Dimensions on which psychological measures of humility vary
Most measures of humility are multi-item, self-report scales. However, the potential approaches to measuring humility are varied and many. Given the varied manifestations of the multiple features of humility summarized in the previous section, alternatives to self-report scales hold promise for strengthening our understanding of the concept and its relevance for behavior and well-being (Davis and Hook, 2014; Davis, Worthington and Hook, 2010).
In this section, we present a general measurement typology that highlights the dimensions on which psychological measures of humility may vary and the types of measures that result from different combinations of those dimensions. The value of considering the full array of alternatives with respect to measuring humility is that (1) research to date has relied almost entirely on trait-level, self-report measurement, and (2) humility may neither evince the stability of a trait nor be subject to valid self-reports (cf. Davis,Worthington and Hook, 2010; Davis et al., 2011;Tangney, 2000). Concretely, researchers must decide whether to measure a particular instance of humility (e.g., ask respondents to recall a particular experience in which they felt humble), humility as a state (e.g., ask respondents to respond about the moment in which they are completing the measure), humility as a trait (respondents’ general levels of humility across situations and contexts), or humility as a personality judgment (how individuals are viewed by others across the situations and contexts in which the rater has observed the individual). Fortunately, the typical approach to measuring humility through self-reports is but one of a number of potential approaches that reflect combinations of distinctions on four characteristics of psychological measures: the source of information on the participants’ level of humility; the form in which that information is expressed; the presumed stability of the participant’s level of humility; and the degree to which the participants’ level of humility is assumed to generalize across all settings, topics, and forms of humility.33.2.1 Source
Although the humility score that is assigned to a research participant is assumed to reflect his or her level of humility, that score does not have to be based on information provided by the participant. Alternatives to self-report, outlined below, when considered alongside self-reports define the first dimension on which approaches to the measurement of humility may vary: source. Essentially, reports on research participants’ humility can come either from the participants themselves in the form of self-reports or from someone else in the form of informant reports.Within the latter category, sources may range from informants who know well the participant on whom they are reporting, to informants who are unacquainted with the participant. The principal question to be addressed with respect to source is who is in the best position to provide valid information about the target’s standing on humility?
A fundamental question to be considered with respect to self-reports of humility is whether people have insight into their levels of humility. Do humble people realize they are humble? Or might humility make people more sensitive to recognizing and admitting instances in which they do not think, feel, or behave humbly (Peterson and Seligman, 2004)? If humility involves less-than-normal amounts of self-focus, humble people may not attend to their humble qualities as much, decreasing their ability to report their own levels of humility (Rowatt et al., 2006). In addition, even if humble people are aware of their humility, being humble may make them modest in reporting their levels of humility (Davis et al., 2010). For all these reasons, people who are more humble may appear less humble on self-reports.
Whereas people who are humble may display a modesty effect, those who are less humble may enhance their report of humility (Rowatt et al., 2006).This reflects a social desirability bias, which is common to the measurement of many constructs in psychology, but is compounded in the case of humility by the fact that the presence and extent of self-enhancement differs on the basis of levels of the construct being measured.
Davis et al. (2010) detail how the combination of modesty and self-enhancement effects within a sample can muddy the waters when using self-reports to assess humility (see their Figure 1). Furthermore, the fact that some research has shown informant reports, described below, to provide higher scores and some research has shown informant reports to provide lower scores on humility in comparison to self-reports, suggests that the modesty and self-enhancement effects on self-reports are present to different extents across samples.In recent years, concerns voiced about the validity of self-reported humility have been somewhat attenuated (e.g., Davis and Hook, 2014;Worthington and Allison 2018). Scores on self-report measures of many constructs studied by psychological scientists can be consciously manipulated by respondents, but this does not necessarily invalidate these measures. Particularly in anonymous or confidential low-stakes research contexts, respondents have little incentive to fake their responses on self-report humility scales (de Vries, Zettler and Hilbig, 2014). Moreover, Ashton, Lee and de Vries (2014) have noted that there is sufficient support for the construct validity of self-reports of humility, including similar score distributions to personality factors; moderately high agreement between self-reports and reports by close acquaintances; and weak loadings on a factor representing social desirability bias. Importantly, shared variance between measures of humility and social desirability may, in fact, result from measures of social desirability validly tapping into substantive traits, including modesty (de Vries et al., 2014). Nonetheless, to be cautious, some researchers have taken measures to minimize potential social desirability bias, such as avoiding the word humility and its derivatives in items on self-report instruments of humility (Kruse et al., 2017).
A separate, though no less troublesome, concern about self-reports of humility is whether respondents are able to judge how humble they are.This concern is not about a motivated bias but rather a lack of adequate insight (Nisbett and Wilson, 1977).
Lacking such insight about their own standing, particularly on intrapersonal features of humility, respondents may rely more on lay theories of humility in people like them as opposed to valid accounts of their own thought, feelings, and motives. Unfortunately, this validity threat is more difficult to detect than threats such as social desirability bias; however, it should be evident in discrepancies between self- and informant-reports on features of humility that are evident to informants.Informants are people who make judgments about the dispositions, preferences, or other characteristics of a target person, in whom evidence of those characteristics is believed to be apparent to others. Informants may base their judgments on direct observation of the target at the time judgments are provided; observation of video recordings of the target; or, if they are acquainted with the target, a history of interacting with target.The advantage of direct observation and observation of video recordings is that they can be staged to give evidence of the characteristic being judged. For example, judges of humility might see the target telling someone about their strengths and weaknesses.The advantage of judgments by informants who know the target is that they will be informed by knowledge of the target's reactions and behaviors across many situations.
A key consideration when using informant reports is who should be enlisted as informants. We have already highlighted the important distinction between informants who know the target versus those who do not.Whereas the former might be able to offer a summary judgment with reasonable validity without significant training or instructions, the latter will need instructions or training that provide clear criteria. If ratings are to be more specific than present-absent, then information about degree will need to be provided for each criterion. For example, Bell et al. (2019) developed a set of criteria on which judgments were made by clinicians based on observations during clinical interviews of clergy candidates.
For each of seven criteria (e.g., “acknowledged and owned personal strengths and limitations”) these researchers provided clinician informants sufficient information to rate targets on 1-to-5 scales.Although the clinician informants did not know the targets, they learned a great deal about them in the clinical interviews and were provided detailed instructions for expressing judgments. In contrast, Meagher et al. (2015) had university students interact with 3-5 fellow students with whom they were not acquainted for about 20 minutes then rate them on four items indicative of intellectual humility (e.g.,“Open to criticism of ideas.”).There was no evidence of consensus between informants ratings. A follow-up study obtained judgments from students who had worked together in groups throughout a semester. Although consensus was reached on ratings of humility, scores were not correlated with self-ratings of humility. These findings suggest that strangers with limited exposure to targets are not good judges of their humility. Moreover, though consensus across informants is evident when informants have even a brief history of interacting with targets, the lack of convergence with self-ratings of humility is a puzzle yet to be solved. Minimally, the findings clearly indicate that source matters in measures of humility.Before turning to other dimensions relevant for measuring humility, we note a special, relatively unexplored research circumstance for which source is likely to matter a great deal— research on children. In such cases, self-reports, unless they focus on explicit behaviors, are unlikely to be useful. For similar reasons, peer informants are not likely equipped to provide useful judgments. Parents, teachers, and adult caregivers may be the only sources available for reliable and valid information about humility in children.
33.2.2 Expression
Regardless of the source of information about a target's level of humility, the means by which that level is expressed or captured may vary.The primary forms of expression are verbal reports of humility, humility as evidenced in behavior, and humility in cognitive or physiological responses to relevant stimuli.
These categories, in theory, cross-cut the source dimension such that both self and informants could be sources of information expressed as verbal reports, behavior, or responses to stimuli; however, informants are likely to have limited access to information about responses to stimuli, which may not be observable.Verbal reports are expressed through responses to questions or statements about one's preferences or tendencies, or narrative responses to open-ended queries designed to elicit descriptions that can be coded for correspondence to different levels of humility. Verbal reports, whether from self or informants, make two assumptions that may not hold for humility and similarly abstract concepts: (1) Valid verbal expressions of the concept can be generated or detected. (2) Respondents or informants are capable of determining whether the feature of the concept targeted by statements or open-ended queries apply. Meeting the first assumption requires that the conceptualization of humility is sufficiently detailed and clear to provide a basis for generating questions or statements or coding open-ended responses. Meeting the second assumption may be possible when the expression is social or behavioral but not when it is cognitive, meta- cognitive, or affective. If either assumption does not hold, then alternative forms of expression are likely to be more valid for use in assessment.
Apart from verbal reports, humility may be expressed as behavior or as responses to stimuli designed to elicit reactions consistent or not with humility. Although humility has implications for behavior, it is rarely defined with explicit reference to behavioral manifestations as is, for example, conscientiousness as measured by the Behavioral Indicators of Conscientiousness (Jackson et al., 2010).The closest conceptual treatments are those that involve an explicit interpersonal focus. Whether motivated by a behavior-oriented conceptualization or assuming that humility should be evident in behavior, several attempts have been made to measure humility as expressed in behavior. For example,Van Tongeren et al. (2014) operationalized humility as the ability to regulate defensiveness in response to being challenged by another person. Other behavioral indicators of humility discussed in the literature include how a person behaves when given the opportunity to share credit (e.g., when accomplishing a task or receiving praise or an award), when vying for power in a group, in interpersonal conflict, when making a mistake or hurting someone, and in relationships involving power differentials (e.g., Chancellor and Lyubomirsky, 2013; Davis et al., 2011;Van Tongeren et al., 2014).The challenge of relying on behavioral measures is that behaviors consistent with humility may or may not be driven by humility. That is, people may engage in behaviors consistent with humility for other reasons, such as following social norms or out of a desire to appear humble to others (Peterson and Seligman, 2004).
Although behavioral expressions of humility would typically be captured by informants (which may include experimenters), research participants may be valid sources of information about their behavior under certain conditions. In particular, ecological momentary assessment (EMA) studies may repeatedly over short periods of time ask participants to indicate whether they have engaged in certain concrete behaviors within the past few minutes.These reports can be elicited in ways that do not involve subjective judgments. Perhaps because most models of humility do not specify specific behavioral indicators of the concept, EMA studies that capture behavior indicative of humility in the moment are not yet evident in the empirical literature on humility.
In addition to behavior, humility may express as reactions to stimuli designed to elicit responses consistent with high or low humility. For example, Rowatt et al. (2006) developed an implicit association test of humility, comparing the reaction times of pairings of self with humility to pairings of self with arrogance. Implicit assessments are difficult to control deliberately, and thereby sidestep potential concerns about response biases. Implicit measures may prove useful, particularly if they are updated in response to progress in the science of humility. However, concerns about a potential lack of awareness about one's levels of humility that have been expressed about self-reports carry through to implicit assessments, as do problems related to self-deceived narcissists or others low in humility who truly view themselves as humble. In such cases, neither self-report nor implicit assessments would prove particularly useful. Physiological correlates of humility have also been explored, but are subject to interpretation and therefore only as valid as their theoretical backing. For example, one might examine physiological responses to ego threat as an indicator of humility (Peterson and Seligman, 2004). However, would a lack of physiological arousal to an ego threat induction be indicative of humility, or would behaving humbly in the face of physiological arousal be a closer approximation of humility? Furthermore, would levels of physiological arousal to ego threat depend on a person's trait level of humility? For example, Worthington has argued that the presence of true ego strain is essential to developing humility (e.g.,Worthington and Allison, 2018). However, if the biomarkers of humility are different depending on the degree to which a person has developed the skill or virtue of humility, then using such physiological data for the purpose of determining levels of humility is problematic.
In summary, verbal reports, for which critical assumptions related to validity may never be met, are far and away the expression of choice in empirical research on humility. Alternative expressions that make fewer assumptions may be viable, but relatively little research to date has attempted to use them. The primary obstacle to consulting potential alternative expressions such as behavior and reactivity is that conceptual models of humility have not yet developed to the point that the relevant behaviors or physiological reactions are clear. Better or more creative approaches to measurement of humility are constrained by underspecified conceptual accounts of the concept.
33.2.3 Specificity
With respect to measurement, specificity concerns whether a concept is reflected in a single score assumed to represent the concept in full, or is reflected in a single score that reflects the concept in part or multiple scores that correspond to specific features or manifestations of the concept. As with the expression dimension, standing on the specificity dimension is, to some degree, dictated by conceptual models of the concept. Those models may provide information about specificity of two types.As noted earlier in the chapter, models of humility often specify particular features of the construct. Examples include being other-oriented and holding an accurate view of oneself.Alternatively, they may suggest different domains across which humility may vary. Examples include relational humility and intellectual humility.These domains may be quite specific. For example, relational humility may be further subdivided into humility with respect to family, friends, and romantic partners. Intellectual humility may be considered with respect to politics or religion or, even more specifically, particular political or religious issues. Humility measures that are specific in focus allow for specificity matching when the outcomes of interest also are specific in focus. Poor specificity matching is typical of individual differences research, including research on humility, typically taking the form of the prediction of specific behaviors or performances from a general or global form of the individual difference (Hoyle and Leary, 2009). Such research likely underestimates the influence of humility on specific outcomes because people are able to function and interact with humility in some settings and not others, or in interactions with some people or with reference to some areas of knowledge but not others.
Although some models of humility are global and unidimensional (e.g., Leary and Banker, in press), most are multifaceted. Indeed, in the seminal treatment of humility as a psychological construct,Tangney (2000) proposed six “elements” of humility—accurate self-assessment, ability to acknowledge mistakes, openness to new ideas, keeping ones' accomplishments in perspective, relatively low self-focus, and appreciation of the value of others' contribution. Although these facets of humility were adjusted, replaced, or supplemented in subsequently articulated conceptual models, nearly all models pose a similar set of dimensions. Reflecting this feature of conceptual models, most current measures of humility include between three and six subscales corresponding to posited features of the construct.
The other form of specificity with implications for measurement—domain specificity— has been less evident in conceptual treatments of humility. In fact the most prominent form of domain-specificity with respect to humility, intellectual humility, has spawned its own set of models (e.g., Leary et al., 2017) and a research literature that has little connection to the literature on general humility. Unlike hierarchical models of personality and other individual differences, wherein facets (e.g., personality traits) or domains (e.g., self-concept) are assumed to be specific variants of higher-order concepts, it is not clear, for example, that intellectual humility can be fully understood as a specific variant of general humility as conceptualized by most models. Nevertheless, specificity matching, by which specific forms of humility, such as intellectual humility, are used instead of general humility to predict reactions or behavior in the same domain (e.g., willingness to consider political perspectives that differ from one's perspective), increases the likelihood of detecting effects of humility on multiply determined behaviors. Unlike individual differences such as self-concept (Marsh and Shavelson, 1985) and self-esteem (Harter, Whitesell and Junkin, 1998), for which domains are well-specified, the full array of humility domains and the nature of their relations with general humility remain unspecified. Nonetheless, some attention has been given to humility in selected domains, including intellectual, spiritual, relational, and cultural humility (e.g., Davis and Hook, 2014; Davis, Rice et al., 2016; Worthington and Allison, 2018).
In the same way that a widely accepted conceptualization of humility would likely hasten scientific progress in understanding the concept and its implications for behavior and wellbeing, an agreed-upon set of domains would promote progress. In developing a more comprehensive set of domains, researchers might look to the outcomes for which humility might be relevant as has been done with self-concept and self-esteem. That kind of analysis gave rise to the limited set currently considered, but a more systematic and wide-ranging analysis would no doubt point to others. Development on this front would strengthen both conceptual models and predictive validity of humility.
33.2.4 Stability
A relatively unexplored feature of humility with relevance for how it is measured is its stability across time and place. Humility is typically assessed in a manner that assumes it is a trait; that is to say, peoples' level of humility varies little from one time or context to another.To the extent that humility demonstrates such stability it might reasonably be positioned in personality space as a facet of a broader personality domain (e.g., Ashton et al., 2014) or as a cross-cutting trait that manifests features of multiple personality domains (like, for example, sensation seeking). Alternatively, it might be the case that humility reliably varies across contexts and short periods of time, in which case referencing structural models of personality as a means of embellishing models of the concept might not prove useful. Rather, conceptualizing the concept as a state that characterizes the person with respect to context and time may prove truer to the concept as it manifests “in the wild.” Humility might evidence properties of both traits and states (as, for example, do anxiety and other forms of affect), suggesting the need for models and measures that capture both stable and varying manifestations of the concept. Indeed, at the outset of contemporary empirical research on humility,Tangney (2000) referred to “two levels of interest, two levels of questions,” and discussed the merits of conceptualizing and measuring humility as “dispositional” and “situational.”
Two prominent models of personality process offer a theoretical rationale for developing conceptual models and measures of humility that characterize it as both trait and state. Fleeson (2001) proposed a resolution of the apparent contradiction between state and trait forms of individual differences in the form of density distributions of behaviors reflective of those individual differences. Fleeson observed that, when personality-relevant behaviors are assessed using EMA methods across several days, most participants report behaviors that span the continuum of typical personality traits, thus showing evidence of variable personality states.Yet the within-person density distributions, characterized by mean and standard deviation of behavior over time and place, are stable, suggesting trait-like consistency.Applied to humility, Fleeson's approach suggests value in research that identifies behaviors indicative of humility, assesses them at the momentary level using EMA methods, and examines within-person density distributions. Such work is likely to reveal that (1) expressions of humility vary across situation, (2) the central tendency and range of people's distributions are stable across situations, and (3) people vary in how variable their expressions of humility are across situations. This pattern would support an inference that humility has both trait and state manifestations, each of which warrants measurement and study.
Cognitive-affective processing system (CAPS) theory offers an alternative perspective that accounts for both variability and consistency in personality (Mischel and Shoda, 1995). The theory asserts that, although people's behavior is contingent on features of situations and therefore state-like, they develop patterns of responses to situational features such that they behave similarly in a given situation as they encounter it repeatedly across time. Consistency in behavior is found in the if-then contingencies that develop between situations and behaviors. Coherent sets of stable if-then contingencies reflect personality, which is both consistent and variable. Operating from a set of principles specified in CAPS theory, researchers have identified the personality signatures for narcissism, characterizing the trait as a stable set of cognitive, affective, and interpersonal processes associated with specific features of situations (Morf and Rhodewalt, 2001). A similar analysis of humility would likely highlight the person-situation contingencies that underlie trait-level humility and contribute to its situated expression.
The preponderance of research on humility has used trait measurement, resulting in relatively little information about the dynamics of humility in daily life. A promising line of research by Kruse and colleagues resulted in the development of a measure of state humility and a set of studies suggesting that variability in state humility is associated with variability in other states over short periods of time (Kruse et al., 2017; Kruse et al., 2014). Kruse et al. (2017) identified a dozen themes that emerged from statements provided by research participants' asked to reflect on what humble people are like, including things they might say or believe. They incorporated themes most clearly reflective of humility in six self-report items, to which respondents were instructed to indicate “how you feel right this moment.” A key validation strategy for the measure was an examination of its covariation with a “stress test” of humility, which involved imagining someone angry with them and providing an attribution and indicating their likely responses. Responses were rated by two sets of judges, one referencing their own definition of humility and the other using the state humility scale items. Both sets of ratings were moderately to highly correlated with self-reports of state humility. In a separate set of studies using their conceptualization and measure of state humility, Kruse et al. (2014) showed that state humility is responsive to experiences of gratitude and, when both are assessed daily, state gratitude and humility are mutually reinforcing. These findings suggest that humility states can be measured with reliability and validity, and provide hints at processes that influence state humility and may, over time, contribute to stability and change in levels of humility.
33.3