Blinded By the Light: The Dark Side of Traditionally Desirable Personality Traits


Guenole (2014) argues that organizational scholars should embrace the newly developed personality model from the Diagnostic and Statistical Manual of Mental Disorders, 5th ed. (DSM-5; APA, 2013). We agree with Guenole that organizational psychologists would benefit from further assimilation of maladaptive traits into the personality spectrum, as it would create a more thorough and nuanced representation of personality. However, whereas Guenole's intent is to steer our focus towards DSM personality traits, we propose an alternative perspective to the study of maladaptive traits. Specifically, we extend Guenole's argument that maladaptive traits are deserving of additional scholarly work by proposing that it is also useful to examine whether personality traits that are traditionally considered to be desirable (e.g., Conscientiousness) become maladaptive when one has “too much of a good thing” (e.g., a conscientious individual who is so perfectionistic that he/she cannot complete a task; Pierce & Aguinis, 2013).

We differ from Guenole by focusing our commentary on the often overlooked maladaptive extreme positive end of the personality spectrum (whereas Guenole's focus on DSM-based personality traits tends to center on maladaptive personality at the negative end of the personality spectrum). In this commentary, we first review research that suggests that the extreme positive poles of the Big Five personality traits can be dysfunctional. We then summarize work on “the dark side” of two non-Big Five personality traits (i.e., emotional intelligence [EI] and self-efficacy) to illustrate that this phenomenon is not unique to the Big Five. Finally, we comment on the implications of investigating “too much of a good thing” in personality measurement, including issues related to the item content of personality measures, the analysis of curvilinear relationships, and ideal point response processes.

The “Darkness” in Positive Personality Traits

The Big Five personality traits (Agreeableness, Conscientiousness, Emotional Stability, Extraversion, and Openness), EI, and self-efficacy are all traits for which high scores are typically considered to be desirable. However, recent research suggests these traits can also be detrimental if manifested at extreme levels (i.e., there is evidence for curvilinear relationships between these traits and organizational outcomes, where both the extreme low pole and the extreme high pole are maladaptive). Although we could have reviewed any number of personality constructs here, we chose to focus on the aforementioned desirable traits because a fair amount of existing research suggests these traits do indeed have a “dark side” at the extreme positive end of the latent trait spectrum (e.g., Côté, Decelles, McCarthy, Van Kleef, & Hideg, 2011; Del Líbano, Llorens, Salanova, & Schaufeli, 2012; Samuel & Gore, 2012). We note that future work in this area could involve a wide array of additional traits (e.g., core self-evaluations, proactive personality).

Big Five Personality Traits

Big Five personality traits have shown small to moderate positive relationships with task performance (Barrick & Mount, 1991; Barrick, Mount, & Judge, 2001; Hurtz & Donovan, 2000), organizational citizenship behavior (OCB; Chiaburu, Oh, Berry, Li, & Gardner, 2011), leadership (Judge, Bono, Ilies, & Gerhardt, 2002), and countless other organizational outcomes of interest (for a review, see Hough & Oswald, 2008). As a result of this research, it is typically considered to be beneficial to have high levels of Conscientiousness, Agreeableness, Extraversion, Emotional Stability, and Openness. However, as we illustrate below, recent work suggests that extreme scores on the positive end of the latent trait spectrum may actually be maladaptive.

For example, very conscientious individuals can be overly cautious and rigid (Le et al., 2011), which may cause these individuals to be less adaptable to change. Moreover, highly conscientious individuals' narrow-minded focus on obeying rules may distract them from acquiring new skills (Tett, 1998). Another negative aspect of obsessive rule following is that it can prevent individuals from performing OCBs because it may seem inappropriate to go above and beyond formal responsibilities as outlined in their job requirements (Le et al., 2011). It is also possible that highly conscientious individuals could derail team creativity when they hyper focus on minute, irrelevant details rather than the big picture, particularly if task instructions are not clear (Judge & LePine, 2007). Ultimately, the rigidity that often characterizes those high on Conscientiousness may result in a decrease in life satisfaction during periods when the obsessive personality has nothing to distract it, such as when one is unemployed (Boyce, Wood, & Brown, 2010).

Consistent with the idea that very high Conscientiousness has negative outcomes, Le et al. (2011) found that Conscientiousness exhibited a curvilinear relationship with task performance, OCB, and counterproductive work behavior (CWB). A curvilinear relationship between Conscientiousness and job performance was also reported by Lahuis, Martin, and Avis (2005) who found that job performance was highest for moderate levels of Conscientiousness compared to when it was very high or very low. At the same time, the inverted U-shaped relationship between Conscientiousness and job performance has proven to be controversial, as some researchers find little evidence of statistically significant quadratic effects (e.g., Robie & Ryan, 1999). However, as we discuss in a later section, the failure to detect a curvilinear relationship may be associated with inadequate measurement of the full content domain of the construct. In sum, whereas Guenole suggests that research should focus on disinhibition, or the maladaptive variant of low Conscientiousness, we recommend that industrial–organizational (I–O) scholars should also further explore the destructive potential of high levels of Conscientiousness.

Similarly, whereas Guenole concentrates on antagonism as the maladaptive dimension corresponding to low Agreeableness, we suggest that high levels of Agreeableness can also be dysfunctional. Agreeable individuals are typically pleasant to be around, but there are also maladaptive aspects of Agreeableness such as gullibility and submissiveness that could lead to an agreeable individual being perceived as an ingratiating pushover (Samuel & Gore, 2012). Taking this idea further, the desire to please everyone may result in agreeable individuals being overly lenient managers (Bernardin, Cooke, & Villanova, 2000). In addition, these individuals' motivation to avoid conflict (Suls, Martin, & David, 1998) could lead to lower pay, fewer promotions, and decreased extrinsic career success (Ng, Eby, Sorensen, & Feldman, 2005). Finally, when highly agreeable individuals work in teams, they may be less constructive and effective team members because they refuse to voice dissenting opinions. This is a problem if poor decisions are argued for most vociferously by other group members (LePine & Van Dyne, 2001).

In regard to the Big Five trait of Extraversion, previous research indicates that Extraversion is ideal for sales and managerial positions (Barrick & Mount, 1991). However, the warm, gregarious nature of extraverts can backfire if manifested at extremely high levels in which case these individuals appear to be self-aggrandizing, long-winded, and egotistic (Coker, Samuel, & Widiger, 2002). In addition, an extravert's social behavior can be disadvantageous for nonsales jobs or autonomous jobs that require long attention spans or routinization due to the high energy and activity level that is characteristic of individuals who are extremely extraverted (Beauducel, Brocke, & Leue, 2006). There is also evidence that people who are too dominant/assertive (a facet of Extraversion; Costa & McCrae, 1992) can be perceived as bullies (Ames & Flynn, 2007). For example, when examining leadership effectiveness, a series of studies by Ames and Flynn (2007) found an inverted U-shaped relationship between leadership and assertiveness such that leaders who were very high in assertiveness were seen as tyrannical, whereas leaders who were very low in assertiveness were seen as pushovers (i.e., both extremes were ineffective). In contrast, moderate levels of assertiveness were optimal for leadership. In summary, it appears that although low levels of Extraversion (perhaps corresponding to the dimension of Detachment described by Guenole) can be maladaptive, we propose extremely high Extraversion can be dysfunctional as well.

Emotionally stable individuals are known to be calm and level headed in comparison to their anxious, neurotic counterparts (e.g., individuals high in negative emotionality). However, the overly emotionally stable person may give the impression of being cold, emotionless, and inhuman (Coker et al., 2002). This notion of high levels of Emotional Stability as maladaptive is supported by Le et al. (2011), who demonstrated a curvilinear relationship between Emotional Stability and task performance, OCB, and CWB. The authors theorized that people who are very high in Emotional Stability are characterized by emotional over control and hence, are less able to connect with others. In addition, they suggested that the over control of emotions in individuals who are high on Emotional Stability may deplete the cognitive resources needed for task performance.

Finally, although individuals high on Openness to Experience are typically considered to be creative and artistic (Goldberg, 1993), high levels of Openness are also associated with being rebellious and unconventional, components that could (in some situations) lead to CWB and a disregard for company rules (Hough, 1992). In addition, behaviors associated with extremely high Openness can have negative organizational consequences because open individuals are more accident prone (Clarke & Robertson, 2005) and less committed to their jobs (Erdheim, Wang, & Zickar, 2006). Very high Openness is even associated with odd or eccentric behavior, bizarre thinking, and distorted cognitions that can result in hostility towards conventionality (Piedmont, Sherman, & Sherman, 2012). This suggests that Openness may be beneficial for nontraditional, innovative organizations/jobs but may also be detrimental beyond a certain threshold.

Emotional Intelligence

Recent evidence suggests emotional intelligence (i.e., the ability to reason about emotions and use emotions and emotional knowledge to enhance thought processes; Mayer, Roberts, & Barsade, 2008) is positively correlated with task performance (Joseph & Newman, 2010; O'Boyle, Humphrey, Pollack, Hawver, & Story, 2011) and leadership effectiveness (Harms & Crede, 2010). Thus, EI is generally considered to be a desirable set of abilities; however, it can also be misinterpreted (as apathy in individuals who may be too efficient at regulating their own emotions) or misused (when one manipulates others' emotions with malicious intent). For example, initial work in this area indicates that those who are very high in EI may seem cold and aloof. In particular, leaders who regulate their emotions too much can appear distant and depersonalized from their subordinates, which can negatively affect subordinate job satisfaction (Kafetsios, Nezlek, & Vassiou, 2011). Further work on the “dark side” of EI demonstrates that while most individuals who are high in EI are also high in moral identity and low in Machiavellianism, there are individuals who are high in EI and Machiavellianism, and these individuals can cause harm through their ability to control and manage others' emotions (Côté et al., 2011). Evidence suggests EI is also moderately correlated with emotional manipulation (r = .29), or purposeful manipulation of someone's feelings for personal gain, particularly for men (Grieve & Panebianco, 2013). In other words, it appears that the (typically positive) trait of EI can sometimes be misunderstood or intentionally used for “dark” or maladaptive purposes.


Finally, although research on self-efficacy has generally supported a positive relationship between self-efficacy and work-related performance (Stajkovic & Luthans, 1998), within-person studies have found that self-efficacy has a negative relationship with performance (e.g., high self-efficacy prompts less planned and actual study time and subsequently poorer exam performance; Vancouver & Kendall, 2006). This is because high self-efficacy can lead to over-confidence in one's abilities. That is, when individuals believe that they are sufficiently equipped to face a challenge, they are less motivated to invest resources to prepare for that challenge (Vancouver, Thompson, Tischner, & Putka, 2002; Vancouver, Thompson, & Williams, 2001).

Measuring the “Darkness” of Positive Personality Traits

The literature reviewed above tends to support curvilinear effects for the relationship between personality and outcomes—where moderate levels of a given trait can be more beneficial than high levels of the trait. It is possible that support for curvilinear effects may be even stronger than the literature suggests, as it is somewhat difficult to detect curvilinear relationships. The difficulty arises because curvilinear effect sizes tend to be relatively small, and studies frequently do not have sufficient statistical power (i.e., large enough sample sizes) to detect small effects—a problem that is likely exacerbated because popular personality measures often do not include enough items to cover the extreme high and low ends of the trait continuum. For example, in an attempt to explain puzzling findings that were inconsistent with theoretical predictions (e.g., the lack of an association between Conscientiousness and obsessive–compulsive personality disorder), Haigler and Widiger (2001) examined how well the NEO-PI-R tapped into the extreme ends of the trait continuum for the Big Five. The authors concluded that few items addressed the maladaptive aspects accompanying high levels of Conscientiousness, Agreeableness, and Openness, but when items were modified to measure very high-levels of these traits, the expected relationships between the Big Five and personality disorders emerged. Upon personal inspection of the item content of two popular Big Five measures (i.e., the 50-item IPIP, 2012 and the Big Five Inventory [BFI]; John, Naumann, & Soto, 2008), it appears that none of the 44 items of the BFI capture the extreme positive end of the latent trait continuum (e.g., Conscientiousness is assessed with items such as “perseveres until the task is finished” and “does a thorough job” rather than “perfectionistic” or “picky about details”). Similarly, the International Personality Item Pool (IPIP) Big Five scale does not seem to include items that address the extreme positive pole of each Big Five personality trait (e.g., Extraversion is captured with items such as “Don't mind being the center of attention” rather than items such as “I have to be the center of attention”). In other words, it appears that popular personality measures typically do not tap into the extreme positive pole of common personality traits, which may limit the extent to which research can successfully examine the phenomenon of having “too much of a good thing” for these traits. This lack of extreme items in nonclinical measures is another reason why we agree with Guenole that it would be useful for I–O psychologists to blend research exploring the new DSM-5 variants of the Big Five with existing research.

For optimal measurement, both moderate and extreme personality items should be combined into a comprehensive measure that captures the very high/low end of the trait distributions in addition to the more moderate aspects of the traits. As such, items that are adjusted to tap into the extreme ends of these traits may display stronger relationships with maladaptive behaviors. For example, previous work suggests that when Conscientiousness items are changed from “I keep my belongings neat and clean” to “I keep my belongings excessively neat and clean,” the correlation between Conscientiousness and obsessive compulsive disorder increases from .03 to .62 (averaged among three OCD scales; Samuel & Gore, 2012). Agreeableness and dependency disorder showed the same pattern of results when items were reworded to be more extreme (Samuel & Gore, 2012). Similar issues are found with EI and self-efficacy scales; EI items typically do not tap potential misuse (i.e., manipulation) of emotions or misinterpreted emotions, and self-efficacy scales do not address maladaptive overconfidence or workaholism. For example, the EI item “I present myself in a way that makes a good impression on others” could easily be adapted to tap into manipulation (e.g., “I present myself in a way that allows me to deceive others when necessary”; Schutte et al., 1998). For self-efficacy, the Occupational Self-Efficacy Scale item, “I am confident that I could deal efficiently with unexpected events in my job” could be amended to address overconfidence: “I am extremely confident that I could deal efficiently with any unexpected events in my job” (Schyns & von Collani, 2002).

An issue related to the paucity of personality items that capture the extreme poles of the latent trait spectrum involves the appropriateness of the assumed response process associated with the measurement of personality. Typically, measures of individual differences are based on a dominance model in which the probability that an individual will choose a response option increases as the individual's standing on the latent trait increases (e.g., as Conscientiousness increases, an individual becomes more likely to respond “strongly agree” to an item such as “I tend to keep my belongings neat and clean”). However, research suggests that an ideal point model may be more appropriate for purposes of personality assessment than a dominance model (Chernyshenko, Stark, Drasgow, & Roberts, 2007; Drasgow, Chernyshenko, & Stark, 2010). Ideal point models describe a response process in which an individual is more likely to choose a particular response option when the item is close to the person's level of the latent trait. In other words, where a dominance model would predict that an extremely conscientious individual would choose the response option “strongly agree” for the item “I tend to keep my belongings neat and clean,” an ideal point model assumes that this individual would not agree with this statement as strongly, perhaps choosing “somewhat agree” or even “disagree” (because the item does not reflect the individual's actual standing on the latent trait, which would be more accurately captured by an extreme item such as “I always keep my belongings neat and clean”). Unfortunately, research suggests that previous reliance on dominance models may have led to current difficulties in detecting the aforementioned curvilinear effects. Specifically, dominance models incorrectly score those who are extremely high on the latent trait as moderately high on the latent trait, and the resulting rank-order of respondents misrepresents curvilinear effects as linear (Carter et al., in press). Therefore, we recommend consideration of an ideal point model in future personality research in order to appropriately scale the latent trait and allow for the detection of “too much of a good thing” personality effects.

In addition to revising item content to tap the more extreme ends of the latent trait continuum and considering the adoption of an ideal point model, there are other steps researchers could take to improve the detection of curvilinear relationships between personality and work outcomes. This includes conducting well-designed research studies with larger sample sizes (e.g., > 500 participants) and taking advantage of meta-analytic techniques that allow researchers to cumulatively examine curvilinear effects (e.g., Verhaeghen & Salthouse, 1997; Williams & Livingstone, 1994). It is also possible that specific contextual factors determine whether a curvilinear effect will be present (e.g., job complexity moderates the curvilinear relationship between Conscientiousness and job performance; Le et al., 2011), so an a priori consideration of these contextual factors is an important step in designing productive studies. In other words, researchers must ask themselves in what jobs or occupations high levels of a trait are likely to be harmful and when will there be enough variance in a sample to detect a curvilinear effect.

In summary, we firmly agree with Guenole that “dark” traits deserve additional research attention from organizational scholars. However, we suggest that personality researchers should not be “blinded by the light” (i.e., we should not overlook the potential for maladaptive characteristics in positive personality traits). Instead we need to focus our efforts on exploring the potential negative outcomes associated with traditionally desirable personality traits. Additional research is needed to better understand curvilinear relationships of personality and organizational outcomes and to develop scales that tap into the extreme ends of trait continuums.