of goals was labeled “reduce peer pres- sure” (21%), which included goals such as getting others to leave him alone. Goals coded as “stop the guy from drinking” tended to indicate a more specific, achievable action (“get him away from the table”), while responses that mentioned safety were more abstract (“keep him safe”).

Types of Strategies Students Use When They Intervene (RQ 2)

The intervention strategies found in the data are presented in Table 2, along with example participant messages for each category. To avoid small cells, we created overarching cate- gories that were used for analyses. For the guy–girl hookup scenario, we identified three overarching categories: inquire, insert, and warn. The first strategy, inquire, is reflected in “inquiry of girl” and “inquiry of girl plus offering an alter- native.” The focus in this approach seemed to be gathering information about what the girl wants by either simply seek- ing information or by combining inquiry with an offer to do something to help the girl leave the situation (such as going to the bathroom). The inquire strategy accounted for 26.6% of the responses. Two of the other approaches (engage girl/disengage guy) centered on participants inserting them- selves into the dyad as a way to gather information and stop what was happening and accounted for 44.7% of the responses. Finally, the strategy of warn focused on delivering information about the woman being too impaired to consent or about the situation being risky—in this case, giving infor- mation to the guy, since no “warn” messages were directed to the female. This type of response was used by 21.1% of participants.

We also identified three overarching categories for anal- yses in the drinking game scenario and labeled the three categories substitution, warn the group, and escape. Two approaches, “speaker offers to play/drink” and “speaker encourages others,” involve substitution—someone else (either the bystander or another in the game) plays for the drunken player. Finding a replacement may reflect the impor- tance of keeping the game going, may indicate an attempt to help the guy “save face” among peers, or may repre- sent the fastest way to withdraw someone from the game. These types of responses accounted for 26.1% of responses. Alternately, the categories of “statement regarding limits”

and “consequences” are focused on reminding the group that members need to respect the guy’s wishes; they seek to reduce peer pressure by calling attention to the inappro- priateness or consequences of pushing someone beyond his or her limits. These approaches illustrate the strategy of warn the group and accounted for 56.3% of responses. A third type of response, which we called escape, focused on actively helping the drunken player leave the game. In these cases, the bystander did not address the group but, interestingly, took on a more active role by offering to escort the player elsewhere. This response was used by 7% of participants.

Relationship Between Goal and Intervention Strategies Used (RQ 3)

Analyzing each scenario separately, we used chi-squared analyses to determine whether the strategy used differed across goals. Results for the guy–girl hookup scenario indi- cated a significant difference in strategy by goal, χ2(6) = 21.47, p < .002, φc = .27. Results for this analysis are pre- sented in Table 3. When individuals had a primary goal of separating the man and woman, they were very likely to use messages that inserted themselves into the conversation with the pair. When individuals saw their goal as “deter- mining the girl’s intentions,” they tended to use messages related to inquiry. Both of these findings suggest a logical relationship between the goal and the type of message used. However, for the goals of “prevent isolation” and “safety,” the connection between the goal and message used appeared less evident. Participants used a variety of messages when the goal was safety, and it is interesting to note that safety messages were distributed relatively equally across the three strategies: Some inquired (35.5%), some inserted (37.1%), and others warned (27.4%). When the goal was to prevent isolation, on the other hand, the most frequently used strat- egy was one that inserted the individual into the conversation between the man and woman (47.6%). “Prevent isolation,” unlike the broader notion of “safety,” identifies a desired outcome (e.g., “keep them from being alone together”).

The goals by message chi-squared analysis for the drinking game scenario revealed no significant difference, χ2(4) = .66, ns. Unlike the guy–girl hookup scenario that

TABLE 3 Goals by Messages for Guy–Girl Hookup Scenario

Message

Goals Inquire Insert Warn

Safety (%) 22 (35.5%) 23 (37.1%) 17 (27.4%) Separation (%) 4 ( 8.9%)a 34 (75.6%)a,b 7 (15.6%)b

Prevent isolation (%) 6 (28.6%) 10 (47.6%) 5 (23.8%) Determine girl’s intention (%) 10 (50.0%)a 7 (35.0%) 3 (15.0%)a

Note. Proportions in the same row that share subscripts are significant at p < .05.

100 WHITE AND MALKOWSKI

presented a somewhat contained and private intervention context, the drinking game scenario likely introduced addi- tional factors, like audience and peer pressure. This may result in a situation where secondary goals are highly rel- evant and thus primary goals have a less direct impact on strategies used.

Gender Differences in Primary Goals and Intervention Strategies (RQ 4 and RQ5)

Chi-squared analyses for gender by goals revealed no sig- nificant gender differences in goals for the guy–girl hookup scenario, χ2(3) = 3.60, ns, or the drinking game scenario, χ2(2) = 2.09, ns. These findings suggest men and women have similar primary goals when seeking to intervene in each type of situation.

Recall that for the guy–girl hookup scenario, we asked participants to indicate who they would speak to in this scenario. Participants reported a higher likelihood of say- ing something to the girl (62.8%) than to the guy (37.2%). However, gender of the participant had a significant impact on whom they would speak to, χ2(1) = 43.40, p = .001. Men were more likely than women to speak to the guy (73% of men indicated they would speak to the guy), while women were much more likely to speak to the girl (83% of women indicated they would speak to the girl). This find- ing reveals that men and women clearly see differences in whom they should address when intervening in this type of situation.

We used chi-squared analyses to examine gender by strat- egy effects. We found a significant difference between men and women in strategy used for the guy–girl hookup sce- nario, χ2 (2) = 24.76, p < .001, φc = .37. Percentages for this analysis are presented in Table 4. Men were much more likely to warn than women. Women’s responses tended to involve inserting themselves in the situation; women were also more likely than men to inquire about what the girl wanted.

No significant difference between men and women in strategy use emerged for the drinking game scenario, χ2(2) = 2.77, ns. This suggests that gender roles may not be as clearly relevant for intervention in this situation.

TABLE 4 Frequencies and Percentages of Intervention Strategy by

Participant Sex

Sex

Intervention Strategy Male (%) Female (%)

Inquire 16 (21.6%) 37 (34.6%) Insert 28 (37.8%)a 60 (56.1%)a

Warn 30 (40.5%)a 10 (9.3%)a

Note. Proportions in the same row that share subscripts are significant at p < .05.

TABLE 5 Frequencies and Percentages of Message Design Logic by

Participant Sex

Sex

Message Design Logic Male (%) Female (%)

Expressive 27 (34.2%)a 8 (7.9%)a

Conversational 37 (46.8%) 57 (56.4%) Rhetorical 15 (19%)a 36 (35.6%)a

Note. Proportions in the same row that share subscripts are significant at p < .05.

Gender Differences in MDL (RQ 6)

Men and women did differ in message design logic in this study; chi-squared analyses revealed a significant difference in distribution of MDL by gender, χ2(2) = 20.84, p < .001, φc = .34. The distribution for MDL across men and women is presented in Table 5. More men than women used expres- sive logic, and women produced messages that displayed rhetorical design logic more frequently than men.

MDL and Bystander Efficacy (RQ 7)

Results revealed a difference in bystander efficacy for indi- viduals with different message design logics, F(2, 178) = 3.79, p = .02, partial η2 = .04. Examination of the means showed that individuals with rhetorical logic reported the highest level of efficacy (7.8); those with conventional logic were slightly lower (7.4) and those with expressive logic reported the lowest efficacy (6.97). Post hoc analyses with Bonferroni correction revealed a significant difference between individuals with rhetorical logic and those with expressive logic. No significant difference emerged between expressive logic and conventional logic or between conven- tional logic and rhetorical logic.

MDL Relationship to Intervention Strategies (RQ 8)

To answer this research question, we examined whether indi- viduals with different design logics used different types of intervention strategies. For the guy–girl hookup scenario, results revealed a significant difference in strategy used for individuals with different MDL, χ2(4) = 13.57, p < .009, φc = .20 (see Table 6). Individuals with an expressive logic were much more likely to use the warn strategy (43.4%) than the insert strategy or inquire strategy. Individuals with rhetorical logic tended to use insert as an intervention strat- egy (64.4%), with a much smaller percentage using an inquire strategy (22%). Individuals with conventional logic favored the insert strategy (44.4%), but many also used the inquire strategy (34.4%). Analyses for the drinking game scenario revealed no significant differences in strategy used for individuals with different design logics, χ2(4) = 3.71,

BYSTANDERS AND COMMUNICATION 101

TABLE 6 Frequencies and Percentages of Intervention Responses by

Participant Message Design Logic for Guy–Girl Hookup Scenario

Response

Message Design Logic Inquire Insert Warn

Expressive (%) 6 (20.0%) 11 (36.7%)a 13 (43.3%)a

Conventional (%) 31 (34.4%) 40 (44.4%) 19 (21.1%) Rhetorical (%) 11 (22.0%) 32 (64.0%)a 7 (14.0%)a

Note. Proportions in the same row that share subscripts are significant at p < .05.

ns. This finding mirrors other results in this study indicating differential effects for the two interventions scenarios.

DISCUSSION

Drawing from a multiple goals perspective, we examined the primary goals students identified as relevant in two interven- tion situations. Results revealed that participants did vary in the primary goals they would purse when intervening. For the guy–girl hookup scenario, participants’ primary goals included safety, separating the guy and girl, preventing the guy and girl from isolating themselves, and determining the girl’s intention. For the drinking game scenario, goals included keeping the peer from drinking, safety, and reduc- ing peer pressure. These differences in primary goals are important because primary goals reflect approach motiva- tions, which shape the frames individuals bring to an influ- ence situation (Sabee & Wilson, 2005). “Safety” emerged as a frequently identified goal for both scenarios, but it is inter- esting to note that safety is a rather abstract goal as compared to other goals like determine the girl’s intention or keep the peer from drinking. As a relatively abstract goal, “safety” may not provide much guidance about what an individual should do to intervene. Following Dillard’s notion of goals– plans–actions (1990), it seems possible that goal specificity may influence an individual’s ability to determine what to do communicatively to achieve their goal. Our analysis of the relationship between goals and intervention strategies provides some support for this notion.

For the guy–girl hookup scenario we found that the specificity of a goal (“determine intent” or “separate the man/woman”) is related to what individuals say when they intervene. For example, participants with the goal of sepa- rating the man and woman tended to insert themselves into the conversation; individuals who wanted to gather informa- tion from the woman tended to inquire by asking if she was okay. Thus, in these cases, the intervention strategy seems clearly aligned with the stated primary goal. But, when indi- viduals had the goal of “safety” or of “preventing isolation,” they pursued less clear goals and the strategies they used were more varied. In many ways, it is encouraging that

students are thinking about keeping others safe; however, what one ought to do to ensure safety is not readily evi- dent and an abstract notion like safety may make intervention seem complex. Further research that examines the relation- ship between goal and intervention strategies could help us to better understand what types of goals lead to more effective interventions.

The findings of this study also indicate that gender dif- ferences matter in intervention situations. Men and women varied in the intervention strategy they anticipated using in the guy–girl hookup scenario. Women were more likely to talk to the woman and to insert themselves into the conver- sation, while men were more likely to talk to the man and to warn about consequences. This makes sense, given research on gender differences in perceptions of hookup behavior (Menegatos et al., 2011), and it suggests that it is essen- tial to explore gender norms with students when helping them consider options for intervening. One possible expla- nation for this finding is that men and women have different secondary goals. Recall that secondary goals often include identity or relational concerns that shape what individuals say or how they approach the situation. Women may, for instance, be more concerned about interaction management (smooth interaction), which could lead to the decision to insert themselves into the conversation. Men, on the other hand, might more strongly embrace identity goals related to appearing authoritative or protective. Prior research has indicated that men are more likely to intervene when doing so reveals their strength or competence (Becker & Eagly, 2004), which might explain their tendency to intervene by warning. Because we were not able to assess secondary goals directly, we can only speculate about what might be happening; further research that explores secondary goals in intervention could provide important insight about how individuals approach intervention.

Finally, we found that beliefs about communication, as indicated by message design logic, impact how individuals approach the task of intervention. Message design logic pre- dicted bystander efficacy. Individuals with rhetorical logic felt significantly more bystander efficacy than those with expressive logic. This makes sense because rhetorical design logic is grounded in the belief that communication involves the negotiation of social situations: a belief that likely con- vinces individuals they are able to do something to impact the outcome of social situations. Additionally, we found that MDL was related to the type of strategy individuals would use to intervene in the guy–girl hookup scenario. Rhetorical individuals used an insert message/strategy where they engaged one or both of the individuals in communication. Expressive individuals tended to warn others about risks of the situation. Conventional individuals were more evenly distributed across the different types of messages used to intervene. These findings align with the MDL framework in that individuals with more sophisticated logic (rhetori- cal) used messages that approached the bystander situation

102 WHITE AND MALKOWSKI

as one to be negotiated communicatively, while those with the least sophisticated logic (expressive) used an approach that emphasized giving others information. Because of the nature of our data, we do not know whether these approaches differ in effectiveness, but this would be a beneficial area for further research.

It is important to note that although participant gender, primary goals, and MDL predicted intervention strategies in the guy–girl hookup scenario, they did not do so in the drinking game scenario. The importance of gender to the guy–girl hookup scenario, but not the drinking game sce- nario, likely has to do with social expectations around sexual encounters. Prior research indicates that gender norms play an important role in hookup situations. Hookups are per- ceived as more acceptable for men than women (Garcia, Reiber, Massey, & Merriwether, 2012), and both men and women are more likely to intervene with a female friend who is headed toward a drunken hookup than with a male friend (Menegatos et al., 2011). Thus, gender was more relevant to the hookup scenario than the drinking game scenario. It is less clear why primary goals and MDL predicted intervention strategies for the guy–girl hookup scenario but not for the drinking game scenario. One pos- sible explanation for this is that the two scenarios differed in clarity about what the “at-risk” student wanted. In the drinking game scenario description, it was clear that the guy did not want to drink anymore while in the hookup scenario, it was unclear what the girl wanted. It seems pos- sible that the clarity of the situation influences how goals shape intervention—which is relevant both to primary goals and to MDL. In situations where what the at-risk student wants is clearer, primary goals may shape what individu- als do, but when what is needed is less clear, secondary goals, such as how the intervening student wants to be seen by others, may come to the fore. With regard to MDL, O’Keefe (1990) has noted that in situations where the con- ventional response is evident, differences in design logic may not emerge. Further research is needed to more fully understand how primary and secondary goals, as well as beliefs about communication, shape intervention in different contexts.

Limitations

There are several limitations to our study. First, we examined goals and messages in response to hypothetical scenarios. Although we took care to construct situations that students were likely to recognize, we instructed them to assume they had decided to intervene. Research such as this provides a useful foundation for thinking about the types of messages that students see as reasonable/workable; however, further research that looks at whether students choose to intervene or not and what they do when they actually intervene is clearly needed. Second, we examined only primary goals in this study, and these were operationalized by having participants

indicate which goal they saw as most important. As Wilson (2002) and Dillard (1990) have noted, primary goals frame the interaction and typically reflect what the individual is try- ing to accomplish in the interaction. This may or may not align with what the individual sees as most important to the interaction. Additionally, although we had hoped to examine secondary goals, procedural problems with how participants listed goals made these data unusable. Secondary goals may play an important role in shaping how individuals approach intervention. For instance, it seems possible that a student who is concerned about relational issues would be more conscious of the face concerns of the other per- son and, therefore, would be more likely to inquire than to warn. Future research that examines secondary goals is needed.

Third, we used the Ron Task to assess message design logic, treating it as an indicator of communication beliefs. However, it would be useful to determine if differences in design logic can be coded from intervention messages. Doing so would require development of an intervention scenario that specifies features of the situation known to clearly reveal design logic differences (O’Keefe, 1990); this would be an important goal for future research. Finally, there were limitations given the population of students who responded to the survey. We targeted first-year college stu- dents; however, it seems possible that students’ ideas about how to intervene may change across their college careers. Additionally, students on this campus had received one training session about bystander intervention at summer ori- entation. Because the training had occurred several months before the data collection and because the training empha- sized having students do what seemed reasonable to them, we are comfortable viewing this data as providing founda- tional information about what students expect to say when they intervene. However, further research with more diverse student populations is important.

Conclusion

Despite these limitations, this study provides a foundation for understanding the communicative challenges of interven- tion. From a communication perspective, intervention can be seen as a compliance-gaining situation whereby individ- uals are encouraged to act on the needs of others largely for the sake of communal wellbeing. Wilson (2002) noted that “people do not seek or resist compliance in a vacuum, but rather as participants in specific relationships embedded in larger institutions and cultures” (p. 293). He argued that by situating our study of compliance-gaining in specific con- texts, we can better understand what people are doing as they seek to influence others. Our results suggest that goals for intervention can impact how students intervene. Programs that teach students about risk and intervention may need to help students identify goals that are more specific than “safety of others.” Additionally, our research indicates that

BYSTANDERS AND COMMUNICATION 103

beliefs about communication, as reflected in message design logic, are relevant to bystander intervention situations. Thus, it may be useful to consider whether training could be tai- lored to address differences in individuals’ beliefs about communication.

In sum, bystander intervention is a complex communicative situation. Research that explores how individuals understand the process of intervention—the goals and logics they bring to the situation—as well as actual messages used to intervene can greatly increase our understanding of bystander intervention. It can also guide practitioners as they think about what students need to know when they attempt to intervene. This project provides a foundation for understanding the communicative nature of intervention and the challenges individuals face during intervention.

ACKNOWLEDGMENTS

The authors thank the Department of Community Health at CU-Boulder, especially Julie Volckens and Teresa Wroe, for their insight and support of this project.

REFERENCES

Banyard, V. L., Moynihan, M. M., & Plante, E. G. (2007). Sexual violence prevention through bystander education: An experimental evaluation. Journal of Community Psychology, 35, 463–481.

Banyard, V. L., Plante, E. G., & Moynihan, M. M. (2004). Bystander edu- cation: Bringing a broader community perspective to sexual violence prevention. Journal of Community Psychology, 32, 61–79.

Banyard, V. L., Plante, E. G., & Moynihan, M. M. (2005). Rape prevention through bystander education: Bringing a broader community perspective to sexual violence prevention. Retrieved from National Criminal Justice Reference Service, U.S. Department of Justice website https://www.ncjrs. gov/pdffiles1/nij/grants/208701.pdf

Becker, S. W., & Eagly, A. H. (2004). The heroism of women and men. American Psychologist, 59, 163–178.

Berger, C. R. (2005). Interpersonal communication: Theoretical perspec- tives, future prospects. Journal of Communication, 55, 415–447.

Caughlin, J. P. (2010). A multiple goals theory of personal relationships: Conceptual integration and program overview. Journal of Social and Personal Relationships, 27, 824–848.

Caughlin, J. P., Brashers, D. E., Ramey, M. E., Kosenko, K. A., Donovan- Kicken, E., & Bute, J. J. (2008). The message design logics of responses to HIV disclosures. Human Communication Research, 34, 655–684. doi:10.1111/j.1468-2958.2008.00336.x

Cross, S. E., & Madson, L. (1997). Models of the self: Self-construals and gender. Psychological Bulletin, 122, 5–37.

Dillard, J. P. (1990). A goal-driven model of interpersonal influence. In J. P. Dillard (Ed.), Seeking compliance: The production of interpersonal influence messages (pp. 41–56). Scottsdale, AZ: Gorsuch-Scarisbrick.

Edwards, A. P., & Graham, E. E. (2009). The relationship between individ- uals’ definitions of family and implicit personal theories of communica- tion. Journal of Family Communication, 9, 191–208.

Edwards, A. P., & Shepherd, G. J. (2007). An investigation of the relation- ship between implicit personal theories of communication and commu- nity behavior. Communication Studies, 58, 359–375.

Garcia, J. R., Reiber, C., Massey, S. G., & Merriwether, A. M. (2012).

Sexual hook-up culture: A review. Review of General Psychology, 16, 161–176.

Howard, D. E., Griffin, M., Boekeloo, B., Lake, K., & Bellows, D. (2007). Staying safe while consuming alcohol: A qualitative study of the protec- tive strategies and informational needs of college freshmen. Journal of American College Health, 56, 247–254.

Johnston K. L., & White K. (2003). Binge-drinking: A test of the role of group norms in the theory of planned behaviour. Psychology and Health, 18, 63–77.

Lambert, B. L., & Gillespie, J. L. (1994). Patient perceptions of pharmacy students’ hypertension compliance gaining messages: Effects of message design logic and content themes. Health Communication, 6, 311–325.

Levine, M., & Crowther, S. (2008). The bystander response: How social group membership and group size can encourage as well as inhibit bystander intervention. Journal of Personality and Social Psychology, 95, 1429–1439.

Menegatos, L., Lederman, L. C., & Hess, A. (2010). Friends don’t let Jane hook up drunk: A qualitative analysis of participation in a simu- lation of college drinking-related decisions. Communication Education, 59, 374–388.

Menegatos, L., Savage, M., & Roberto, A. (2011, November). When do friends prevent friends from hooking up drunk?: The effects of gen- der and intoxication on college students’ use of interpersonal protective behaviors. Paper presented at the National Communication Association convention, New Orleans, LA.

O’Keefe, B. (1988). The logic of message design: Individual differences in reasoning about communication. Communication Monographs, 55, 80–103.

O’Keefe, B. (1990). The logic of regulative communication: Understanding the rationality of message designs. In J. P. Dillard (Ed.), Seeking compli- ance: The production of interpersonal influence messages (pp. 87–104). Scottsdale, AZ: Gorsuch-Scarisbrick.

Oster-Aaland, L., Lewis, M. A., Neighbors, C., Vangsness, J., & Larimer, M. E. (2009). Alcohol poisoning among college students turning 21: Do they recognize the symptoms and how do they help? Journal of Studies on Drugs and Alcohol, 16, 122–130.

Sabee, C. M., &Wilson, S. R. (2005). Students’ primary goals, attribu- tions, and facework during conversations about disappointing grades. Communication Education, 54, 185–204.

Thomas, R. W., & Seibold, D. R. (1995). College students’ decisions to intervene in alcohol-related situations. Journal of Studies of Alcohol, 56, 580–588.

Trost, M. R., & Yoshimura, S. (2006). The importance of competing rela- tional goals in resisting offers of alcohol. In R. M. Dailey & B. A. LePoire (Eds.), Applied interpersonal communication matters: Family, health, and community relations (pp. 175–196). New York, NY: Peter Lang.

Wilson, S. R. (2002). Seeking and resisting compliance: Why people say what they do when trying to influence others. Thousand Oaks, CA: Sage.

Wilson, S. R., & Kunkel, A. W. (2000). Identity implications of influ- ence goals: Similarities in perceived face threats and facework across sex and close relationships. Journal of Language and Social Psychology, 19, 195–221.

Wilson, S. R., & Sabee, C. M. (2003). Explicating communicative com- petence as a theoretical term. In J. O. Greene & B. R. Burleson (Eds.), Handbook of communication and social skills (pp. 3–50). Mahwah, NJ: Lawrence Erlbaum Associates.

APPENDIX: SURVEY INTERVENTION SCENARIOS

Situation 1: The Guy–Girl Hookup Scenario

You are at a party with several of your friends. You’ve been there for a while and can tell that there are a lot of people

104 WHITE AND MALKOWSKI

who are really drunk. A guy at the party is hitting on a girl who seems fairly drunk and is sitting on the couch across from you. You don’t know either of them.

You then watch as he gets her another drink and he starts to rub her leg and leans his face in close to hers. She looks pretty out of it and appears to be uncomfortable, but she is also giggling nervously. He takes her by the hand and pulls her up off the couch. You hear him say, “Come on, let’s get out of here.” She seems reluctant. Assume that you decide to say something.

Situation 2: The Drinking Game Scenario

You are at a party playing beer pong with a group of peo- ple. You know several people at the party. One of the guys on the opposing team is getting completely trashed. The last time he took a drink, he spilled beer all over himself and fell down. He keeps getting back up and playing, but seems really drunk. It is his turn to drink again and he says, “No way, I’m out.” People start pressuring him to drink and cheering him on. Assume that you decide to say something.

Copyright of Health Communication is the property of Taylor & Francis Ltd and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.