Thanks for Asking,but Let’s Talk About Something Else

Abstract

When people engage in explicit topic avoidance, they risk offending or hurting others. Using a goals theoretical framework, we examined how goal-relevant message features influenced participants’ reactions to hypothetical topic-avoidance messages. Participants in this experiment (N = 536) responded to messages that friends might use to deflect a question about a stressful situation. Five types of messages reflected conventionally relevant interaction goals. Results indicated that message features influenced reactions (e.g., judgments of competence, and emotional and verbal responses) and that messages were generally perceived as more competent if they featured secondary relational and identity goals while simultaneously expressing the desire to avoid the subject. Certain secondary goals (e.g., expressing appreciation for the friend’s concern) were especially well received compared with others.

Keywords

avoidance goals interpersonal relationships coping

Topic avoidance is a strategic communication behavior that involves an interlocutor either implicitly or explicitly attempting to prevent or terminate the discussion of a topic (Dailey & Palomares, 2004; Guerrero & Afifi, 1995b). Numerous investigations have demonstrated that topic avoidance is important to understand because it is associated with a variety of markers of individual and relational well-being, including anxiety (Afifi, Afifi, Morse, & Hamrick, 2008), uncertainty (Afifi & Schrodt, 2003b), relational turbulence (Knobloch & Carpenter-Theune, 2004), feeling caught between family loyalties (Afifi & Schrodt, 2003a), closeness (Dailey & Palomares, 2004), and relationship satisfaction (Caughlin & Afifi, 2004).

The collection of scholarship on topic avoidance has established an excellent foundation for further research, yet important questions remain unanswered. First, more work is needed to examine potential variation in terms of how topic avoidance behaviors are actually communicated. In other words, not all efforts to say “I don’t want us to talk about that” are necessarily equal. Second, relational partners employ topic avoidance for numerous reasons, even though it risks being hurtful, offensive, or otherwise dissatisfying. We know little, however, about the discursive features of topic-avoidance messages that reflect people’s objectives and could mitigate or exacerbate the risks of the avoidance. Third, through systematically analyzing theoretically derived features of topic-avoidance messages, we can generate knowledge about communicative aspects of topic avoidance that are associated with more or less competent avoidance efforts while simultaneously testing theoretical propositions about communication.

Designed to address these three gaps in the literature, the present study examined topic avoidance in response to unwanted questions about stressful circumstances. Our objective was to observe how theoretically derived discursive features of topic-avoidance messages might influence reactions to the avoidance. Given the range of motivating forces pertinent to strategic topic avoidance, we contend that an important issue to consider is how goal-relevant features of topic-avoidance messages are meaningful to recipients of the messages. We begin with a review of relevant research on topic avoidance and an overview of the study’s theoretical framework, hypotheses, and research questions.

Topic Avoidance in Close Relationships

An extensive body of scholarship in recent years has advanced our understanding of a range of issues regarding topic avoidance, including what topics people tend to avoid (e.g., Baxter & Wilmot, 1985; Golish & Caughlin, 2002), why people avoid (e.g., Afifi, Caughlin, & Afifi, 2007; Afifi & Guerrero, 2000), how avoidance tendencies are perceived by others, and how avoidance is associated with relational factors (e.g., Afifi & Caughlin, 2006; Caughlin & Golish, 2002). Generally speaking, people avoid talking about sensitive topics because they perceive that open communication entails a risk to themselves, to their relational partners, or to their relationships, or because openness would violate social norms or their partners’ wishes (Afifi & Steuber, 2009; Afifi & Guerrero, 2000; Caughlin & Afifi, 2004; Lepore, 1997; Tesser & Rosen, 1975). Furthermore, people sometimes avoid discussing traumatic events or stressful circumstances in an effort to regulate distressing cognitions or difficult emotions (Folkman & Lazarus, 1985) or to harness a sense of control during a crisis (Arrington, 2005; Donovan-Kicken, Tollison, & Goins, 2011).

The Active Performance of Topic Avoidance

Researchers have found that people use a range of diverse tactics to avoid topics, from subtle or even invisible strategies to behaviors that are actively avoidant like changing the subject of a conversation, deflecting questions, challenging the validity of the topic and the need to discuss it, and explicitly agreeing to engage in topic avoidance (Afifi et al., 2008; Caughlin & Malis, 2004b; Roloff & Ifert, 2000; Rubin, 1994). Recent efforts by communication researchers to classify topic-avoidance behaviors have pointed to the variety of strategies that people report using to enact explicit topic avoidance, such as topic shifts, postponement, deception, humor, crying, and hostility (Afifi et al., 2008; Dailey & Palomares, 2004; Mazur & Ebesu Hubbard, 2004). For example, several of Mazur and Ebesu Hubbard’s (2004) categories of how adolescents responded to unwelcome questions from parents included active strategies that were declarations of a desire to change the conversation.

Topic Avoidance as a Response to Questions About Sensitive Subjects

Topic-avoidance messages are sometimes produced in response to questions about a difficult subject (e.g., Mazur & Ebesu Hubbard, 2004; Suter & Ballard, 2009). For instance, people may be confronted with unwanted requests for information when they are coping with fertility problems (Bute, 2009), when they have adopted children (Suter & Ballard, 2009), or when they are voluntarily childless (Durham, 2008). When people are asked specific questions about a stressful topic, they may state clearly and directly that they wish to postpone the conversation or decline to discuss it altogether. It is important to examine this type of communication because knowing how to engage in topic avoidance strategically and effectively could assist people in accomplishing a number of aims. These include maintaining ownership over their personal information in satisfying ways (Petronio, 2002), coping with the stressful situation in the manner that they desire (e.g., by keeping distressing thoughts at bay; Zahlis & Shands, 1993), and preserving favored relational climates by not hurting people’s feelings or causing conflict (Knudson, Sommers, & Golding, 1980; Roloff & Ifert, 2000). Such goals reflect some underlying dialectics that are reflected in much of the research on information management—for instance, that people want to be open in their relationships while wanting to protect their privacy, or that people want to accomplish the goal of avoiding a stressful topic while wanting to cooperate with the relational partner’s goal of inquiry and achieving intimacy (Baxter & Montgomery, 1996; Petronio, 2002).

The active performance of direct topic avoidance in response to an unwelcome question could prove challenging to enact, because topic avoidance risky, and direct, explicit topic avoidance can be especially treacherous (Caughlin & Malis, 2004a; Dailey & Palomares, 2004; Heffner et al., 2006; Roloff & Ifert, 1998). In fact, people may wish to avoid social situations altogether, thus risking social isolation, if they do not feel capable of deflecting unwanted inquiries in effective and appropriate ways (Garel & Blondel, 1992). Knowing how to enact topic avoidance competently may be a valuable tool for people who are coping with stressful circumstances such as a family crisis or a health problem, when avoiding communication about the distressing issue is relatively common (e.g., Afifi & Schrodt, 2003b; Goldsmith, Miller, & Caughlin, 2008). One pertinent reason why people who are dealing with problems could benefit from research like the present study is that these are the very individuals whose social networks may be shrinking in the wake of crisis (Lyons, Sullivan, Ritvo, & Coyne, 1995). Therefore, the social ties that remain are especially valuable and possibly tenuous, making their preservation much more important. To the extent that relational partners can be competent in their delivery of avoidance messages, they may be able to steer clear of the risks of enacting avoidance, such as alienating supportive others. Yet research on topic avoidance provides virtually no information about how it can be accomplished competently.

Interaction Goals Theoretical Framework

Recognizing that topic avoidance is a strategic communication behavior, researchers have suggested a need for more research that conceptualizes topic avoidance as a behavior characterized by various message features. A meaningful way to categorize discursive features of topic-avoidance messages in response to unwanted questions, and to systematically analyze potential variance in message effects, is to use a goals framework (Afifi et al., 2007; Caughlin & Scott, 2010). In interpersonal communication research, there are a number of perspectives that fall under the classification of goals theories; the common assumptions of goals perspectives center around the purposeful nature of interpersonal communication and the relevance of multiple, sometimes conflicting, goals in every social interaction (Clark & Delia, 1979; Dillard, 1990; Dillard, Segrin, & Harden, 1989; Goldsmith, 2004; Wilson, 2002). Scholars have argued that studying how interaction goals are enacted communicatively, including how goals are prioritized and achieved (or not achieved), is essential for understanding interaction (Tracy, 1991) and that goals are managed in important ways that we are only beginning to understand (Wilson, 2002). Typologies of goals distinguish among instrumental goals that focus on accomplishing tasks, relational goals that emphasize maintaining or changing relationships, and identity goals that attend to impression management and the mutual negotiation of roles (Clark & Delia, 1979). Goals frameworks also identify primary goals, which define communicative action, and secondary goals, which constrain the pursuit of primary goals (Dillard et al., 1989).

A goals framework is useful when analyzing reactions to topic-avoidance messages because research and theory have demonstrated that goal-relevant discursive features are associated with message quality as well as emotional and verbal responses to messages (e.g., Caughlin et al., 2008, 2009; O’Keefe & Shepherd, 1987; Wilson, 2002). When identified in the word choices of speakers, these features offer insight into the secondary goals that are relevant to the situation and salient to the communicators—for example, offering advice without seeming judgmental (Goldsmith, 2004) or disclosing the diagnosis of a stigmatized illness while simultaneously seeking support (Caughlin et al., 2009). Relational partners’ perceptions about each others’ interaction goals are implicated in the judgments that they make about communication behaviors and in their assessments of how satisfied they are with the communication in their relationships (Caughlin, 2010). Moreover, the pursuit of goals becomes increasingly significant to interactions the more that a communication situation is challenging (O’Keefe, 1988), a premise that relates directly to the theoretical foundation of the present investigation: Responding to an unwanted question is a challenging conversational task because in order to do it well, a person may have to prioritize certain goals at the expense of others, for example, reassuring a friend who asks a nosy question that the concern is appreciated, even though one does not wish to discuss the matter.

Using the goals framework to guide this study, we conceptualized topic avoidance as a primary goal that was explicitly enacted in response to a question about a sensitive topic, while the relational and identity considerations that supplemented the topic avoidance were secondary goals that were simultaneously present in some topic-avoidance messages. We focused on secondary goals that pursued conventional prosocial aims that would conceivably make the topic avoidance less risky. Dillard and colleagues (1989) described how primary goals provide the push behind communicative action—in this case, prompting a topic-avoidance message—and explained how secondary goals shape and often constrain the behavior—which, in our study, would entail handling the challenge of avoiding someone’s question in a competent manner. This might be accomplished, for example, by attending to the relational goal of reassuring a friend that the relationship is valued even if the question is unwelcome.

Investigating Multifaceted Reactions to Topic-Avoidance Messages

Based on the extant literature, we propose that topic avoidance may be communicatively enacted via messages that are more or less successful and that it is worth learning more about what constitutes higher quality topic-avoidance messages. The main objective of this study was to examine how pursuing various relevant interaction goals during topic avoidance might influence the extent to which the topic avoidance might engender different reactions. We focused on three facets of friends’ reactions to topic-avoidance messages: judgments about the avoiders’ communication competence, emotional responses, and verbal responses.

Communication Competence

A challenge in enacting explicit topic avoidance is that in pursuing the goal of protecting his or her privacy or autonomy (Afifi, 2003; Petronio, 2002), the avoider is taking a risk by engaging in a speech act that could be interpreted as rude, hurtful, uncooperative, or face threatening (Brown & Levinson, 1987; Goffman, 1967; Grice, 1989). Thus, a central empirical question about direct topic avoidance is how it might be enacted in ways that allow the speaker to self-present as a more socially competent communicator. Interpersonal communication competence is traditionally conceptualized as the ability to produce messages that are effective, appropriate, and generally high in quality (Canary & Spitzberg, 1987; Goldsmith & MacGeorge, 2000). In general, competent communicators are able to achieve their instrumental goals in ways that allow them to preserve valued identities and maintain or enhance their relationship with the person to whom they are speaking. According to goals theories, communication competence would involve the ability to pursue a primary goal (e.g., avoiding a sensitive topic) while concurrently accomplishing a relevant secondary goal (e.g., seeming like a private person or attempting to reassure a friend that he is held in high esteem).

Thus, the fundamental logic of the present study is that one way to mitigate the offense that may be incurred through topic avoidance is to construct a message that conveys the avoidance while simultaneously accomplishing a secondary goal that softens its severity. Based on previous research (e.g., Dailey & Palomares, 2004), we anticipated that the interaction goals in the avoidance messages would be associated with message recipients’ reactions. According to the goals framework, messages that attend to more goals are theorized to be more sophisticated and are often rated as more successful (Caughlin et al., 2008; O’Keefe, 1988); thus, we expected that messages would be more competent if they featured goals that reduced the off-putting nature of the avoidance. Whereas some messages would be relatively plain and straightforward, focusing on a singular primary goal (sometimes referred to as “core” or “kernel” messages), other messages could be more intricate and pursue secondary goals on top of primary goals. It is of particular note that literature on message production and interpretation suggests that messages that adhere to normatively relevant aspects of a social situation are viewed as higher in quality (e.g., Caughlin et al., 2008; Goldsmith, 2004). In this study, the normatively relevant secondary goals included conveying that the speaker was a courteous and considerate friend and that there was a reasonable explanation for the avoidance. We hypothesized that messages that attended to the potential risks of avoidance would be more successful than those that did not. Stated more formally, our expectation was as follows:

Hypothesis 1: Messages attending to secondary goals that mitigate the risk of avoidance will be rated as more competent than messages attending only to the primary goal of avoidance.

Beyond this general expectation of a main effect of message feature, there was little evidence from the existing literature to enable specific predictions about which secondary goals would be viewed as more competent ways of pursuing topic avoidance, compared with other goals. For instance, it may be normatively appropriate for people to express how much they value their friend’s concern (e.g., Caughlin et al., 2008) or they may provide a reason for engaging in avoidance, such as that the topic is too personal (Roloff & Ifert, 1998). As a directional relationship about the relative success of such messages is unclear, a follow-up research question was asked as follows:

Research Question 1: How do topic-avoidance messages featuring different secondary goals compare in terms of recipients’ ratings of the speaker’s communication competence?

Goals perspectives assume that the quality of communication depends on the extent to which it accomplishes the primary interaction task (in this case, avoiding the topic) while successfully managing the implications that the message has for identities and relationships (Goldsmith, 2004; Goldsmith & MacGeorge, 2000; O’Keefe & McCornack, 1987). Evaluations of message quality include the perception that a message could be stated more competently (consistent with the saying, “It’s not what you said, it’s the way you said it”; Goldsmith & Donovan-Kicken, 2009). Therefore, one way to triangulate judgments about competence is by asking message recipients to consider whether they can think of a better way to perform the topic avoidance. In line with the logic that led to our first hypothesis, we theorized that participants would interpret a friend’s hypothetical avoidance message as less competent if it did not include secondary goal features that mitigated the potential for offense. This was an empirical question with two dimensions: (a) whether message recipients would word the topic-avoidance message differently to accomplish it better and (b) how they would do so. First, we expected the following:

Hypothesis 2: Compared with messages that attend to secondary goals, messages attending only to the primary goal of avoidance will be more likely to be rephrased.

As a follow-up to this hypothesis, we planned to compare the effects of the messages that attended to relevant secondary goals:

Research Question 2a: How do topic-avoidance messages featuring different secondary goals compare in terms of recipients’ intent to rephrase the topic-avoidance message?

Finally, in addition to asking whether participants would reword the message, we inquired about the strategies that people would use when phrasing the topic avoidance in an effort to perform it more competently:

Research Question 2b: How do topic-avoidance messages featuring different secondary goals compare in terms of how people would rephrase the topic-avoidance message?

Emotional Responses

In addition to judgments of competence, our rationale thus far includes the premise that topic avoidance may evoke negative reactions such as unhappiness or hurt feelings. Explicit topic avoidance is a face-threatening act (Brown & Levinson, 1987; Dailey & Palomares, 2004) that may imply that the speaker does not trust or value his or her conversational partner enough to be open about a topic; this social affront may be quite hurtful to the recipient of the topic-avoidance message (Goldsmith & Donovan-Kicken, 2009). From previous research grounded in multiple-goals theories, there is some evidence that messages that attend to secondary goals and provide context or relational reassurance are less likely to evoke negative emotions such as sadness (Caughlin et al., 2009). If people engaging in topic avoidance attend to secondary goals like reinforcing the strength of the partners’ friendship, then perhaps they can mitigate the face threat of direct avoidance by reinforcing the esteem in which they hold the other. Some topic-avoidance messages may thus engender different levels of hurt, confusion, or even solidarity. We hypothesized the following:

Hypothesis 3: Messages attending to secondary goals that mitigate the risk of avoidance will be less likely to evoke negative emotional responses (such as hurt) than messages attending only to the primary goal of avoidance.

Verbal Responses

Finally, it is certainly of interest to communication scholars to glean an understanding of how the conversation might continue to unfold after one person has enacted explicit topic avoidance. It is useful to rely on the goals framework to think about the individual motivations and the conventionally relevant properties of this situation: The successful performance of topic avoidance should presumably entail effectively putting an end to the discussion of the sensitive subject, given that the primary goal of the avoidance message is to avoid the topic. In addition, if the exchange centers around a stressful situation with which a friend is coping, then it would seem normative for the other person to offer some support to the friend (Caughlin et al., 2008; Goldsmith, 2004). A number of verbal responses are possible in the present context, then, such as allowing the change of subject, offering support, or apologizing for asking in the first place. There is conflicting evidence about whether messages that attend to secondary goals during conversations about stressful situations elicit relevant responses. For example, Caughlin and colleagues (2009) found that hypothetical HIV disclosure messages that attended to the secondary goal of seeking support were met with higher ratings of recipients’ intent to comfort the speaker compared with more basic, primary-goal-only messages. However, recipients’ open-ended response messages were actually less likely to contain emotionally supportive elements in response to disclosure messages featuring this secondary goal. Consequently, the following research question was put forth:

Research Question3: How (if at all) do different goal-relevant message features influence verbal responses to the topic-avoidance messages?

Method

Research Design and Development of Stimulus Materials

The design of this study involved experimentally manipulating topic-avoidance messages to reflect different interaction goals that are conventionally relevant to an explicit topic-avoidance situation. Five levels of the experimental condition factor, corresponding to five interaction goals, were tested. Condition I was the Kernel message (Jackson, 1992) or the basic core statement which expressed the speaker’s wish not to discuss something. This Kernel explicitly pursued the primary goal of avoiding the topic. From the research literature on topic avoidance, privacy management, and coping, we identified four secondary goals that are relevant to a conversation where one person inquires about a difficult situation and the other person explicitly avoids the topic. The remaining four conditions are as follows: Expressing Gratitude for Concern (Condition II), Reinforcing the Relational Bond (Condition III), Seeming Polite (Condition IV), and Invoking Privacy (Condition V). The Kernel condition represented the instrumental goal of enacting topic avoidance, a goal that appeared in every message. Conditions II and III were examples of relevant relational goals because they focused (in different ways) on maintaining relational harmony, and Conditions IV and V were illustrations of relevant identity goals because of their emphasis on self-presentation as either a polite or private person. The complete list of messages is in the appendix.

A multiple-message design was employed to enhance the generalizability of findings because even within the conceptual boundaries of each relevant secondary goal, there is a virtually infinite number of ways to articulate one’s desire to avoid a topic. For each of the five conditions, four message instantiations were developed to represent different ways of accomplishing the same interaction goal or combination of goals. For example, in the Privacy condition, we created a message that said, “I don’t want to talk about that because it’s such a personal matter,” which conveyed the message that the speaker both (a) desired to avoid the topic and (b) felt a need for privacy. Conditions II through V were all designed as replications of the Kernel condition (Jackson & Brashers, 1994), such that the Kernels were core messages repeated within each condition and layered with the relevant secondary goal that corresponded to the condition. Subsequently, in analyzing the data, the interaction goal conditions were treated as fixed factors, whereas the core message instantiations were treated as random factors. This design was similar to that used in previous experimental research on the features of disclosure messages (Caughlin et al., 2008, 2009). The message manipulations were pilot tested in structured, one-on-one interviews with a small sample of undergraduate students for clarity and comprehensiveness.

Participants

Study participants were undergraduate students at a large southwestern university in the United States who received course extra credit for their involvement. Five hundred thirty-six individuals participated; 377 (70.3%) participants were female, 148 (27.6%) were male, and 11 (2.1%) did not provide data on sex. Participants were instructed to choose a particular friend and think about the same person throughout the study; 32% (n = 173) of participants imagined conversations with male friends, whereas 67% (n = 359) imagined talking with female friends, and <1% (n = 4) of the sample did not report the sex of the friend with whom they imagined interacting. The mean age of participants was 20.27 years (SD = 2.61). The sample was composed of respondents who described themselves as African American/Black (5.2%), American Indian (<1%), Asian/Pacific Islander (12.5%), Hispanic/Latino(a) (13.6%), Middle Eastern (2.4%), Multiracial (3.4%), White (60.6%), and Other (1.5%). Two individuals did not provide information on ethnicity.

Procedures

Participants completed the questionnaire via an online survey system. A survey flow randomization command was employed to randomly assign participants to different message instantiations from a single survey link. The survey response rates were monitored to ensure that the number of people in each condition was approximately equal. After the informed consent page, the survey began by describing the following scenario: “Imagine you are alone talking with a good friend. Think of something difficult that your friend has been going through. If you cannot think of something, either focus on a different friend or think of a hypothetical situation that your friend could be dealing with. Imagine that this happens to be a problem, a difficult situation, or some sort of stressful circumstance. You are aware that your friend is having a tough time with it, but you do not know much else about it. After several minutes of conversation, you decide to bring it up. You ask your friend about the difficult situation that he or she has been going through.” We focused on avoidance in one particular type of relationship because of evidence that topic-avoidance patterns are associated with relational types and characteristics (e.g., mother vs. significant other; Caughlin & Afifi, 2004; Dailey & Palomares, 2004). We situated the topic avoidance in the context of friendship because our design asked participants to imagine that the topic was a relational partner’s stressful problem, and we wanted them to think of a conversation that could occur in person. Friends are important confidants and local sources of support, especially as people mature and gain independence from their families (Furman & Buhrmester, 1992).

After the introductory passage, participants were told that in response, their friend said something to avoid the topic; in other words, each participant was randomly shown one of the topic-avoidance messages that had been crafted for the study (e.g., “In return, your friend says, ‘I don’t want to talk about that.’”). To enhance the validity of this procedure, at specified points within the aforementioned script, participants were directed to type their friend’s initials so that they were thinking of a particular person throughout the entire questionnaire, and they were asked to write the specific problem the friend was experiencing. (All participants provided their friend’s initials and 99% of respondents wrote a description of the problem.) Following the presentation of this scenario, participants were asked to write answers to a series of open-ended questions, which asked them (a) exactly what they would say in response, (b) how their friend’s statement would have made them feel, and (c) whether and how they would have worded the message differently, if they were in their friend’s position. In the ensuing section of the questionnaire, respondents completed scaled measures of communication competence and answered a question about the realism of the scenario, then provided demographic information.

Measures

Communication competence

A 7-point Semantic Differential Scale was adapted from conceptualizations and operationalizations of communication competence that have been used in previous research to assess message quality in communication situations such as disclosure (Caughlin et al., 2009) and social support (Goldsmith & MacGeorge, 2000). The six items that were used focused on standard dimensions of communication competence: effectiveness and appropriateness (e.g., Canary & Spitzberg, 1987). Anchors were effective–ineffective, appropriate–inappropriate, competent–incompetent, sensitive–insensitive, sophisticated–unsophisticated, and polite-rude. Interitem reliability was satisfactory, α = .87.

Realism

As a means of checking the plausibility of the hypothetical conversation, a single semantic differential item assessed realism (“How realistic was this scenario?” 1 = unrealistic, 7 = realistic). The mean score for this item was significantly above the midpoint, M = 5.40, SD = 1.81, t(535) = 17.28, p < .001, providing evidence that participants found the scenario to which they responded plausible. The descriptions of the problems that were provided by participants, and their general level of detail, bolstered our impression that the situation was easy to imagine and report on.

Coding of open-ended responses

Five types of information, gathered from participants’ responses to open-ended questions, were coded: (a) the type of problem the friend was experiencing (i.e., the avoided topic), (b) the emotions that participants would feel if the friend avoided the topic in the way described, (c) whether and (d) how the participants would have rephrased the topic-avoidance message, and (e) how the participants would respond verbally to the topic avoidance. The first three judgments involved assigning participants’ answers to discrete categories. The last judgment involved observing the presence of any of the following in the participant’s response: (a) agree to drop the subject (e.g., “Okay, no problem, it’s fine if you don’t want to talk about it”), (b) express available support (e.g., “I’m always here if you need someone to listen”), or (c) apologize for asking (e.g., “I’m sorry, I didn’t mean to intrude”). In keeping with the nature of our data, these three categories were not mutually exclusive.

To develop the codebook for the open-ended data, all four authors independently analyzed participants’ responses. Using inductive thematic analysis (Bulmer, 1979; Creswell & Plano Clark, 2007), we identified categories that captured the nature of the data. We used the entire response as the unit of analysis. We subsequently met to discuss our observations, to settle disagreements (which pertained mainly to nomenclature and degree of differentiation), and to agree upon a comprehensive yet parsimonious set of categories that best encapsulated the data. We were blind to the experimental message and condition to which participants were exposed when creating the coding scheme and when coding the data. Two authors coded the entire set of responses and discrepancies were resolved through discussion. A random subsample (20% of responses) that comprised an approximately equal number of cases per experimental condition was examined for intercoder agreement using Cohen’s kappa (Cohen, 1960), which was high and as follows: problem type, κ = .96; emotional reaction, κ = .93; rephrasing, κ = .94; agree to drop the subject, κ = .81; express available support, κ = .90; and apologize, κ = .90.

Results

We set α at .05 for our statistical tests. In between-subjects designs, relatively large sample sizes are required to attain adequate statistical power, and our particular study involved balancing the power trade-offs between random factors and covariates (Algina & Olejnik, 2003). Given our sample size of N = 536 participants, power to detect a small effect (d = .20; Cohen, 1989) was .97 and power to detect both a medium (d = .50) and large (d = .80) effect exceeded .99.

Coding and Controlling for Problem Type

In light of previous research suggesting that perceptions of information management can be situation- and context-specific (Baxter & Wilmot, 1985; Cloven & Roloff, 1993; Dailey & Palomares, 2004; Guerrero & Afifi, 1995a, 1995b), we considered the possibility that the nature of the avoided topic itself might influence people’s reactions to the topic avoidance. Nine different problem types were identified from the descriptions participants provided of the difficult situations with which their friends were coping. In descending order of frequency, the problem types were as follows: romantic relationship issues (33%), education or career challenges (15%), health problems (14%), family crises (9%), feeling generally overwhelmed or experiencing low quality of life (9%), financial troubles (6%), death or grief (5%), friendship conflict (4%), and identity turmoil, mainly pertaining to sexuality (3%). An additional 4% of responses were coded as “other” because they appeared infrequently (e.g., trouble with the law).

Interaction Goal Conditions and Communication Competence

To test the main hypothesis (Hypothesis 1) of the study, a 5 (Interaction goal) × 4 (Core message) between-subjects mixed model analysis of covariance (ANCOVA) was conducted. Communication competence was the dependent variable, Interaction Goal (i.e., the experimental treatment) was a fixed factor, and Core Message was included as a random factor. As sex was not equally distributed among cells, we included sex of respondent and sex of friend as covariates. Problem type was entered as a covariate to rule out or control for any potential influence of the nature of the friend’s stressful circumstances. The ANCOVA revealed a significant and sizable main effect of condition, F(4, 12) = 12.86, p < .001, partial η² = .80, after adjusting for the three covariates, which accounted for virtually no variance in competence; indeed, neither sex of respondent, F(1, 499) = 0.39, p = .53, sex of friend, F(1, 499) = 0.32, p = .57, nor problem type, F(1, 499) = 0.31, p = .58, were significant predictors of respondents’ communication competence ratings. The Kernel messages were rated as less competent than the messages that attended to secondary goals, with the exception of the Privacy messages. Thus, Hypothesis 1 was partially supported.

Table 1 displays the means for each condition and the results of planned contrast analyses that examined which conditions significantly differed from each other in terms of competence, providing details about Hypothesis 1 and Research Question 1. The Kernel messages were rated as less competent than all others except for messages referring to Privacy. Messages expressing Gratitude, which was the condition with the highest mean communication competence score, were rated significantly higher than messages that invoked a need for Privacy, messages that expressed a desire to seem Polite, and the Kernel messages. Messages that reinforced the Bond of friendship were rated as more competent than messages that invoked Privacy and the Kernel messages. Messages that were intended to seem Polite were rated as more competent than messages that requested Privacy. Stated differently, messages invoking Privacy were generally viewed as less conversationally competent than all other types of messages, except for the Kernel.

Table 1.

Competence Scores by Condition.

Condition	Communication competence
Kernel (avoid topic)	4.38_a,b,c (1.10)
Gratitude	5.25_a,d,e (1.09)
Bond	5.06_b,f (1.18)
Polite	4.83_c,d,g (1.04)
Privacy	4.49_e,f,g (1.17)

Note: Adjusted means (with standard deviations in parentheses) for communication competence by condition. Higher scores indicate greater perceptions of friends’ competence. Cell means in the same column that share a subscript are significantly different at p < .05.

Consistent with other experimental multiple-message designs (e.g., Caughlin et al., 2009; Jensen, 2008), we took steps to examine whether the effect of the treatment was stable across messages. In terms of communication competence, neither was there evidence of a significant main effect for individual messages, F(3, 499) = 0.33, p = .81, nor was there evidence of a significant condition by message interaction, F(12, 499) = 0.84, p = .61. These analyses provided important evidence that it was the experimental conditions (i.e., types of goal-relevant message), rather than particular core message instantiations or interactions between conditions and individual messages, that influenced respondents’ judgments of their friends’ communication competence.

Interaction Goal Conditions and Open-Ended Response Outcomes

Rephrasing of the topic-avoidance message

To assess whether participants believed that there would have been a better way to articulate the topic-avoidance message (Hypothesis 2 and Research Question 2a), we first created a dichotomous variable according to whether participants affirmatively or negatively answered the question about whether they would have rephrased the message. We assigned a code of 1 to participants who indicated that they would have worded the topic avoidance differently and a 0 to participants who said “no” or wrote that they would have said it the same way. Tests of differences between independent proportions (Bruning & Kintz, 1997) were conducted to compare the proportion of participants who would have rephrased the topic-avoidance message across condition. Tests of differences of proportions are distributed as z; in all instances, two-tailed tests were conducted. (To conserve space, the individual test statistics are not included here; however, they are available upon request from the first author.) As shown in Table 2, limited support was found for Hypothesis 2 because the Kernel was only significantly different from the Gratitude condition. People were less likely to say that they would rephrase the topic-avoidance message in the Gratitude condition than in the Kernel, Polite, or Privacy conditions. Participants whose friends’ messages emphasized the relational Bond were less likely to reword the message as compared with participants whose friends’ messages emphasized a need for Privacy.

Table 2.

Proportions of Respondents Who Would Have Stated the Topic-Avoidance Message Differently, by Condition.

	Kernel	Gratitude	Bond	Polite	Privacy
Would rephrase	.72_a	.59_a,b,c	.68_d	.73_b	.80_c,d
n	116	106	103	111	100

Note: Proportions in the same row that share subscripts differ significantly at p < .05.

In addition to asking participants whether they would rephrase the message were they in their friend’s position, we also asked them to write out specifically how they would have said the message themselves (Research Question 2b). To analyze this, we excluded respondents who reported that they would not have rephrased it, n = 157, and the one person who did not respond to this question. This left an n of 378. By far, the most common manner in which participants indicated that they would have rephrased the topic-avoidance message was by postponing the conversation rather than cutting it off outright (54% of participants denoted that this would have been their preferred strategy). This option was followed in frequency by people who said that they would do the following: provide a reason or excuse for wanting to avoid the topic (25%), completely change the subject (8%), be more polite (5%), be more blunt or less polite (3%), or request privacy (1%). Five percent of respondents said that they would have discussed the problem rather than avoid the topic, and 8% of answers to this question were uncodable into any of the above categories. Among those participants who would have said the topic-avoidance message differently (70% of respondents), there was little variation between conditions with respect to how people felt that the topic-avoidance message ought to be rephrased. Only two ways of handling the topic avoidance differently exhibited significantly different proportions: The proportion of participants who indicated that they would actually decide to talk about the stressful problem was significantly higher in the Kernel and Gratitude conditions compared with the Polite condition.

Emotional reaction to the topic-avoidance message

According to participants’ open-ended responses, the most common emotions they would likely experience as a result of this exchange with their friends were feelings of understanding, mixed feelings, and hurt feelings: 23% of participants reported that they would be understanding about this exchange with their friends, whereas 19% said that they would experience mixed emotions (e.g., “At first I would feel upset and shut out, but I would realize that it’s hard to talk about things that hurt and I would understand”) and 18% indicated that they would feel hurt. Other participants reported that they would feel concerned or sad for their friends (8%), annoyed or defensive (7%), guilty or awkward (5%), or confused (4%). One percent of participants indicated that they would feel happy or relieved not to have to delve more deeply into the matter. The remaining 15% of responses were coded as “other” because they were left blank, they were incomprehensible, or, most commonly, they pertained to the participants’ perceptions of how their friends were feeling rather than their own emotions (e.g., “I would feel like she was hurt and hiding it”).

Tests of differences between independent proportions were conducted to investigate variation in the occurrence of emotional responses by condition. (We excluded cases in the “other” category.) According to Hypothesis 3, we expected that negative emotional responses would be more common in response to the Kernel messages. To test this, we first classified negative responses according to literature on affect and hurt feelings (Vangelisti, 2009; Watson, Clark, & Tellegen, 1988). The negative emotional responses were as follows: hurt, concerned, annoyed, and guilty. The proportions of responses containing each emotional reaction by condition are displayed in Table 3. There was little evidence that recipients of topic-avoidance messages were more likely to experience negative emotional responses in the Kernel condition compared with conditions where secondary goals were addressed. One exception was that the proportion of respondents who would be concerned and feel helpless about the friend’s situation was significantly higher in response to the Kernel condition as compared with the Polite condition.

Table 3.

Proportions of Emotional Responses to Topic-Avoidance Message, by Condition.

	Kernel	Gratitude	Bond	Polite	Privacy
Understanding	.29	.22	.31	.33	.20
Mixed emotions	.21	.29	.20	.25	.19
Hurt	.26	.18	.19	.13_a	.28_a
Concerned	.12_a	.10	.10	.04_a,b	.12_b
Annoyed	.05	.09	.13	.07	.05
Guilty	.05	.08	.06	.07	.06
Confused	.01_a	.03	.01_b	.09_a,b	.06
Happy	.01	.01	.00	.02	.02
n	97	90	88	89	93

Note: Proportions in the same row that share subscripts differ significantly at p < .05.

Verbal response to the topic-avoidance message

To investigate Research Question 3, participants’ responses were coded and categorized according to the presence of three particular elements that would be conventionally relevant if one friend expressed to another friend a desire to avoid discussing a stressful problem: (a) agreeing to drop the subject, (b) expressing available support to the friend, and (c) apologizing for intruding. Table 4 displays the proportions of those response elements for each condition. The Kernel condition differed from all four other conditions in terms of the percentage of participants’ responses that indicated their willingness to discontinue discussion of the subject; participants seemed less likely to abandon the topic when the topic-avoidance message was quite basic rather than when it was accompanied by a secondary goal. Expressing support occurred more frequently in response to messages in the Gratitude and Privacy conditions than in response to Kernel, Bond, or Polite messages.

Table 4.

Proportions of Verbal Responses to Topic-Avoidance Message, by Condition.

	Kernel	Gratitude	Bond	Polite	Privacy
Agreed to drop	.52_a,b,c,d	.72_a	.71_b	.72_c	.75_d
Expressed support	.38_a,b	.57_a,c,d	.42_c,e	.40_d,f	.56_b,e,f
Apologized	.09	.10	.06	.10	.07
n	116	106	102	109	100

Note: Proportions in the same row that share subscripts differ significantly at p < .05.

Discussion

It has been nearly two decades since Guerrero and Afifi (1995a) noted that it would be necessary to advance research on topic avoidance by examining how sensitive subjects are successfully avoided. The present investigation is among the few studies that have addressed the relative success of topic-avoidance strategies, and to our knowledge, it is the only effort that has examined the competence of avoidance messages that were systematically altered along theoretical dimensions. The main finding of this experiment was that the presence of relevant secondary goals appear to increase ratings of the competence of topic-avoidance messages above and beyond factors such as the nature of the avoided topic or the sex of interlocutors. Furthermore, some emotional and verbal reactions to topic-avoidance messages were attributable to goal-relevant message features. We begin our discussion by synthesizing the findings to identify key patterns, contextualizing them against other scholarship to describe the contributions of this project. We then discuss the opportunities for future research that have emerged from this study, along with the limitations of its design.

Overview of Main Findings

Grounded in the logic of interaction goals perspectives on communication, the experimental conditions in this study involved messages that could be compared based on how the presence of goal-relevant features in the topic-avoidance messages might influence reactions to the topic avoidance. While the primary goal of all messages was to avoid the topic, the secondary goals varied by condition and were specifically intended to mitigate the risk of offense that explicit, direct topic avoidance can entail. In general, messages that attended to secondary goals were viewed as more competent than the comparison Kernel messages that only conveyed the primary goal of avoidance. In addition, the planned contrasts of the secondary goal conditions suggested some interesting trends.

One secondary goal that stood out in the results was that of expressing Gratitude while engaging in the topic avoidance. Messages that conveyed the friend’s Gratitude for a participant’s concern were rated the most competent (i.e., effective and appropriate) overall, and they were rated as significantly more competent than the Kernel, the Polite messages, and the messages that appealed to a need for Privacy. Participants in the Gratitude condition were also least likely to say that they would reword the topic avoidance, indicating a certain level of approval for this type of wording. The proportion of individuals in the Gratitude condition who would rephrase was significantly lower than the proportions in the Kernel, Polite, and Privacy conditions. In addition, participants were more likely to allow the subject to drop in the Gratitude condition than they were in the Kernel condition, and they were as likely to do so in the Gratitude condition as in all three other conditions. Plus, the Gratitude condition exhibited the highest proportion of individuals who expressed available support with their verbal responses to the topic-avoidance messages. Taken together, these findings indicate the potential advantages of conveying appreciation for an unwanted question when engaging in explicit topic avoidance, if it is desirable to the person enacting the avoidance to manage his or her impression as a courteous friend.

Politeness theory (Brown & Levinson, 1987) may offer an explanation for why expressing gratitude would be a linguistic cue associated with messages of relatively high perceived quality. As we have already argued, explicit topic avoidance is a face-threatening act because it may imply a lack of regard for the person who is being brushed off and because it constrains the other person’s conversational autonomy to a certain extent. Other research has documented that topic-avoidance strategies are viewed by others as varying in directness and rudeness (Dailey & Palomares, 2004). Expressing gratitude for the inquirer’s concern may serve as a form of positive redress, which conveys a sense of solidarity with the inquirer. The positive redress also communicates approval for the identity of the person who asked the question as someone who is caring and did the right thing by requesting information about a friend’s difficult situation (rather, perhaps, than implying that the person is thoughtless or a busybody). As a result, the topic-avoidance message may be rendered more polite and more palatable (Brown & Levinson, 1987; Goldsmith, 2006). Expressing gratitude for a friend’s question may serve multiple purposes by simultaneously achieving the change in topic and being courteous, which is satisfying to the avoider, and reinforcing a positive regard for the friend, which is satisfying to the questioner.

This interpretation is, of course, consistent with premises of politeness theory that linguistic resources can be marshaled to convey esteem of others, thereby preserving positive face. This is not to say, however, that face and politeness concerns lie at the root of understanding how topic-avoidance messages can be constructed more or less successfully. We contend that a goals framework provides essential nuance for digesting the findings of the present study. One can see from the range of messages that we tested that some secondary goals relevant to topic avoidance involve not just concerns for face but also justifications for personal preferences and social norms (such as the case for invoking privacy). Reinforcing the bond of friendship demonstrates some conceptual overlap with positive facework; however, it also serves an actual relationship-maintaining purpose and not just a wish to make someone feel liked. All of the messages that we tested would be classified as on-record, and the basic tenets of the theory focus on messages that accomplish either positive or negative redress and not both simultaneously. (In fact, some theorizing has proposed revisions of politeness theory so that the complexity of interpersonal communication messages can better be captured; e.g., Wilson, Aleman, & Leatham, 1998.) We propose that a multiple-goals perspective specifically addresses the potential for multilayered, sophisticated messages to accomplish numerous purposes in a competent manner. In terms of goals theories, one should bear in mind that secondary goals and primary goals co-occur and that, in some cases, the secondary goals may override the desire to pursue the primary goal. In other words, there may be times when friends do not engage in explicit topic avoidance because the relevant secondary goals are quite powerful. Dillard and colleagues (e.g., Dillard et al., 1989) have made the point that “primary” ought not to be mistaken for “most important” to the speaker. For some individuals, during some situations, it will be of paramount importance for them to avoid risking any damage to their relationships or their identities, and topic avoidance may seem too risky to pursue. Ultimately, the goals framework that was used here provided a worthwhile structure for this experiment and also offers some intriguing opportunities for continued research.

For instance, an additional secondary goal that was rather prominent in the findings was invoking Privacy, conceptualized as the speaker indicating a preference to retain ownership of the information because it was personal (Petronio, 2002). Messages that appealed to Privacy were rated as less competent than messages expressing Gratitude, bolstering the relational Bond, or attempting to be Polite. Likewise, participants who received the Privacy topic-avoidance message were more likely to think that the message should be rephrased than participants in the Gratitude or Bond condition. Privacy messages were significantly more likely to prompt hurt feelings and concern than were Polite messages, unlike the Kernel condition. This initial pattern of results comports somewhat with what Caughlin and colleagues (2009) observed in their examination of message features of HIV disclosures, which was that disclosing information and then asking that the information be kept private was viewed negatively by recipients of the disclosure. Yet a somewhat contradictory finding in the present study emerged when the Privacy condition evoked more responses conveying available social support than all other conditions, with the exception of Gratitude. The comparison of findings from these two studies raises some questions about the strategic value of explicitly asking for privacy when managing sensitive information because doing so may signify alarming circumstances.

Another interesting finding in the current study involved the way that participants said they would have performed the topic avoidance themselves, if the roles had been reversed. Although some notable variations occurred in four outcomes—competence judgments, intent to rephrase, and emotional and verbal responses—there were virtually no significant findings by condition for how people would have rephrased the topic avoidance. In essence, there was a lack of evidence that goal-relevant message features predicted participants’ own subsequent rephrasing of the topic-avoidance message, despite the fact that 70% of participants indicated that they personally would have enacted the topic avoidance differently. It is possible that this lack of evidence is a methodological artifact, that the coding scheme derived from the data to categorize the rephrasing options was somehow flawed. Given that our categories overlapped conceptually with some other topic-avoidance strategies (e.g., Dailey & Palomares, 2004; Mazur & Ebesu Hubbard, 2004), however, we are inclined not to believe that the coding was problematic. Rather, this finding may provide some theoretical insight into the message production aspects of topic avoidance. It may be that the manner in which someone would reword a statement is not the same sort of reactionary judgment as evaluating competence or believing that the statement should have been said differently. Theorizing on message production has pointed to important individual differences that account for how people communicate, such as cognitive complexity and message design logics (Clark & Delia, 1979; O’Keefe, 1988, 1990). Although in this study we would expect any confounding individual differences to be randomly spread out across conditions, it is possible that goal-relevant message features are not as robust a predictor of how people would correct each other’s conversational missteps as other cognitive forces, behavioral habits, or personality differences might be.

Implications

Overall, the results of this study indicate that the way explicit topic avoidance is communicatively negotiated predicts the way that message recipients both respond to the topic avoidance and perceive the person who is avoiding. These findings build on the existing body of research on topic avoidance in a number of ways. This study helps to advance the literature on topic avoidance beyond overall extent of avoiding certain topics or tendencies to use particular avoidance tactics by focusing on actual messages. In addition, unlike most research, this project addressed the understudied person in topic-avoidance interactions—the person who is being avoided—rather than the avoider. Moreover, a notable attribute of this study is that it tested theoretically based manipulations of avoidance messages. This study represents progress in topic-avoidance research because it does not rely on empirically derived topic-avoidance strategies, which is a limitation of some previous work (e.g., Dailey & Palomares, 2004). A focus on the goal-relevant message features of avoidant communication can enrich scholarship on topic avoidance and information management processes in a manner similar to how research on the features of disclosure messages (Caughlin et al., 2008, 2009), supportive messages (e.g., Goldsmith, 1995), and persuasive communication (e.g., O’Keefe & Shepherd, 1987) has improved our understanding of what constitutes effective and appropriate interpersonal communication.

This research also forges a new connection between topic-avoidance research and literature on social support and coping, by situating topic avoidance as a response to an unwanted request for information about a stressful situation. This is a divergence from much of the previous work on topic avoidance and taboo topics, which has highlighted why and how relational partners will implicitly or explicitly avoid controversial topics that cause problems and conflict within their relationships (e.g., Roloff & Ifert, 1998). As such, the results of our study may offer some opportunities for practical application. Theorizing about information management has indicated that people are compelled by dialectical tensions in that they want to be open in their relationships but they also want to protect their privacy; they want to be connected to their loved ones but they also need autonomy (e.g., Afifi & Guerrero, 2000; Petronio, 2002). As discussed previously, people who are struggling with a stressful situation may find it challenging to extricate themselves from questions, even well-meaning ones, that feel intrusive. Yet it may be beneficial for people to do so; for example, individuals who are coping with cancer seem to feel a sense of personal control to the extent that they are able to determine how much or how little to talk to others about their illness (Donovan-Kicken et al., 2011). The findings of the present investigation offer some tangible ideas for how people might be able to avoid an unwanted conversation about a distressing issue while not being rude, so that they do not alienate potential sources of support.

Limitations and Directions for Future Research

Several promising directions for future research are evident at the close of this study, all of which have the potential to extend or challenge the findings of the present investigation. Some areas for future research derive from the limitations of the project, which, as in many experiments, often amounted to trade-offs between internal validity and external generalizability. This research is limited to the context of friendships among young adult college students, and our sample was primarily composed of women. Although this study is a useful first step in examining reactions to avoidance messages, we recognize that the act of avoidance itself is more complex than the exchanges represented here. When we asked participants to reflect on a stressful situation with which a friend was coping and to react to an imagined interaction about that topic, neither did we take into account whether participants had actually discussed or avoided this subject with the friend, and to what end, nor did we factor in other aspects of the problem, such as how long ago it has occurred or started, how distressing it was for the friend, and how the participant knew about the situation in the first place. Participants’ descriptions of their friends’ problems indicated that most were real, but a small percentage of the problems may have been hypothetical. Although some of these issues should only introduce random error into the design, they should all be considered in future research that builds on the present study.

There are obvious and inherent limits to the use of imagined conversations, although important findings about communication have been garnered with research employing hypothetical scenarios (e.g., Burleson, 1984; Burleson & MacGeorge, 2002; Dillard, Weber, & Vail, 2007; Greene, 2000) and this type of methodology has been widely used and validated in research grounded in goals perspectives (for reviews, see Dillard, Anderson, & Knobloch, 2002; Wilson, 2002). However, it is likely that the level of involvement in the imagined conversation may not have been as strong as it would have been in a real-life situation when the participant was talking with a dear friend; thus, for example, our data may inflate the extent to which people actually feel understanding about a friend’s direct topic avoidance. To augment confidence in the validity of the findings, we followed an approach that other researchers have employed, which was to measure scenario realism (e.g., Caughlin et al., 2009; Feng, 2009). Although participants indicated that the conversations they imagined were realistic, this step is not a substitute for observing real interactions. Many aspects of the present study could be replicated and extended through the collection of interaction data. Such a strategy might be especially illustrative in terms of understanding the verbal responses to topic avoidance, which in the present investigation were only gathered through written questionnaires. Clearly, a naturalistic extension of topic-avoidance messages as responses to questions about sensitive issues would be logistically challenging and would lack the theoretical message manipulations of the present study; a laboratory replication, on the other hand, would need to carefully weigh the important ethical considerations that lie in encouraging participants to ask friends about distressing subjects and would suffer from threats to external validity. Another limitation is that the topic-avoidance message stimuli developed for this study emphasized the language that people would use to explicitly avoid a topic, and thus we cannot account for variation that might occur as a result of nonverbal components of the messages. A replication of this study could examine nonverbal cues in isolation or in concert with the verbal features of avoidance strategies.

One interesting avenue of future research, certainly, would be to examine other types of secondary goals that could enhance or detract from a topic-avoidance message’s effectiveness and appropriateness. A possible limitation of the present study is that all of the pertinent secondary goals, and therefore the message manipulations, we examined were meant to be relatively prosocial in nature. We did not investigate some other goals that are potentially relevant or their message features, for example, admonishment (e.g., “You should know better than to ask about something so personal and stressful”) or de-escalation of the relationship (e.g., “I don’t feel comfortable talking to you about this anymore”). Furthermore, we did not create stimulus messages that were outright rude (e.g., “Leave me alone,” “You’re the last person in the world I would discuss that with”), even though such messages are plausible.

Although we acknowledge that not all topic-avoidance messages in real life will be gracious, research on close relationships suggests that partners’ self-interest is served by relationship-maintaining behaviors (e.g., Reis & Knee, 1996; Rusbult & Zembrodt, 1983). Likewise, the body of literature on disclosure and avoidance indicates that one of the primary reasons why managing sensitive information is so important and challenging to people is that they are concerned about damaging or destroying their relationships (Canary & Dainton, 2006; Derlega, Metts, Petronio, & Margulis, 1993; Zhang & Siminoff, 2003). Furthermore, it is our belief that designing conditions to reflect uncharitable or offensive rhetorical strategies is likely to yield unsurprising results, namely, that rude messages tend to result in lower ratings of communication competence, greater negative affect, and less understanding from partners.

In raising this aspect of our research design as a potential limitation, we wish to recognize that the messages we tested did not necessarily represent an exhaustive collection of topic-avoidance message types. However, at the root of our choice of interaction goals was the logic that explicit topic avoidance is a potentially hurtful, rude, and face-threatening act. The rationale in our study was that the highly relevant secondary goals during enacted topic-avoidance situations were goals that would soften the blow of the topic avoidance; thus, we focused on identity and relational goals that could reasonably satisfy that criterion. We thought it especially informative to examine treatment effects among types of messages that were all reasonably civil, where variation in respondents’ perceptions of topic-avoidance message quality and competence would be harder to predict from the face value of the messages and where main effects would inform our understanding of normative conversations. That being said, improvements could be made in replicating this research in the future. In our opinion, a wise avenue for a follow-up study would be to include some less prosocial message instantiations that are replicated within each condition.

Finally, it would be worthwhile to extend this research to examine topic-avoidance messages in which the secondary goals are communicated more subtly. As we have already argued, people who are coping with something stressful may not always wish to discuss it openly with others, but presumably want to avoid the topic without being inappropriate or hurtful. This line of reasoning is evocative of the literature on invisible support—meaning, efforts that are classified by support providers as intended to be helpful but that go unnoticed by support recipients. Such behaviors can be particularly effective because social support acts of which recipients are aware entail emotional costs (Bolger & Amarel, 2007; Bolger, Zuckerman, & Kessler, 2000). When recipients are aware that they are being helped and they regard the acts as implying that they are inefficacious, then they are no better off. Previous research on information management has suggested that topic avoidance can occur in ways that go undetected by relational partners and that people can be open and avoidant at the same time (e.g., Kirkman, Rosenthal, & Feldman, 2005; Mazur & Ebesu Hubbard, 2004). Like the consequences of support, the consequences of avoidance may depend on the extent to which it is detected by relational partners. This may involve communicating secondary goals that are off-record and could be interpreted in numerous ways or enacting topic avoidance using far more sophisticated messages than were tested here.

Conclusion

This study builds on previous research indicating that the meaning assigned to topic avoidance is important and that interaction goals are relevant to people’s reasons for avoidance and perceptions of avoidance. We applied a multiple-goals perspective to elucidate the interpretation of explicit topic-avoidance messages, one process within the larger conceptual milieu of information management, and found evidence that the presence of certain secondary goals during the enactment of direct, explicit topic avoidance affected message recipients’ judgments of speakers’ communication competence. The findings presented here suggest that the multiple-goals perspective demonstrates continued relevance and utility for examining information management processes, indicate that the performance of topic avoidance may be differentially effective depending on the presence of certain message features, and imply that the way one verbally avoids topics does matter.

Footnotes

Appendix

Acknowledgements

The authors thank Caroline Ashmore, Ashley Teykl, and Christian Torres-Lopez for their assistance with the pilot phase of this research. An earlier version of this manuscript was presented at the Interpersonal Communication Division’s Top Four Paper Panel at the 2010 National Communication Association convention in San Francisco, California.

Authors’ Note

The order of the second and third authors was determined randomly.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Author Biographies

Erin Donovan-Kicken (PhD, University of Illinois at Urbana–Champaign) is an assistant professor in the Department of Communication Studies at the University of Texas at Austin.

Trey D. Guinn (MA, Baylor University) is a doctoral candidate at the University of Texas at Austin.

Lynsey Kluever Romo (MA, American University) is a doctoral candidate at the University of Texas at Austin.

Lea D. L. Ciceraro (MA, San Diego State University) is an adjunct faculty member at Austin Community College. She was a doctoral student at the University of Texas at Austin when this research was conducted.

References

Afifi

T. D.

(2003). “Feeling caught” in stepfamilies: Managing boundary turbulence through appropriate privacy coordination rules. Journal of Social and Personal Relationships, 20, 729-756.

Afifi

T. D.

Afifi

W. A.

Morse

C. R.

Hamrick

(2008). Adolescents’ avoidance tendencies and physiological reactions to discussions about their parents’ relationship: Implications for postdivorce and nondivorced families. Communication Monographs, 75, 290-317.

Afifi

T. D.

Caughlin

J. P.

Afifi

W. A.

(2007). The dark side (and light side) of avoidance and secrets. In Spitzberg

B. H.

Cupach

W. R.

(Eds.), The dark side of interpersonal communication (2nd ed., pp. 61-92). Mahwah, NJ: Erlbaum.

Afifi

T. D.

Schrodt

(2003a). “Feeling caught” as a mediator of adolescents’ and young adults’ avoidance and satisfaction with their parents in divorced and non-divorced households. Communication Monographs, 70, 142-173.

Afifi

T. D.

Schrodt

(2003b). Uncertainty and the avoidance of the state of one’s family in stepfamilies, postdivorce single-parent families, and first-marriage families. Human Communication Research, 29, 516-532.

Afifi

Steuber

(2009). The revelation risk model (RRM): Factors that predict the revelation of secrets and strategies used to reveal them. Communication Monographs, 76, 144-176.

Afifi

W. A.

Caughlin

J. P.

(2006). A close look at revealing secrets and some consequences that follow. Communication Research, 33, 467-488.

Afifi

W. A.

Guerrero

L. K.

(2000). Motivations underlying topic avoidance in close relationships. In Petronio

(Ed.), Balancing the secrets of private disclosures (pp. 165-180). Mahwah, NJ: Erlbaum.

Algina

Olejnik

(2003). Conducting power analyses for ANOVA and ANCOVA in between-subjects designs. Evaluation & the Health Professions, 26, 288-314.

10.

Arrington

M. I.

(2005). “She’s right behind me all the way”: An analysis of prostate cancer narratives and changes in family relationships. Journal of Family Communication, 5, 141-162.

11.

Baxter

L. A.

Montgomery

B. M.

(1996). Relating: Dialogues & dialectics. New York, NY: Guilford.

12.

Baxter

L. A.

Wilmot

W. W.

(1985). Taboo topics in close relationships. Journal of Social and Personal Relationships, 2, 253-269.

13.

Bolger

Amarel

(2007). Effects of social support visibility on adjustment to stress: Experimental evidence. Journal of Personality and Social Psychology, 92, 458-475.

14.

Bolger

Zuckerman

Kessler

R. C.

(2000). Invisible support and adjustment to stress. Journal of Personality and Social Psychology, 79, 953-961.

15.

Brown

Levinson

(1987). Politeness: Some universals in language usage. Cambridge, UK: Cambridge University Press.

16.

Bruning

J. L.

Kintz

B. L.

(1997). Computational handbook of statistics (4th ed.). New York, NY: Longman.

17.

Bulmer

(1979). Concepts in the analysis of qualitative data. Sociological Review, 27, 651-677.

18.

Burleson

B. R.

(1984). Age, social-cognitive development, and the use of comforting strategies. Communication Monographs, 51, 140-153.

19.

Burleson

B. R.

MacGeorge

E. L.

(2002). Supportive communication. In Knapp

M. L.

Daly

J. A.

(Eds.), Handbook of interpersonal communication (2nd ed., pp. 374-424). Thousand Oaks, CA: Sage.

20.

Bute

J. J.

(2009). “Nobody thinks twice about asking”: Women with a fertility problem and requests for information. Health Communication, 24, 752-763.

21.

Canary

D. J.

Dainton

(2006). Maintaining relationships. In Vangelisti

A. L.

Perlman

(Eds.), The Cambridge handbook of personal relationships (pp. 727-743). New York, NY: Cambridge University Press.

22.

Canary

D. J.

Spitzberg

B. H.

(1987). Appropriateness and effectiveness in the perception of conflict strategies. Human Communication Research, 13, 93-118.

23.

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.

24.

Caughlin

J. P.

Afifi

T. D.

(2004). When is topic avoidance unsatisfying? Examining moderators of the association between avoidance and dissatisfaction. Human Communication Research, 30, 479-513.

25.

Caughlin

J. P.

Brashers

D. E.

Ramey

M. E.

Kosenko

K. A.

Donovan-Kicken

Bute

J. J.

(2008). The message design logics of responses to HIV disclosures. Human Communication Research, 34, 655-684.

26.

Caughlin

J. P.

Bute

J. J.

Donovan-Kicken

Kosenko

K. A.

Ramey

M. E.

Brashers

D. E.

(2009). Message features influence reactions to HIV disclosures. Health Communication, 24, 270-283.

27.

Caughlin

J. P.

Golish

T. D.

(2002). An analysis of the association between topic avoidance and dissatisfaction: Comparing perceptual and interpersonal explanations. Communication Monographs, 69, 275-295.

28.

Caughlin

J. P.

Malis

R. S.

(2004a). Demand/withdraw between parents and adolescents as a correlate of relational satisfaction. Communication Reports, 17, 59-71.

29.

Caughlin

J. P.

Malis

R. S.

(2004b). Demand/withdraw communication between parents and adolescents: Connections with self-esteem and substance use. Journal of Social and Personal Relationships, 21, 125-148.

30.

Caughlin

J. P.

Scott

A. M.

(2010). Toward a communication theory of the demand/withdraw pattern of interaction in interpersonal relationships. In Smith

Wilson

S. R.

(Eds.), New directions in interpersonal communication (pp. 180-200). Thousand Oaks, CA: Sage.

31.

Clark

R. A.

Delia

J. G.

(1979). Topoi and rhetorical competence. Quarterly Journal of Speech, 65, 187-206.

32.

Cloven

D. H.

Roloff

M. E.

(1993). The chilling effect of aggressive potential on the expression of complaints in intimate relationships. Communication Monographs, 60, 199-219.

33.

Cohen

(1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20, 37-46.

34.

Cohen

(1989). Statistical power analysis for the behavioral sciences (2nd ed.). San Diego, CA: Academic Press.

35.

Creswell

J. W.

Plano Clark

V. L.

(2007). Designing and conducting mixed methods research. Thousand Oaks, CA: Sage.

36.

Dailey

R. M.

Palomares

N. A.

(2004). Strategic topic avoidance: An investigation of topic avoidance frequency, strategies used, and relational correlates. Communication Monographs, 71, 471-496.

37.

Derlega

V. J.

Metts

Petronio

Margulis

S. T.

(1993). Self-disclosure. Newbury Park, CA: Sage.

38.

Dillard

J. P.

(1990). A goal-driven model of interpersonal influence. In Dillard

J. P.

(Ed.), Seeking compliance: The production of interpersonal influence messages (pp. 41-56). Scottsdale, AZ: Gorsuch Scarisbrick.

39.

Dillard

J. P.

Anderson

J. W.

Knobloch

L. K.

(2002). Interpersonal influence. In Knapp

M. L.

Daly

J. A.

(Eds.), Handbook of interpersonal communication (pp. 425-474). Thousand Oaks, CA: Sage.

40.

Dillard

J. P.

Segrin

Harden

J. M.

(1989). Primary and secondary goals in the production of interpersonal influence messages. Communication Monographs, 56, 19-38.

41.

Dillard

J. P.

Weber

K. M.

Vail

R. G.

(2007). The relationship between the perceived and actual effectiveness of persuasive messages: A meta-analysis with implications for formative campaign research. Journal of Communication, 57, 613-631.

42.

Donovan-Kicken

Tollison

A. C.

Goins

E. S.

(2011). A grounded theory of control over communication among individuals with cancer. Journal of Applied Communication Research, 39, 310-330.

43.

Durham

W. T.

(2008). The rules-based process of revealing/concealing the family planning decisions of voluntarily child-free couples: A communication privacy management perspective. Communication Studies, 59, 132-147.

44.

Feng

(2009). Testing an integrated model of advice giving in supportive interactions. Human Communication Research, 35, 115-129.

45.

Folkman

Lazarus

R. S.

(1985). If it changes it must be a process: Study of emotion and coping during three stages of a college examination. Journal of Personality and Social Psychology, 48, 150-170.

46.

Furman

Buhrmester

(1992). Age and sex differences in perceptions of networks of personal relationships. Child Development, 63, 103-115.

47.

Garel

Blondel

(1992). Assessment at 1 year of the psychological consequences of having triplets. Human Reproduction, 7, 729-732.

48.

Goffman

(1967). Interaction ritual: Essays in face-to-face interaction. Garden City, NJ: Doubleday.

49.

Goldsmith

D. J.

(1995). The communicative microdynamics of support: Interactive coping in close relationships. In Burleson

B. R.

(Ed.), Communication yearbook 18 (pp. 414-433). Thousand Oaks, CA: Sage.

50.

Goldsmith

D. J.

(2004). Communicating social support. Cambridge, UK: Cambridge University Press.

51.

Goldsmith

D. J.

(2006). Brown & Levinson’s politeness theory. In Whaley

B. B.

Samter

(Eds.), Explaining communication: Contemporary theories and exemplars (pp. 219-236). Mahwah, NJ: Erlbaum.

52.

Goldsmith

D. J.

Donovan-Kicken

(2009). Adding insult to injury: The contributions of politeness theory to understanding hurt feelings in close relationships. In Vangelisti

A. L.

(Ed.), Feeling hurt in close relationships (pp. 50-72). New York, NY: Cambridge University Press.

53.

Goldsmith

D. J.

MacGeorge

E. L.

(2000). The impact of politeness and relationship on perceived quality of advice about a problem. Human Communication Research, 26, 234-263.

54.

Goldsmith

D. J.

Miller

L. E.

Caughlin

J. P.

(2008). Openness and avoidance in couples communicating about cancer. In Beck

C. S.

(Ed.), Communication yearbook 31 (pp. 62-115). Hillsdale, NJ: Erlbaum.

55.

Golish

T. D.

Caughlin

J. P.

(2002). “I’d rather not talk about it”: Adolescents’ and young adults’ use of topic avoidance in stepfamilies. Journal of Applied Communication Research, 30, 78-106.

56.

Greene

(2000). Disclosure of chronic illness varies by topic and target: The role of stigma and boundaries in willingness to disclose. In Petronio

(Ed.), Balancing the secrets of private disclosures (pp. 123-135). Mahwah, NJ: Erlbaum.

57.

Grice

(1989). Studies in the way of words. Cambridge, MA: Harvard University Press.

58.

Guerrero

L. K.

Afifi

W. A.

(1995a). Some things are better left unsaid: Topic avoidance in family relationships. Communication Quarterly, 43, 276-296.

59.

Guerrero

L. K.

Afifi

W. A.

(1995b). What parents don’t know: Topic avoidance in parent–child relationships. In Socha

T. J.

Stamp

G. H.

(Eds.), Parents, children, and communication: Frontiers of theory and research (pp. 219-245). Mahwah, NJ: Erlbaum.

60.

Heffner

K. L.

Loving

T. J.

Kiecolt-Glaser

J. K.

Himawan

L. K.

Glaser

Malarkey

W. B.

(2006). Older spouses’ cortisol responses to marital conflict: Associations with demand/withdraw communication patterns. Journal of Behavioral Medicine, 29, 317-325.

61.

Jackson

(1992). Message effects research: Principles of design and analysis. New York, NY: Guilford.

62.

Jackson

Brashers

D. E.

(1994). Random factors in ANOVA. Thousand Oaks, CA: Sage.

63.

Jensen

J. D.

(2008). Scientific uncertainty in news coverage of cancer research: Effects of hedging on scientists’ and journalists’ credibility. Human Communication Research, 34, 347-369.

64.

Kirkman

Rosenthal

D. A.

Feldman

S. S.

(2005). Being open with your mouth shut: The meaning of “openness” in family communication about sexuality. Sex Education, 5, 49-66.

65.

Knobloch

L. K.

Carpenter-Theune

K. E.

(2004). Topic avoidance in developing romantic relationships: Associations with intimacy and relational uncertainty. Communication Research, 31, 173-205.

66.

Knudson

R. M.

Sommers

A. A.

Golding

S. L.

(1980). Interpersonal perception and mode of resolution in marital conflict. Journal of Personality and Social Psychology, 38, 751-763.

67.

Lepore

S. J.

(1997). The social context of coping with chronic stress. In Gottlieb

(Ed.), Coping with chronic stress (pp. 133-160). New York, NY: Plenum.

68.

Lyons

R. F.

Sullivan

M. J. L.

Ritvo

P. G.

Coyne

J. C.

(1995). Relationships in chronic illness and disability. Thousand Oaks, CA: Sage.

69.

Mazur

M. A.

Ebesu Hubbard

A. S.

(2004). “Is there something I should know?” Topic avoidant responses in parent–adolescent communication. Communication Reports, 17, 27-37.

70.

O’Keefe

B. J.

(1988). The logic of message design: Individual differences in reasoning about communication. Communication Monographs, 55, 80-103.

71.

O’Keefe

B. J.

(1990). The logic of regulative communication: Understanding the rationality of message designs. In Dillard

J. P.

(Ed.), Seeking compliance: The production of interpersonal influence messages (pp. 87-104). Scottsdale, AZ: Gorsuch Scarisbrick.

72.

O’Keefe

B. J.

McCornack

S. A.

(1987). Message design logic and message goal structures: Effects on perceptions of message quality in regulative communication situations. Human Communication Research, 14, 68-92.

73.

O’Keefe

B. J.

Shepherd

G. J.

(1987). The pursuit of multiple objectives in face-to-face persuasive interactions: Effects of construct differentiation on message organization. Communication Monographs, 54, 396-419.

74.

Petronio

(2002). Boundaries of privacy: Dialectics of disclosure. Albany: State University of New York Press.

75.

Reis

H. T.

Knee

C. R.

(1996). What we know, what we don’t know, and what we need to know about relationship knowledge structures. In Fletcher

G. J. O.

Fitness

(Eds.), Knowledge structures in close relationships: A social psychological approach (pp. 169-191). Mahwah, NJ: Erlbaum.

76.

Roloff

M. E.

Ifert

D. E.

(1998). Antecedents and consequences of explicit agreements to declare a topic taboo in dating relationships. Personal Relationships, 5, 191-206.

77.

Roloff

M. E.

Ifert

D. E.

(2000). Conflict management through avoidance: Withholding complaints, suppressing arguments, and declaring topics taboo. In Petronio

(Ed.), Balancing the secrets of private disclosures (pp. 151-163). Mahwah, NJ: Erlbaum.

78.

Rubin

J. Z.

(1994). Models of conflict management. Journal of Social Issues, 50, 33-45.

79.

Rusbult

C. E.

Zembrodt

I. M.

(1983). Responses to dissatisfaction in romantic involvements: A multidimensional scaling analysis. Journal of Experimental Social Psychology, 19, 274-293.

80.

Suter

E. A.

Ballard

R. L.

(2009). “How much did you pay for her?” Decision-making criteria underlying adoptive parents’ responses to inappropriate remarks. Journal of Family Communication, 9, 107-125.

81.

Tesser

Rosen

(1975). The reluctance to transmit bad news. In Berkowitz

(Ed.), Advances in experimental social psychology (Vol. 8, pp. 193-232). New York, NY: Academic Press.

82.

Tracy

(Ed.). (1991). Understanding face-to-face interaction: Issues linking goals and discourse. Hillsdale, NJ: Erlbaum.

83.

Vangelisti

A. L.

(2009). Hurt feelings: Distinguishing features, functions, and overview. In Vangelisti

A. L.

(Ed.), Feeling hurt in close relationships (pp. 3-11). Cambridge, UK: Cambridge University Press.

84.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070.

85.

Wilson

S. R.

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

86.

Wilson

S. R.

Aleman

C. G.

Leatham

G. B.

(1998). Identity implications of influence goals: A revised analysis of face-threatening acts and application to seeking compliance with same-sex friends. Human Communication Research, 25, 64-96.

87.

Zahlis

E. H.

Shands

M. E.

(1993). The impact of breast cancer on the partner 18 months after diagnosis. Seminars in Oncology Nursing, 9, 83-87.

88.

Zhang

A. Y.

Siminoff

L. A.

(2003). Silence and cancer: Why do families and patients fail to communicate? Health Communication, 15, 415-429.