Honesty as truthfulness: New trait and state measures of truthfulness to advance research and theorizing on when and why people are honest

Honesty is a diverse construct

The term honesty captures a wide variety of seemingly disparate traits and behaviors. For example, Miller’s (2021) conceptual account of honesty includes actions such as telling the truth, being appropriately forthcoming, refraining from stealing, refraining from cheating, and keeping one’s promises. Similar diversity is found in experimental studies of honesty, which often operationalize honesty as specific behaviors in the research setting, such as returning a “lost” wallet without removing money, overreporting one’s performance on a task, paying for goods taken from unattended “honor stands,” or reporting overpayments (see Rosenbaum et al., 2014). Even more diversity may be found in dispositional measures related to honesty (Fleeson et al., 2022). For example, the Values in Action’s Honesty scale includes content focusing on promise-keeping and value-adherence; the SAPA Personality Inventory’s Honesty scale (Condon, 2017) focuses on lying, cheating, and exploitation; the Workplace Productivity Questionnaire includes Deceiving and Cheating facets that focus on lying, taking undeserved credit, and shifting blame (Nicol, 1999; Nicol & Paunonen, 2002); and the HEXACO-PI-R’s Honesty-Humility factor (Lee & Ashton, 2018) includes content related to generosity, modesty, law-abidingness, and (not engaging in) exploitative lying.

The diversity captured by the term honest raises conceptual and empirical challenges. Conceptually, it is unclear what, if anything, unites such disparate behaviors and qualities (i.e., the unification challenge, see Miller, 2021), or what is at the core of honesty. Among other empirical issues, the diversity means that ostensible measurements of honesty may reflect meaningfully different phenomena and that ostensible measures of honesty might reflect phenomena that are actually peripheral to honesty. Such ambiguity, in turn, hinders clarity in interpreting individual studies and in accumulating understanding across studies.

Conceptual and empirical evidence for the centrality of truthfulness

Recent philosophical and empirical work has begun resolving the ambiguity of honesty, revealing that truthfulness is an especially important facet of honesty, perhaps even its central core. First, Miller (2021) proposes and evaluates various conceptualizations of the virtue of honesty, but all of them frame it in terms of a disposition “to not intentionally distort the facts as the agent sees them.” Similarly, Um (2024) provides several examples of honest behavior as his motivating cases, and they involve either telling the truth or avoiding deception. This conceptualization then highlights the “respect for the right not to be deceived” as honesty’s fundamental motivation. Such strong emphasis on avoiding deception and on truth appears in many recent conceptual discussions of honesty (Carr, 2014; Roberts & West, 2020; Wilson, 2018).

Reinforcing this, a recent set of empirical studies (Reynolds et al., 2025) suggests that truthfulness may be the most central feature to how laypeople apply the term honest to both people and behaviors, at least in the United States. In this work, participants freely generated characterizations of honest people/behaviors (over 6000 in total). Qualitative and quantitative analyses revealed the frequency with which various themes of honesty/dishonesty (e.g., truthfulness, forthrightness, cheating, stealing) emerged and the priority in which the themes emerged in participants’ lists (e.g., beginning, middle, or end). Separate samples of participants then judged these free responses on their importance to honesty and indicated whether each response was an example of honesty versus other moral categories. Across five studies and all metrics of conceptual centrality, truthfulness consistently emerged as the most central feature of lay honesty concepts.

Given its nature and its importance within the scope of honesty, we see three nested reasons that personality psychologists should focus significant attention on truthfulness. First, even beyond its status as a part of honesty, truthfulness is an important psychosocial phenomenon in its own right, featuring as an essential element of human coordination (e.g., T. R. Levine, 2014). Although truthfulness has likely always been an important psychosocial issue, understanding how people respond to truth, deception, and “alternative facts” may be important now more than ever in a world overflowing with information access, much of which is inaccurate (e.g., van der Linden, 2022). Second, if one grants that truthfulness is an important component of honesty, even if not the core of honesty, fully understanding the psychological processes underlying honesty means understanding the processes of truthfulness. So, attention is needed to the extent that we care about understanding honesty. Finally, if truthfulness is indeed at the center of honesty, understanding honesty in a way that resonates with both lay peoples’ views and philosophical conceptualizations of honesty is impossible without understanding truthfulness.

Defining and conceptualizing truthful communication

To advance understanding of honesty, we focus on the core personality manifestation of truthfulness, truthful communication. At the dispositional level, we define truthful communication as the tendency toward expressing one’s beliefs accurately and faithfully. People with high levels of truthful communication tend toward expressing their beliefs accurately and in a way that avoids being misleading (about those beliefs) when communicating with others. Even in situations that might make such truthfulness difficult, uncomfortable, or costly, they are oriented toward truthful expression. People with lower levels of truthful communication are less oriented toward expressing their beliefs accurately and truthfully, perhaps even tending toward deception.

Similar to most other traits, we see truthful communication as a multidimensional construct (see, e.g., Wilt & Revelle, 2015). Its cognitive/evaluative component is the belief that truthful/non-misleading communication is “correct” and “the right thing,” morally speaking. Its affective component is the experience of guilt and/or shame in instances where one fails to express one’s beliefs faithfully. This component also includes anticipating those moral emotions if one were to fail to express beliefs faithfully. Its motivational component is the desire that others know one’s actual beliefs when those beliefs are expressed. Finally, its behavioral component is the tendency to actually express one’s beliefs faithfully and non-misleadingly, that is, not intentionally misleading others. These components are consistent with contemporary theories of personality that articulate the causal processes underlying personality traits (e.g., Fleeson & Jayawickreme, 2021; Furr & Funder, 2021). These components also reflect a virtue-oriented conceptualization of truthful communication (Miller, 2021; Wright et al., 2021), as they refer to a value-oriented type of cognition, morally relevant emotions, and a morally oriented type of motivation (i.e., as opposed to self-enhancing or self-protective motivation to be truthful).

In terms of “expressing one’s beliefs accurately and faithfully,” we intend “beliefs” to reflect a variety of cognitive representations. These include: (a) one’s understanding of objective physical or temporal facts (e.g., whether the earth is flat), (b) one’s interpretation/evaluation of physical or temporal facts (e.g., whether the weather is “warm”), (c) one’s recollection of events (e.g., what happened during an argument), (d) one’s understanding of other people’s psychological states and traits (e.g., whether one thinks that Columbus believed Earth was flat), and (e) one’s opinions or attitudes (e.g., whether one enjoyed a dinner). Note that, on this view, if a person’s beliefs are wrong (objectively speaking) but they express them accurately, then the person is “truthful” according to this definition. ¹ Put another way, we use “truthful” as a description of people and their communicative behaviors, not as a description of beliefs or claims themselves, which can be said to be “truthful” to the extent that they are consistent with reality. ²

Although our definition is articulated at a dispositional level (i.e., “a tendency toward…”), truthful communication is also a situational phenomenon. As one moves from situation to situation in daily life, there can be variability in the degree to which one expresses one’s beliefs accurately and faithfully. One’s truth-related cognitions, emotions, motivations, and behaviors can vary across situations. A full understanding of the psychological causes, correlates, and consequences of truthful communication in daily life requires examination from both dispositional and situational levels. Ideally, such examination employs a measurement approach that operationalizes truthful communication similarly across those levels of examination.

Understanding truthful communication in daily life: Incorporating trait and state perspectives

To understand truthful communication and its role in individuals’ daily life experiences, researchers should examine it from both state and trait perspectives. From a trait perspective, truthful communication is a variable that differs between people, in terms of a relatively stable tendency to act, think, feel, and be motivated toward (or away from) truthfulness. From a state perspective, recent advances in personality theory have also highlighted the importance of examining how people’s manifestations of personality systematically vary across contexts (Furr & Funder, 2021; Jayawickreme et al., 2021). Notably, this dynamic perspective on personality has recently been extended to understanding moral traits and virtues (Fleeson & Jayawickreme, 2021; Jayawickreme et al., 2021; Jayawickreme & Fleeson, 2017; Mendonça et al., 2023; Wright et al., 2021). Dynamic accounts of personality can thus explain both variations in people’s manifestations of truthful communication from context to context, as well as long-term stability within a person.

Whole Trait Theory (WTT; Fleeson & Jayawickreme, 2015, 2021) is one such dynamic theory that also fosters engagement with the philosophical literature on virtue and character development and has emerged as an important model of moral traits (Doris, 2022; Jeffrey & Beary, 2023; Wright et al., 2021). According to the principles of WTT, while an individual’s behavior varies rapidly and frequently across different occasions, their behavior is both stable and distinct from other individuals when averaged across a larger period of time such as a few days or a week. Extending from this account, moral traits such as truthful communication are stable in the sense that there is reliable between-person variation, and flexible in the sense that there is also substantial within-person variation in an individual’s behavior depending on both situational cues and the specific goals being pursued in that context.

Furthermore, WTT proposes that the descriptive side of an individual’s character traits can be represented as a distribution of states (Fleeson & Jayawickreme, 2017; Jayawickreme & Fleeson, 2017) measuring the extent to which an individual manifests a given moral traits at a specific point in time. According to WTT, possessing the descriptive side of a specific trait—truthful communication in this instance—is simply akin to communicating in a more truthful manner more often. Such states are similar to traits in affective, cognitive, and behavioral content, but states are manifested for a few minutes or hours rather than for months or years (Cattell et al., 1947; Fleeson, 2001). Past research has used experience sampling methodology (ESM) in which participants are asked to report to the extent to which they have manifested moral behavior consistent with a given moral trait at regular time intervals across the duration of several days or weeks (e.g., Meindl et al., 2015). These reports can then be aggregated into a distribution to convey the frequency with which an individual enacted each level of a given state. By examining the full distribution of an individual’s behaviors, research testing the predictions of WTT has found that reliable between-person differences in individuals’ aggregated moral behaviors exist, and that also individuals also respond flexibly in different situational contexts.

A further insight of WTT is the functional nature of traits—individuals enact trait manifestations in order to fulfill specific goals and thereby achieve specific outcomes (McCabe & Fleeson, 2016) and in response to specific opportunities afforded by the situation (Zachry et al., 2018). In other words, people intentionally modulate their levels of truthful communication to fulfill specific goals and manage distinct environments. Examining how truthful communication functions and systematically varies in daily life should therefore be an important task for personality scientists interested in the dynamics of truthful communication in daily life.

Measurement of truthful communication

Unfortunately, the field’s ability to study truthful communication is limited by a lack of relevant precise measures. Existing measures of honesty (or other similar constructs, such as authenticity) are inadequate for assessing truthful communication, as they commonly suffer from one of two problems (see Fleeson et al., 2022 for an overview of scale content in trait honesty measures). First, many trait measures of honesty incorporate content beyond truthfulness or possibly not directly related to honesty. For example, the Values in Action honesty scale based on the International Personality Item Pool (IPIP, n.d.) includes content related to promise-keeping and adherence to one’s values, making it imprecise as a measure of truthfulness. More broadly, the HEXACO-PI-R Honesty-Humility scale incorporates items related to generosity and modesty, and its truthfulness-related items focus specifically on lying to manipulate or take advantage of others. ³ This inclusion of other features in measures muddies interpretation to the extent that criterion variables should be expected to differentially relate to different aspects of honesty. ⁴

Second, many trait honesty measures focus solely on overt behaviors such as telling the truth or lying. ⁵ Though these behavioral manifestations are an important aspect of trait honesty/truthfulness, these scales miss other mechanisms and processes that make up moral traits, such as beliefs, emotions, and motivational processes (Jayawickreme & Fleeson, 2017). For example, the Honesty subscale of the Externalizing Spectrum Inventory-100 asks participants to assess how frequently they lie but makes no reference to any psychological components beyond overt behaviors. By missing other psychological components, such measures miss important elements of traits and fail to capture cases where the elements may not be congruent with each other. For example, a person who speaks truthfully in a situation only after a difficult struggle with a temptation to lie is meaningfully different from a person who speaks truthfully without experiencing such temptation (see Ng & Tay, 2020, for a broader discussion of the importance of including multiple psychological components when assessing moral traits).

These two problems also exist with the very few examples of state assessments of honesty that exist in the current literature. One study (Sherman et al., 2015) assessed state manifestations of Honesty-Humility with a set of adjectives combining both honesty and humility (“humble, honest—arrogant, dishonest”). Such assessments provide an imprecise assessment of honesty in daily life. Meindl and colleagues (2015) developed a measurement tool for assessing state manifestations of moral thoughts and behaviors in daily life. This tool included a multi-item assessment of honesty. While this measure included items relevant to truthful communication (“Did you try to make someone believe something that you know is false?”), it focused on behaviors, as opposed to other characteristics of the trait of truthful communication (e.g., motivations, emotions).

The present studies

To advance the field’s ability to understand a crucial component of honesty, we present six studies in which we develop and psychometrically evaluate both trait and state measures of truthful communication. We see these new measures as important complements to other measures that assess broader honesty-relevant personality traits. Studies 1–4 developed a reliable trait measure of truthfulness with good convergent and discriminant validity (Study 2), incremental validity in predicting a behavioral criterion differently than a related measure (Study 3), and predictive validity in predicting people’s honesty throughout the events of their day (Study 4). Studies 5 and 6 developed a reliable measure of state manifestations of truthfulness and used it to examine how moment-to-moment truthfulness varies between and within people in cross-sectional (Study 5) and ESM designs (Study 6). Moreover, Studies 5 and 6 examined how such variability is not random, but systematic as a function of both trait truthfulness and momentary goals, evidencing the validity of the state measure.

In developing these measures, we frequently compare the novel trait measure of truthfulness to measures of the Honesty-Humility factor of the HEXACO personality framework, for two reasons. First, HEXACO Honesty-Humility scales are widely used measures related to trait honesty, and comparing our new measure to familiar and common scales helps clarify its nomological network. Second, the authors often see Honesty-Humility scales used as a proxy for honesty (or more specifically, for truthfulness) in probing various truthfulness-relevant behaviors. ⁶ To the extent that researchers use broad Honesty-Humility measures to assess truthfulness specifically, they likely introduce unrelated variance, potentially obscuring understanding about truthfulness. However, if broad Honesty-Humility measures already capture the same variance as a precise truthfulness measure, it would make the more precise measure somewhat redundant. Thus, a useful truthfulness measure would complement Honesty-Humility measures by capturing additional unique variance.

All studies were approved by the Wake Forest University IRB. Studies 2b (osf.io/xeauj) and 3 (osf.io/9ug6y) were preregistered.

Item generation and refinement

Seven team members (faculty, postdocs, and graduate students) generated 148 items based on the working definition of truthful communication:

Truthful Communication (TC) is the tendency toward expressing one’s beliefs accurately and faithfully. People with high levels of TC tend toward expressing their beliefs accurately, truthfully, and in a way that avoids being misleading (about those beliefs) when communicating with others. Even in situations that might make such truthfulness difficult, uncomfortable, or costly, they are relatively oriented toward truthful expression. People with lower levels of TC are less oriented toward expressing their beliefs accurately and truthfully, perhaps even tending toward deception.

Project leaders independently reviewed those items, evaluated the conceptual fit and clarity of each, discussed their reviews, and retained 75 items.

Next, ten members of the interdisciplinary team (psychology and philosophy faculty, postdocs, and graduate students) independently reviewed these 75 items, rating each item’s conceptual fit to the definition and likely clarity to respondents. They also noted which psychological component (i.e., behavior, motivation, etc.) each item reflected. Using strict criteria based on these ratings ⁷ , we retained, revised, and added items, resulting in 49 items that we administered to Study 1 respondents (see Supplement Table S1 for full list).

Method

Participants and procedure

For Studies 1 and 2, we aimed for N = 500 to facilitate robust factor analytic work, which offers power >.80 for small to medium individual difference correlations (r > |.13|, see Gignac & Szodorai, 2016). We requested 500 participants from Qualtrics Panels. Initially, 548 participants met the standard data quality criteria used by Qualtrics Panels. We retained only participants above 17 years of age, for N = 532. Table 1 provides detailed demographic information for this and all subsequent samples.

OPEN IN VIEWER

Table 1.

Overview and demographics for all studies.

Context	Study 1	Study 2a	Study 2b	Study 3	Study 4	Study 5	Study 6
Design/Task	Item selection & exploratory factor structure	Confirmatory factor structure, convergent/ discriminant validity		Behavioral criterion study, incremental validity	Day reconstruction, predictive validity	State item selection, cross-sectional, incremental validity	Experience sampling, incremental validity
Sample (location)	Qualtrics panels (US)	Qualtrics panels (US)	Prolific (UK)	Prolific (US)	Qualtrics panels (US)	Community (SE US)	Qualtrics panels (US)
Valid N (recruited N)	532 (548)	525 (530)	522 (522)	295 (316)	543 (544)	199 (215)	181 (200)
Age M (SD)	35.8 (12.6)	55.0 (19.5)	46.0 (15.5)	45.4 (15.8)	48.5 (17.1)	46.8 (14.1)	36.2 (10.9)
Gender (% women)	70.9	62.3	50.4	51.9	70.0	69.3	59.1
Race (%)
Asian-American	7.5	2.9	8.2	6.4	2.2	1.0	7.2
Black/African-American	14.7	8.0	4.2	13.6	10.5	10.1	14.4
White	71.8	85.3	83.9	60.7	83.1	69.3	60.8
Other	4.7	1.6	1.1	8.5	3.6	0.5	0.6
Multi-racial	2.8	1.1	1.9	10.5	4.8	1.0	--
Declined	0.8	1.0	0.6	0.3	0.2	17.6	0.0
Hispanic (%)	15.0	5.3	-	-	5.3	1.5	17.1
Highest education
Doctorate	1.0	4.0	-	-	2.4	3.0	4.4
Master’s	6.6	14.3	20.7	-	11.6	27.6	16.0
Bachelor’s	18.4	22.7	44.8	-	26.8	24.6	24.3
Some college	36.8	31.2	23.4	-	34.0	22.1	28.2
High school diploma/Equivalent	28.6	23.4	10.3	-	22.1	3.5	22.7
Didn’t finish high school	8.6	4.4	0.2	-	1.8	0.0	3.3
Other	0.0	0.0	-	-	1.1	1.5	1.1
Median household income	$40–50K	-	-	-	$40–50K	$60K–70K	$50–60K

Note. SE US = Southeastern United States. For race/ethnicity items, participants in studies 1–5 could select multiple options or explicitly indicate multiracial identity. Study 6 did not allow multiple selections. For UK education data, we list post-graduate degrees as “Master’s” (not differentiating between masters and doctorates), college or university degree as “Bachelor’s,” higher/secondary/further education as “Some college,” secondary school up to 16 years of age as “High school diploma/equivalent,” and primary school as “Didn’t finish high school.” These education equivalencies are for convenience of description and do not factor into any analyses.

Participants provided demographic information and responded to the 49 items using a 5-point agreement scale (strongly disagree to strongly agree). We conducted a series of exploratory factor analyses (EFAs) to evaluate internal structure and determine the final scale.

Results and discussion

We began with an EFA of all 49 candidate items. An initial EFA’s scree plot suggested a strong common factor, and all but two items loaded ≥.32 on that factor. However, parallel analyses suggested six factors. We thus conducted a series of EFAs (using promax rotation) to gauge the meaning and relevance of different structures. When several factors were extracted, they largely reflected the intended components of motivation, behavior, affect, and cognition.

Based on their sizable cross-loadings and because they were at a different level of abstraction from the other items (which specifically targeted motivations, cognitions, etc.), we dropped several items reflecting broad self-perceptions/reputations, including identity concerns (i.e., seeing oneself as a truthful person). To refine the scale further, we retained items that (a) strongly represented a general factor common to all items, (b) strongly represented several more subtle factors reflecting behavior, motivation, cognition, and affect, and (c) collectively included positively keyed and negatively keyed items.

Ultimately, we retained 15 items. EFA of these items revealed a strong first factor, with all items loading >= |.45| in a one-factor structure (see Table 2 for item content and loadings) and with a first eigenvalue 5.5 times larger than the second. Because we systematically included variance related to different psychological processes (motivation, etc.) and selected items reflecting these processes, we also explored a 5-factor structure within these 15 items. Results revealed clearly interpretable factors with 3 items per factor and no meaningful cross-loadings. Factor 1 reflected truth-speaking beliefs (negatively-keyed items, e.g., “White lies are sometimes OK”), factor 2 reflected truthfulness motivations (e.g., “If someone asks for my opinion, I want that person to know what I truly believe”), factor 3 reflected moral emotions (e.g., “If I lie (or if I were to lie), I feel some shame or guilt”), factor 4 reflected truthful behaviors (e.g., “Even if the truth might be hard for someone to hear, I tell the truth”), factor 5 reflected lying or misleading behaviors (negatively keyed items, e.g., “I often deceive people about my real beliefs”). Underscoring the fact that all items strongly share a common core, the factors inter-correlated at |.45| to |.59|. Estimated reliability for the 15-item scale was α = .89.

OPEN IN VIEWER

Table 2.

Trait truthful communication scale (T-TCS) items with factor loadings, Study 1.

	Item	Psychological Component	Factor Loading
1	In some situations, lying is the right thing to do. (N)	Belief	−.687
2	In some situations, it is OK to tell a lie. (N)	Belief	−.734
3	White lies are sometimes OK. (N)	Belief	−.718
4	If someone asks for my opinion, I want that person to know what I truly believe	Motivation	.606
5	It is important to me that I share my honest beliefs with others when I’m asked about them	Motivation	.617
6	In situations where I’m asked for information, I want other people to know the information as I truly see it	Motivation	.621
7	If I were to successfully lie or mislead someone intentionally, I would feel guilty	Emotion	.576
8	Even if I got away with it, I would feel bad about lying to someone	Emotion	.575
9	If I lie (or if I were to lie), I feel some shame or guilt	Emotion	.486
10	Even if the truth might be hard for someone to hear, I tell the truth	Truthful behavior	.669
11	I tell the truth even if it might hurt someone’s feelings	Truthful behavior	.608
12	I tell the truth, even when it could cause problems for me	Truthful behavior	.629
13	I often deceive people about my real beliefs. (N)	Deceptive behavior	−.446
14	I often say something that’s not true. (N)	Deceptive behavior	−.474
15	I occasionally misrepresent my thoughts and beliefs when speaking with people. (N)	Deceptive behavior	−.501

Initial examination suggests the T-TCS predominantly reflects a single general factor and thus could be scored as a single, reliable score reflecting truthful communication. The presence of five more subtle factors is conceptually important as it attests to the breadth of content of the T-TCS and its convergence with conceptual considerations. In Study 2, we conducted confirmatory tests to examine the appropriateness of scoring the T-TCS as a single truthful communication score.

Method

Results and discussion

Internal structure

We conducted confirmatory factor analysis (CFA) of the T-TCS items (after reverse-scoring negatively-keyed items). To account for subtle but intended systematic differences across sets of items representing different psychological processes (e.g., items for motivation, items for emotion, etc.), we evaluated a bifactor model. All items loaded on a single common factor ostensibly reflecting trait truthful communication. Each item also loaded on a second factor common only to the items within each set, representing what those items share beyond trait truthful communication (e.g., factor-specific features of items, systematic variance in reporting different types of psychological content). In total, this model includes six orthogonal factors—a general trait truthful communication factor on which all items loaded and five three-item factors for cognition/beliefs, motivation, emotion, truthful behavior, and deceptive behavior.

This bifactor model fit well based on general cutoffs (see West et al., 2023) ¹⁰ in both Sample 2a, χ² (75) = 295.39, p < .001; CFI = .96, TLI = .94, RMSEA = .075, SRMR = .09, and Sample 2b, χ² (75) = 157.29, p < .001; CFI = .98, TLI = .97, RMSEA = .046, SRMR = .051. ¹¹ Importantly, the bifactor model fit better than a unidimensional model (Sample 2a: ΔBIC = −2618.0; Sample 2b: ΔBIC = −2279.4), which fit poorly in both samples: Sample 2a, χ² (90) = 3007.33, p < .001; CFI = .43, TLI = .33, RMSEA = .25, SRMR = .18; Sample 2b, χ² (90) = 2530.52, p < .001; CFI = .46, TLI = .35, RMSEA = .23, SRMR = .15. ¹²

Standardized factor loadings for the bifactor model are presented in Table 3. Loadings on the general truthful communication factor ranged from .27 to .72 whereas loadings on the psychological process factors ranged from .33 to .89. Not surprisingly, negatively keyed items (1–3, 12–15) loaded lower on the general factor (and higher on their specific factors) than did the positively-keyed items (4–11). This type of slight factorial separation between positively keyed and negatively keyed items is common (e.g., Biderman et al., 2011). Overall, this pattern of model fit and factor loadings suggests that the T-TCS items share a robust commonality (i.e., truthful communication), while also systematically tapping subtle but detectable differences between different types of psychological manifestations (behavior, motivation, etc.).

OPEN IN VIEWER

Table 3.

Standardized factor loadings for bifactor models of the T-TCS, Study 2.

Item	Truthful communication	Belief	Motivation	Emotion	(Truthful) behavior	(Deceptive) behavior
Sample 2a (US)
1	.37	.80
2	.34	.89
3	.38	.78
4	.56		.59
5	.66		.62
6	.52		.48
7	.56			.55
8	.63			.69
9	.66			.50
10	.72				.41
11	.62				.63
12	.71				.33
13	.28					.80
14	.27					.82
15	.32					.77
Sample 2b (UK)
1	.50	.70
2	.47	.82
3	.43	.56
4	.50		.56
5	.56		.47
6	.39		.26
7	.38			.73
8	.45			.82
9	.41			.78
10	.65				.39
11	.58				.84
12	.71				.36
13	.53					.59
14	.57					.51
15	.55					.45

Note. All factor loadings were significant at p < .001. Loadings not represented in the table were constrained to zero.

Method

Results & discussion

Although the direction of its correlation with prosocial lying was consistent with preregistered predictions and previous studies (r = .14 in both Paul et al., 2022; Reinhardt et al., 2024), Honesty-Humility did not significantly predict prosocial lying, r = .11, CI₉₅ = [−.01, .22], p = .061. Similarly, though directionally consistent with preregistered predictions, the T-TCS did not significantly predict prosocial lying, r = −.07, CI₉₅ = [−.19, .04], p = .201.

However, as predicted, Honesty-Humility more positively predicted prosocial lying than did T-TCS, Z = 3.17, p = .001. That is, though neither trait measure reached statistical significance in directly predicting prosocial lying, the two showed different relationships with prosocial lying, with Honesty-Humility showing a more positive correlation than did T-TCS. Moreover, we examined a (preregistered) regression model predicting prosocial lying as a function of Honesty-Humility and T-TCS simultaneously. Honesty-Humility significantly predicted more prosocial lying, b (SE) = 0.15 (0.05), p = .003, whereas T-TCS predicted less, b (SE) = −0.14 (0.05), p = .009, and adding T-TCS to a model where Honesty-Humility predicted prosocial lying significantly increased variance accounted for, ΔR² = .02, p = .009 (though overall variance accounted for was still low, R² = .035).

Neither Honesty-Humility nor truthfulness predicted third party ratings of the feedback the participant gave. Inconsistent with Paul et al. (2022), Honesty-Humility did not predict feedback kindness, r < .01, p = .985. As in both Paul et al. and Reinhardt et al. (2024), Honesty-Humility did not predict feedback truthfulness, r = .01, p = .85. Similarly, T-TCS did not significantly predict ratings that the feedback was truthful, r = −.07, p = .247, or kind, r = .09, p = .123. Unlike with prosocial lying, Honesty-Humility and TCS did not show different relationships with ratings of how truthful, Z = 1.34, p = .089, or kind, Z = −1.53, p = .063, the feedback was. The same pattern held in multiple regression models.

In summary, the T-TCS showed a different relationship with prosocial lying than did an Honesty-Humility measure. Put simply, when looking at prosocial lying, Honesty-Humility was about the “prosocial,” truthfulness was about the “lying.” This suggests that the two measures capture different variance in conceptually meaningful ways, providing some evidence of construct differentiation. Moreover, T-TCS predicted variance in prosocial lying above and beyond that shared with Honesty-Humility, evidencing incremental validity for the T-TCS.

However, truthfulness did not significantly predict fewer prosocial lies (though it was directionally consistent) and did not predict the rated kindness or truthfulness of the feedback. Notably, results for Honesty-Humility in this sample were slightly weaker than in prior studies we were replicating (Paul et al., 2022; Reinhardt et al., 2024) though magnitude of effects for prosocial lying was comparable (r = .11 here, .14 in prior work), indicating weak relationships in this study were not a feature unique to the T-TCS. This may be in part because of low reliability due to the use of difference scores (but see Gollwitzer et al., 2014) or single instances of behavior (e.g., Epstein, 1979), or it may have to do with participants being less convinced by the paradigm in our study (especially compared to its original use by Lupoli et al. (2017), who used the grad school essay cover story among undergraduate students).

Method

Results and discussion

Honesty affordances

Based on both a priori and participant-identified affordances, participants experienced honesty-relevant situations quite frequently. About half of all episodes offered at least one of the researcher-defined honesty affordances (49.3%, see Table 4). Participant honesty ratings across episodes corroborates this frequency: Participants indicated their own honesty was not relevant in around 1/3 of episodes (32.9%), instead rating their honesty for 2/3 of episodes. This high frequency of honesty-relevant experiences suggested to us that an interval-contingent approach to experience sampling (i.e., sampling every X amount of time, rather than whenever participants had an honesty-relevant experience) would likely be sufficient, as most participant days contained (as a lower bound) one or more relevant experiences (Bolger & Laurenceau, 2013).

OPEN IN VIEWER

Table 4.

Percent of episodes for which participant indicated the affordance was present, Study 4.

	Present (%)	Mean honesty (SD)	Corr. btw DRM honesty and T-TCS
Social interaction	42.1	4.60 (0.88)	.22***
Asked to express	29.3	4.55 (0.94)	.20***
Opportunity to express	39.2	4.59 (0.90)	.22***
Difficult for self	29.5	4.58 (0.89)	.23***
Difficult for other	28.2	4.51 (0.97)	.25***
Any of above	49.3	4.55 (0.93)	.24***

Note. ***p < .001.

Item generation and refinement

With the working definition of truthfulness and the trait items in mind, several members of the research team (3 psychology PhDs, 2 graduate students) generated items intended to capture state manifestations of truthfulness. Taking the T-TCS as a starting point, we wrote items sampling a breadth of psychological components that characterize state experiences: behaviors, beliefs, motivations, emotions, and identity concerns.

In addition to capturing multiple elements of state manifestations of truthfulness, a broad set of items has the added benefit of potentially increasing variability in state measures of truthfulness. Several studies have indicated that most people report very few lies in day-to-day life (DePaulo et al., 1996; DePaulo & Kashy, 1998; Serota et al., 2022), which suggests that we should expect relatively low variability in truthfulness behaviors (see also the high mean ratings for honesty in Study 4, Table 4). However, including other situation-specific psychological elements such as evaluations (how right or wrong truthfulness is in that particular situation) or motivation (how much one wants the truth to be known at that moment) might capture meaningful differences absent in overt behaviors. For example, a person who struggles with the temptation to lie but tells the truth anyway is meaningfully different from a person who tells the truth without temptation. The two would look identical on a behavior-focused state measure but would look different on a state measure that captured momentary motivation in addition to overt behavior.

We sought a set of state items that exhibited several desiderata: (1) Content coverage of several trait domains—relevant behaviors, beliefs, motivations, emotions, and experiences of identity by which truthfulness would manifest in specific experiences, (2) Internal consistency as a single-factor measure (mirroring the observed structure for the trait form) (3) Variability in responses within and across persons and situations, (4) a relatively small number of items, to facilitate use in intensive longitudinal studies such as ESM or diary studies.

After several iterations of combining, adding, eliminating, and editing items, we identified twenty candidate items to present to participants (see Supplement Table S8 for full pool). Items each consisted of four statements of increasing endorsement (Table 5).

OPEN IN VIEWER

Table 5.

State truthful communication scale (S-TCS) items with factor loadings.

	Item	Psychological component	Study 5		Study 6
			λ Within	λ Between	λ Within	λ Between
1 (N)	(1) I believed that it was absolutely not OK to tell a lie	Belief	−.696	−.638	−.581	−.952
	(2) I believed that it was slightly OK to tell a lie
	(3) I believed that it was somewhat OK to tell a lie
	(4) I believed that it was absolutely OK to tell a lie
2	(1) I absolutely did not want someone to know my true beliefs about something	Motivation	.731	.887	.525	.780
	(2) I mostly did not want someone to know my true beliefs about something
	(3) I mostly wanted someone to know my true beliefs about something
	(4) I absolutely wanted someone to know my true beliefs about something
3	(1) I strongly did not want someone to know information as I truly saw it	Motivation	.752	.809	.517	.797
	(2) I mostly did not want someone to know information as I truly saw it
	(3) I mostly wanted someone to know information as I truly saw it
	(4) I strongly wanted someone to know information as I truly saw it
4 (N)	(1) I did not say (or otherwise communicate) anything untrue	Behavior	−.743	−.744	−.652	−.977
	(2) I said (or otherwise communicated) something that was slightly untrue
	(3) I said (or otherwise communicated) something that was mostly untrue
	(4) I said (or otherwise communicated) something that was totally untrue
5 (N)	(1) I told (or otherwise communicated) what I thought was the truth	Behavior	−.798	−.935	−.627	−.988
	(2) I told (or otherwise communicated) what I mostly thought was the truth
	(3) I told (or otherwise communicated) what I somewhat thought was the truth
	(4) I told (or otherwise communicated) what I did not think was the truth
6 (N)	(1) I completely believed that the truth had to be told	Belief	−.839	−.905	−.674	−.982
	(2) I mostly believed that the truth had to be told
	(3) I slightly believed that the truth had to be told
	(4) I believed that it was OK to not tell the truth

Method

Results and discussion

Goals

Dynamic theories of personality imply that a valid measure of state truthfulness should systematically vary as a function of different goals that people pursue, as social cognitive mechanisms such as goals determine how a person manifests a trait at a given time. To test this, we examined whether and how state truthfulness varied based on the situation participants were assigned to report on.

First, a multilevel model predicting S-TCS as a function of assigned goals indicated that state truthfulness varied as a function of goals (see Figure 2), which accounted for ∼20% of variance in state reports, F (11, 1322) = 31.24, p < .001. Thus, the goals that people were pursuing were a significant contributor to how truthful they were. Coefficients for the various situational goals (Figure 3) indicated that participants were less truthful when trying to leave conversations, avoid doing something, or avoid blame and were more truthful when trying to deliver unpleasant news or negative feedback, tell a story, explain instructions, complete a task, or change someone’s mind. Thus, not only did state manifestations of truthfulness vary based on situation, but they did so in reasonable, systematic ways.OPEN IN VIEWER

Figure 2.

State TCS Scores by Assigned Goals, Study 5. Note. Goals are ordered by median then by mean truthfulness.

OPEN IN VIEWER

Figure 3.

Fitted Means of State Truthfulness by Assigned Goals, Study 5. Note. The vertical line denotes the grand mean, error bars denote 95% confidence intervals.

Method

Results and discussion

Test-retest for trait truthfulness scores

Test-retest correlations for the T-TCS indicated that test-retest reliability three weeks later was good (r = .75, p < .001), and comparable to domains of the HEXACO (Table 6). This constitutes further evidence for the reliability of the T-TCS as a measure of a stable personality feature.

OPEN IN VIEWER

Table 6.

Test-retest correlations for truthful communication and HEXACO scales, correlations with aggregated state honesty, Study 6.

Time 1 variable	Test-retest	Time 1 T-TCS	Aggregated S-TCS scores	Aggregated S-TCS behavior scores	Single S-TCS scores	Single S-TCS behavior scores
T-TCS	.75***		.53***	.52***	.37***	.38***
Honesty-humility	.85***	.57***	.30***	.35***	.23***	.27***
Emotionality	.66***	.02	−.10	−.06	−.07***	−.05***
Extraversion	.67***	.36***	.26***	.27***	.17***	.19***
Agreeableness	.71***	.47***	.27***	.32***	.20***	.24***
Conscientiousness	.75***	.65***	.35***	.38***	.25***	.28***
Openness	.59***	.43***	.33***	.30***	.20***	.19***

Note. T-TCS = Trait Truthful Communication Scale; S-TCS = State Truthful Communication Scale.

State manifestations

Factorial structure and variance components

As in Study 5, we fit multilevel confirmatory factor analysis models to examine the factor structure for the S-TCS, modeling a single truthfulness factor both within and between persons. We again constrained loadings to be equal across levels. As in Study 5, initial fit was mediocre—χ² (23) = 704.69, p < .001; CFI = .92, TLI = .90, RMSEA = .074, SRMR = .107—but allowing correlated residuals for the pairs of related items (i.e., the two items for each psychological component) dramatically improved fit to excellent levels—χ² (17) = 87.84, p < .001; CFI = .99, TLI = .99, RMSEA = .028, SRMR = .067. See Table 5 for factor loadings. About half of the variance in state TCS scores was within participants (48.0%), and reliability was good at the between-person level (ω = .94) but questionable at the within-person level (ω = .74). See Supplement for comparable results for the behavior-focused state measure.

Situational features and goals

Participants indicated that one or more of the 21 researcher-defined situational features were applicable in 86.2% of the state reports (median = 1 feature, mean = 1.90). Because situational features were not exclusive, we analyzed them as a series of independent indicator variables. Figure 4 presents fixed effect coefficients for the effect of each situational feature on state truthfulness, with reports nested within participants. The strongest positive effects were for telling a story or joke, having a conversation about something another person did not want to do, fulfilling a desire, and disagreeing with another person. The strongest negative effects were for having to provide negative feedback, dealing with undesirable responsibilities, or having an uncomfortable or awkward conversation. ²³ When modeled simultaneously, these situational features only accounted for 3.1% of the variance in state truthfulness. ²⁴

What did substantially predict peoples’ state manifestations of truthfulness were the goals they said they were trying to achieve in that moment. We modeled state truthfulness as a function of fixed effects of momentary goals, with a random intercept by participant, and found that the four goals we included accounted for nearly a quarter of variance in state truthfulness (23.4%). Sensibly, goals to avoid disapproval and protect others’ feelings predicted less truthful communication, whereas goals to treat someone with respect or convey needed information predicted more truthful communication (Table 7). This strong and reasonable pattern of relations provides further evidence for the utility of the S-TCS, inasmuch as participants are neither constantly reporting maximal truthfulness, nor are they reporting random deviations from truthfulness. Instead, participants report systematic deviations from truthfulness that align with reasonable assumptions about how truthfulness/untruthfulness will serve momentary goals.

OPEN IN VIEWER

Table 7.

Fixed effects of goals on state truthfulness, Study 6.

Predictor	Estimate	CI	p
(Intercept)	2.95	2.85–3.05	<0.001
Avoiding disapproval	−0.09	−0.10–−0.08	<0.001
Protecting another’s feelings	−0.02	−0.03–−0.00	0.021
Treating someone with respect	0.03	0.02–0.05	<0.001
Conveying needed information	0.17	0.16–0.19	<0.001

Note. Effects are unstandardized regression coefficients indicating the change on a 1–4 scale for a one unit change on a 1–5 scale for that goal.

The Trait Truthful Communication Scale

To assess truthful communication at a trait level, we developed a 15-item scale reflecting the construct in terms of its dynamic components and in terms of a virtue. Intended to tap into dynamic processes underlying personality traits (e.g., Fleeson & Jayawickreme, 2021), the T-TCS includes content related to behavior, emotion, cognition, and motivation. Reflecting truthful communication as a virtuous characteristic, the T-TCS focuses on specific forms of emotion, cognition, and motivation (Miller, 2021; Wright et al., 2021). That is, it focuses on morally related emotions (i.e., feelings of guilt or shame related to being untruthful), morally related cognition (i.e., believing that truthful communication is good or correct), and morally related motivation (i.e., wanting others to know one’s true beliefs, rather than wanting to avoid being caught lying).

Evidence indicates that the T-TCS has good psychometric quality. Analysis of bifactor models (Study 2) revealed that all items loaded significantly on a general TCS factor and on five specific factors, but that variance in total scores (e.g., overall item mean scores) was primarily attributable to the general factor. Thus, the items as a whole reflect truthful communication in general, but also systematically tap into subtle but detectable differences between different types of psychological manifestations (behavior, motivation, etc.). These findings suggest that the entire T-TCS can be used to reflect trait-level truthful communication in general.

Analyses also reveal that the T-TCS has good to very good reliability. Its omega reliability estimate was .95 (and .94 in a preregistered replication), and its 3-week retest correlation was .75. Notably, the latter value was equal to or greater than the retest correlations for five of the six HEXACO scales in the same sample, suggesting that stability over a few weeks is comparable to well-established trait measures.

Evaluation of T-TCS item content reveals conceptual precision in indicating truthfulness. In a supplemental Study 7 (see Supplement pp. 32–42), participants rated the clarity with which each T-TCS item would reveal a person’s level/standing on several psychological dimensions related to truthfulness and honesty/humility facets. T-TCS items were rated as significantly more revealing of a truthful-versus-deceptive dimension than any other dimension (e.g., greed avoidance), suggesting conceptual precision. Participants also rated each HEXACO Honesty-Humility item, and they saw T-TCS items as being more revealing of a truthful-versus-deceptive dimension than were any HEXACO Honesty-Humility facets. The latter finding is not only consistent with our suggestion that the broad HEXACO-PI-R’s Honesty-Humility scale (nor any of its facets) does not precisely reflect truthfulness specifically but it also suggests that the T-TCS item content reflects truthfulness better than does HEXACO Honesty-Humility item content. In other words, while Honesty-Humility scales are extremely useful for targeting the broad Honesty-Humility trait, the TCS complements them by targeting the narrower trait of truthfulness.

Finally, analyses revealed good convergent, discriminant, and predictive validity and incremental validity over Honesty-Humility for predicting state truthfulness manifestations and potentially for predicting decreased prosocial lying. When examined alongside a diverse range of other personality trait measures, the pattern of correlations between T-TCS scores and those measures was strongly consistent with expert judges’ predictions for a hypothetical measure of truthfulness, in both an initial study (2a) and a preregistered replication in a different population (2b). For example, people with relatively high T-TCS scores tended to have relatively high scores on measures such as the CIVIC Honesty/Authenticity scale, the Realness scale, and the (IPIP version of the) VIA Integrity scale, as predicted. They also tended to have relatively low scores on measures such as the SD3 Machiavellianism scale, Bullshitting scales, and the Lie Acceptability scale, again as predicted. Demonstrating discriminant validity, T-TCS scores showed weak or null associations with scales such as the Intellectual Humility scale, HEXACO Openness, and CIVIC Humor, once again as predicted. ²⁵ As revealed by our QCV analyses, the overall pattern of convergent and discriminant validity correlations was strongly consistent with what our experts predicted should be found for a valid measure of truthful communication. In terms of predictive validity, T-TCS scores were significantly positively correlated with contextualized reports of truthful behavior in daily life, assessed in multiple ways (Furr, 2009)—in a day-reconstruction study of truthful behavior from the previous day (Study 4), retrospective reports of behavior in specific situations (Study 5), and experience sampling measurements of behavior in-situ from the last few hours (Study 6). Finally, in directly contrasting the T-TCS with what is likely the most commonly used trait honesty construct—HEXACO Honesty-Humility—the T-TCS showed incremental validity over measures of Honesty-Humility, predicting unique variance in behavioral assessments of prosocial lying (Study 3) ²⁶ , retrospective reports of truthfulness (Study 5), and ESM reports of recent truthfulness (Study 6).

In sum, we believe that the psychometric evidence indicates that T-TCS scores are reliable and interpretable as indicating individual differences in the tendency toward truthful communication. Moreover, we believe the evidence also indicates that the measure is not redundant with the preeminent honesty construct but has unique utility for understanding truthfulness. Given the importance of truthfulness and its centrality within the scope of honesty, the T-TCS should be a valuable tool for researchers interested in individual differences in general and for those interested in honesty specifically.

The State Truthful Communication Scale

We also designed a 6-item scale to assess truthful communication at a state level, again reflecting the construct in terms of multiple dynamic components. With the notable exception of emotion, the S-TCS captured the same components of truthful communication as the T-TCS: behaviors, motivations, and beliefs. While we included items assessing emotion as part of the pool of S-TCS items in Study 5, we excluded them from the final measure since experiencing emotions related to truthful communication is contingent on manifesting high or low levels of truthful communication in a given situation. This makes such items distinctive from the other components, which directly ask about manifestations of truthful communication. Given the need to be clear about the function of this state measure (i.e., assessing the extent to which people manifested truthful communication; Horstmann & Ziegler, 2020), and to ensure that the final measure had good psychometric properties, we decided to exclude emotion assessment from the S-TCS. However, assessing such contingent emotional reactions when manifesting truthful communication in daily life is an important goal for future research.

In designing the S-TCS, we were mindful to use best practices in state measure development (e.g., Horstmann & Ziegler, 2020). In particular, we developed a set of state items that would both ensure that the S-TCS would successfully capture the core features of truthful communication and be brief enough to be easily used in intensive longitudinal assessment. We sought content coverage of several characteristics of truthful communication manifestations in specific experiences (i.e., momentary behaviors, beliefs, and motivations), a single-factor measure (similar to the trait measure), good variability in responses within and across persons and situations, and a final set of brief items that would lend itself to use in ESM and other intensive designs.

We found that the S-TCS has good psychometric properties—there was decent fit for a single truthfulness factor at both the within- and between-person level. We also found that T-TCS scores strongly predicted aggregated state S-TCS scores. Furthermore, consistent with the predictions of dynamic personality theories, S-TCS scores exhibited substantial between- and within-person variability. Indeed, the observed within-person variability when assessed via ESM was comparable to that observed in the honest behavior ESM studies reported by Meindl and colleagues (2015), suggesting that people vary meaningfully in their manifestations of truthful communication from situation to situation (a fact further suggested by the moderate correlation between single-state S-TCS and T-TCS scores).

Furthermore, and also in line with dynamic personality theories, variability in S-TCS was systematically associated with specific goals and situational features. Specifically, WTT predicts that S-TCS scores will vary as a function of different goals that people pursue, and indeed, people were less truthful than average when trying to leave conversations, avoid doing something, or avoid blame. Conversely, they were more truthful than average when trying to deliver unpleasant news or negative feedback, tell a story, explain instructions, complete a task, or change someone’s mind. We also found that specific goals predicted variability in S-TCS scores across different situations. Goals to avoid disapproval and protect others’ feelings predicted lower levels of truthful communication, whereas goals to treat someone with respect or convey needed information predicted higher levels of truthful communication.

Implications

Strengths of current approach

Our trait and state measures of truthful communication break new ground by providing morality researchers with a conceptually grounded assessment of the central psychological characteristics of a core feature of honesty (Reynolds et al., 2025). Although the data reported here use self-report assessments (although we note that trait and state assessments represent different forms of self-report; Jayawickreme et al., 2023), these measures can be easily adapted as informant assessments, similar to other work from our lab (see also Sun & Goodwin, 2020, who adapted assessments from Furr et al., 2022). While limitations exist with self-report measures, we note that self-report assessments are likely of primary importance when assessing specific psychological characteristics of truthful communication that are particularly subjective in nature (e.g., motivation). Nevertheless, we fully endorse a multi-method approach to understanding morality (Helzer et al., 2014), and expect that self-informant agreement on T-TCS scores will be significant (we are currently collecting data to test this prediction).

Truthfulness is an important construct at the heart of honesty, and as such, it deserves special attention from researchers wishing to understand honesty. To advance such research, the current work presents a new measure focused on this important construct; however, other facets of honesty also deserve greater measurement attention. Many constructs lie within the scope of honesty, including (not) cheating, (not) stealing, (not) promise-breaking, truth-seeking, and fostering accurate understanding in others (e.g., B. Cooper et al., 2023; Miller, 2021). Although such constructs tend to co-occur within people (thus producing the Honesty/Humility factor), they are conceptually separable and may have distinct psychological causes, correlates, and consequences. A full understanding of honesty as a personality characteristic requires psychometrically strong and focused examination of each narrower construct.

Unfortunately, the observation that existing honesty scales fail to reflect truthfulness precisely likely also applies to those constructs. Most honesty scales measure multiple narrower constructs (Fleeson et al., 2022), making it unclear which distinct facets of honesty are operative in criteria behaviors. This is an inherent feature of the tradeoff between broad traits that can predict a wide array of behaviors and narrow traits that can strongly predict precise behaviors (e.g., Buss, 1989), so it is not specific to honesty. Nonetheless, we believe that future study of honesty is facilitated by more precise conceptualizations and measurements of specific constructs within honesty. We thus encourage future measurement-oriented work to produce conceptually well-grounded and psychometrically strong measures of many specific constructs within the broad scope of honesty.

Limitations and future directions

Although based on eight samples representing multiple methods of recruitment and sampling, the current studies are geographically and culturally limited. Although the US samples are diverse in age and location within the US, the studies include relatively few persons of color. Moreover, all samples are from Western countries, so they reflect Western, industrialized perspectives and experiences. Future research should evaluate the TCS’s properties within a wider range of participants.

More generally, empirical work centering truthfulness at the core of honesty (Reynolds et al., 2025) has focused on Western (especially US) samples. To the extent that conceptualizations of moral traits may differ cross-culturally (e.g., Shi et al., 2021), truthfulness may not feature as prominently to honesty outside of Western populations. Thus, though TCS may capture trait truthfulness outside of Western populations, that feature might not be as central to studying honesty in these contexts. We invite future cross-cultural research on the relative importance of various facets of honesty.

To keep the T-TCS relatively brief but inclusive in its coverage of key psychological processes, we included sets of three items assessing several processes—motivation, emotion, and cognition, and both positive and negative behaviors. This ensures that the overall scale is balanced and representative of important processes underlying dynamic personality characteristics. We anticipate that most researchers will wish to use the T-TCS in this way, relying on a total score across all items. However, some researchers might want to focus even more deeply on one or more of those processes, and they might generate sub-scale scores based on the relevant T-TCS items. We recommend that such work be done with caution. As noted in Study 2 and the Supplement (pp. 7–10), scores on subfactors predominantly reflect the common truthfulness factor, making specific subscale scores than if the subscales were more distinct. Again, we anticipate that most researchers will opt for total T-TCS scores, and our current analysis suggests such scores have robust psychometric properties. We suspect that the 3-item sub-clusters within the T-TCS may thus be good foundations for more narrowly focused measures of truthfulness-related motivation, emotion, and cognition. However, the development and evaluation of such narrower measures is a task for further work.