Impression management attenuates the effect of ability on trust in economic exchange

Significance

It is often assumed that competent actors are trusted, but is this always true? In this report, we test the position that impression management decreases the amount of trust competent actors receive. Employing a variety of experimental paradigms and measures and confirming predictions based on attribution theory, we demonstrate that impression management can substantially backfire, at least for competent actors, and that this effect can be explained by decreases in perceived benevolence and integrity. Our findings make important contributions to the literatures on trust, status, and impression management.

Abstract

Are competent actors still trusted when they promote themselves? The answer to this question could have far-reaching implications for understanding trust production in a variety of economic exchange settings in which ability and impression management play vital roles, from succeeding in one’s job to excelling in the sales of goods and services. Much social science research assumes an unconditional positive impact of an actor’s ability on the trust placed in that actor: in other words, competence breeds trust. In this report, however, we challenge this assumption. Across a series of experiments, we manipulated both the ability and the self-promotion of a trustee and measured the level of trust received. Employing both online laboratory studies (n = 5,606) and a field experiment (n = 101,520), we find that impression management tactics (i.e., self-promotion and intimidation) can substantially backfire, at least for those with high ability. An explanation for this effect is encapsuled in attribution theory, which argues that capable actors are held to higher standards in terms of how kind and honest they are expected to be. Consistent with our social attribution account, mediation analyses show that competence combined with self-promotion decreases the trustee’s perceived benevolence and integrity and, in turn, the level of trust placed in that actor.

Trust—understood as the willingness to make oneself vulnerable to the actions of another party (1)—is indispensable for building and maintaining economic relationships (2, 3). Trust has been associated with a wide range of desirable outcomes, ranging from the career success of individuals (4) and the cooperation among group members (5) and organizations (6), to the wealth of nation states (7). As a result, scholars across the social sciences have endeavored to identify relevant antecedents of trust in economic exchange (8–10). One stream of this research zooms in on specific characteristics of trustees that constitute their trustworthiness: that is, their propensity not to exploit the trustor’s trust (11, 12). A key dimension of trustees’ trustworthiness is their ability (or competence): that is, their capacity to accomplish a specific task at hand effectively (1, 13, 14). For example, a prospective customer may perceive a vendor as high in ability if that vendor has experience in successfully accomplishing sales transactions, including providing high levels of customer service and delivering the product on time. It is commonly argued that competent trustees will enjoy high levels of trust because of increased confidence in their proficient performance (15, 16). Although ability and trust are undoubtedly distinct concepts (17), they are directly related. Expectations of “technically competent role performance … involve some of the fundamental meanings of trust” (13), such that much theorizing has “considered ability an essential element of trust” (1). A metaanalysis of research on trust in job task contexts identified ability as one of the strongest predictors of trust (18). In contemporary research, the link between ability and trust therefore appears to be virtually taken for granted, to the extent that it is more often assumed than put to an empirical test.

However, we challenge the prevailing assumption that ability necessarily translates into increased trust in economic actors. We argue that prior trust research has treated ability as an uncontested asset and fact-like property, while failing to account for how ability is socially constructed as trustees present themselves during interactions (19). Especially in economic exchanges with strangers in which trust needs to be constructed swiftly (20, 21), the trustee’s level of ability is largely unknown before it is communicated. To incorporate this self-presentational element, we bring theory on impression management (22–26)—and in particular the notion of self-promotion (27, 28)—to the literature on trust. This integrative approach allows us to demonstrate that ability may fail to foster trust when combined with self-promotion.

Symbolic interactionism (29) emphasizes how anticipated responses by other people influence actors’ behaviors (30, 31). This theory draws attention to the important role of impression management—that is, the processes by which actors attempt to control how others perceive them (32)—in how interactions unfold (22). Economic actors, such as commercial vendors or organizational employees, may adopt a variety of self-presentational techniques with the goal of being viewed in a positive light (26, 33). Here, we primarily focus on self-promotion, not only because it is directly related to managing ability perceptions but also because it is particularly common (34). In an exploratory study in which we asked participants to assess their own ability (35) and the extent to which they promote themselves (36), we found that self-promotion is a pervasive behavior among both high-ability individuals (mean = 2.716) and low-ability individuals [mean = 2.778, t(818) = 1.193, P = 0.233, Cohen’s d = 0.085] (see SI Appendix for study details). By definition, self-promotion entails efforts to highlight one’s own accomplishments, strengths, and talents (37). Prior research on the antecedents of self-promotion has highlighted the important effects of role ambiguity, job involvement, need for power, shyness, and emotional stability as relevant driving conditions (38, 39). While highlighting one’s virtues can have favorable consequences, self-promotion may also have detrimental effects, because it is often regarded as a deviation from norms of modesty (28, 40, 41). In many societies, economic actors are expected to be selfless and modest (42, 43), and deviance from this norm risks social rejection (44, 45). Self-promotion is thus a double-edged sword, which can ultimately create more harm than good for the actor employing it.

We argue that self-promotion functions as an important contingency in the link between ability and trust. In the absence of self-promotion, we expect a strong positive effect of ability on trust, in line with prior literature (e.g., refs. 1, 18, 46, and 47). However, this positive effect will be substantially attenuated or even muted when trustees use self-promotional impression management. Accordingly, highly capable partners will receive less trust when they engage in self-promotion.

Attribution theory (48–50), which attempts to understand people’s perceptions and judgments of other individuals and collectives, serves as our conceptual framework for explaining this difference in assessments of high- vs. low-ability actors applying self-promotion. In many settings, very capable evaluation targets are held to higher standards (51), with the expectation that they will act competently, displaying their knowledge of and conformity to social norms (52, 53). To the degree that high-ability actors engage in deviance and thus fail to meet the high expectations placed on them, they may garner criticism and even penalties (51, 54, 55). Highly capable actors “really ought to know better” than to ignore role-appropriate behaviors (56). In contrast, an exchange partner of lesser competence triggers lower expectations (53). Compared to norm violations committed by more competent actors, the deviating conduct of low-ability partners may be appraised as less problematic since it displays less deviation from expected behavior.

Because self-promotion is associated with deviance from social norms of modesty (28, 40, 41), high-ability (vs. low-ability) partners employing this impression management technique will be regarded as less trustworthy. If a partner is highly competent, self-promotion produces a discrepancy from the high standards to which capable actors are held and thus a violation of expectations that erodes the trust placed in them (51, 57). Less capable actors, in contrast, do not suffer from the “liability of competence” in the form of heightened expectations, and their deliberate attempts at being perceived in a more favorable light may in fact be viewed as understandable. This asymmetric pattern leads us to expect a negative interactive effect of ability and self-promotion on trust, such that the positive effect of ability on trust will be considerably weaker in the presence (vs. absence) of self-promotion.

In particular, we anticipate a mediated-moderation pattern, whereby the interaction of ability and self-promotion will influence assessments of the trustee’s character, which in turn shape trust. The trustworthiness literature addresses two distinct dimensions of character or goodwill that we suggest act as mediating mechanisms: benevolence and integrity (1, 58, 59). First, benevolence can be defined as the trustor’s perception that the trustee has his or her best interests at heart (1). Benevolence assessments are often based on perceptions of the trustee’s caring and warmth. Prior impression management research has shown that actors engaging in self-promotion tend to be rated lower on these traits (60), and we expect this benevolence penalty to be particularly pronounced for trustees high in ability. Second, integrity refers to the trustor’s perception that the trustee is committed to an acceptable set of principles (1). Self-promoters are often viewed as manipulative, unsociable, and prompted by ulterior motives (61), resulting in perceptions of low integrity (62), particularly if the self-promoter is high in ability. Such detrimental consequences of self-promotion can occur even in settings in which economic motivations prevail, such as when customers are wary of vendors’ persuasion attempts (63) or penalize businesses that lack in perceived sincerity (64).

Across seven studies, we show that impression management can substantially backfire, at least for those endowed with high ability. Detailed materials and methods are provided in SI Appendix. Study 1A demonstrates that self-promotion attenuates the positive effect of ability on trust ratings of an online vendor. Study 1B generalizes these results to an organizational setting in which trust ratings of a superordinate actor are examined. Study 2 replicates the interactive effect of ability and self-promotion in the context of entrusted down payments for an electronics product. Study 3, a field experiment among users of a global social media platform, demonstrates that present (vs. absent) self-promotion leads to fewer purchase attempts. Study 4 sheds light on the underlying mechanisms of our focal interactive effect, showing that the combination of ability and self-promotion decreases benevolence and integrity perceptions and, in turn, reduces trust. Study 5 generalizes this moderated mediation effect to another impression management tactic: intimidation. Study 6, a prediction market study, shows that a majority of participants bet a monetary endowment on our focal interactive effect and are able to correctly forecast the outcome of a replication study.

Study 1A: Self-Promotion Attenuates the Positive Effect of Ability on Trust Ratings in a Commercial Setting

Study 1A begins our examination of the interactive effect of ability and self-promotion on trust. Similar to prior research on trust in economic exchange (e.g., refs. 65–67), the study focuses on buyer–seller relationships as an empirical context. Specifically, our vignette experiment is situated in an online shopping context, in which trust is known to play a central role (68–70). Participants were asked to imagine they were shopping for a new television and had found a relevant offer on the internet. We further told participants that the website appeared a bit suspicious, and it was unclear whether they could trust the vendor; thus, we suggested they consult the vendor’s customer reviews. Participants were randomly assigned to one of four conditions in this 2 (ability: low, high) × 2 (self-promotion: absent, present) between-subjects experimental design.

To manipulate the vendor’s ability, we used customers reviews (i.e., star levels and quantity of reviews), which have previously been established as a crucial and stable proxy of a trustee’s ability. For example, a recent experimental study demonstrates that customer reviews are a strong indicator of vendor competence, to a much greater extent than, for example, guarantees (71). Indeed, in an independent pretest we conducted, two-thirds of participants indicated that star levels serve as the first indicator they consider when judging an online vendor’s ability. In contrast, for benevolence and integrity, other indicators (such as likes and certifications from independent institutions) seemed to play a more important role when assessing these two seller characteristics. Moreover, employing both star levels and quantity of reviews in our manipulation of ability underscores the usefulness of customer reviews for ability judgments, because having a large quantity of online customer reviews makes vendors look more capable (72). Finally, customer reviews are known to be particularly relevant for ability perceptions of smaller and unknown vendors (73), such as those in our vignette.

To manipulate the vendor’s self-promotion, we either did or did not present participants with a self-promoting message from the vendor. Manipulating self-promotion through communication (e.g., messages) of the actor being judged is a common approach in experimental impression management research. For example, self-promoting one’s skills and accomplishments has been used to examine how job applicants and employees are evaluated by their employers (74, 75) and how vendors are perceived by buyers (76). Our manipulation of a vendor’s self-promotion directly builds on these methods and presents a boasting message targeted at potential customers, which is a common and realistic occurrence in advertising and buyer–seller relationships (77). In the present self-promotion condition, the self-promotion message included the vendor saying, for example, that there had never been a single situation in which he did not deliver as promised. This message is certainly an inflated form of self-promotion that infringes on norms of modesty; nonetheless, it closely resembles previous conceptualizations of self-promotion as efforts to highlight one’s own accomplishments, strengths, and talents (37). In the absent self-promotion condition, no message was displayed; instead, participants were asked to solve an anagram by rearranging a meaningless set of letters (HACIR) to form an actual word (i.e., CHAIR). This filler task, which has been widely used in previous research (78, 79), ensured that our experimental conditions were comparable in terms of cognitive load. Next, trust, ability, and self-promotion were measured. An exemplary trust item is “I trust this seller” (80, 81), an exemplary ability item is “I feel very confident about the seller’s skills” (35), and an exemplary self-promotion item is “This seller was bragging” (82). SI Appendix, Supplementary Text provides all measures. Finally, to assess the nature of the trust relationship (interpersonal vs. institutional), participants indicated whether they perceived the vendor as an individual or organization.

Manipulation checks revealed that our manipulations of ability and self-promotion were successful. Participants in the high-ability condition rated the vendor as having higher ability (mean = 3.399) than did those in the low-ability condition [mean = 1.921; t(594) = 13.641, P < 0.001, Cohen’s d = 1.118]. Furthermore, participants in the present self-promotion condition rated the vendor as more self-promoting (mean = 6.451) than did those in the absent self-promotion condition [mean = 2.599, t(594) = 38.333, P < 0.001, Cohen’s d = 3.140].

We next tested for the negative interactive effect of ability and self-promotion on trust. As predicted, an ANOVA with ability (low, high) and self-promotion (absent, present) as fixed factors and trust as the dependent variable revealed an interactive effect of ability and self-promotion on trust, F(1, 592) = 29.379, P < 0.001, partial η² = 0.047. The difference in trust between low ability and high ability was considerably higher when the vendor did not self-promote (mean_{High ability, absent self-promotion} = 3.977; mean_{Low ability, absent self-promotion} = 1.997; post hoc Tukey honest significant difference [HSD] test: P < 0.001) than when the vendor did self-promote (mean_{High ability, present self-promotion} = 2.415; mean_{Low ability, present self-promotion} = 1.606; P < 0.001). Fig. 1A illustrates the interactive effect. Whether participants perceived the vendor as an individual (83.2%) or organization (16.8%) did not influence the interactive effect of ability and self-promotion on trust at statistically significant levels (three-way interaction: P = 0.173), indicating that our focal effect holds in both interpersonal trust relationships between two individuals and institutional trust relationships between an individual and an organization. Our analysis also revealed main effects of ability on trust, F(1, 592) = 166.682, P < 0.001, partial η² = 0.220, and self-promotion on trust, F(1, 592) = 81.676, P < 0.001, partial η² = 0.121.

Fig. 1.

Interactive effects of ability and self-promotion on trust, as found in studies 1A and 1B. Self-promotion attenuates the positive effect of ability on trust ratings in both a commercial setting (A) and an organizational setting (B).

Study 1B: Self-Promotion Attenuates the Positive Effect of Ability on Trust Ratings in an Organizational Setting

Study 1B had three objectives. The first objective was to generalize the negative interactive effect of ability and self-promotion on trust in an additional context adopted from previous research (83): namely, an organizational setting between an employee and their boss. The second objective was to use a different ability manipulation (83) and a well-established trust scale (35) to increase the validity of our measures. The third objective was to rule out the possibility that our focal interactive effect of ability and self-promotion on trust is driven by a floor effect by carefully pretesting the trust ratings in the low ability and absent self-promotion condition and ensuring that these ratings can, in fact, further decrease.

We adapted the experimental procedures from previous research (83). Participants were instructed to picture themselves as a procurement manager of a defense and aerospace company and to read a work scenario involving a vice president of procurement, Jamie Smith, who had been recently hired to develop a cost-cutting policy for the company (83). We manipulated the vice president’s ability by denoting this individual’s industry experience in years. In the low- (high-) ability condition, the vice president had 0 (6) years of prior experience in the industry. Next, participants read an email sent by the vice president to all procurement managers providing information about a new cost-cutting policy (83). To manipulate the vice president’s self-promotion, the email either did or did not include a self-promoting message. In the present self-promotion condition, the self-promotion message stated that the vice president is the best person for this job. In the absent self-promotion condition, no such message was displayed. Next, participants’ trust in the vice president was measured. An exemplary (reversed) trust item was “If I had my way, I wouldn’t let Jamie Smith have any influence over issues that are important to me” (35). SI Appendix, Supplementary Text provides all measures.

Manipulation checks (conducted in separate pretests) revealed that our manipulations of ability and self-promotion were successful. Participants in the high-ability condition rated the vice president as having higher ability (mean = 4.612) than did those in the low-ability condition [mean = 2.670; t (93) = 7.698, P < 0.001, Cohen’s d = 1.580]. Furthermore, participants in the present self-promotion condition rated the vice president as more self-promoting (mean = 6.130) than did those in the absent self-promotion condition [mean = 2.840, t (98) = 11.969, P < 0.001, Cohen’s d = 2.394].

In the main study, we next tested for the negative interactive effect of ability and self-promotion on trust. As predicted, an ANOVA with ability (low, high) and self-promotion (absent, present) as fixed factors and trust as the dependent variable revealed an interactive effect of ability and self-promotion on trust, F(1, 473) = 6.951, P = 0.009, partial η² = 0.014. The difference in trust between low ability and high ability was higher when the vice president did not self-promote (mean_{High ability, absent self-promotion} = 4.088; mean_{Low ability, absent self-promotion} = 3.058; post hoc Tukey HSD test: P < 0.001) than when the vice president did self-promote (mean_{High ability, present self-promotion} = 2.969; mean_{Low ability, present self-promotion} = 2.523; P = 0.025). Fig. 1B illustrates the interactive effect. Finally, our analysis again revealed main effects of ability on trust, F(1, 473) = 44.301, P < 0.001, partial η² = 0.086, and self-promotion on trust, F(1, 473) = 55.758, P < 0.001, partial η² = 0.105.

Study 2: Self-Promotion Attenuates the Positive Effect of Ability on Entrusted Down Payments

The objective of study 2 was to replicate the negative interactive effect of ability and self-promotion on trust while employing a different dependent measure that allows for capturing greater variance than the one anchored on a seven-point scale, thereby further alleviating concerns that the focal interaction may be purely driven by a floor effect (84). Specifically, participants were instructed to imagine that they recently received a $300 bonus, which they wanted to use as a down payment toward a new television. They were further told that the vendor expected them to make this down payment prior to shipping the television, and participants could choose their down payment amount between $0 and $300. Analogous to the previous study designs, participants were randomly assigned to one of four conditions in this 2 (ability: low, high) × 2 (self-promotion: absent, present) between-subjects experimental design. The manipulations of ability and self-promotion resembled those used in study 1A (see SI Appendix, Supplementary Text for a detailed description).

Results again revealed that our manipulations of ability and self-promotion were successful. Participants in the high-ability condition rated the vendor as having higher ability (mean = 3.143) than did those in the low-ability condition [mean = 1.861; t(1,606) = 20.297, P < 0.001, Cohen’s d = 1.012]. Participants in the present self-promotion condition rated the vendor as more self-promoting (mean = 6.336) than did those in the absent self-promotion condition [mean = 2.917; t(1,606) = 53.517, P < 0.001, Cohen’s d = 2.670].

Next, we again tested for the negative interactive effect of ability and self-promotion on trust. Due to highly skewed data for the dependent variable (down payment in dollars), we applied a logarithmic transformation before running the analysis (85). As predicted, an ANOVA with ability (low, high) and self-promotion (absent, present) as fixed factors and trust as the dependent variable revealed a negative interactive effect of ability and self-promotion on trust, F(1, 1,604) = 6.470, P = 0.011, partial η² = 0.004. Results also showed that the difference in trust between low ability and high ability was higher when the vendor did not self-promote (mean_{High ability, absent self-promotion} = 1.534; mean_{Low ability, absent self-promotion} = 0.624; post hoc Tukey HSD test: P < 0.001) than when the vendor did self-promote (mean_{High ability, present self-promotion} = 1.104; mean_{Low ability, present self-promotion} = 0.620; P < 0.001), providing convergent evidence in support of the hypothesis that the positive effect of ability on trust is attenuated by self-promotion. As expected, the analysis also revealed main effects of ability on trust, F(1, 1,604) = 69.197, P < 0.001, partial η² = 0.041, and self-promotion on trust, F(1, 1,604) = 6.712, P = 0.010, partial η² = 0.004.

Study 3: Self-Promotion Lowers Unique Click-Through Rate on Social Media

The objective of study 3 was to increase external validity in the testing of our hypothesis that self-promotion disproportionally hurts competent actors by conducting an online field experiment on a global social media platform (86). Specifically, we ran an advertisement on Facebook promoting a new coffee machine from a coffee gear website. The advertisement encouraged users to click on the advertisement link in order to buy the new coffee machine now. Clicking on the advertisement serves as an adequate behavioral measure of trust because: 1) the adoption of a new product from an internet seller requires some degree of trust in the remote merchant (87); 2) clicking on a social media ad makes participants vulnerable to potential data security threats, which are known to be a problem on social media platforms (88, 89); and 3) our call to action to “shop now”—unlike other calls to action (e.g., “learn more” or “contact us”)—reflects concrete purchase intentions (90). These three points are in line with the common definition of trust, which states that people trust others when they demonstrate the willingness to make themselves vulnerable to the actions of those others (1).

Using Facebook’s advertising manager software tool, we followed pertinent recommendations and budgeted an advertising campaign to reach over 100,000 Facebook users (91). The users were assigned to one of two conditions in this single-factor (self-promotion: absent vs. present) between-subjects design, while holding constant the ability of a website featuring coffee gear at a high level and ensuring consistent audience, placement, and delivery settings (see SI Appendix, Supplementary Text for the advertisement shown to users and a detailed description of the advertisement specifications). In line with our theoretical account, we expected a lower unique click-through rate (i.e., the percentage of people who saw our advertisement and performed a unique click [all]) (92, 93) for the advertisement that depicts a self-promoting coffee gear website.

As predicted, the advertisement featuring the self-promoting website generated a lower unique click-through rate (0.056%) compared to the one featuring the nonself-promoting website [0.091%; χ²(1) = 3.989; n = 101,520; P = 0.046, 95% CI for odds ratio = (0.390, 0.995)] (see SI Appendix, Supplementary Text for a robustness check and additional metrics).

Study 4: Self-Promotion Attenuates the Positive Effect of Ability on Trust Via Benevolence and Integrity Perceptions

The objective of study 4 was to shed light on the underlying mechanisms explaining the interactive effect of ability and self-promotion on trust. We hypothesized that the negative interactive effect is explained by decreased levels of both benevolence and integrity; that is, high-ability trustees who self-promote (vs. do not self-promote) are viewed as lower in benevolence and integrity, which in turn affect levels of trust. Study 4 builds on the experimental design of study 1A while also introducing measures for perceived benevolence and integrity (see SI Appendix, Supplementary Text for a full report on how benevolence and integrity were measured). An exemplary benevolence item was “This seller has my interests in mind” (94), and an exemplary integrity item was “I thought this seller has integrity” (94).

Results again revealed that our manipulations of ability and self-promotion were successful. Participants in the high-ability condition rated the vendor as having higher ability (mean = 2.381) than did those in the low-ability condition [mean = 1.969; t(399) = 3.242, P = 0.001, Cohen’s d = 0.324]. Participants in the present self-promotion condition rated the vendor as more self-promoting (mean = 6.179) than did those in the absent self-promotion condition [mean = 2.582; t(399) = 26.060, P < 0.001, Cohen’s d = 2.607].

Turning to the proposed interactive effect, an ANOVA with ability (low, high) and self-promotion (absent, present) as fixed factors and trust as the dependent variable revealed an interactive effect of ability and self-promotion on trust, F(1, 397) = 6.555, P = 0.011, partial η² = 0.016. Results also revealed that the difference in trust between low ability and high ability was higher when the vendor did not self-promote (mean_{High ability, absent self-promotion} = 2.793; mean_{Low ability, absent self-promotion} = 2.074; post hoc Tukey HSD test: P < 0.001) than when the vendor did self-promote (mean_{High ability, present self-promotion} = 1.662; mean_{Low ability, present self-promotion} = 1.605; P = 0.988). As expected, the analysis also revealed main effects of ability on trust, F(1, 397) = 9.019, P = 0.003, partial η² = 0.022, and self-promotion on trust, F(1, 397) = 38.331, P < 0.001, partial η² = 0.088.

To test the underlying mechanisms of benevolence and integrity, we fitted a moderated-mediation model with two parallel mediators, using the standard PROCESS Model 8 script (95) with 5,000 bootstrap samples and 95% CIs. Ability was modeled as the independent variable, self-promotion as moderator, benevolence and integrity as mediators, and trust as the dependent variable. The moderated-mediation model revealed that ability was a positive predictor of both benevolence (B = 0.607, SE = 0.189, t = 3.212, P = 0.001) and integrity (B = 0.685, SE = 0.188, t = 3.644, P < 0.001) but not trust (P = 0.133), suggesting full mediation (96). Both benevolence (B = 0.512, SE = 0.046, t = 11.181, P < 0.001) and integrity (B = 0.369, SE = 0.046, t = 8.002, P < 0.001) were positive predictors of trust, suggesting that benevolence and integrity are parallel mediators. Self-promotion was a negative moderator of both the link between ability and benevolence (B = −0.615, SE = 0.260, t = −2.366, P = 0.019) and the link between ability and integrity (B = −0.509, SE = 0.259, t = −1.969, P = 0.050) but not the link between ability and trust (P = 0.260). In further support of moderated mediation, there was an indirect effect of ability through benevolence on trust conditional on self-promotion (Index = −0.315, SE = 0.139, CI [−0.590; −0.048]) and an indirect effect of ability through integrity on trust conditional on self-promotion (Index = −0.188, SE = 0.104, CI [−0.408; −0.003]). Fig. 2 illustrates the moderated-mediation model.

Fig. 2.

Moderated-mediation model of study 4. Benevolence and integrity both mediate the effect of ability on trust, and self-promotion represents a negative moderator of these mediation effects (dashed lines denote statistically nonsignificant effects).

Study 5: Intimidation Also Attenuates the Positive Effect of Ability on Trust Via Benevolence and Integrity Perceptions

The objective of study 5 was to further extend our investigation by generalizing our theoretical account to another frequently employed impression management tactic: intimidation, whereby actors let others know that they can make things difficult for them (97). Similar to self-promotion, intimidation represents an assertive impression management tactic (98), is often considered counternormative (97), and has been linked to an actor’s desire to be viewed as more capable (99). Given these similarities, we have reason to believe that our expectations concerning the role of self-promotion will also be applicable to intimidation. In particular, we expect that, like self-promotion, intimidation will attenuate the positive effect of ability on trust by decreasing perceived benevolence and integrity. Study 5 employed an experimental design similar to that of study 4 (see SI Appendix, Supplementary Text for a detailed description), except that we manipulated intimidation instead of self-promotion. For this purpose, a message from the vendor put pressure on the buyer to make their purchase quickly, stating that the product was in high demand and might be sold out soon, a well-known intimidation tactic in sales contexts (100) (see SI Appendix, Supplementary Text for a full report on how intimidation was manipulated).

Results revealed that our manipulations of ability and intimidation were successful (Ps < 0.001). Results also revealed an interactive effect of ability and intimidation on trust (P < 0.001) as well as main effects of ability on trust (P < 0.001) and intimidation on trust (P < 0.001). Moderated-mediation analyses further revealed a highly similar role of intimidation to that reported under study 4 for self-promotion, suggesting moderated mediation by intimidation (see SI Appendix, Supplementary Text for detailed statistics).

Study 6: A Prediction Market Forecasts that Self-Promotion Attenuates the Positive Effect of Ability on Trust

The objective of study 6 was to set up a prediction market (101, 102) in which participants predicted the trust other people would have in different trustees. As such, study 6 took a third-party perspective, in which participants were asked to provide a social judgment about our focal interactive effect. This approach follows prior work on social judgments, which argues that many people are reliable judges of other people’s attitudes and behaviors (103). To create incentive compatibility, participants were endowed with a monetary incentive (in addition to their study compensation), which they were asked to use in a bet on four different television vendors. As in previous studies, these vendors were either high or low in their ability and either self-promoting or not. We asked participants to predict which of the four vendors would most likely be trusted by other consumers. Participants were also told that they must bet their full monetary endowment, but it was up to them whether they wanted to bet the full endowment on one vendor or divide up the money among multiple vendors. To determine participants’ winnings, we then conducted a separate study replicating a design similar to study 1A, in which we asked a sample of other participants about their trust in the four vendors. Participants of the prediction market study were told that the amount of money that they bet on the most trusted vendor in the replication study would be doubled and the total would be paid to them as a bonus.

In this 2 (ability: low, high) × 2 (self-promotion: absent, present) repeated-measures within-subjects experimental design, participants were shown the customer reviews and, depending on condition, the self-promoting message of four different television vendors, in random order. After studying the four different vendors, participants were endowed with $1 and were asked to place their bets on the four vendors (any fraction of $1 per vendor but totaling $1 across the four vendors).

We next tested whether prediction market participants correctly predicted the negative interactive effect of ability and self-promotion on trust. A repeated-measures ANOVA with ability (low, high) and self-promotion (absent, present) as within-subject factors and monetary bets as the dependent variable revealed an interactive effect of ability and self-promotion on bets, F(1, 527) = 776.149, P < 0.001, partial η² = 0.596. Results also showed that the difference in bets between low ability and high ability was higher when the vendor did not self-promote (mean_{High ability, absent self-promotion} = $0.78; mean_{Low ability, absent self-promotion} = $0.03; paired-sampled t test: P < 0.001) than when the vendor did self-promote (mean_{High ability, present self-promotion} = $0.17; mean_{Low ability, present self-promotion} = $0.02; P < 0.001). As expected, the analysis also revealed main effects of ability on bets, F(1, 527) = 4122.565, P < 0.001, partial η² = 0.887, and self-promotion on bets, F(1, 527) = 781.326, P < 0.001, partial η² = 0.597.

We then conducted a replication of an experimental design similar to the one in study 1A on a separate sample of participants. Results again revealed a negative interactive effect of ability and self-promotion on trust (P < 0.001) (see SI Appendix, Supplementary Text for detailed results of the replication study). Prediction market participants were then paid double the amount of their bet on the vendor with high ability and absent self-promotion.

Discussion

This article addresses the question of just how pervasive the widely presumed positive effect of ability on trust truly is. Addressing this question has major implications for understanding trust production in a variety of economic exchanges in which actors’ abilities play a prominent role. Across our seven experiments, we found convergent evidence for a negative interaction effect of ability and impression management on trust, such that the ability–trust relationship is impaired or even suppressed when trustees engage in self-promotion or intimidation. Our investigation contributes to several literatures and gives rise to important avenues for future research.

First and foremost, our inquiry contributes to research on the antecedents of trust in economic exchange. Contrary to the taken-for-granted assumption that the trustee’s ability constitutes an unconditional driver of trust (1, 18, 46, 47), we demonstrate that this effect is more conditional than previously recognized. By presenting impression management as a relevant boundary condition, we challenge the assumption that ability invariably increases trust. Specifically, our investigation acknowledges that a trustee’s ability is not always readily apparent but often needs to be socially constructed through interaction. In this process, the trustor relies not only on objective indicators of ability but also on the subjective impressions they form of the trustee, which can be substantially shaped by impression management techniques employed by the trustee. Importantly, we argue that the trustee’s level of ability and the appraisal of their impression management attempts are noninterdependent. Building on attribution theory, we suggest that self-promotion and intimidation can be viewed as forms of norm deviation that are considered particularly inappropriate when the deviant is highly competent. As a result, we propose a negative interaction effect, indicating that impression management reduces the trust that high-ability actors receive, whereas the impression-management–induced trust discount is less pronounced for low-ability actors. The irony of this pattern is striking in light of the fact that self-promotion is defined as seeking the attribution of competence (34). We show that this attribution seeking can substantially backfire when an actor’s high level of competence is already apparent. This insight opens avenues for future research into additional moderators of the ability–trust effect, including other self-presentational techniques but also going beyond impression management. For example, it would be interesting to further examine whether, in repeated exchange settings, ability interacts with perceived long-term commitment, such that trust in high-ability trustees is highly sensitive to their anticipated (dis)loyalty (104).

Second, our investigation sheds light on the interplay among dimensions of trustworthiness. Prior work largely assumed that ability, benevolence, and integrity are positively interrelated, with one dimension having positive spillovers on the others and hence all three being highly congruent (80, 105). Our mediation analyses, however, reveal that ability (when paired with impression management) does not always have a significant association with benevolence and integrity. This finding is consistent with a recent study hinting at possible tensions between trustworthiness components (58). Future research on trustworthiness may build on this insight and extend the inquiry into the conditions that affect the degree of (non)correspondence between trustworthiness dimensions. Particularly useful would be a set-theoretic approach (106) to identify the antecedents and consequences of diverse ability, benevolence, and integrity configurations.

Third, because ability is closely tied to prestige (107, 108), our article also has implications for the literature on status. There is an ongoing debate on whether high-status (vs. low-status) actors tend to receive more favorable treatment (109). The traditional view, popularized by research on the Matthew effect (110), suggests that status comes with a variety of benefits for those who possess it (111). However, recent scholarship points to the possibility of status liabilities, such as when high-status actors are held to higher standards and thus are judged more harshly (112). Our investigation offers support for both positions. Ceteris paribus, highly able partners enjoy the advantage of superior trustworthiness assessments and consequently greater trust being placed in them; at the same time, however, our findings suggest that transgressions of social norms are punished to a greater extent when the deviant is high in ability. Our article thus contributes to the development of a more nuanced understanding of status benefits and liabilities (53) and serves as a springboard for future research uncovering contingencies other than impression management, ranging from relationship tenure to power differences between trustor and trustee to the specific type of norm violation.

Fourth, our findings inform research on impression management. Today’s meritocratic society appears to incentivize people to proactively present themselves and their competencies in the best possible light (113). Aggressively self-promoting one’s strengths is often portrayed as a sine qua non in the quest for successful social and economic exchange in this type of environment (114). The viewpoint that actors engaging in self-promotion will fare better than those who do not appears to be widely accepted, especially within popular discourse, to the extent that practical guidelines for overcoming modesty concerns in favor of blatant bragging are hitting best seller lists (e.g., refs. 115 and 116). However, our empirical findings provide little support for this apparent enthusiasm surrounding self-promotion. Our studies point to a negative main effect on trust, lending further credence to more cautious scholarly assessments of the merits of self-promotion (37, 40). Interpreting our focal interaction term from a different perspective, we offer competence as an important boundary condition to the effectiveness of self-promotion (as well as intimidation). Clearly, more work is needed to understand the contingent effects of impression management techniques on a variety of outcomes (117) before any well-grounded practical recommendations for or against the use of impression management can be made.

Finally, this article makes an integrative contribution by bringing together two ever-expanding but largely separate bodies of social science research: those on trust and on impression management. Trust scholars have rarely adopted an impression management perspective (see ref. 118 for an exception). This is surprising because trustees are known to make substantial (conscious or unconscious) efforts to be perceived as trustworthy (119, 120); thus, impression management techniques are commonly encountered but infrequently studied. We therefore see significant potential for future research to follow up on our investigation and examine the complex relationships between impression management and trust building.

To conclude, in this article we contest the assumption that ability will always lead to trust. By embracing a symbolic interactionist approach and separating ability from how it is conveyed through impression management, we demonstrate that the relationship between ability and trust is more conditional than previously assumed. Even in meritocratic societies, highly competent actors may experience greater trust from their fellow citizens if they are humble and refrain from efforts to manipulate others’ impressions of themselves.

Materials and Methods

A series of experiments provides convergent support for a negative impact of two impression management tactics—self-promotion and intimidation—on the effect of ability on trust. Following pertinent recommendations for constructive replication (121), our experiments were situated in different contexts, employed a variety of manipulations and dependent measures, and included data collected in both online laboratory and field settings. Furthermore, we set up a prediction market in which participants bet a monetary endowment, increasing both validity and incentive compatibility. The research was approved by the Institutional Review Board of the University of Arizona. After reading an online disclosure form for research participation, consent was obtained from participants in the exploratory study and studies 1A, 1B, 2, 4, 5, and 6. Because of the aggregated nature of the data obtained in study 3 from Facebook, the Institutional Review Board provided a waiver of informed consent. Materials and data collection plans, including hypotheses and data exclusion criteria, were stored on the Open Science Framework before data collection began. Sample sizes were either exploratorily determined or formally calculated based on previous studies’ effect sizes (see SI Appendix, Supplementary Text for sample size calculations). Data and code are available for download on the same repository.

Data Availability

Experimental data have been deposited in the Open Science Framework, https://osf.io/QMZCY/ (122).