The Quest for (Eliciting) Hubristic Pride: Nomological Shockwaves/Networks, Attributions, and Social Desirability Across Different Experimental Manipulations

Self-conscious emotions such as pride result from evaluation of different aspects of the self. For instance, negative self-conscious emotions, guilt and shame, arise from distinct evaluations of the self: guilt involves negative attributions of behavior, whereas shame involves a focus on negative attributions of the global self (Tangney et al., 1996). Similarly, the two-facet theory of pride argues that authentic pride and hubristic pride emerge from distinct attributions to the self (Tracy & Robins, 2004, 2007). A core proposition of their theory of pride is:

Both types of pride arise from success and give rise to status. However, the nature of the event that sparks pride does not necessarily predict which type of pride is elicited. What differentiates the two is in which manner individuals attribute success to the self. People feel authentic pride when success is judged to originate from something unstable about the self (e.g. “I passed the exam because I worked hard on it”). This type of success involves efforts, skills, and competence. By contrast, hubristic pride arises when success is attributed to something stable about the self (e.g., “I passed the exam because I am talented”), and it does not require earned skills or hard work. In short, just as the underlying attributions represent a central ingredient in whether guilt or shame are elicited, the underlying attributions distinguish the two facets of pride (Tracy & Robins, 2007).

Theoretically, the two facets of pride produce independent outcomes and have distinct correlates. At the trait level, authentic pride and hubristic pride are only weakly correlated (meta-analytic r = .13, Dickens & Robins, 2020). Individuals with higher disposition to feel authentic pride exhibit desirable traits such as lower shame-proneness, higher self-esteem, higher agreeableness, and higher conscientiousness; individuals with higher disposition to feel hubristic pride exhibit maladaptive traits such as lower self-esteem, lower agreeableness, and higher narcissism (Shi et al., 2015; Tracy et al., 2009; Tracy & Robins, 2007). Another line of research has proposed that the two facets of pride are associated with two distinct status strategies concerning prestige and dominance (Henrich & Gil-White, 2001; Tracy et al., 2020). Authentic pride tends to occur in the context of prestige—status attainment through respect, demonstration of knowledge, mastery, skills, communal traits, and competency within culturally valued domains. The experience of hubristic pride tends to occur in the context of dominance—status attainment through coercive tactics, induction of fear, aggression, and intimidation (Cheng et al., 2013).

Tracy and Robins’ (2007) two-facet theory of pride is to date the dominant framework of pride research, if not the only psychological theory that has explicitly theorized the multi-faceted nature of pride (e.g., Lewis, 2000). Yet, there exists an emerging controversy about the two-facet model of pride, especially concerning the construct validity of hubristic pride. The ongoing controversy comes down to two issues: (1) whether data accumulated by Tracy and Robin’s AP-HP scales support the status of hubristic pride as a pride emotion (Holbrook et al., 2014); and (2) whether adequate evidence warrants the two-facet model of pride over a more parsimonious, a single-factor view of pride (Williams & DeSteno, 2010).

Holbrook et al. (2014) raised questions about the validity of AP-HP scales for capturing the two-facet concept of pride. The authors argued that the HP scale items (e.g., pretentious) represent the state in which one is aware of the gap between the ascribed status and actual competency, and one is overclaiming success that is not justified. In this sense, the HP scale measures nothing like pride—an emotion arising from success. Across three studies, these authors challenged the assumptions of the two-facet theory of pride: (1) the AP scale correlated with internal attributions including effort, ability, and stable causes; and (2) the HP scale correlated with external attribution such as luck. These findings indicate that people’s attributions of success do not align with the attributional antecedents (i.e., stable vs. unstable) proposed by Tracy and Robins (2007). Internal attributions of personal success, which presumably produce authentic pride, were not divorced from stable causes like natural talent or privilege (Holbrook et al., 2014). This evidence suggests that authentic pride arises whenever a person takes credit for success regardless of stable or unstable causes. By contrast, the highly pejorative terms like pretentious or arrogant of the HP scale represent one’s negative evaluation of unmerited success—accepting the fact that the person received admiration without demonstrating the admirable behavior. Perhaps for this reason, the HP scale correlated with self-deprecating tendencies such as a person’s readiness to accept blame for failure and lower self-esteem (Holbrook et al., 2014).

Holbrook et al. (2014) concluded that something akin to hubristic pride may still exist, but Tracy and Robins’ HP scale does not measure anything about pride at all. If anything, the AP scale measures a more general concept of pride focused on not only effort attribution but also stable attribution including natural talent. By contrast, the HP scale fails to measure a general concept of pride defined by many as: “a positively valenced emotion that occurs in response to success” (Mercadante et al., 2021, p. 130).

Another major criticism against the two-facet theory of pride relates to parsimony. Williams and DeSteno (2010) argued that there is no adequate evidence to warrant the two-facet model of pride over a single-factor view of pride. For instance, work on non-verbal expression of pride suggests that authentic pride and hubristic pride constitute the same facial expression; as a result, people readily recognize authentic pride (Tracy et al., 2005; Tracy & Robins, 2008), but contextual information (the cause of the expression) is critical for people to label it as hubristic pride (Tracy & Prehn, 2012). Williams and DeSteno (2010) argued that this type of evidence does not support the status of hubristic pride as a discrete emotion. They concluded that, if contextual cues are necessary to assist the distinction between authentic and hubristic pride, the two-facet model of pride is inappropriate. They further rejected the view that authentic and hubristic pride represent natural kinds which evolved separately to serve distinct social functions (Tracy et al., 2010).

There is ample evidence linking authentic/hubristic pride to distinct psychological outcomes at the trait level (see Dickens & Robins, 2020). However, there are much less data on the construct validity of state-level hubristic pride, and a lack of experimental evidence on eliciting authentic and hubristic pride. Here, we take on the issue of manipulation checks, which should be used as the foundation to prove causal mechanisms of the very construct a theory conceptualizes (Ejelöv & Luke, 2020; Fiedler et al., 2021). If the two-facet theory of pride is correct, one’s perception of success based on unstable traits (e.g., effort) lead to only authentic pride, and attribution to stable traits (e.g., talent) lead to only hubristic pride. A successful manipulation would demonstrate that manipulation of success to the distinct attributions would elicit the two facets of pride. We do not know of such evidence.

There are only two published series of studies that claim to have manipulated the two facets of pride (Experiment 1, 2, & 3, Ashton-James & Tracy, 2012; Experiment 1 & 2, Ho et al., 2016).¹ Both sets of studies relied on the so-called Relived Emotion Task, in which participants recall their past emotional episodes (Ekman et al., 1983). We highlight several challenges to the method and interpretation of the observed data. First, self-reported levels of authentic pride were always higher than the levels of hubristic pride regardless of the type of recall instructions. In a similar vein, the hubristic-pride manipulation sometimes induced authentic pride just as strongly as did participants asked to recall authentic-pride episodes (e.g., Ho et al., 2016, Experiment 1). These findings imply that, based on participants’ self-reports, hubristic pride was never the more prominent emotion even among participants who were told to recall hubristic-pride episodes.

The authors attributed the low frequency of hubristic pride to social desirability, which presumably suppressed ratings of hubristic pride. To avoid this problem, the authors explicitly encouraged participants to report feelings such as pretentious or stuck-up, admitting that these were socially aversive emotions. Arguably, these instructions may increase demand characteristics. When the experimenter explicitly tells participants to feel certain emotions, it confounds true effects caused by the target construct with an effect caused by the experimenter’s instruction. In the case of hubristic pride, the experimenter’s encouragement to feel “pretentious” might seemingly show an increase in hubristic pride, but the observed effect may be disconnected with any attributional antecedents expected by the theory. Others raised a similar criticism against the Velten procedure in research on the effects of mood (e.g., Larsen & Sinnett, 1991).

Overall, the current evidence of the construct validity of hubristic pride contains a serious limitation. More importantly, the existing manipulation fails to differentiate authentic pride from hubristic pride. Whereas this may be the reason why there are so few studies that have arguably manipulated hubristic pride, this can be also interpreted as favoring the parsimonious view that there is only authentic pride (Williams & DeSteno, 2010).

Good theories should not be abandoned by bad methods. We believe that the two-facet theory of pride merits further scrutiny, but there has not been a systematic investigation to assess the construct validity of experimental inductions of authentic and hubristic pride. We provide such an investigation with two specific goals.

The first goal was to consider various manipulation methods other than the Relived Emotion Task for inducing hubristic pride independently from authentic pride, establishing an alternative method that spontaneously manipulates specific attributional processes as proposed by the two-facet theory of pride (Tracy & Robins, 2007). Study 1 aimed to replicate and extend previous findings obtained using the Relived Emotion Task (Ashton-James & Tracy, 2012; Ho et al., 2016). Given the inconsistent findings and limitations of Study 1, we subsequently conducted three studies to explore alternative manipulation methods. Study 2 adopted a false-feedback paradigm of pride induction (Williams & DeSteno, 2008), in which we varied the cause of participants’ success in an interpersonal perception task to be either unstable (effort) or stable (talent). Study 3 and 4 modified Study 2 by including a leadership aptitude test (Fast & Chen, 2009) and examined the presence of social desirability by asking participants to imagine how others would react to the same emotion-induction test. Study 2-4 were exploratory in the hope of finding the strongest manipulation of hubristic pride; however, at every step, our method of induction was informed by theory with a priori predictions.

A successful manipulation based on theory-driven induction method should represent a test of the theory itself (Fiedler et al., 2021). A liberal standard for a successful manipulation—adopted by prior research—requires that both manipulations cause an increase in the target facet of pride at least relative to the control condition. That is, the hubristic pride manipulation should increase hubristic pride relative to the control condition, while the same manipulation may increase authentic pride (Figure 1a). A more conservative standard—which we aimed to discern—requires that an experimental manipulation increases only one facet of pride relative to the other pride: e.g., the hubristic-pride manipulation should only increase hubristic pride but should not increase authentic pride relative to the authentic-pride condition (Figure 1b).

The second goal of our investigation was to document validity evidence of state-level authentic/hubristic pride. The ongoing debate revolves around the status of hubristic pride as a pride emotion (Holbrook et al., 2014; Williams & DeSteno, 2010). We contribute to the debate by documenting the “nomological shockwaves” of the manipulation of authentic/hubristic pride—the degree to which an experimental manipulation causes expected changes in relevant constructs relative to comparison conditions (Chester & Lasko, 2021). In our research, we provide such investigations by following the recommendation by Chester and Lasko (2021). In addition, we conducted standard nomological network analyses of state-level pride by computing correlations with other relevant emotion items.

Our first study directly pitted Tracy and Robins’ (2007) two-facet theory of pride against Holbrook’s alternative account of hubristic pride (2014) using the Relived Emotion Task. All manipulations in Study 1 followed the same format of the Relived Emotion Task by Ashton-James and Tracy (2012) and Ho et al. (2016) and manipulated the attributions of success episodes that participants were supposed to recall. We then subjected all manipulations to the assessment of the nomological shockwaves (Chester & Lasko, 2021) by assessing changes in the state-level authentic/hubristic along with other emotion variables. All data, analysis codes, and Appendices are available via OSF (https://osf.io/9rfby/).

We aimed to detect an effect size of d = 0.49, the smallest effect size we were able to find concerning the experimental induction of hubristic pride (versus authentic pride) in previous research (Ashton-James & Tracy, 2012; see also Appendix, Table A5). Assuming α = .05, and β = .80, the required sample size was 134 for a two-group comparison. Allowing for a small number of unusable cases, we aimed to collect at least 70 participants for each condition. Whereas this calculation was merely based on a two-group comparison (hubristic pride condition vs. authentic pride condition), when applied to our four-cell design of Study 1, we needed a sample of 180 participants to detect an effect of Cohen’s f = 0.25.

Participants were undergraduate students at a large Western U.S. public university. We collected a total of 337 participants, eight of which were excluded for not providing open-ended responses in the Relived Emotion Task. Participants (70% female) were randomly assigned to either authentic-pride (n = 83), Tracy and Robins’ hubristic-pride (n = 84), Holbrook’s hubristic-pride (n = 78), or control (n = 84) conditions. Students received partial course credit upon participating in the online survey. Participant demographics are summarized in the Appendix (Table A1).

The Relived Emotion Task. The original study by Ekman et al. (1983) asked professional actors and emotion scientists to relive a past emotional experience for 30 seconds. This procedure has been adopted as a common emotion induction method in emotion research. All of our manipulations followed the same format used by Ashton-James and Tracy (2012) and Ho et al. (2016). The Appendix provides full descriptions of the manipulations used in Study 1 (Table A3). The manipulation of authentic pride asked participants to write down an authentic-pride inducing episode in which they were doing really well as a result of their efforts and felt good without feeling superior to others. Our manipulation of hubristic pride asked participants to write down a hubristic-pride inducing episode in which they accomplished something while not having to work hard and felt pretentious or stuck-up. A newly generated manipulation of hubristic pride inspired by Holbrook et al. (2014) asked participants to write down an episode in which they accomplished something but they found that they received credit because of something other than themselves—not due to their talent, ability, nor effort. Instructions in the control condition asked participants to share their typical day as a college student and write down everything they had done that day.

Self-rated Pride Feelings. We assessed authentic pride by aggregating seven items—accomplished, successful, achieving, fulfilled, self-worth, confident, and productive $(α=$ .94, $ωTotal$ = .96); hubristic pride by aggregating seven items—snobbish, pompous, stuck-up, conceited, egotistical, arrogant, and smug $(α=$ .92, $ωTotal$ = .95) (Tracy & Robins, 2007). We also conducted a Confirmatory Factor Analysis (CFA) on this scale using R’s lavaan package (Rosseel, 2012). The CFA result supported the two-factor model of pride (see the Appendix Study 1 for the model fit indices).

Filler Items. We included other filler items to quantify “the nomological shockwaves” (Chester & Lasko, 2021) of the manipulation of authentic/hubristic pride. Following previous research on theoretical predictions on the expected correlations with authentic/hubristic pride (see the Appendix Table A2), we added other relevant items. Specifically, we added items from the Negative Affect and the Positive Affect scales of the PANAS-X (Watson & Clark, 1999) and several items that were applicable yet absent from the PANAS-X (aggressive, intellectual, creative, and curious). All ratings were assessed on a 7-point Likert-type scale. The complete list of items is detailed in the Appendix Table A3. For data reduction, we first ran an exploratory factor analysis with oblimin rotation across all conditions. Consistent with Watson and Clark (1999), our solution suggested a two-factor solution replicating PANAS-X: “Positive Affect (enthusiastic, interested, determined, excited, inspired, alert, active, powerful, creative, $α=$ .92, $ωTotal$ = .94)” and “Negative Affect (guilty, ashamed, aggressive, hostile, $α$ = .75, $ωTotal$ = .85).”

Personality Measures. For exploratory and validation purposes, we assessed individual differences using the ten-item Big Five Inventory (Rammstedt & John, 2007) and the 13-item Narcissistic Personality Inventory (Gentile et al., 2013) in Study 1 (and in all studies reported here, and all results were highly similar). We found pattens consistent with previous research: individuals who felt more authentic pride in this experimental setting were more narcissistic (r = .29, p < .001), conscientious (r = .36, p < .001), extraverted (r = .21, p < .001), agreeable (r = .11, p = .045), and less neurotic (r = -.30, p < .001); individuals who felt more hubristic pride were more narcissistic (r = .28, p < .001), extraverted (r = .12, p = .034), and less agreeable (r = -.12, p = .024). We do not report further analyses regarding individual differences in the present article, but we made personality data available to interested readers (see our OSF repository).

Upon consenting to the online survey, all participants were assured that (1) the purpose of the study was to assemble an extensive catalog of emotional experiences and (2) their responses would remain anonymous. Participants then received one of the manipulation instructions. Participants were encouraged to describe the event “as vividly as possible in at least 100 words.” After that, they rated their current feelings, answered a battery of demographics and personality questionnaires, and were debriefed.

We conducted a 2 within (Pride Rating: Authentic vs. Hubristic) x 4 between (Instruction Condition: Authentic vs. Hubristic [Tracy & Robins] vs. Hubristic [Holbrook et al.] vs. Control) mixed-factorial analysis of variance (ANOVA) on the rating of pride. Both the main effects of Pride Rating and Instruction Condition were significant, F(1, 325) = 817.12, p < .001, $ηp2$ = .72, and F(3, 325) = 3.36, p = .019, $ηp2$ = .03, respectively. In general, participants reported far greater authentic pride than hubristic pride regardless of condition, t(328) = 28.32, p < .001, M_diff = 2.55, 95% CI [2.38, 2.73] (see Figure 2).

We also found a significant Pride Rating x Instruction Condition interaction, which suggested that the main effect of Instruction Condition was qualified by the type of pride, F(3, 325) = 3.05, p = .029, $ηp2$ = .03. For the rating of authentic pride, there was no significant difference across conditions, F(3, 325) = 1.41, p = .241. Notably, participants in the two Hubristic conditions reported feelings of authentic pride just as strongly as those in the Authentic condition and in the Control condition. By contrast, we found significant differences across conditions on the rating of hubristic pride, F(3, 325) = 7.86, p < .001, $η2$= .07. Participants in the Hubristic condition adapted from Ashton-James and Tracy (2012), Ho et al. (2016), and Tracy and Robins (2007) showed the highest ratings of hubristic pride (M = 2.04) compared to the other three conditions: the Authentic condition (M = 1.53), the Hubristic condition (M = 1.55) inspired by Holbrook et al. (2014), and the Control condition (M = 1.42), ps < .001.

These findings provide mixed support for the effectiveness of the Relived Emotion Task as a tool to induce the two facets of pride. On the one hand, the method differentiated hubristic pride between conditions. Yet, this outcome may have depended on instructions explicitly asking participants to feel “pretentious” or “stuck-up,” with the same or similar terms being part of the dependent variables. Despite this explicit instruction, mean levels of hubristic pride were generally low across all conditions. On the other hand, and perhaps more critically, the Relived Emotion Task failed to induce different levels of authentic pride across conditions; moreover, authentic pride was always greater than hubristic pride regardless of condition. This general pattern means that the instruction to feel “pretentious” and “stuck-up” inadvertently induced authentic pride, and that the method failed to distinguish the two facets of pride—with the exception that in the hubristic pride condition self-reports of hubristic pride were slightly higher than in the authentic pride condition. See Table 1 for descriptive statistics for all conditions (and for all studies).

Nomological Shockwaves. We next assessed the nomological shockwaves of our manipulation of authentic/hubristic pride in two ways. First, we compared means and effect sizes of Positive Affect and Negative Affect in the target experimental conditions with those in the Control condition (see Table 2). Second, we compared item-level effect sizes of all filler items and visualized the patterns in one figure (see Figure 6 for all studies). Any increase or decrease observed in the target experimental conditions relative to the control condition can infer “shocks” caused by the manipulations (Chester & Lasko, 2021).

Though none of the observed effect sizes reached statistical significance (ps > .05), they were generally in the expected directions. The Authentic condition increased Positive Affect (d = 0.32) and decreased Negative Affect (d = -0.17). The Hubristic condition adapted from Ashton-James and Tracy (2012) and Ho et al. (2016) slightly increased both Positive Affect (d = 0.27) and Negative Affect (d = 0.22), suggesting that the recalled experience of hubris generated mixed emotions. More specifically, at the item level we observed that this condition slightly increased ratings of hostile and aggressive. The Hubristic condition inspired by Holbrook et al. (2014) did not result in any effects. Overall, we conclude that the Relived Emotion Task generally produced the expected nomological shockwaves for authentic pride, but the effects were unreliable and small.

Nomological Networks. Next, we assessed correlations between the ratings across all conditions (see Table 3; see Figure 7 for item-level correlations). Consistent with prior research, authentic pride positively correlated with Positive Affect, r = .73, but negatively with Negative Affect, r = -.19. Hubristic pride showed mixed patterns: it positively correlated with Positive Affect, r = .18, but more strongly with Negative Affect, r = .47. Figure 7 more clearly shows that hubristic pride correlated most strongly with negative items such as hostile and aggressive.

Study 1 showed that there were indeed small, but significant differences in the extent to which different manipulations elicited self-ratings of hubristic pride. In the condition with instructions derived from Ashton-James and Tracy (2012) and Ho et al. (2016), there were slightly elevated levels of hubristic pride compared to the other conditions. Yet, this manipulation unexpectedly induced authentic pride as well: there was no difference in levels of self-rated authentic pride across all experimental conditions. This finding fails to support the most liberal standard of successful manipulation illustrated in Figure 1a. In spite of the fact that we replicated some of the earlier findings, the Relived Emotion Task raises a potential concern about the process underlying these ratings—namely an inability to differentiate the intensity of felt authentic pride ratings between experimental conditions.

We found inconsistent support for the nomological networks of state-level hubristic pride suggested by the theory of Tracy and Robins (2007). Authentic pride covaried with positive affect, and this means that authentic pride represents genuine positivity. However, hubristic pride covaried with negative affect such as hostility and aggression more strongly than with positive affect. What emerges is a complex picture of hubristic pride as a pride emotion. If pride is a positively valenced emotion and occurs in response to success, why would one feel hostile and aggressive? The instruction to feel “pretentious” in the Hubristic condition adapted from Ashton-James and Tracy (2012) and Ho et al. (2016) may have introduced unexpected variability in participants’ answers to the instruction for various reasons. We sought to provide more insights into this puzzle in the subsequent studies.

Given the inconsistent findings of Study 1, we turned to developing a standardized task using a false-feedback paradigm, in which participants’ experience is ostensibly manipulated by means of positive/negative/control feedbacks that are independent of participant’s objective performance. Our paradigm was inspired by Williams and DeSteno (2008) who administered a dot estimation task, in which participants counted the number of dots on a screen with manipulation of positive feedback on accuracy. Participants who thought they did well were more likely to feel pride. We sought to make our performance task more engaging by adopting an interpersonal perception task by Anderson et al. (2012), where participants were shown video clips of interpersonal interactions and then asked to predict characteristics of the people in the materials. Anderson et al. (2012) induced overconfidence among participants by presenting false information as to the accuracy of their predictions, generating positive/negative self-evaluations. We supplemented the interpersonal perception task by adopting a feedback protocol used by Mahadevan et al. (2019), who manipulated state narcissism by providing visual information about social rank relative to the normative standard.

Arguably, a standardized task such as the interpersonal perception task compensates for a notable limitation of the Relived Emotion Task. Writing about one’s past experience should not equate to manipulating antecedents of pride. Ideally, an experimenter should manipulate the actual process by which participants achieve something. A critical element of the Relived Emotion Task is that participants report feelings of pride in a variety of contexts, and not all pride-inducing episodes are directly related to status attainment. It is possible that, as we found in our Study 1, participants may write about circumstances where they felt aggressive and hostile in response to achievement. To the extent that pride is associated with status (Henrich & Gil-White, 2001), and to the extent that success in a performance situation does indeed convey success, the Relived Emotion Task produces unwelcome variability in our quest to find valid and effective means of eliciting pride. By contrast, the interpersonal perception task can control specific processes by which participants attain success relative to others with a clear visual aid.

Participants in both the authentic pride condition and the hubristic pride condition received the same performance feedback but were led to attribute their success (a high score compared to other participants) to either malleable or stable personal characteristics. Participants in the control condition received feedback that their performance was average. As in Study 1, we explored the nomological shockwaves of the manipulations and nomological networks with a host of theoretically relevant constructs.

Because this manipulation was novel, we did not have prespecified effect size comparable to previous research. We therefore aimed to detect a medium effect size with α = .05, and β = .80. When applied to a three-cell design, the required sample size was 158. We collected a total of 182 participants, five of whom were excluded due to missing data. Participants (53% female) were undergraduate students at a large Western U.S. public university who were randomly assigned to either authentic (n = 59), hubristic (n = 60), or control (n = 58) conditions. Participants received partial course fulfillment upon participating in the online survey. Participant demographics are summarized in Appendix, Table A1.

Interpersonal Perception Task. The cover story read that the research was about developing an interpersonal perception task to predict students’ career success. Participants were assured that they would receive an accuracy score based on the match between their answers and actual answers we had collected from individuals displayed in the photos. In the first task, participants saw a series of 24 facial photos and predicted the age of each individual. In the second task, they saw a series of eight facial photos and judged the personality of each individual. In the third task, participants saw a series of eight facial photos and rated how these individuals might have responded to ten value ratings.

Face Photos. The Chicago Face Database (Ma et al., 2015) provides a set of high-resolution, standardized photographs of human faces. Photos included in this database have norming data for varying attributes such as race, age, attractiveness, or prototypicality. We selected 24 photos that vary on race (White vs. Black vs. Asian vs. Latino), gender (male vs. female), and percentile on attractiveness (49th vs. 50th vs. 51st). We provide these photographs in the Appendix.

Personality Ratings. Personality items were adopted from a 10-item measure of the Big Five personality dimension (Gosling et al., 2003). The instruction was modified so that item descriptions were directed towards the facial photos. Participants rated the personality of each person the photographs using a 7-point scale.

Value ratings. Value items were adopted from Schwartz (1992). For the sake of brevity, only value contents that scored the highest importance for each dimension in the analysis of Schwartz (1992) were adopted. Participants rated values of each person using a 5-point scale.

False-feedback of Interpersonal Perception. Participants received varying feedback on the interpersonal perception task, depending on the experimental condition to which they were assigned. We manipulated participants’ pride feelings by changing stable vs. unstable descriptions about their strengths assessed by the interpersonal perception task. In both authentic pride and hubristic pride conditions, participants’ achievements were attributed to distinct aspects of their performance. We referenced previous research to identify several key dimensions on which authentic pride should differ from hubristic pride (see Appendix Table A4). Among them, we aimed to tap into four dimensions: (1) antecedent (behavior vs. self), (2) attribution (unstable vs. stable; effort vs. talent), (3) communion (high vs. low), and (4) competency (skill vs. talent). We predicted that success as a result of unstable characteristics would elicit authentic pride, while stable characteristics would elicit hubristic pride. At the end of the feedback, we ostensibly offered participants an extra opportunity to participate in an interview about their performance as an opportunity for them to get involved in future research. This additional message was meant to increase realism and elicit a sense of social approval and a clear signal that others are impressed with their high score. The structure of descriptions for the control condition were kept in line with those of pride conditions to maintain consistency and comparability. Our Appendix provides each feedback in detail.

Self-rated Pride Feelings. Again, we used Tracy and Robins’ (2007) scale of authentic pride (⁠$α$ = .92, $ωTotal$ = .94) and hubristic pride (⁠$α$ =.88, $ωTotal$ = .91). As before, we conducted a CFA, which supported the two-factor model with a good fit similar to Study 1.

Filler Items. We extended a list of filler items in addition to those used in Study 1. We added hubristic-related items and additional items related to negative affect of PANAS-X (Watson & Clark, 1999). We ran an exploratory factor analysis and retained a three-factors solution with oblimin rotation: we replicated the two-factors solution by Watson and Clark (1999) and our own Study 1. But for the present purpose, we deemed a three-factors solution to be more adequate because it allowed us to differentiate an Insecure Affect factor (afraid, scared, nervous, jittery, guilty, ashamed, upset, distressed, $α$ = .89, $ωTotal$ = .92), and a Hostile Affect factor (irritable, hostile, aggressive, $α$ = .77, $ωTotal$ = .80) along with a Positive Affect (gifted, exceptional, enthusiastic, interested, determined, excited, inspired, alert, active, intellectual, creative, curious, $α$ = .92, $ωTotal$ = .93).

Participant first worked on the interpersonal perception task and then received the performance feedback. Subsequently, participants rated their current feelings. Following some demographic questions, they were thanked and debriefed.

We conducted a 2 within (Pride Rating: Authentic vs. Hubristic) x 3 between (Feedback Condition: Unstable vs. Stable vs. Control) mixed-factorial ANOVA on the rating of pride. Both the main effects of Pride Rating and Feedback Condition were significant, F(1, 174) = 585.52, p < .001, $ηp2$ = .77, and F(2, 174) = 8.19, p < .001,$ηp2$ = .09, respectively. In general, participants reported far greater authentic pride than hubristic pride regardless of feedback conditions, M_diff = 2.68, 95% CI [2.45, 2.91].

The significant main effect of Feedback was qualified by a Pride Rating x Feedback Condition interaction, F(2, 174) = 11.27, p < .001, $ηp2$ = .11, where both attribution feedbacks significantly induced authentic pride but not hubristic pride. For the rating of authentic pride, we found significant differences across feedback conditions, F(2, 174) = 14.25, p < .001, $ηp2$ = .14. Participants in the Unstable feedback condition (M = 4.81) and the Stable feedback condition (M = 4.80) showed higher ratings of authentic pride than those in the Control feedback condition (M = 3.72), ps < .001. However, there was no significant difference between the two feedback conditions, t(174) = 0.02, p = .999. For the rating of hubristic pride, however, there were no significant differences across conditions, F(2, 174) = 1.10, p = .336. Across feedback conditions, mean levels of hubristic pride were fairly low (see also Figure 3). In contrast to Study 1, Study 2 was successful in inducing authentic pride but not hubristic pride.

Nomological Shockwaves. We compared effect sizes of all ratings in the experimental conditions with those in the Control condition (see Table 2 and Figure 6). Both the Unstable and Stable feedback conditions produced a significant increase in Positive Affect, d = 0.68 and d = 0.66, respectively. The Unstable feedback produced a significant decrease in Insecure Affect and Hostile Affect, d = -0.76 and d = -0.49, respectively. On the other hand, the Stable feedback produced weaker decreases in Insecure Affect and Hostile Affect, d = -0.49 and d = -0.18, respectively. Figure 6 further corroborates these findings by showing that negative items such as ashamed and guilty had the highest (negative) effect sizes. This suggests that those who received average feedback in the Control condition felt greater negative affect. Taken together, the Unstable manipulation designed to induce authentic pride reliably produced expected changes in positive and negative affect. However, the Stable manipulation designed to induce hubristic pride only increased positive affect but did not reliably increase negative affect, highlighting the difficulty of inducing negative affect by positive feedback.

Nomological Networks. We next assessed correlations between the ratings (see Table 3 and Figure 7). Consistent with Study 1, authentic pride correlated positively with Positive Affect, r = .82, but negatively with both Insecure Affect (r = -.34) and Hostile Affect (r = -.10). By contrast, hubristic pride produced more nuanced correlations. Hubristic pride did show a positive correlation with Positive Affect (r = .21), but more strong correlations with Insecure Affect (r = .37) and with Hostile Affect (r = .51). Notably, hubristic pride correlated more strongly with aggressive and hostile than with guilty and ashamed (see Figure 7). These correlations highlight the socially aversive nature of hubristic pride. Overall, our analyses of the nomological networks of authentic/hubristic pride revealed consistent findings with Study 1.

Study 2 showed that our manipulation was successful with regard to the induction of authentic pride, overcoming the limitation of Study 1. Compared to the Relived Emotion Task in Study 1, the interpersonal perception task generated a greater intensity of authentic pride, perhaps because of the standardized procedure. However, the interpersonal perception task did not induce a strong sense of hubristic pride, and the ratings of authentic pride were not separable across the two attribution manipulations. Again, Study 2 failed to meet the most liberal standard of successful manipulation illustrated in Figure 1. Moreover, the average feedback in the Control condition produced a greater intensity of negative affect, implying that participants’ lower status were made salient with clear visual information of the social rank.

Study 2 enabled us to assess the nature of hubristic pride via nomological shockwaves and networks. A notable insight obtained from Study 2 is the disconnect between the nomological shockwaves and networks with regard to hubristic pride. At the item level of nomological networks, hubristic pride covaried with a range of negative items such as hostile, aggressive, shamed, and guilt (Figure 7). However, it was clear that positive feedback produced a decrease in those negative items: people’s experiences with personal success were clearly positive regardless of attribution. The results of Study 2 imply that, because hubristic pride correlates with both positive and negative emotions, the kind of event that produces strong hubristic pride must increase one’s status but also elicit socially aversive experience, and people must admit to such negative feelings. Although we are not aware of a type of social situation that would typically induce both status and negative feelings, Study 3 and 4 sought to strengthen and replicate findings of Study 2.

Study 3 aimed to address three points: (1) intensity of hubristic pride rating; (2) inability to differentiate authentic pride from hubristic pride; and (3) social desirability of reporting hubristic pride. To address the first two points, we modified the interpersonal perception task used in Study 2 to be more explicit about social rank in the hope of increasing intensity of hubristic pride and even greater aggression and hostility through dominance-related feedback. We did so by adopting multiple tasks in the context of a bogus leadership aptitude test—a task frequently used to manipulate sense of power (e.g., Fast & Chen, 2009, Study 3). In Study 3, participants performed tasks similar to the interpersonal perception task in Study 2, but the feedback related to participants’ leadership aptitude. We linked this manipulation to the literature on prestige and dominance, assuming that dominance (prestige) gives rise to hubristic (authentic) pride (Witkower et al., 2021). We reasoned that more grandiose feedback would be necessary to make participants admit feeling a high level of hubristic pride and some degrees of hostile emotions than the previous procedures. We also inserted multiple feedbacks between tasks to ensure that participants did feel competent in the experimental conditions. We assumed that feedback on dominant leadership in this way would elicit greater hubris (and greater hostility and aggression) than would a mild compliment on one’s stable characteristics as in the case of Study 2. We expected that participants who were shown to have a dominant-leadership to experience hubristic pride and that participants who were shown to have prestige-leadership to experience authentic pride.

To address the third point, Study 3 included a novel element to detect social desirability against reporting hubristic pride. We built upon a prior work that participants were generally willing to infer hubristic-pride expressions of others (Wubben et al., 2012). In Study 3, we asked our participants in one condition to predict how others would feel in the same situation irrespective of how they themselves feel. If social desirability is unique to ratings of hubristic pride, we should expect to see differences between self-ratings and others-ratings of hubristic pride but not of authentic pride.

Participants were recruited at a large Western U.S. public university and were randomly assigned to either Prestige x Self-rating (n = 54), Prestige x Others-rating (n = 52), Dominance x Self-rating (n = 75),² Dominance x Others-rating (n = 49), Control x Self-rating (n = 49), or Control x Others-rating (n = 49) conditions. With this total sample size of 328 (70% female) participants, we were able have .99 power to detect an effect of f = .25, assuming an α = .05. Participants received partial course fulfillment upon participating in the online survey. Participant demographics are summarized in Table A1 of the Appendix.

Interpersonal Perception Task. We used the same photo materials and personality scales as Study 2 for the interpersonal perception task.

Leadership Competency Task. Using the same photos as Study 2, participants judged leadership competency of each photo based on the leadership style questionnaire by Van Vugt and Ahuja (2011) and two seven-point bipolar scales: “Likely to be a follower–Likely to be a leader,” and “Cooperative–Competitive.”

Emotion Recognition Task. We also added a novel task, which ostensibly assessed participants’ emotional recognition skills using facial photos provided by FACES (Ebner et al., 2010). Participants were presented with a sequence of six facial photos, each with one of six emotional expressions (anger, disgust, fear, happy, sad, control), an order randomized for each individual participant. For each of the six emotions, we had available six different facial images, which varied by age (young, middle, old) and gender (male, female), all showing a white individual. For each of emotional display, we selected one of these six images at random. With each photo, participants were presented with a scale that read, “Which emotion best describes the expression?” and selected their answer from one of the six response alternatives among Angry, Disgusted, Fearful, Happy, Sad, Surprised, and Neutral.

False-feedback of Leadership Aptitude. Similar to Study 2, participants received performance-independent feedback on the leadership aptitude, depending on the experimental condition to which they were assigned. We sought to manipulate participants’ pride feelings by changing descriptions of the type of leadership style on which participants had a high score. As in Study 2, we referenced previous research to identify several key dimensions on which prestige should differ from dominance in relation to pride experiences (see Table A3 of Appendix). The prestige feedback emphasized leadership associated with respect from others, wisdom, skill, and great interpersonal skill; the dominance feedback emphasized assertiveness, directiveness, authority, talent, and intimidation; the control condition emphasized average performance on these traits. See the Appendix, Study 2 for complete descriptions of each feedback.

Self-rated Pride Feelings. Again, we used Tracy and Robins’ (2007) scale of authentic pride $(α$ = .97, $ωTotal$ = .98) and hubristic pride $(α$ = .96, $ωTotal$ = .96).

Filler Items. Items were identical to those used in Study 1. We ran an exploratory factor analysis and again obtained a two-factor solution with oblimin rotation: Positive Affect (enthusiastic, interested, determined, excited, inspired, alert, active, powerful, creative, $α=$ .93, $ωTotal$ = .96) and Negative Affect (guilty, ashamed, aggressive, hostile, $α$ = .84, $ωTotal$ = .90).

Our cover story read that the research was about assessing a skill to accurately read others, a highly critical skill in today’s diverse society. Participants were told that their performance score would be provided to them at the end, based on three tasks: judging the personality of others, judging the leadership competency of others, and emotion recognition ability.

In the first task, participants saw a series of eight facial photos and predicted the personality of each individual shown in the photos. In the second task, they saw a sequence of eight facial photos and judged leadership competency by guessing how the individuals in the photos had rated themselves using the same scales. Once participants had completed each of these two tasks, they received brief feedback which was positive (or neutral) and unrelated to their actual performance. In the third task, participants saw a series of eight facial photos and rated how the individuals were feeling. At the end of the third task, participants were received false feedback: “Your overall accuracy for emotion recognition: Excellent (Average).”

We then explained to participants that the next page would display their leadership aptitude in comparison with the last three participants who took the survey. After a pause of 60 seconds, allegedly for the system to compile a summary report, participants received a one-page report of their leadership aptitude (see Appendix for details).

Immediately after receiving this feedback, participants rated their feelings. In the self-rating condition, participants rated their current moods in the same way as Study 1 and 2. In the other-rating condition, participants read the following instruction:

They then answered a battery of demographics and personality questionnaires and were debriefed.

We conducted a 2 within (Pride Rating: Authentic vs. Hubristic) x 3 between (Leadership Condition: Prestige vs. Dominance vs. Control) x 2 between (Reference: Self vs. Others) mixed-factorial ANOVA on the rating of pride. Similar to Study 1 and 2, participants reported greater authentic pride than hubristic pride regardless of leadership conditions and references, F(1, 322) = 820.20, p < .001, $ηp2$ = .72, M_diff = 2.81, 95% CI [2.59, 3.02]. This main effect was qualified by a Pride Rating x Leadership Condition interaction, F(2, 322) = 17.25, p < .001, $ηp2$ = .10. The difference between authentic pride and hubristic pride was significant in the Control condition, t(97) = 11.94, p < .001, M_diff = 1.93, 95% CI [1.61, 2.25], but the difference was much greater in the Prestige and Dominance conditions, t(105) = 16.99, p < .001, M_diff = 3.21, 95% CI [2.84, 3.59], and t(123) = 17.72, p < .001, M_diff = 3.17, 95% CI [2.81, 3.52], respectively.

We found a main effect of Leadership Condition on the rating of pride, F(2, 322) = 73.07, p < .001, $ηp2$ = .31. The Prestige condition induced greater authentic pride than the Control condition, M_diff = 2.01, p < .001, 95% CI [1.57, 2.44], but not greater than the Dominance condition, M_diff = 0.05, p = .942, 95% CI [-0.35, 0.47]. Similarly, the Prestige condition induced greater hubristic pride than the Control condition, M_diff = 0.72, p < .001, 95% CI [0.26, 1.19], but not greater than the Dominance condition, M_diff = 0.01, p = .997, 95% CI [-0.43, 0.45]. Overall, the Prestige and the Dominance conditions did not differ from each other in inducing both types of pride. These results suggest that the current feedback manipulations were successful for inducing both types of pride relative to the Control condition even though the distinction between Prestige and Dominance did not distinguish the two types of pride.

Our central concern in Study 3 is how reference of rating changes the reporting of pride experience. The main effect of Reference was significant, F(1, 322) = 51.63, p < .001, $ηp2$ = .14, M_diff = 0.66, 95% CI [0.35, 0.98]. This main effect reveals that when participants rated how others would feel in response to the leadership feedback they received, they reported greater pride regardless of the types of pride and conditions. This effect was not moderated by the Leadership Condition effect, F(2, 322) = 1.32, p = .268. More importantly, we found a significant Pride Rating x Leadership Condition x Reference interaction effect, F(2, 322) = 8.05, p < .001, $ηp2$ = .05. This three-way interaction suggests that the Reference x Leadership Condition interaction effect depended on the type of pride being evaluated. For the rating of authentic pride, the Leadership Condition x Reference interaction was not significant, F(2, 322) = 1.37, p = .255, implying that the nature of reference did not depend on the experimental condition. However, the Leadership Condition x Reference interaction was significant for the rating of hubristic pride, F(2, 322) = 7.98, p < .001, $ηp2$ = .05. In both the Prestige and the Dominance conditions, participants’ estimates of others’ hubristic pride were significantly higher than those who rated their own hubristic pride in response to the same leadership feedback: M_diff = 1.39, p < .001, 95% CI [0.82, 1.95], and M_diff = 1.69, p < .001, 95% CI [1.17, 2.21], respectively. However, the Reference effect on the rating of hubristic pride was minimal in the Control condition, M_diff = 0.36, p = .058, 95% CI [-0.01, 0.73]. In contrast with the rating of authentic pride, we observed a stronger effect of reference on the rating of hubristic pride. That is, participants’ ratings of hubristic pride systematically differed based on the type of leadership feedback they received and whom they were evaluating. The fact that we did not find such a stark effect of reference on the rating of authentic pride supports our prediction: hubristic pride carries a more negative connotation, which participants did not readily report, than does authentic pride.

Nomological Shockwaves. As with Study 1 and 2, we compared effect sizes of all self-ratings in the target leadership conditions with those in the Control condition (see Table 2 and Figure 6). Patterns of effect sizes were highly similar across the experimental conditions. Both the Prestige and Dominance conditions produced a significant increase in Positive Affect: d = 1.54; and d = 1.61, respectively. Both conditions produced a small decrease in Negative Affect, d = -0.14; and d = -0.18, but these effects did not reach statistical significance (ps > .27). This pattern suggests that those who received average feedback in the Control condition did not feel particularly negative such as ashamed and guilty. Figure 6 shows that the observed effect sizes in positive items in Study 3 were generally greater than those found in Study 2, supporting a greater intensity of the current manipulation. More specifically, the Dominance condition produced slightly greater effect sizes than the Prestige condition across all items. At the item level, the Dominance condition increased aggressive slightly, but this effect was not significant. Taken together, the manipulation of status, whether it was prestige or dominance, produced intense positive affect but not negative affect.

Nomological Networks. The analysis of nomological networks generated clear convergence with Study 1 and 2 (see Table 3 and Figure 7). Authentic pride correlated with Positive Affect, r = .83, but not with Negative Affect, r = -.13. Hubristic pride correlated only weakly with Positive Affect, r = .24, but more strongly with Negative Affect, r = .45. Item-level correlations more clearly illuminated the nature of hubristic pride as a mixed emotion as it correlated with items such as hostile, aggressive, ashamed, guilty, and superior (Figure 7).

Study 3 sought to develop a strong manipulation of hubristic pride by giving participants blatant feedback on their dominance-leadership aptitude. This manipulation did not produce a strong distinction between self-ratings of hubristic pride between conditions but induced a greater level of authentic pride than hubristic pride. Apparently, participant were flattered whether their high performance indicated dominance characteristics or not. This result again fails to meet the most liberal standard of successful manipulation illustrated in Figure 1.

The analyses of nomological shockwaves and networks showed patterns similar to those in Study 1-2. When participants received positive feedback on leadership aptitude, they felt more positive than negative, with the dominance-leadership feedback showing slightly stronger effects than the prestige-leadership feedback. By means of positive feedback, it was unlikely to observe any increases in negative feelings, which should covary with hubristic pride.

Though our manipulation was not successful, we clarified a potential roadblock to finding hubristic pride: participants’ ratings differed systematically depending on whether they rated how others would react to the same feedback and the type of feelings they rated. When participants were told to indicate their own feelings, they did not necessarily report intense levels of hubristic pride; yet, participants predicted that others would experience much greater hubristic pride in response to the same leadership feedback. This differential responding was stronger for the leadership feedback conditions compared to the control condition. This type of interaction effect was not observed for the ratings of authentic pride, where participants were more willing to report greater ratings of authentic pride regardless of the feedback conditions. Our results potentially illuminate social desirability of suppressing socially aversive feelings such as hubristic pride.

Study 4 aimed to replicate Study 3 to further confirm robustness of social desirability of reporting hubristic pride. Study 4 was identical to Study 3 except that Study 4 was a 2 x 2 mixed-model design: a between-subject condition (Leadership Condition: Prestige vs. Dominance) and a within-subject condition (Reference: Self vs. Others). Because our central focus was on differentiating prestige and dominance (and attendant forms of pride), unlike Study 3, the present study did not include a control condition. All participants first rated their own feelings and subsequently imagined how others would react to the same leadership aptitude feedback.

We had a total of 114 participants, seven of whom were excluded due to missing data. Study 4 was based on a 2 between (Leadership Condition: Prestige vs. Dominance) x 2 within (Reference: Self vs. Others) mixed design: participants were randomly assigned to either a Prestige (n = 57) or Dominance (n = 50) condition. Assuming an α = .05, this allowed us to have .99 power to detect an interaction of f = .25. Participants (69% female) received partial course credit upon participating in the online survey. Participant demographics are summarized in Appendix, Table A1. The study was conducted at a large Western U.S. public university.

All materials were identical to those used in Study 3.

The procedure was identical to Study 3 with one exception. Whereas in Study 3 self and other-ratings were manipulated between participants, in the present investigation participants rated both their own feelings and, subsequently, imagined how hypothetical others would feel. Immediately after receiving the leadership aptitude feedback, participants rated their own feelings first. Participants did not receive any advance warning that they would be rating others’ feelings as well, which followed immediately afterwards. We did not counterbalance the order of rating because we did not want their other-ratings to influence their own ratings.

We performed a 2 within (Pride Rating: Authentic vs. Hubristic) x 2 within (Reference: Self vs. Others) x 2 between (Leadership Condition: Prestige vs. Dominance) mixed-factorial ANOVA on the rating of pride (see Table 1). The main effect of Leadership Condition was not significant, F(1, 105) = 1.03, p = .314. Both the main effects of Pride Rating and Reference were significant: F(1, 105) = 314.50, p < .001, η_p² = 0.75, and F(1, 105) = 25.36, p < .001, η_p² = 0.19, respectively. Participants’ ratings were greater for authentic pride compared to hubristic pride, M_diff = 3.05, p < .001, 95% CI [2.78, 3.33], and when they imagined ratings of others compared to themselves, M_diff = 0.53, p < .001, 95% CI [0.33, 0.73]. These pattens replicated results found in Study 3 (see Figure 5).

The central goal of Study 4 was to replicate participants’ apprehension to report hubristic pride. As before, the Reference main effect was qualified by a significant Pride Rating x Reference interaction, F(1, 105) = 11.74, p < .001, η_p² = 0.10. The Reference effect was primarily found for the rating of hubristic pride, where participants imagined that others would experience hubristic pride more strongly than themselves, M_diff = 0.84, p < .001, 95% CI [0.57, 1.11]. But this differential responding was not significant for the rating of authentic pride, M_diff = 0.22, p = .114, 95% CI [-0.05, 0.49]. Taken together, participants’ ratings of hubristic pride systematically varied: participants were more reluctant to admit their own feelings of hubristic pride compared to admitting their own feelings of hubristic pride.

Nomological networks. Results of nomological networks were highly similar to those of Study 3 (See Table 3 and Figure 7).

Study 4 replicated and corroborated Study 3. The within-subject design of Study 4 makes it clear that participants may be reluctant to express their feelings of hubristic pride, whereas they have little problem expressing feelings of authentic pride.

We present two mini meta-analyses of differences between authentic vs. hubristic pride manipulations for the self-ratings of authentic/hubristic pride across four studies. We draw on the two-facet theory of pride by Tracy and Robins (2007) and evaluate the most stringent criteria suggested by the theory: (1) authentic pride is independent of hubristic pride; and (2) an experimental manipulation should increase only one facet relative to the other (see Figure 1b). We examined whether authentic (hubristic) pride manipulation increased ratings of authentic (hubristic) pride compared to the other pride manipulation. We computed fixed effects, weighting the mean effect size by sample size (Goh et al., 2016; see Table 4 for effect sizes). Overall, the combined effect for authentic pride was negligible, Mean d = 0.011, 95% CI [-0.161, 0.184]. This means that the manipulations designed to increase hubristic pride also evoked authentic pride just as strongly as did the manipulations designed to induce authentic pride.

We next meta-analyzed the effect sizes of hubristic-pride manipulation relative to authentic-pride manipulation on the self-rating of hubristic pride across our four studies using fixed effects, weighting the mean effect size by sample size. The overall effect was in the predicted direction, mean d = 0.252, 95% CI [0.079, 0.425], such that the hubristic-pride manipulations induced participants’ self-ratings of hubristic pride more strongly than the authentic-pride manipulations.

For comparison, we also meta-analyzed the previously published research that reported an experimental manipulation of hubristic vs. authentic pride (Ashton-James & Tracy, 2012; Ho et al., 2016). Results are summarized in Table A5 of the Appendix for self-rated authentic pride and self-rated hubristic pride. As is evident there, the five studies reported in Ashton-James and Tracy (2012) and Ho et al. (2016) reported a combined effect of d = 0.377, 95% CI [0.149, 0.605] in the induction of self-rated authentic pride. Likewise, they reported a combined effect of d = 0.974, 95% CI [0.744, 1.205] in the induction of self-rated hubristic pride. Clearly, these effect sizes are much larger than the ones we observed across our four studies.

Our series of four studies aimed to address the ongoing controversy about the construct validity of hubristic pride as conceptualized by Tracy and Robins’ Authentic Pride and Hubristic Pride (AP-HP) Scales (2007). Beside the direct attempt to replicate previous manipulation using the Relived Emotion Task, we drew on relevant literature in the domains of overconfidence (Anderson et al., 2012), power (Fast & Chen, 2009), status (Witkower et al., 2021), and state narcissism (Mahadevan et al., 2019) to develop a strong manipulation of hubristic pride. Yet, none of our manipulations induced the desirable levels of authentic and hubristic pride consistent with the standard we set up at the beginning (Figure 1). Whereas the hubristic-pride manipulations in our studies generally increased ratings of hubristic pride relative to the manipulations of authentic pride (meta-analytic effect size, d = 0.252), we did not observe that participants experienced hubristic pride with any intensity. Most problematic is that the hubristic-pride manipulations generally induced strong levels of authentic pride; as a result, authentic pride always overshadowed hubristic pride to a significant degree even in the control conditions. These results do not support the even more liberal standard of successful manipulation (Figure 1a).

Our systematic analyses of experimental manipulations in terms of nomological shockwaves, nomological networks, and social desirability illuminated potential problems that may have been previously overlooked. Below, we evaluate Tracy and Robins’ (2007) two-facet theory of pride and offer specific avenues for future research in the light of the observed findings. We then discuss our work in the context of reproducible and open science.

Our answer is no. The observed results should remind the reader of the controversy about the nature of hubristic pride as raised by Holbrook et al. (2014). The authors argued that Tracy and Robins’ HP Scale is a flawed measurement of pride emotion, as HP items (e.g., pretentious, arrogant) correlate with self-deprecating tendencies. Our analyses of nomological networks of hubristic pride support the criticism by Holbrook et al. (2014). We found that hubristic pride covaried more strongly with negative affect (e.g., hostile, aggressive, ashamed, guilty) than with positive affect. These findings contradict the general view that pride is a positive emotion that arises from success.

Perhaps due to the negativity associated with hubristic pride, it was difficult to find intense levels of hubristic pride in people’s experience of success. Our analyses of the nomological shockwaves (Chester & Lasko, 2021) revealed that any positive feedback produced significant increases in authentic pride and positive affect but did not significantly produce any expected changes in hubristic pride and negative affect (Study 2-3, Figure 6). Participants felt a great deal of authentic pride, even when they were asked to write about a hubristic episode (Study 1), when their success was based on talent (Study 2), or when their superior performance was related to dominance characteristics (Study 3-4).

We conclude that people feel authentic pride regardless of how they attain success. This may correspond to the reality of most domains of achievement, in which success may be driven by a combination of both ability and effort. Consider world-class athletes who work hard but who must also have a high degree of dispositional advantage to get to the very top. It is unlikely that these athletes attribute their success only to hard work. Even among non-athletes, we found that people felt genuinely good about their achievement regardless of its cause, or they were unable to evaluate the relative contributions of effort versus talent.

In sum, a clear picture emerging from our analyses is that hubristic pride is distinct from pride and has little relevance with success. Recall that, according to the two-facet theory of pride (Tracy & Robins, 2007), both types of pride arise from success, but the difference between authentic and hubristic pride is in how success is attributed to different aspects of the self (e.g., effort vs. talent). It was unlikely that any personal success based on talent would induce a strong sense of hubristic pride, which our analyses discovered to be linked with negative emotions. We struggle to imagine everyday events that grant one a status but induce negative feelings in the person. Indeed, the theory is mute about why stable attributions of success should elicit hostility, aggression, shame, and guilt, which we used to assess negative affect linked with hubristic pride. The nature of hubristic pride emerging from our experiments may align better with Holbrook et al.’s (2014) account of hubristic pride as overclaiming one’s unmerited success. In this definition, hubristic pride is nothing like a pride emotion associated with personal success.

Study 3-4 confirmed that the difficulty of inducing hubristic pride was in part due to social desirability, with participants readily predicting feelings of hubristic pride in others, which they did not report for themselves to the same extent. However, we note that, while participants expected stronger reactions of others than their own feeling of hubristic pride in response to the leadership feedback, their estimates tended to be still low (M = 3.09, Study 3-4) compared to their estimates of authentic pride (M = 5.67, Study 3-4). That is, hubristic pride may never be more salient as an emotion than authentic pride. To our knowledge, our research is the first to detect the possibility of social desirability associated with expressing hubristic pride.

An alternative interpretation of our data is to take the observed reference effects on the ratings of hubristic pride as a general part of actor-observer asymmetry in emotion recognition. One’s success may be based on a myriad of factors, but one’s first-person account of success may deviate from a third-person account; accordingly, one’s self-conscious emotions may look different to others. For example, think of Maria, an American high school senior who just enrolled at a prestigious university. Maria attributes her success to her own hard work getting high grades and having an excellent record of extracurricular activities; she feels a great deal of authentic pride. But her peers attribute Maria’s success to her parents giving her a tremendous financial advantage to extracurricular activities or consider her talent to be a product of luck. In the eyes of a beholder, Maria’s success is less based on hard work but more on privilege; hence, a display of pride on the part of Maria is viewed as hubris. Yet, Maria is indifferent to social desirability of feeling hubristic pride: she is genuinely proud of her accomplishment! This type of actor-observer asymmetry may apply to almost all emotions. Your smile represents your internal feeling of happiness, but it may elicit pity, envy, anger, disgust, or contempt in observers for various reasons. Our findings of reference effects on the ratings of hubristic pride may speak to this potential nature of emotion recognition in the interpersonal dynamics.

Though our findings are by no means the definitive answers to settle the ongoing controversy, we are well positioned to offer specific directions for future research on pride. We think that there are two major directions that researchers can take with unique premises and challenges.

One direction is to defend the two-facet theory by Tracy and Robins (2007) and continue pride research under this framework. However, research built upon the two-facet theory may revise parts of the theory and overcome a few challenges. First, future research needs to develop a method to induce and measure hubristic pride over and above social desirability. Without such a method, taking the two-facet theory of pride is tantamount to having too much faith in the unknown possibility of social desirability: hubristic pride really exists, but inability to observe it is a mere reflection of social desirability. To be sure, our mini meta-analytic estimate yielded a small effect of experimental manipulation for hubristic pride (d = 0.252), but this estimate is largely driven by the Relived Emotion Task in Study 1. Compared to the other methods that failed to induce hubristic pride, the task’s effectiveness depended on its explicit encouragement to report a hubristic-inducing episode, and not to worry about social desirability concerns. However, we do not know to what extent this method obscures the true amount of hubristic pride people internally feel. Thus, any self-report measures of hubristic pride without an effective way to amend social desirability does not allow us to falsify the claim that hubristic pride exists independently of authentic pride, at least in the way conceptualized by Tracy and Robins (2007). Key is that researchers need to develop a standardized task to manipulate specific attributions of success, and a pride-inducing manipulation needs to evoke a cluster of negative feelings such as guilt, shame, hostility, and aggression along with positive emotion.

The second challenge is to dissociate the accumulated evidence on trait-level pride with state-level pride. We acknowledge that trait-level studies have shown utility of psychometric properties of the AP-HP scales (Dickens & Robins, 2020). The kind of correlations found between trait-level hubristic pride and various outcomes have been taken for granted as supporting the nature of hubristic pride as an emotion (e.g., Tracy et al., 2009). However, our analyses reveal this assumption to be unwarranted. Our experimental work showed that state-level hubristic pride correlated with a range of negative feelings, yet people clearly felt positive about their success. Thus, correlations found at the trait level may imply that people who admit to label themselves arrogant or smug tend to have heightened self-awareness of low esteem and imperfection of their characteristics. In this view, hubristic pride measured by the HP scale correlates with socially maladaptive outcomes such as lower self-esteem and higher depression (Dickens & Robins, 2020). Of course, this line of evidence should not be used to defend the status of hubristic pride as an emotion. This conceptualization of hubristic pride diverges from a general concept of pride, which should be a positive emotion associated with elevated status (Holbrook et al., 2014). Therefore, the accumulated evidence on personality correlates of hubristic pride may tell us little about the nature of hubristic pride as conceptualized by the original two-facet theory. But if the two-facet theory conceptualizes hubristic pride as a part of general pride, then future research needs to show how exactly one feels negative emotions in response to achievement and success.

After all, one may interpret our findings as consistent with the two-facet theory, specifically the claim that hubristic pride still exists. We did find that hubristic pride correlated with positive emotions to some extent, and one manipulation was successful in inducing a small amount of hubristic pride. However, a challenge is that authentic pride was always a more salient emotion than hubristic pride across all experiments. A possible interpretation of our data is that hubristic pride may exist, but it may represent a small portion of one’s pride experience when one finds reasons to be aggressive, hostile, guilty, and shameful about one’s success. At times, one may find hubristic sides of success, whether the success was obtained unfairly or without much effort. We, therefore, acknowledge that hubristic pride may exist, but it might be a complex emotion, one that is very hard to manipulate experimentally. However, we believe that any personal success would elicit greater authentic pride than hubristic pride. From this view, it is an overstatement that hubristic pride is a discrete emotion independent of authentic pride, and that both facets reliably arise from distinct attributions to the self.

Another direction, perhaps a more straightforward interpretation of our results, is to favor a parsimonious view of pride—a single-facet view that there is only authentic pride (Williams & DeSteno, 2010). This view assumes that the observed social desirability effect on hubristic pride (Study 3-4) is just one type of actor-observer asymmetry in emotion recognition. One may feel and express authentic pride, but the same expression may be interpreted as hubris by others. In this view, there is only one (authentic) pride on the part of actor, but the same expression is readily interpreted as hubris by observers, depending on the context. What we may call hubristic pride is an observer’s cognitive evaluation of others who express (authentic) pride expression. Thus, the single-facet model of pride assumes that the reference effect on self-report of hubristic pride found in our Study 3-4 is something inherently natural about pride (and any emotions), not a mere artifact that prevents people from reporting true internal states of hubristic pride.

There is a virtue in parsimony. The single-facet model of pride makes great synergies with other areas of research related to status without necessarily taking a more complex view. Indeed, we adopted the present manipulation contents from experiments of overconfidence (Anderson et al., 2012), power (Fast & Chen, 2009), status (Witkower et al., 2021), and state narcissism (Mahadevan et al., 2019). Our analyses of nomological shockwaves of the current manipulations indicated that any kinds of positive feedback would elicit authentic pride and positive feelings. In other words, manipulation of these constructs via positive feedback will likely elicit authentic pride as an emotional marker of manipulation, and pride may potentially be used as a mediator of the observed associations. This possibility inspires a bridge with other relevant areas, as pride may be an important intra-personal motivation in leadership emergence (De Cremer & Van Dijk, 2008), cooperation in a social dilemma (Dorfman et al., 2014), or altruism (Hardy & Van Vugt, 2006). The single-facet model of pride will benefit future research by encouraging research integration with fewer administrative costs. This possibility is especially relevant for researchers who collect data from a crowdsourcing platform in which brevity of instrument is critical.

We believe that our work complements recent attention to the issue of manipulation check. Chester and Lasko (2021) coded 348 manipulations in the Journal of Personality and Social Psychology in 2017 and found that the majority (80%) of these published works used the “on-the-fly” manipulation without any systematic pilot testing. In short, researchers in social psychological research spend little time in testing the validity of experimental manipulation, though such work should precede replicability and reproducibility (Fiedler et al., 2021). The same problem may apply to the manipulation of authentic/hubristic pride. Although seemingly successful manipulation checks were used to justify the effect of authentic/hubristic pride, no prior work had scrutinized the construct validity of state-level authentic/pride via systematic manipulation checks. Our work highlights importance of validating manipulations prior to main experiments. As noted by many in recent years, there remains much ambiguity as to how researchers view usefulness of manipulation checks (Fayant et al., 2017; Hauser et al., 2018). However, one cannot guarantee the validity of the proposed causal relationship without evidence of successful manipulation itself. As we have shown in this article, lack of evidence of valid manipulations renders reported results hard to interpret. As in the case of prior research on authentic/hubristic pride, it is unclear what really drove the observed difference between authentic/hubristic and control conditions (Ashton-James & Tracy, 2012; Ho et al., 2016). Manipulation checks should be central to theory building in psychology (Fiedler et al., 2021), but we add that inserting a manipulation check between a treatment and an outcome measure in a main experiment—and subsequently show successful manipulation checks—is insufficient. Researchers need to conduct more pre-test studies devoted to validating manipulations.

Our results have several limitations. One might speculate that the online format of the manipulation methods may have introduced elements that were not previously presented in the work by Ashton-James and Tracy (2012) and Ho et al. (2016). However, some of our studies did overcome the limitations of these previous studies. Our Study 1 showed consistent patterns with regard to hubristic manipulation as did previous studies. Our Study 2 was also conducted online and yielded the expected effect with regard to a manipulation of authentic pride relative to a control group. Hence, it is extremely unlikely that the online format itself is to blame.

Due to limited sample sizes, some of our studies may suffer from low power of detecting a true effect of experimental manipulations, assuming those true effects are very small. We mitigated this limitation partially by conducting a series of mini meta-analyses. However, we suspect that a greater sample size would not eliminate the fundamental difficulty of observing a strong sense of hubristic pride. We believe that our studies as a whole provide sufficient information to suggest constructive criticisms and future direction. Reporting the null results, albeit with limited sample sizes, would help future research create a more innovative manipulation.

As with much of previous research on pride, our sample consisted of highly homogenous college students. Though we think it is necessary to establish valid manipulations with any samples first, future research may investigate more diverse pride experience of other populations.

Our work contributes to the ongoing controversy about the two-facet theory of pride, but much remains to be investigated. More research is necessary to defend the validity of two-facet theory of pride over a single-facet model of pride. Future research needs to clarify the nature of hubristic pride and accordingly develop a stronger manipulation of hubristic pride without confounding it with authentic pride and social desirability. A successful manipulation of authentic/hubristic pride will inspire many research areas related to status, power, leadership, and hierarchy. Yet, more work on validating experimental manipulation of state-level hubristic pride should take priority. The single-facet (parsimonious) model of pride still stands as a fruitful model, until pending evidence otherwise supports the two-facet theory of pride.

The author(s) declared no potential conflicts of interest with respect to the authorship and/or publication of this article.

Contributed to conception and design: KK, MK

Contributed to acquisition of data: KK, MK

Contributed to analysis and interpretation of data: KK, MK

Drafted and/or revised the article: KK, MK

Approved the submitted version for publication: KK, MK

We would like to thank Hirotaka Imada for his comment on the earlier version of the manuscript.

Materials, data, and code are publicly available on: https://osf.io/9rfby/?view_only=54dde4c60f694cbeaa90c0788410c1d2