JOURNAL OF CONSUMER PSYCHOLOGY, 15(2), 141–148 Copyright © 2005, Lawrence Erlbaum Associates, Inc.

CONSUMER TRUST AND ADVICE ACCEPTANCE WHITE

Consumer Trust and Advice Acceptance: The Moderating Roles of Benevolence, Expertise, and Negative Emotions Tiffany Barnett White University of Illinois at Urbana—Champaign

I explored advice acceptance for high-stakes decisions (i.e., those with subjectively important and risky outcomes), focusing on the relative influence of two components of consumer trust— benevolence and expertise—as well as perceived emotional decision difficulty. Participants solicited advice from experts when their decisions were low in perceived emotional difficulty but favored the advice of predominantly benevolent providers when making highly emotionally difficult decisions. Although consumers who faced emotionally difficult decisions were willing to trade off expertise for benevolence, they did not perceive this non-normative trade-off to influence decision quality. Instead, the results support a “stress buffering” effect whereby consumers were more confident in the accuracy of predominantly benevolent providers’ advice.

Imagine that you must choose whether or not to have a new and experimental optical procedure that could substantially and permanently enhance your vision. Also, imagine that although your lifestyle stands to be greatly improved if the procedure is successful, there is a chance that it could fail and cause permanent damage to your eyes. How likely is it that you would opt to have the procedure? What factors would you consider when making your decision? Under what conditions would you take the advice of a benevolent provider rather than an expert? This example describes a decision environment characterized by high stakes and substantial outcome uncertainty. This type of situation is increasingly common. Consumers have to make more complex decisions than ever before with more alternatives and less certainty about product and service performance. In addition to medical treatments, consumers must make decisions about insurance, investments, and legal issues. When faced with high-stakes (i.e., risky and important) decisions, consumers often engage in especially effortful and systematic processing (e.g., planful problem solving and other forms of problem-focused coping; Luce, Bettman, & Payne, 1997). Furthermore, when such consumers solicit adRequests for reprints should be sent to Tiffany B. White, University of Illinois at Urbana—Champaign, Department of Business Administration, 350 Wohlers Hall, 1206 South Sixth Street, Champaign, IL 61820. E-mail: [email protected]

vice (e.g., from physicians, financial advisors, etc.), they expect these agents to make recommendations that are based on careful and systematic deliberation as well (Kahn & Baron, 1995). In this research, I explored advice acceptance (i.e., from product and/or service providers) in high-stakes decision making. In doing so, I focused on the relative influence of two components of consumer trust—benevolence and expertise—as well as perceived emotional decision difficulty on advice acceptance. The very nature of high-stakes decisions bespeaks the importance of making the most accurate decision possible (e.g., taking the advice of only the most expert, trusted professional). This research suggests that as high-stakes decisions increase in perceived emotional difficulty, consumers rely less on rational decision criteria and more on emotional criteria when considering whether to accept advice. Individuals often rely on the advice of trusted sources under conditions of high perceived risk (Perry & Hamm, 1969). However, trust is most commonly studied as a global, multidimensional construct—encompassing perceptions of the provider in terms of competence (e.g., perceived skill and expertise), benevolence, and integrity. Far less consideration is given to the relative impact of these perceived characteristics on behavior (for exceptions, see McAllister, 1995; Sirdeshmukh, Singh, & Sabol, 2002). Furthermore, trust researchers have largely ignored the influence of negative emotions and the extent to which the goal of coping with these emotions might moderate the rel-

142

WHITE

ative impact of expertise versus benevolence-based trust perceptions on behavior. Drawing from research on coping (e.g., Lazarus & Folkman, 1984; Luce, 1998) and social support (e.g., Cohen & Wills, 1985; Helgeson, 2003), I discuss conditions under which consumers rely on predominantly benevolent versus predominantly expert provider recommendations. This research indicated that consumers rely more heavily on expert providers’ recommendations when high-stakes decisions are low in emotional difficulty. When decisions are emotionally difficult, however, consumers were more likely to take the advice of predominantly benevolent providers rather than experts. Although these latter decision-making criteria are arguably less normative (Bettman, Luce, & Payne, 1998), I found that consumers were actually more confident in the accuracy of their advice givers’ recommendations when these criteria were used. In this article, I first describe the notion of high-stakes decision making, define trust, and discuss the normative influence of trust in high-stakes decision making environments. In doing so, I conceptually disentangle two dominant dimensions of trust and examine the extent to which the relative influence of these dimensions on advice acceptance may be moderated by negative decision-related emotion. Next, I discuss the nature of coping and social support and the influence of both on consumer advice acceptance. I then describe an experiment that evaluated the implications of this discussion.

CONSUMER TRUST IN HIGH-STAKES DECISION MAKING For purposes of this article, a decision is considered to involve high stakes if the outcome of the decision is subjectively important, but the likelihood of its occurrence is unclear (Kahn & Baron, 1995; Kunreuther et al., 2002). What factors might influence advice acceptance when making such decisions? When considering whether to accept the recommendation to adopt a risky service, consumers must assess their confidence in the accuracy of their provider’s recommendation. This assessment entails at least two sources of uncertainty. First, consumers may consider the probability that the service provider will fail to deliver the service due to incompetence. In addition, they may consider the possibility of willful malevolence on the part of the provider. For example, a consumer may factor in the probability that a stockbroker might recommend a stock she or he knows to be overpriced when deciding whether to purchase the stock. Consumer trust entails both the belief that the product or service provider has subject-specific knowledge (expertise) and the belief that she or he is concerned about the welfare and best interests of the consumer (benevolence; Mayer, Davis,

& Schoorman, 1995). Both factors could influence consumers’ reliance on provider recommendations.1 This multidimensional conception of trust is useful and robust (Lewis & Weigart, 1985; Mayer et al., 1995; Moorman, Deshpande, & Zaltman, 1993; Rempel, Holmes, & Zanna, 1985). However, it does not explicitly address the relative impact of these factors. Thus, although consumers often have stronger intentions to purchase from a provider they trust (Sirdeshmukh et al., 2002; Zeithaml, Parasuraman, & Berry, 1990), this could be true because the provider is seen as highly expert, as benevolent, or both. Therefore, I distinguish between two forms of trust. Expert-based trust is predominantly grounded in beliefs about the provider’s skills, competencies, and expertise. In contrast, benevolence-based trust is predominantly grounded in perceived benevolence, the belief that the trustee wishes the trustor well aside from an egocentric profit motive (Mayer et al., 1995; for a similar distinction between cognition-based and affect-based trust, see McAllister, 1995). Although acceptable levels of both perceived expertise and perceived benevolence are necessary for trust (Mayer et al., 1995), trustees may dominate on one dimension or the other. One implication of such a distinction is that consumers’ willingness to accept risky advice may depend on the type of trust in question. As developed next, the influence of benevolence versus expert-based trust on advice acceptance may depend on the extent to which high-stakes decisions are perceived as emotionally difficult.

PERCEIVED EMOTIONAL DIFFICULTY IN DECISION MAKING Drawing from Lazarus’s (1991) theory of emotion elicitation, Luce, Payne, and Bettman (1999) examined the concept of emotionally difficult trade-offs in a consumer decision-making context. By emotional difficulty, Luce et al. (1999) referred to the degree of subjective threat perceived by a decision maker when explicitly trading off personally important attributes (e.g., price and safety). However, emotional difficulty can be associated not only with deciding among choice attributes within a given alternative but also with deciding whether to adopt a single (risky) alternative. Luce et al. (1997, 1999; Drolet & Luce, 2004) and others (Hogarth, 1987) have documented the negative affect that can accompany decisions in which risky and emotionally difficult trade-offs must be made. That is, consumers often make decisions that reflect a trade-off between effort and accuracy (Bettman et al., 1998). However, highly emotionally 1It is important to acknowledge that the components of trust under investigation in this research—expertise and benevolence—represent exemplars of a broader category of competence and character-based considerations that are not explicitly addressed here (e.g., perceived reliability and integrity).

CONSUMER TRUST AND ADVICE ACCEPTANCE

difficult decisions appear to be distinct in that they also trigger the desire to cope with (i.e., minimize) negative decision-related emotions (Lazarus & Folkman, 1987; Luce et al., 1999). Thus, individuals may be motivated to make the best decision. To the extent emotional difficulty is perceived, however, they must also manage the negative emotions that are associated with the decision. Attempts to manage these goals simultaneously can often result in avoidant and/or non-normative decision behaviors (Bettman et al., 1998). For example, Drolet and Luce (2004) noted that in such decisions, consumers often “sacrifice decision accuracy in order to minimize negative emotions, even (and perhaps especially) in consequential decisions” (p. 63; also see Kahn & Baron, 1995). The Role of Emotional Support Individuals who anticipate stressful or negative outcomes often seek the social support of others in an attempt to cope with such stress (e.g., through advice seeking; Cohen, Gottlieb, & Underwood, 2000; Helgeson, 2003; MacGeorge, Feng, Butler & Budarz, 2004). Researchers have conceptualized social support as a process through which tangible (e.g., financial), emotional (e.g., caring, understanding), and/or informational (e.g., advice) resources are provided to or exchanged with others in an attempt to facilitate one or more adaptational goals (Cohen et al., 2000). To the extent that the goal of minimizing negative decision-related emotions may dominate the goal of maximizing decision accuracy when decisions are emotionally difficult (Drolet & Luce, 2004), the motivation to seek the advice and support of providers who can help to minimize these negative emotions may be preferred. Thus, although consumers may value perceived expertise, they may nevertheless favor the advice of providers they perceive to be predominantly benevolent over those they consider to be knowledgeable but relatively less able or willing to provide emotional support. In contrast, the minimization of negative emotions should not be a concern for those who experience low levels of emotional decision difficulty. Consumers making such decisions are therefore “free” to concentrate on the goal of maximizing decision accuracy that is logically associated with high stakes decision making. As such, they are expected to seek informational support from those they perceive as most qualified to provide accurate advice, namely, predominantly expert providers.

143

predominantly benevolent providers where advice acceptance is concerned, do such trade-offs reflect a conscious choice? That is, do consumers actually believe they are sacrificing accuracy in such conditions? Theory and research on social support has suggested that the positive effect of emotional support on support seekers’ well being is driven by the “psychological buffer” it provides against real or imagined stressful outcomes (Cohen & Wills, 1985; Sarason, Sarason, Brock, & Pierce, 1996). Rather than altering attitudes toward potentially negative outcomes, this buffering occurs because emotional support seekers are optimistic that they can avoid potentially negative outcomes as a result of this support (Helgeson, 2003). To the extent that such buffering occurs in the emotionally difficult decision-making domain, consumers receiving advice from emotionally supportive (i.e., predominantly benevolent) providers should have greater confidence that such advice acceptance will lead to positive outcomes. Thus, although consumers may trade off perceived expertise for perceived benevolence, such a trade-off does not reflect a motivated decision to accept less accurate advice. Rather, consumers may actually perceive the advice of predominantly benevolent providers to be more accurate. Finally, the stress-buffering hypothesis also suggests that the influence of emotional support is directly related to the perception of stress (e.g., that resulting from negative decision-related emotions). When support seekers experience little or no stress, the effect of social support on anticipated and actual well-being is greatly diminished (Helgeson, 2003). Thus, the stress-buffering explanation implies that consumers who encounter an emotionally difficult decision will have greater confidence in the advice of predominantly benevolent providers, whereas those who encounter a less emotionally difficult decision will have greater confidence in the advice of providers who are high in expertise. Moreover, such confidence should underlie the effect of provider benevolence and expertise on advice acceptance. Because consumers are expected to rely on the advice of predominantly expert providers only under conditions of low (but not high) perceived emotional difficulty (i.e., conditions under which stress buffering does not occur), confidence in the advice of high (vs. low) perceived expertise was not expected to vary between levels of emotional decision difficulty.

METHOD

Stress Buffering and Decision Confidence

Design and Participants

Of additional theoretical interest in this research is the psychological mechanism underlying the effect of perceived benevolence on advice acceptance. When consumers trade off the accuracy-enhancing skills and abilities of predominant experts for the emotion-enhancing caring and empathy of

The experiment consisted of 16 cells of a 3 (benevolence: high/low/no) × 3 (expertise: high/low/no) × 2 (emotional difficulty: high/low) between-subject design. That is, because the “no benevolence, no expertise” conditions (for high and low levels of emotional difficulty) would not include recom-

144

WHITE

mendations, these two cells were omitted. Three hundred seven undergraduate students enrolled in an introductory marketing course at the University of Illinois participated for extra course credit.

and it continues to do poorly, he will have to leave campus by the end of the academic year [he won’t have enough money to pay for the local day camp]. Tom has decided that he should seek professional advice.

Procedure and Stimuli Participants were informed that the study aimed to assess students’ decision-making styles. After reading a scenario describing a decision faced by Tom, a university sophomore, participants were asked to respond to a series of questions regarding Tom’s decision and asked to put themselves “in Tom’s place” as they responded to the questions that followed. In the scenario, Tom was confronted with a decision of whether to keep a risky investment fund in which his money was currently invested. The description of the circumstances surrounding Tom’s decision was varied over two levels of emotional difficulty. The scenarios were as follows, with the alternative phrases used under high and low emotional difficulty conditions indicated in italics and with verbiage for the low emotional difficulty scenarios in brackets: Tom is a University of Illinois sophomore. Two years ago, he inherited money from his parents [aunt] whose wishes were to see Tom attend and complete college, which he decided to use to attend an expensive summer band camp for young musicians]. Without this money, Tom would not be able to pay for his tuition, room and board during his time at the University of Illinois [camp]. Because the money was not enough to cover all of Tom’s college expenses, he decided to invest it in an aggressive stock fund, Fund K, which he learned about while browsing an Internet website. Tom realized that investing in Fund K involved considerable risk—it could perform well enough for him to meet his goal of paying for college through graduation [paying for the camp]. However, this alternative also carries a chance that Tom’s savings might actually decrease—he could lose some or even all of his money. Though Tom’s investment returns started out strong, his money has recently declined in value since he invested it two years ago. Because time is of the essence, Tom is deliberating whether he should keep his money in Fund K or take it out of market completely. Taking his money out of the market now would mean that he could not afford to stay here at the University of Illinois [go to the band camp]. Instead, he would have to continue college [his music lessons] at a much less expensive university or community college [day camp in his town]. If he keeps the money in Fund K and it does well, he will have more than enough money to finish college here at Illinois [pay for his entire camp experience]. However, if he leaves the money in Fund K

Manipulation of Perceived Expertise Respondents then read scenario descriptions about a financial advisor who varied on perceived benevolence and/or expertise and from whom Tom sought advice. Under high-expertise conditions, the scenario read as follows: Because Tom purchased shares of Fund K over the Internet (thereby avoiding the need to pay hefty commissions) he does not have access to a financial advisor at the company through which he purchased the shares. Instead, Tom visits Pat, Chief Advisor for mutual funds at an internationally successful investment firm in Tom’s area. Pat has managed some of the firm’s largest and most successful funds over the years. Under low-expertise conditions, the last paragraph was replaced with the following: Instead, Tom visits Pat, who works at a small investment firm in Tom’s area. Among other functions, Pat manages mutual funds for the investment firm. Under no-expertise conditions, these paragraphs were omitted. Manipulation of Perceived Benevolence Under high-benevolence conditions, the scenario continued as follows: Over the course of his interactions with Pat, Tom has found him to be especially caring and honest. Pat always provides fair and objective information. Tom is confident that Pat does not seek to protect his own interests, but is instead always concerned about Tom’s welfare and best interests. In Tom’s opinion, Pat really wants to see him finish school at the University of Illinois. Pat advises Tom to keep his money in Fund K. Under low-benevolence conditions, however, the preceding paragraph was replaced by the following: As he would during an initial meeting with any financial advisor, Tom wonders about how caring and honest Pat is. Tom seeks fair and objective information, but feels he doesn’t know Pat well enough to decide whether he seeks to protect his own interests, or is instead concerned about Tom’s welfare and best inter-

CONSUMER TRUST AND ADVICE ACCEPTANCE

145

160) = 3.03, p < .09. Similarly, respondents reported significantly higher levels of perceived expertise in high-expertise (M = 5.47) than in low-expertise conditions (M = 3.33), F(1, 160) = 165.12, p < .0001. These perceptions were not significantly influenced by the benevolence manipulation, F(1, 160) = 2.95, p < .09. Further, when benevolence and expertise judgments were combined with two more direct measures (“I would trust this provider,” and “Based on what you read, how much would you trust Pat’s opinion”) to create a global measure of trust (α = .88), the results yielded significant benevolence and expertise main effects, Fs(1, 160) = 105.48 and 37.15, respectively, ps < .0001. Benevolence and expertise did not interact to influence the global trust measure, F(1, 160) = 0.08, p < .80. Finally, participants perceived the decision to be significantly more emotionally difficult in the high emotional difficulty conditions than in the low emotional difficulty conditions (6.18 vs. 4.33, respectively), F(1, 160) = 138.36, p < .0001. No other effects were significant in this analysis.

ests. Tom wonders whether Pat really wants to see him finish school at The University of Illinois. Pat advises Tom to keep his money in Fund K. Dependent Variables Respondents then indicated how likely they believed Tom would follow the provider’s recommendation along a 7-point scale ranging from 1 (very unlikely) to 7 (very likely). In addition, they reported their confidence that the provider’s advice was accurate and the likelihood that it would lead to a positive outcome along scales ranging from 1 (very unlikely) to 7 (very likely; r = .93). Finally, they provided responses to the relevant manipulations checks. Perceived benevolence was measured using a five-item scale composed of items such as “How much would you say Pat is concerned about Tom’s welfare?” and “Pat is likely to be concerned about Tom’s best interests” (α = .87). Perceived expertise was assessed using a three-item scale composed of items such as “Pat is likely to be a highly skilled investor,” and “Based on what you’ve read, do you consider Pat an expert?” (α = .82; Mayer et al., 1995). Perceived emotional difficulty was measured using a five-item scale that included items such as “Tom would feel the decision is an emotionally difficult one to make,” and “Tom would feel the decision is stressful” (α = .88). Responses to all scale items were reported along scales ranging from 1 to 7, the endpoints of which depended on the particular item involved.

Advice Acceptance I predicted that the impact of expertise would be greater when decisions were perceived as low (vs. high) in emotional difficulty, whereas the effect of benevolence should be greater in decisions perceived as high (vs. low) in emotional difficulty. Because I did not use a full factorial design, I tested these predictions in an analysis of variance in which advice acceptance was modeled as a function of emotional difficulty (high vs. low) and trust, a single variable reflecting the eight manipulated combinations of high, low, and absent levels of benevolence and expertise. All main, interactive, and simple effects were tested using the pooled error term from this model. Table 1 summarizes the least squares adjusted means for the effects tested for advice acceptance. Unless otherwise noted, the simple effects reported here and subsequently are significant at p < .05 or better and where indicated, nonsignificant at p > .1 or higher. As expected, the impact of trust on advice acceptance was significantly moderated by perceived emotional decision difficulty, F(7, 291) = 3.57, p < .0001. Before assessing the relative impact of benevolence and expertise on advice acceptance under conditions of high and low perceived emotional

RESULTS Manipulation Checks Because benevolence and expertise judgments were not measured in all conditions (i.e., they were not assessed in the no benevolence and/or no expertise conditions), we assessed the strength of the manipulations via a 2 (benevolence: high/low) × 2 (expertise: high/low) × 2 (emotional difficulty: high/low) analysis of variance. As expected, participants reported significantly higher levels of perceived benevolence in high-benevolence conditions (M = 5.41) than in the low-benevolence conditions (M = 3.79), F(1, 160) = 163.10, p < .0001. The expertise manipulation did not significantly influence perceived benevolence, F(1,

TABLE 1 Least Squares Adjusted Means for Advice Acceptance by Benevolence, Expertise, and Emotional Decision Difficulty High Emotional Difficulty Benevolence High Low None Note.

Low Emotional Difficulty

High Expertise

Low Expertise

No Expertise

High Expertise

Low Expertise

No Expertise

5.96a 4.24b 5.24a

5.60a 4.23b 3.92b

5.41a 3.82b

5.50a 5.47a 5.60a

4.18b 4.01b 4.36b

5.30a 3.30b

Means with different subscripts are significantly different at p < .05.

146

WHITE

difficulty, I first established the effects of these variables in isolation. When respondents were not presented with expertise information, acceptance was greater when the advisor was benevolent (M = 5.36) than when she or he was less so (M = 3.56), F(1, 291) = 14.29, p < .005, and this was true regardless of whether the decision was emotionally difficult (5.41 vs. 3.82) or not (5.30 vs. 3.30). Correspondingly, the advisor’s expertise had a significant effect in the absence of benevolence information (5.42 vs. 4.98 when expertise was high vs. low, respectively), F(1, 291) = 8.27, p < .01, and this was also true regardless of whether the decision was emotionally difficult (5.74 vs. 3.92) or not (5.60 vs. 4.36). These findings highlight the importance of both factors as significant drivers of advice acceptance. When benevolence and expertise information were both available, however, their effects were contingent on emotional difficulty as expected. When the emotional difficulty of the decision was low, the advisor’s expertise had as strong an effect on advice acceptance when benevolence information was presented (5.49 vs. 4.10), F(1, 291) = 9.24; p < .002 as it had in the absence of this information (5.60 vs. 4.36), whereas benevolence itself had no effect (4.84 vs. 4.74 when benevolence was high vs. low, respectively). When the decision was emotionally difficult, however, the effect of the advisor’s benevolence on advice acceptance was as strong when expertise information was available (5.78 vs. 4.24), F(1, 291) = 18.15, p < .005 as when it was not (5.24 vs. 3.92), and the effect of the advisor’s expertise under these conditions was negligible (5.10 vs. 4.92 when expertise was high vs. low, respectively). Thus, although participants gave priority to the advisor’s expertise when the emotional difficulty of the decision was low, they based their acceptance on the advisor’s benevolence alone when the decision was emotionally difficult. Advice Confidence In line with my predictions, the impact of trust on advice confidence was significantly moderated by perceived emotional decision difficulty, F(7, 291) = 4.41; p < .05. Data bearing on this contingency are summarized in Table 2. In the absence of benevolence information, confidence was marginally greater when providers’ expertise was high (M = 4.23) than when it was low (M = 3.54), F(1, 291) = 2.59, p < .10, and this was

true regardless of whether emotional difficulty was high (4.28 vs. 3.58, respectively) or low (4.18 vs. 3.50, respectively). When expertise information was unavailable, however, an effect of benevolence was only evident when decisions were low in difficulty (4.00 vs. 3.30 when benevolence was high vs. low, respectively); it had no influence at all when emotional difficulty was high (3.64 vs. 3.77). This suggests that the presence of information about provider expertise may have been a necessary but not sufficient condition for stress buffering. As expected, when benevolence and expertise information were both available, the effects of benevolence on advice confidence were contingent on emotional difficulty, whereas the effects of expertise were not. When the emotional difficulty of the decision was low, the advisor’s expertise had as strong an effect on advice confidence when benevolence information was presented (4.75 vs. 3.74), F(1, 291) = 5.18, p < .05 as it had in the absence of this information (4.23 vs. 3.54). Further, this effect of expertise in the presence of benevolence information persisted across high (4.25 vs. 3.88) and low (4.75 vs. 3.74) levels of emotional decision difficulty. In contrast, the effect of benevolence on advice confidence was evident only when the decision was emotionally difficult (4.56 vs. 3.58 when benevolence was high vs. low, respectively), F(1, 291) = 7.84, p < .01. When emotional difficulty was low, the impact of benevolence on advice confidence was negligible (4.21 vs. 4.27). I also expected confidence judgments to mediate the influence of perceived benevolence on advice acceptance for decisions of high (but not low) emotional difficulty. I investigated several conditions to establish this pattern of mediation (Baron & Kenney, 1986; James & Brett, 1984). First, greater advice confidence was associated with higher levels of advice acceptance for decisions of high (r = .28, p < .0001) emotional difficulty (this correlation was also significant for low difficulty decisions; r = .45, p < .0001). Next, having established that the effect of benevolence on advice acceptance and advice confidence was contingent on emotional difficulty, I added respondents’ confidence ratings as a covariate to an analysis of covariance modeling the effects of emotional decision difficulty and trust on advice acceptance. Consistent with the requirements for moderated mediation (James & Brett, 1984), introducing confidence as a covariate under conditions of high emotional difficulty reduced the ef-

TABLE 2 Least Squares Adjusted Means for Advice Confidence by Benevolence, Expertise, and Emotional Decision Difficulty High Emotional Difficulty Benevolence High Low None

Low Emotional Difficulty

High Expertise

Low Expertise

No Expertise

4.46a 4.04a 4.28a

4.65a 3.11b,c 3.58b

3.64b 3.77b,d

High Expertise 4.75a 4.74a 4.18a,d

Low Expertise

No Expertise

3.67b 3.80b 3.50b,c

4.00a,d 3.30b

Note. Means with different subscripts a and b are significantly different at p < .05. Means with different subscripts c and d are significantly different at p < .05, one-tailed.

CONSUMER TRUST AND ADVICE ACCEPTANCE

fect of benevolence by 27%. A Sobel test of the mediation further confirmed its significance (Sobel test statistic = 2.72, p < .001, and Goodman I test statistic = 2.68, p < .001; Preacher & Leonardelli, 2001). On the other hand, confidence could be logically ruled out as a mediator of benevolence on advice acceptance under conditions of low decision difficulty because benevolence significantly affected neither advice acceptance nor confidence under these conditions.

DISCUSSION The results of this research show that consumers favor the advice of experts when high-stakes decisions are low in perceived emotional difficulty but favor the advice of predominantly benevolent providers for highly emotionally difficult decisions. This latter finding is consistent with the growing body of evidence on emotionally difficult decision making that highlights consumers’ tendencies to cope with negative decision-related emotions by engaging in non-normative decision strategies (Bettman et al., 1998; Drolet & Luce, 2004; Luce et al., 1999). Specifically, the results suggests that as decisions become perceived as high in emotional difficulty, consumers assign greater weight to the goal of minimizing negative decision-related emotions than on decision accuracy, which results in greater reliance on characteristics of providers that are more aligned with the former than the latter decision goal. As noted earlier, the literature on emotionally difficult decision making has mainly focused on consumers’ individual choice strategies and shown that consumers often trade off accuracy maximization for negative emotion minimization. My findings supplement and extend this research, drawing from the social support literature to explore one reason why such a phenomenon might occur in the domain of advice acceptance. Although consumers seem consciously willing to trade off advice givers’ perceived level of expertise for benevolence (i.e., they assign little or no weight to the former in the presence of the latter), they do not perceive the trade-off as a potential sacrifice in accuracy. Instead, they seem to believe that following the advice of predominantly benevolent (vs. predominantly expert) providers will lead to better—that is, more accurate—outcomes. The literature on social support in interpersonal relationships has provided a theoretical explanation for this result. The availability of emotional support (e.g., from predominantly benevolent providers) serves as a psychological buffer against the imagined outcomes of stressful decisions and seems to result in an overarching feeling of optimism and comfort regarding the stressful event (or decision) at hand. This article also contributes to the literature on consumer trust, which should benefit from understanding more about the relative impact of benevolence and expertise components of trust on behaviors. Much research in the trust literature has grappled with the inconsistent relation between global trust

147

beliefs and behaviors that imply trust (e.g., personal information disclosure, choice, and/or purchase; Schlosser, White, & Lloyd, 2004; Sirdeshmukh et al., 2002; White, 2004). The findings in this research imply that the strength of the relation between trusting beliefs and trusting behaviors such as advice acceptance may be better understood by understanding the extent to which disparate provider characteristics are instrumental in helping consumers achieve their decision-related goals. Indeed, these findings supplement a growing body of evidence in support of this notion. For example, Schlosser et al. (2004) found evidence that consumers’ online purchase intentions were more readily influenced by provider characteristics that they perceived to reflect expertise (e.g., Web site design) than by characteristics that communicated benevolence (e.g., privacy and security statements). In this situation, perceived expertise was a stronger driver of online purchase intentions than was perceived benevolence. In this sense, Schlosser et al.’s results suggest that consumers were more readily influenced by provider information that was aligned with the goal of choosing the most competent provider than with the goal of choosing the one highest in perceived character. Caution should be taken in generalizing the conclusions drawn from this research to situations outside the laboratory. The scenario-based experimental paradigm I employed has inherent limitations, and confirmation of these findings in a more externally valid context would clearly be desirable. I also limited my investigation to a single decision category (i.e., financial decisions). The consideration of a broader cross-section of decisions (e.g., medical, legal, etc.) would also increase confidence in the generalizability of my findings. Such investigations might also seek an understanding of how robust the findings demonstrated here are across other types of emotionally difficult decision contexts. In this research, I investigated advice acceptance and have considered advice about whether to maintain a risky course of action. Whether these results would change as a function of the type of decision (and the corresponding risks in question) merits additional consideration. Finally, it is worthwhile to note that classic conceptualizations of trust are not bidimensional but tridimensional and include perceived benevolence and expertise as well as a third consideration—integrity/honesty (Mayer et al., 1995). For conceptual clarity, I have limited this analysis to benevolence and expertise as affect- and cognition-based considerations. Although it seems reasonable to assume that the integrity component of trustworthiness judgments is conceptually related to more character-based considerations, this is a matter for empirical consideration. Are there aspects or categories of consumer decisions that would result in the disproportionate weighting of integrity perceptions relative to benevolence and expertise? Such a question invites further research on trust mechanisms operating in various consumer decision-making contexts.

148

WHITE

ACKNOWLEDGMENTS This research was funded, in part, by funds from Duke University under an award from the General Electric Fund. The findings, opinions, and recommendations expressed in this article are mine and are not necessarily those of Duke University or the General Electric Fund. I am grateful to Jim Bettman, Julie Edell, and Ashok Lalwani for their insights and comments on earlier drafts of this article. I am also grateful to Linda Tuncay for outstanding research assistance. Finally, a special thanks is due the anonymous Journal of Consumer Psychology reviewers for their guidance and suggestions.

REFERENCES Baron, R. M., & Kenney, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Bettman, J. R., Luce, M. F., & Payne, J. W. (1998, December). Constructive consumer choice processes. Journal of Consumer Research, 25, 187–217. Cohen, S., Gottlieb, B. H., & Underwood, L. G. (2000). Social relationships and health. In S. Cohen, L. G. Underwood, & B. H. Gottlieb (Eds.), Social support measurement and intervention: A guide for health and social scientists (pp. 1–25). New York: Oxford University Press. Cohen, S., & Wills, T. A. (1985). Stress, social support, and the buffering hypothesis. Psychological Bulletin, 98, 310–357. Drolet, A., & Luce, M. F. (2004). The rationalizing effects of cognitive load on emotion-based trade-off avoidance. Journal of Consumer Research, 31, 63–77. Helgeson, V. (2003). Social support and quality of life. Quality of Life Research, 12, 25–31. Hogarth, R. (1987). Judgment and choice. New York: Wiley. James, L. R., & Brett, J. M. (1984). Mediators, moderators and tests for mediation. Journal of Applied Psychology, 69, 307–321. Kahn, B. E., & Baron, J. (1995). An exploratory study of choice rules favored for high-stakes decisions. Journal of Consumer Psychology, 4, 305–328. Kunreuther, H., Meyer, R., Zeckhauser, R., Slovic, P., Schwartz, B., Schade, C., et al. (2002). High stakes decision making: Normative, descriptive and prescriptive considerations. Marketing Letters, 13, 259–269. Lazarus, R. S. (1991). Cognition and motivation in emotion. American Psychologist, 46, 352–367. Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.

Lazarus, R. S., & Folkman, S. (1987). Transactional theory and research on emotions and coping. European Journal of Personality, 1, 141–169. Lewis, J. D., & Weigart, A. (1985). Trust as social reality. Social Forces, 63, 967–985. Luce, M. F. (1998). Choosing to avoid: Coping with negatively emotion-laden consumer decisions. Journal of Consumer Research, 24, 409. Luce, M. F., Bettman, J. R., & Payne, J. W. (1997). Choice processing in emotionally difficult decisions. Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 384–405. Luce, M. F., Payne, J. W., & Bettman, J. R. (1999). Emotional trade-off difficulty and choice. Journal of Marketing Research, 36, 143–159. MacGeorge, E. L., Feng, B., Butler, G. L., & Budarz, S. K. (2004). Understanding advice in supportive interactions: Beyond the facework and message evaluation paradigm. Human Communication Research, 30, 42–70. Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20, 709–734. McAllister, D. (1995). Affect- and cognition-based trust as foundations for interpersonal cooperation in organizations. Academy of Management Journal, 38, 24–59. Moorman, C., Deshpande, R., & Zaltman, G. (1993). Factors affecting trust in market research relationships. Journal of Marketing, 57, 81–102. Perry, M., & Hamm, B. C. (1969). Canonical analysis of relations between socioeconomic risk and personal influence in purchase decisions. Journal of Marketing Research, 6, 351–354. Preacher, K. J., & Leonardelli, G. J. (2001). Calculation for the Sobel test: An interactive calculation tool for mediation tests. Retrieved June 1, 2004, from http://www.unc.edu/~preacher/sobel/sobel.htm Rempel, K. J., Holmes, J. G., & Zanna, M. P. (1985). Trust in close relationships. Journal of Personality and Social Psychology, 49, 95–112. Sarason, I. G., Sarason, B. R., Brock, D. M., & Pierce, G. R. (1996). Social support: Current status, current issues. In C. D. Spielberger & I. G. Sarason (Eds.), Stress and emotions: Anxiety, anger, and curiosity (Vol. 16, pp. 3–27). Washington, DC: Taylor & Francis. Schlosser, A. E., White, T. B., & Lloyd, S. (2004). The role of website investment in increasing trusting beliefs and online purchase intentions. Unpublished manuscript, University of Washington, Seattle. Sirdeshmukh, D., Singh, J., & Sabol, B. (2002). Consumer trust, value and loyalty in relational exchange. Journal of Marketing Research, 66, 15–37. White, T. B. (2004). Consumer disclosure and disclosure avoidance: A motivational framework. Journal of Consumer Psychology, 14, 41–51. Zeithaml, V. A., Parasuraman, A., & Berry, L. L. (1990). Delivering quality service: Balancing customer perceptions and expectations. New York: Free Press.

Received: September 11, 2003 Revision received: June 28, 2004 Accepted: October 18, 2004