The Relative Effectiveness of Active Listening in Initial Interactions

Abstract

Although active listening is considered an important communication skill in a variety of occupational and therapeutic fields, few experiments compare dyadic partners' perceptions of active listening with other types of listening responses. This study involves 115 participants engaged in interactions with 10 confederates trained to respond with active listening messages, advice, or simple acknowledgements. Results indicate that participants who received active listening responses felt more understood than participants who received either advice or simple acknowledgements. Further, participants who received either active listening responses or advice were more satisfied with their conversation and perceived the confederate to be more socially attractive than participants who received simple acknowledgements, although the effect sizes for these differences were small. Conversational satisfaction and social attractiveness did not differ between participants receiving active listening responses and participants receiving advice, however.

Among the listening skills academics and practitioners identify as valuable, active listening (also called empathic listening, speaker-listener technique, reflected listening, dialogic listening, etc.) continues to garner the lion's share of attention. Active listening involves restating a paraphrased version of the speaker's message, asking questions when appropriate, and maintaining moderate to high nonverbal conversational involvement. Practitioners and researchers from fields such as social work (Rogers & Welch, 2009), nursing (Bryant, 2009), education (McNaughton, Hamlin, McCarthy, Head-Reeves, & Schreiner, 2007), physician-patient communication (Fassaert, van Dulmen, Schellevis, & Bensing, 2007), leadership (Hoppe, 2007), public administration (Stein, 2009), sales (Boe, 2008), and crisis negotiation (Royce, 2005), specifically identify active listening as an important communication skill during initial interactions.

In the communication field, the value placed on active listening makes it de rigueur in the listening chapters of many popular communication textbooks (e.g., Adler & Proctor, 2011; Canary, Cody, & Manusov, 2008; Devito, 2007; Pearson, Nelson, Titsworth, & Harter, 2010; Verderber, Verderber, & Berryman-Fink, 2009; Wood, 2012). Yet few studies produce convincing evidence that active listening actually produces comparatively better outcomes than other response types in early interactions. The purpose of our investigation is to compare perceptions of interaction partners' who use active listening, unsolicited advice, or simple acknowledgements (back-channel cues) during an initial interaction.

ACTIVE LISTENING

The International Listening Association (ILA; 2012) defines listening as, “the process of receiving, constructing meaning from and responding to spoken and/or nonverbal messages”. In this study, we focus on the response component of listening. Active listening (see also, speaker-listener technique, Stanley, Bradbury, & Markman, 2000) was developed by Gordon (1975) and has roots in Rogers' (1951) conceptualization of empathic listening (Orlov, 1992). Rogers formulated empathic listening as a psycho-therapeutic technique, which demonstrates unconditional acceptance and unbiased reflection of a client's experience through message paraphrasing. Levitt (2001) identifies active listening as a therapeutic micro-skill involving listening attentively and responding empathically so a client feels heard.

With slight variations across scholars, most treatments of active listening include at least three elements. The first element involves expressing interest in the speaker's message by displaying nonverbal involvement in the form of back channeling (McNaughton et al., 2007). The second element includes refraining from judgment and paraphrasing the speaker's message (e.g., “what I'm hearing you say is …”; Garland, 1981). For example, if the speaker expresses frustration, the active listener might respond, “I can understand how that situation could cause frustration.” Third, active listening may also include asking questions to encourage the speaker to elaborate on his or her beliefs or feelings (Paukert, Stagner, & Hope, 2004). An active listening response theoretically communicates empathy and builds trust by indicating unconditional regard and by confirming the other's experience (Lester, 2002; Orlov, 1992; Rogers, 1951). We conceptualize active listening as having three parts: 1) demonstrates moderate to high nonverbal involvement, 2) reflects the speaker's message using verbal paraphrasing, and 3) may include asking questions that encourage speakers to elaborate on his or her experiences.

LISTENING IN INITIAL INTERACTIONS

Initial interactions constitute a significant context for the study of listening skill. Every relationship begins with a first encounter. Further, a great deal of our instrumental goals are accomplished in first-time conversations with sales people, customers, doctors, therapists, potential employers, yoga instructors, and the like. In first encounters, people attempt to solve a variety of information seeking, relationship, and impression management problems. For example, people are motivated to gather information to reduce uncertainty in novel situations (e.g., Afifi & Weiner, 2004; Berger & Calabrese, 1975). In first encounters people also work to create a favorable first impressions and form accurate impressions of others (e.g., Bodie, Cyr, Pence, Rold, & Honeycutt, 2012; Bachman & Zakahi, 2000). These first impressions are particularly important because they shape future interactions and can determine whether interactions will occur in the future (e.g., Ramirez, 2007). The rewards realized from early enjoyable conversations lay the foundation for predictions of rewards and costs in the future (e.g., Burleson & Denton, 1992; Sunnafrank & Ramirez, 2004).

Research points to the role of listening in producing positive interaction outcomes. For example, listening impacts uncertainty reduction and information management via comprehension and retention of a speaker's message (Janusik, 2007). Additionally, effective listeners generally project more positive impressions than ineffective listeners (e.g., Drollinger, Comer, & Warrington, 2006; Haas & Arnold, 1995) and are perceived to be more trustworthy (Ramsey & Sohi, 1997), friendly (Bodie et al., 2012), understanding (Cahn & Frey, 1989), and socially attractive (Weger, Bell, & Emmett, 2010; Young & Cates, 2010). Finally, good listeners produce more satisfying (i.e., rewarding) interactions between patients and their physicians (Henry, Fuhrel-Forbis, Rogers, & Eggly, 2012), real estate clients and their agents (Amba-Rao, 1991), protégés and their mentors (Young & Cates, 2010), and between wives and husbands (e.g., Pasupathi, Carstensen, Levenson, & Gottman, 1999).

More specifically, Bodie et al. (2012) identify both attributes and behaviors related to perceptions of listening competence in initial interactions. People characterize competent listeners in initial interactions as attentive, friendly, understanding, responsive, and able to manage the flow of conversation (Bodie et al., 2012). Additionally, particular listening behaviors are linked to different attributes. For example, verbal paraphrases are associated with attentiveness and responsiveness while questions are linked to conversation management, attentiveness, and responsiveness. Nonverbal behaviors such as eye contact and composure appear related to attentiveness, friendliness, and conversation management skills.

The structure of active listening seems well suited to helping people accomplish initial interaction goals; verbal paraphrasing, asking appropriate questions, and nonverbal involvement all associated with good listening practice. One might assume a healthy amount of research examining this listening practice, however, little empirical evidence exists demonstrating active listening results in more positive interactions when compared to other types of listening responses. Moreover, only a portion of that research involves observations of actual behavior and research comparing active listening to other types of responses does not consistently demonstrate an advantage for active listening. Below, we identify and summarize essentially all of the extant research examining active listening in initial interactions and a small part of the research on active listening in marital interaction.

COMPARATIVE ADVANTAGES OF ACTIVE LISTENING

Several studies point to positive outcomes associated with active listening. Gearhart and Bodie (2011) found higher scores on the Active Empathic Listening scale (AEL) associate positively with Riggio's (2005) measures of social sensitivity, social expression, emotional sensitivity, and social control, all of which are related to a host of positive interaction and relationship outcomes (e.g., Riggio & Reichard, 2008). In addition, ratings of target others on the AEL associate positively with targets' communication skill and discriminate between people judged to be “good” or “bad” listeners (Bodie, 2011). Further, Drollinger et al. (2006) reported sales people's engagement in active listening responses were positively associated with customers' ratings of the sales professionals' listening skill. An organizational study in Japan indicates that employees' with lower levels of psychological job stress have supervisors who self-report regular use of active listening in discussions with employees (Mineyama, Tsutsumi, Takao, Nishiuchi, & Kawakami, 2007). Expert hostage negotiators at the Federal Bureau of Investigation use paraphrasing, reflecting, and other active listening behaviors more often than novice negotiators in role play exercises (Van Hasselt et al., 2005). Finally, Reznik, Roloff, and Miller's (2012) study of romantic couples' serial arguments report that active listening associates positively with problem solving, relationship stability and perceived resolvability of the problem, and associates negatively with intrusive thoughts during arguments.

In addition, scholarship from the training literature reports active listening training improves trainees' listening skill. For example, active listening training increased education students' confidence in their own listening skills (McNaughton et al., 2007). Federal law enforcement crisis negotiators received higher ratings on communication skill during mock hostage negotiation exercises after receiving active listening training than before training (Van Hasselt et al., 2006). Similarly, supervisors rated college student helpline volunteers more skillful six weeks following active listening training (Paukert et al., 2004). Likewise, mental health counseling students received higher scores in counseling effectiveness after learning active listening (Levitt, 2001). Miller, Hendrick, and Orlofsky (1991) also successfully improved crisis intervention counselors' empathic listening skill by training them to deliver active listening responses. Finally, D'Augelli and Levy (1978) report that crisis volunteers responding with advice and problem solving generated shorter, less effective, conversations than when employing empathic (active listening) responses. Overall, these studies suggest active listening training improves perceptions of trainees' listening skill.

Unfortunately, not all of the research on active listening produces favorable results. Besides contradictory studies in the marital research (see below), two studies contradict D'Augelli and Levy's (1978) study comparing active listening and advice. In one study, volunteer telephone crisis counselors offering advice were rated more helpful and likeable than counselors responding with empathic (or active) listening (Libow & Doty, 1976). In a second study, Scholl (2002) reports college academic counseling clients preferred receiving advice rather than active listening responses early in counselor-client relationships. Additional research points to the possibility that people either prefer advice to active listening in initial interactions, or simply do not favor one over the other. For example, in Bodie et al. (2012, Study 2), giving advice was mentioned more often by participants than verbal paraphrasing when asked to identify preferred listening behaviors. Although participants' agreement about rank ordering of behaviors is highly variable, a study by Bodie et al. (2012, Study 3) reports participants ranked advice received higher (mean ranking 5th out of 19) than paraphrasing (15th) in their perceptions of behaviors characteristic of good listening.

EVALUATING ACTIVE LISTENING OUTCOMES IN INITIAL INTERACTIONS

One of the limitations in our understanding of active listening in initial interactions is that few studies compare active listening to other response types directly. Most observational research compares judgments of listening skills by outsiders (often the trainers themselves) or compares partners' perceived interaction/relationship outcomes pre and post active listening training. The majority of the remaining studies report associations between self and/or other reports of using active listening and some criterion variable (e.g., Drollinger et al., 2006; Young & Cates, 2010). To date, only four studies directly compare participants' judgments of interaction outcomes following confederates deployment of active listening versus another type of listening response. Three studies compare active listening to giving advice (i.e., D'Augelli & Levy, 1978; Libow & Doty, 1976; Scholl, 2002). One limitation of these studies is that participants were seeking out professional or volunteer counselors, so the reception of advice may have been a more goal-relevant outcome than validation communicated through active listening. The forth study compares active listening to simple acknowledgements in informational interviews that also involved a situation in which the participant and confederate were not on an equal footing (Weger, Bell, & Emmett, 2010). Because none of these studies involve peer-to-peer initial interactions, and because advice seems to be an important and useful comparison point, we attempt to fill this gap by comparing active listening to both advice and simple acknowledgments in initial peer interactions.

Listening Outcomes

In general, competent listening involves communicators' perceived ability to accomplish contextually driven listening goals and produce rewarding interactions (e.g., Bodie, Worthington, Imhof, & Cooper, 2008). Bodie et al.'s (2008) synthesis of listening literature identifies understanding, experiencing positive affect, and relationship building as essential products of the listening process. This approach squares with both theory and research strongly suggesting people prefer conversational partners and interactions that provide real or perceived rewards (e.g., Berscheid, 1985; Burleson & Samter, 1996) and that these rewards guide peoples' interest in future interactions (Sunnafrank, 1986; Sunnafrank & Ramirez, 2004). Following Bodie et al. (2008) and others (e.g., Cahn, 1990) we suggest skilled listening produces rewarding interaction outcomes by expressing understanding of a partner's message, creating positive affect during the interaction, and by creating affiliation between partners.

Hypotheses

Understanding is identified by Bodie et al. (2008) as one of three listening outcomes. Although one goal of listening is to comprehend the message, producing listening responses demonstrating to the speaker that s/he has been understood is a more important dyadic outcome (e.g., Cahn, 1990). Feeling understood by the listener has important implications for the relationship between speaker and listener because it aids in establishing and maintaining satisfying relationships (Cahn, 1990), verifying peoples' self-conceptions (Swann, 1983), and differentiating between good and poor listeners in initial interactions (Bodie et al., 2012). Active listening ought to outperform advice and simple acknowledgements in creating the perception that one has been understood for a few reasons. First, communicating understanding of a speaker forms the primary theoretical rationale for engaging in active/empathic listening (Orlov, 1992). Similarly, the speaker-listener technique was specifically designed, in part, to increase partners' feelings of being understood (Stanley, Markman, Jenkins, & Blumberg, 2008). Second, most treatments of active listening discourage advice because, rather than expressing understanding of the speaker's message, advice refocuses attention from the speaker's to the listener's perspective (Rogers, 1951; Stanley et al., 1997). Simple acknowledgements, although providing acknowledgement of message reception, do not directly communicate the listener's understanding of the speaker. Finally, although two studies found advice to be preferable to active listening, the advice seeking nature of the communication contexts in those studies probably favored receiving advice. We therefore make the following prediction:

H1: People in initial interactions experience higher levels of feeling understood when the interaction partner engages in active listening compared to giving advice or enacting simple acknowledgements.

A second listening outcome identified by Bodie et al. (2008) involves the creation of positive affect. One perceptual outcome of initial interactions related to positive affect involves the positive feelings one receives from engaging in a satisfying interaction. Communication satisfaction, as Hecht (1978) conceptualizes it, results from a person's overall assessment of the positive and negative affect generated in a conversation. We chose communication satisfaction as an important listening outcome in initial interactions because people base their intentions for future interactions on their evaluations of their satisfaction with the current (or most recent) conversation (Sunnafrank & Ramirez, 2004). Hecht (1978) maintains conversations are satisfying to the extent that the interaction meets, or exceeds, the person's expectations. We predict that active listening will generate more communication satisfaction than advice and simple acknowledgements for a few reasons. Zakahi (1985) found that communication satisfaction is a function of communication skill and Gearhart and Bodie (2011) found active/empathic listening associates positively with general social skills. Further, Bodie (2011) reports positive associations among participants' ratings of a partner's active/empathic listening and their communication effectiveness and appropriateness. In addition, advice is somewhat disconfirming because it implicitly denies the legitimacy of the speaker's perspective by suggesting that the listener's understanding of the situation is superior to the speaker's. Research suggests that confirming communication produces more satisfying conversations than disconfirming communication (LeBlanc, 2004). We therefore predict:

H2: People in initial interactions experience higher levels of communication satisfaction when the interaction partner engages in active listening compared to giving advice or enacting simple acknowledgements.

A third listening outcome related to initial interactions involves relationship building (Bodie et al., 2008). One of the impressions people form of others in initial interactions involves the partner's desirability as an interaction partner, or social attractiveness (McCroskey, McCroskey, & Richmond, 2006). People build relationships with people who are rewarding conversational partners and who are predicted to be rewarding partners in the future. Therefore, social attractiveness is an important listening outcome in initial interactions because it motivates the desire for future interactions and forms the foundation for relationship development (e.g., Berscheid & Walster, 1978; Sunnafrank, 1986). One of the few experimental studies examining initial interactions found that participants rated active listeners higher in social attractiveness than confederates enacting simple acknowledgements (Weger et al., 2010). We therefore predict:

H3: People in initial interactions experience more social attraction to an interaction partner when the partner engages in active listening compared to giving advice or enacting simple acknowledgements.

METHOD

Participants

Participants included 115 undergraduates from an introductory communication course for noncommunication majors at a large southeastern university. The course did not include active listening in its curriculum. Participants received extra-credit in exchange for participating in the study. Males represented 39.1% of the sample, and females represented 70.9%. The ethnic makeup of the sample was 3.5% African American/non-Hispanic, 5.2% Asian, 67% White/non-Hispanic, 2.6% Caribbean, 18% Hispanic, and 3.5% answered “other” or “prefer not to say.” The sample ranged in age from 18 to 50 with a mean of 20.01 and a standard deviation of 4.37.

TABLE 1 Example of Active Listening Response

Speaker	Message
Confederate	(Reading the topic card) Please discuss your biggest disappointment or a negative experience with (university) since you have been a student here. Be as specific as possible.
Participant	Well, it's been pretty much the same because so many of the classes offered for majors and minors come at the wrong time so I'm stuck with classes I don't need, or seven-thirty classes, or two classes at the same time stuff like that so I have to pick up classes that I don't necessarily want to take right now so I can't get all my gen-ed classes taken at once.
Confederate	I can see why that's kind of, uhm, hard to, uhm, work out a schedule when you need two classes and they happen to be at the same time, and you wanna get your gen-ed's out of the way before you go ahead and then you have to take a class you don't need, so I would be frustrated too.

Experimental Stimuli

The experimental stimuli consisted of the three types of listening responses. The active listening condition included paraphrasing the participant's message and, when appropriate, asking the participant to elaborate on what they said (see Table 1 for an example from the data). The simple acknowledgement condition consisted of only back channel cues such as head nods, and short verbal statements such as “I see,” “OK,” “That makes sense,” and so on. Simple acknowledgements (or back channel cues) in initial interactions provide a control condition for the experiment. Simple acknowledgements represent the minimum level of listening skill from a developmental perspective (Hess & Johnston, 1988), signal basic agreement with the speaker, and indicate rapport between the speaker and listener (Bennett & Jarvis, 1991). The advice condition involved giving advice to the participant such as suggesting activities the participant should engage in over the weekend, solutions to problems identified by participant, and so on (see Table 2 for an example from the data). Many conceptualizations of active listening prohibit offering advice because the listener's needs are placed above the speaker's (e.g., Rogers, 1951; Stanley, Markman, & Blumberg, 1997). Also, several studies identify advice as a common kind of response strategy for both males and females in naturally occurring conversations (e.g., Cutrona & Suhr, 1994; D'Augelli & Levy, 1978; Mickelson, Helgeson, & Weiner, 1995) making it an interesting comparison point.

TABLE 2 Example of Advice Response

Speaker	Message
Confederate	(Reads topic card): Please describe your plans for the weekend. Be as specific as possible.
Participant	I'm going home actually, probably go out on my boat. Maybe go fishing.
Confederate	Have you ever caught swordfish? You should go deep sea fishing and try to catch some swordfish cuz, I think, as a fisherman, you would like that.
Participant	I'm going to the Gulf, and there aren't really swordfish.
Confederate	Still you should go, it's quite an adventure.

Confederates and Confederate Training

Confederates (N = 12) were recruited from undergraduate interpersonal communication courses. Two of the confederates did not interact with the minimum three participants per condition (at least one in each condition) and were dropped from the analysis. ¹ The remaining confederates self-identify their ethnicity as African American (n = 3), Hispanic (n = 2), and Caucasian or White (n = 5), ranged in age from 19 to 25 (M = 21.4, SD = 2.1) and were divided evenly between males and females. Confederates received about four hours of training in active listening, about one hour training in advice giving and one half hour training in simple acknowledgements. The training time for advice and simple acknowledgement was shorter because these responses were easier for the confederates to learn. Confederates practiced the different response conditions, which included round-robin practice sessions with other confederates, for about five hours. Confederates were instructed to maintain a moderate level of nonverbal involvement across conditions so that only their verbal response varied.

Experimental Procedures

Participants and confederates arrived at the communication lab at an appointed time, sat in the same waiting area, and were instructed not to talk to each other until the experiment began. Once in the lab, the participant/confederate dyad was told the experiment, “Looks at how people talk to each other when they meet for the first time.”

Dyads sat chairs spaced about 63–74 cm from front edge to front edge. Participants received two 3 × 5 cards with the following conversation starter topics printed on them, “Please discuss your plans for this weekend,” and, “What has been your biggest disappointment with (university name) since you began attending school here.” The topics were developed to engage students in conversation. The participant and confederate took turns answering the first question before moving on to the second question and alternated the order in which they took the role of speaker and listener. Each confederate invented and memorized his or her own response to each of the starter questions to ensure that variation in responses to the starter questions were not conflated with the effects of the listening response conditions. Lab assistants used a six-sided die to determine the experimental condition and secretly signaled confederates which condition to enact. The confederates enacted the same listening condition for both topics. The conversations lasted about three to seven minutes (M = 4.9 mins, SD = 1.11 mins). Following the conversation, a lab assistant escorted the participant and confederate to separate rooms where the participants completed the dependent measures along with some demographic information.

Dependent Measures

Perceived Understanding

Perceived understanding was operationalized using the Feelings of Understanding/Misunderstanding Scale (FUMS; Cahn & Shulman, 1984). The FUMS consists of eight affective terms associated with respondents' experience of feeling understood (e.g., satisfaction, relaxation, pleasure), eight affective terms associated with respondents' experience of feeling misunderstood (e.g., annoyance, discomfort, dissatisfaction), and eight distractor terms (e.g., self-reliance, enviousness, humbleness). Participants used a 5-point scale (1 = very little; 5 = very great) to rate the extent to which each term described how they felt immediately following the interaction.

A feelings of understanding scale is computed by summing over the items related to feeling understood, a feelings of misunderstanding scale is computed by summing over the items associated with feeling misunderstood, and distractor items are discarded. The FUMS is then computed by subtracting the feelings of misunderstanding scale from the feelings of understanding scale creating a possible range of scores from −32 to +32. Evidence exists for both the validity and reliability of the instrument (Cahn, 1990). Data gathered for this study also suggests that the instrument is reliable, α = .81, when α is computed across all understanding and (reverse scored) misunderstanding items. Table 3 reports the means and standard deviations for all variables in the analysis.

TABLE 3 Means and Standard Deviations for Dependent Variables Across Conditions

	Listening Response Condition
Dependent Variable	Simple Acknowledgement n = 37	Advice n = 41	Active Listening n = 37	Total n = 115
FUMS	11.79a (7.47)	14.48b (9.07)	17.27c (7.14)	14.90 (7.14)
Social Attraction	3.65a (.39)	3.84b (.55)	4.02b (.36)	3.87 (.44)
Com-Sat	3.27a (.44)	3.68b (.53)	3.81b (.42)	3.59 (.47)
Note. Means in rows that do not share subscripts differ at the p < .05 level with Scheffé correction. Standard deviations are in italics. FUMS = Feelings of Understanding/Misunderstanding Scale; Com-Sat = Communication Satisfaction.

Communication Satisfaction

We measured communication satisfaction using Hecht's (1978) Interpersonal Communication Satisfaction Inventory (Com-Sat). The Com-Sat is a self-report measure indexing a person's perception that his/her experience during an interaction provided positive reinforcement and fulfilled positive expectations (Hecht, 1978). The Com-Sat consists of 19 items using a 7-point scale (7 = strong agreement; 1 = strong disagreement). Sample items include, “I was very satisfied with the conversation,” “The other person let me know that I was communicating effectively,” and “Nothing was accomplished” (reverse coded). Previous research reports coefficient alphas ranging from .73 to .93 (Rubin, Palmgreen, & Sypher, 1994). Evidence exists for both the reliability (Hecht, 1978) and validity of the instrument (Newton & Burgoon, 1990; Spitzberg, 1991). Data gathered for this project also indicate that the measure meets conventional levels of reliability, α = .87.

Social Attraction

Social attraction was measured using McCroskey, McCroskey, and Richmond's (2006) social attractiveness scale. McCroskey et al.'s instrument measures a person's attractiveness as an interaction partner. Social attraction was operationalized by participants' rating the confederate's social attractiveness by marking their agreement with each of 12 statements (e.g., “He/she is easy to get along with,” “He/She is not very friendly,” reverse coded) on a 7-point scale (1 = strongly disagree; 7 = strongly agree). Participants' ratings of confederates across the 12 items were averaged to compute the social attractiveness scores. McCroskey et al. identify several studies pointing to the reliability and validity of the instrument and our data indicate an acceptable level of reliability for the present study, α = .85.

RESULTS

Manipulation Check

A manipulation check was performed to determine whether the confederates' responses conformed to experimental conditions. Watching video of the dyads' interactions, three assistants independently rated the degree to which confederates paraphrased participant's responses, offered advice, and engaged in back-channel responses (5 = strongly agree; 1 = strongly disagree). Raters received definitions of each behavior and trained together using video from a different study on the same topic. Due to inaudibility in some videos, only 107 dyads were observed. Inter-rater reliability was assessed using alpha coefficients and was acceptable for all of the variables, paraphrase α = .77, offering advice α = .95, and back-channel responses α = .79. Ratings were averaged across the raters to compute each variable. Table 4 reports the means and standard deviations across conditions.

TABLE 4 Means and Standard Deviations for Confederate Behaviors in the Manipulation Check

	Listening Response Condition
Confederate Behavior	Simple Acknowledgement n = 32	Advice n = 39	Active Listening n = 36	Total n = 107
Paraphrase	1.27a (.32)	1.93b (.61)	3.57c (.69)	2.31 (1.01)
Advice	1.16a (.33)	3.38b (1.18)	1.27a (.91)	1.96 (1.40)
Back Channel	3.95a (.85)	3.37b (.78)	3.68ab (.86)	3.69 (.85)
Note. Means in rows that do not share subscripts differ at the p < .05 level with Scheffé correction. Standard deviations are in parentheses.

A one-way multivariate analysis of variance was computed with listening response condition as the independent variable and mean ratings of confederates' paraphrasing, advice giving, and back channeling as the dependent variables. The omnibus test indicates a significant multivariate effect for condition, Pillai's Trace = 1.25, F(6, 206) = 57.62, p < .001, η² = .63. Univariate tests indicate confederates significantly differed in their enactment of paraphrasing, F(2, 104) = 96.14, p < .001, η² = .65, advice giving, F(2, 104) = 77.56, p < .001, η² = .60, and to a lesser extent, back channeling, F(2, 104) = 4.19, p = .02, η² = .07.

Post hoc multiple comparisons, computed using the Scheffé correction and an inspection of the means, indicate that confederates in the active listening condition produced more paraphrased messages than confederates in the other two conditions and confederates in the advice condition engaged in more advice giving than confederates in the other two conditions. Post hoc tests indicate that confederates in the advice condition produced somewhat fewer back channeling behaviors than confederates in the simple acknowledgement condition but not fewer than confederates in the active listening condition. Because differences among groups in back channel cues was not consistent with any of the results of hypothesis testing, we ruled out differences in back channel cues as an alternate hypothesis to our findings.

Hypothesis Testing

Hypothesis testing commenced with three 3 (Response Condition) X 10 (Confederates) random factor ANOVAs. Confederates were treated as a random, rather than fixed, factor because they are random instances taken from the population of possible confederates (see Jackson & Brashers, 1994). The first hypothesis predicted that active listening would produce higher perceived understanding than advice or minimum responses. A main effect emerged for response condition, F(2, 20.68) = 3.79, p = .04, η² = .09. Neither the main effect for confederate, F(9, 18.87) = .98, p = .48, η² = .07 nor the Response Condition X Confederate interaction, F(18, 85) = 1.01, p = .45, η² = .10 were significant. Planned comparisons using Dunnett's tests, with active listening as the comparison group, indicate that the active listening condition produced significantly more perceived understanding than the simple acknowledgement condition and the advice condition (See Table 3). ²

The second hypothesis predicted participants would perceive higher conversational satisfaction with partners who engaged in active listening than with partners who engaged in advice giving or simple acknowledgements. Hypothesis testing reveals a significant main effect for response condition, F(2, 19.64) = 5.81, p = .01, η² = .12, but no main effect for confederates, F(9, 18.52) = .97, p = .50, η² = .08 and no Response Condition X Confederate interaction effect, F(18, 85) = 1.65, p = .07, η² = .18. Planned comparisons using Dunnett's tests indicate that active listening produced more conversational satisfaction than simple acknowledgements but not more than advice giving.

The final hypothesis predicted that participants would perceive their partners to be more socially attractive when their partners engaged in active listening than when they gave advice or produced simple acknowledgements. Data analysis revealed a main effect for response condition, F(2, 21.65) = 6.11, p = .008, η² = .09, but no effects for either confederate F(9, 19.01) = 1.14, p = .39, η² = .06, or the Response Condition X Confederate interaction, F(18, 85) = .69, p = .82., η² = .12. Follow up Dunnett's contrasts among the individual conditions revealed a significant difference between active listening and simple acknowledgements but not between active listening and advice.

DISCUSSION

Our experimental study compares active listening to giving advice and simple acknowledgements in producing three interaction rewards. Consistent with the first hypothesis, confederates' active listening resulted in participants' feeling more understood when compared with confederates' giving advice or back channel cue acknowledgment. This result has both theoretical and practical appeal. Theoretically (e.g., Cissna & Sieburg, 1981; Rogers, 1951), message paraphrasing and encouraging speaker elaboration should communicate understanding and interest. Active listening is most often thought of as a social support or counseling skill because it communicates that the listener understands and cares about the speaker's thoughts and feelings. Our findings suggest that active listening appears to accomplish this goal better than either giving unsolicited advice or offering simple verbal and nonverbal acknowledgements. This study corroborates the theoretical assumption that active listening confirms partners by communicating understanding and acceptance of the speaker's message.

Hypotheses 2 and 3 both received partial support. Although active listening resulted in greater conversational satisfaction and social attraction than simple acknowledgements, active listening and advice giving produced statistically equivalent levels. These findings bear some similarity to other studies comparing active listening and advice (e.g., Libow & Doty, 1976; Scholl, 2002). Given the similarity in results for hypotheses 2 and 3, taken together, research is beginning to suggest that people respond to listeners' overall level of responsiveness regardless of the form it takes. Thus, satisfying interactions and the social attractiveness of partners are associated with conversational expectations (Afifi & Burgoon, 2000; Hecht, 1984), which involve at least a moderate degree of involvement from strangers in initial interactions (Burgoon, Newton, Walther, & Baesler, 1989). Active listening and advice giving result in more positive evaluations of conversations and partners than simple acknowledgments because people expect a greater degree of involvement in a conversation than simple backchannel cues. Additionally, paraphrasing the speaker's message and offering suggestions for activities or solutions to common problems might communicate equivalent levels of involvement and interest resulting in equivalent levels of social attraction and conversational satisfaction.

The failure to find a difference between advice and active listening in conversational satisfaction and social attraction might also lie in the relatively humdrum nature of the conversational topics used to elicit conversations. Identity confirmation during talk about weekend activities and disappointments with one's college experience may not have the same salience for communicators' judgments about a partner's social attractiveness or conversational satisfaction as sensitive social support, personal self-disclosures, or conflict situations. Future research should include topics that increase the salience of identity validation to determine whether active listening produces more positive judgments of partners and more pleasurable conversations in less surface-level conversational activities.

Although not the primary focus of our study, our results also bear on the literature on advice provision and receipt (e.g., MacGeorge, Feng, & Thompson, 2008; Goldsmith & Fitch, 1997) Our study reveals that participants perceive strangers who offer unsolicited advice (two qualities associated with “bad advice” in MacGeorge et al., 2008) as socially attractive as partners who engage in active listening. Our participants also rate their interactions with advice givers and active listeners to be equally satisfying. Our study, in combination with others, implies giving advice can play a role in an effective listening response strategy.

First, unsolicited advice might not necessarily translate into negative judgments about the advice giver's value as an interaction partner or imply that the listener puts his/her needs above the speaker's (e.g., Rogers, 1951; Stanley et al., 1997). Our results comport with research by Bodie et al. (2012) indicating many people perceive advice as a sign of listener attention and interest. Perhaps unsolicited advice can result in an enjoyable interaction as long as the person refrains from insisting that the partner take the advice or as long as the advice appears as a natural response given the sequence of moves in the conversation.

In addition, our conversation starters asked participants to identity some activity for the upcoming weekend and then asked participants to discuss some problem they have experienced while attending the university. This procedure might have made advice a natural contribution to the discussion because, although not explicitly solicited, offering advice following the identification of a problem or a course of action can appear to the advice recipient like a relevant response and carries a low level of face threat compared to other sequences of advice provision (Goldsmith, 2000). The example of advice in Table 2 depicts such a sequence. In this example, the confederate's advice to go deep-sea fishing requires attention to the speaker's intention to go fishing the following weekend and consists of a relevant contribution to the general topic of conversation. The response further confirms the speaker's choice of activities by nominating another activity of the same type. Although a great deal of research examines the influence of message qualities and contextual cues on the reception of advice, few, if any, studies examine perceptions of an advice giver following an actual conversation. Our research suggests that, at least under the conditions in our study, perceptions of advice givers can be as positive as those of active listeners.

Our research has potential implications for research involving studies assessing the effectiveness of active listening dyads with long relationship histories, such as marital intervention research (e.g., Cole & Cole, 1999; Schilling, Baucom, Burnett, Allen, & Ragland, 2003). Recognizing a romantic partner's attempts to use a newly learned relationship skill might elicit positive feelings toward the partner no matter the particular skill or how well the partner performs it. Relationship dynamics influence how people perceive and value their partners' behavior (e.g., Waldinger & Schulz, 2006), so perceptions of newly acquired communication skills might be colored by romantic partners' or family members' relationship history and satisfaction. Additionally, these studies rarely employ control groups who learn alternate listening response strategies. Our study suggests family/marital skills assessments should include potentially valuable alternative response strategies to form a basis of comparison.

Finally, although our research is still in the early stages, it has implications for the literature on communication design. Over the last decade, several communication scholars have called for recognizing communication research as a design enterprise (e.g., Aakhus & Jackson, 2005). Active listening is an excellent example. It was designed first to increase trust between patients and their psychotherapists and later adapted to regulate negativity in marital conflict (e.g., Gordon, 1975; Markman, Stanley, & Blumberg, 2001). Further, Gottman, Coan, Carrere, and Swanson (1998) report that few married couples naturally engage in active listening during conflict. Markman et al. (2001) confirm that the speaker-listener technique (i.e., active listening) feels unnatural to couples because it is designed to help couples avoid natural conversation pitfalls.

Considering communication as a design enterprise enables communication scholars to help identify design issues and to make recommendations for increasing communication quality through redesign. The present study functions, in part, as an assessment of active listening's functionality in initial interactions. First dates, meeting co-workers for the first time, face-to-face initial sales calls, job interviews that include lunch or dinner conversations, talking at an airport bar between flights, sitting next to a stranger on a bus/plane/train trip, meeting a co-worker's spouse at a Christmas party, and meeting a local while out of town are among several types of situations in which people meet and talk for the first time. This study suggests little harm, and some potential benefit, to using active listening paraphrasing in such interactions, although additional research is needed to confirm this. Like active listening, several other designed communication “skills” commonly taught in interpersonal communication classes, human resource trainings, and marital enrichment courses receive little, if any, empirical support in the literature (e.g., research does not support the effectiveness of “I” messages; Bippus & Young, 2005). One significant contribution communication science can deliver to the common good involves both designing and testing designed communication tools in a variety of situations.

LIMITATIONS AND DIRECTIONS FOR FUTURE RESEARCH

This study has a few limitations. First, this study did not vary nonverbal immediacy. An interesting direction could examine how much influence the nonverbal behavior of an active listener matters in perceptions of the communicator and communication situation. It is likely that an active listening response delivered simultaneously with nonverbal signs of disinterest or detachment may result in perceptions that the listener is disingenuous, thereby reducing any benefit. Another potential limitation involves the contrived nature of interacting with a stranger in a lab, which may have limited generalizability. In our interactions, participants had no clear direction other than “getting to know” the other person. Participants probably did not expect to interact with the person again. Active listening, compared to advice, might make more difference in judgments about a person's attractiveness as an interaction partner if participants expect future interactions (e.g., Sunnafrank & Ramirez, 2004). Active listening may be less significant in situations driven by instrumental goals rather than relational or self-presentation goals.

Another limitation involves the relative homogeneity of the participants' ages, ethnic background, and cultural context. Our sample consisted of mostly young, White, female college students. People with more life experience, from different ethnic backgrounds, and without shared educational experiences may react differently to active listening. Research suggests that behaviors associated with “good” and “bad” listeners differ between cultures (e.g., Imhof, 2003). Future research should include these elements of sample variability. The homogeneity of our study limits our ability to clearly generalize to more diverse populations. However, we are somewhat confident in generalizing to different populations of active listening providers since our confederates were a fairly diverse group of students and our results were not dependent on ethnic or sex differences among them.

CONCLUSION

This project is among the first to use experimental methods to compare active listening responses to other types of listening responses during initial interactions between peers. Our results suggest that active listening responses in initial interactions increase the recipient's perception of feeling understood compared with other response strategies. We also found that advice giving and active listening did not differ in the degree to which recipients liked their interaction partner or enjoyed the conversation, although both advice and active listening out-performed simple acknowledgements. It appears that active listening, at least in terms of these outcomes, is a low-risk response strategy with some potential dyadic rewards compared to responding with advice and minimal feedback. Future research should continue to look at factors related to the context and the relationship between speaker and listener in tests of the effectiveness of active listening.

ACKNOWLEDGMENTS

An earlier version of this article was presented at the annual meeting of the National Communication Association, November 2008, Chicago, Illinois. The data reported in the article is exclusive to this study.

Notes

¹The focal person in this analysis is the confederate, not the participant—each confederate, therefore, received at least three ratings of their social attractiveness, the degree to which their behavior produced feelings of understanding, and the degree to which interaction with them was satisfying per condition. See Jackson and Brashers (1994) for a discussion of research design using random factors.

²Sex of participant was included in the original analyses. Neither a main effect for sex nor a sex by condition interaction were found in the data for this study and were therefore dropped from further analyses in order to increase cell sizes and power.