Realness refers to the relatively stable tendency to behave on the outside the way one feels on the inside, without regard for proximal personal or social consequences (Hopwood et al., 2021). Previous research suggests that people who are more real generally tend to have many other psychological strengths, including lower neuroticism and higher conscientiousness, extraversion, openness, and social competence. However, people who are real do not tend to be more (or less) agreeable or communal with others. This pattern suggests that sometimes, being real requires saying or doing things others may not like, but that ultimately people who are able to be real even when it is difficult to do so tend to experience greater psychological functioning. Extending previous cross-sectional associations findings that speak to the disposition towards realness, the goal of this study was to examine the co-development of realness and relevant correlates over the course of 18 months in young adults. We aimed to both replicate and extend previous findings regarding the association of realness with personality traits and examine associations with perceived stress and social support.

Realness

Realness can be conceptualized as part of the broader concept of authenticity. From this point of view, authenticity is understood as a dynamic, multidimensional process, whereas realness is a more specific, core feature of that process that is revealed only in social situations in which there is internal or social pressure to be inauthentic. Parsing realness from authenticity is important for two reasons.

First, authenticity measures often use multiple cross-sectional scales to represent different steps in a dynamic temporal process that involves inner values, self-awareness, or various styles of external expression (Kernis & Goldman, 2006; Knoll et al., 2015; Wood et al., 2008). The challenge is that it is not logically possible to capture a temporal process with scales that are all measured at the same time. The application of multidimensional authenticity scales in cross-sectional data to make inferences about an authenticity process has likely led to obscured findings that do not speak directly to the construct (Hopwood et al., 2022). Although this kind of research can show how different parts of authenticity are related to other constructs, it does not show how they fit together to produce authentic behavior. Moreover, some constructs in such scales are more central and specific to authenticity than others. For instance, insight is often included as one feature of authenticity, but insight is a core feature of many psychological constructs beyond authenticity. Thus, realness focuses on a single, unidimensional, and core element of authenticity processes. By narrowing the focus to this core feature, the hope is that a more solid body of research can be generated, onto which studies using experimental and longitudinal methods could be used to build up an evidence-based model of dynamic authenticity processes.

Second, classical theories of authenticity emphasize that being real comes with potential downsides such as social discomfort and reputation loss (Al-Khouja et al., 2022; Jongman-Sereno & Leary, 2019; Maslow, 1968; Rosenblum et al., 2020). Nearly universal and largely indiscriminate correlations between existing measures of authenticity and measures of well-being and adaptivity suggest that these features are mostly absent from modern models of authenticity (Bailey & Iynagar, 2022; Hicks et al., 2019). In particular, whereas previous theories in the existential, psychodynamic, and humanistic traditions have emphasized that being authentic often comes with social risks and costs because it often requires a departure from norms, all existing measures of authenticity correlate positively and often strongly with agreeable or communal behavior. In contrast, realness focuses on being real (i.e., authentic) in situations in which there is potential for negative personal consequences, and thus should correlate neither positively nor negatively with agreeable behavior.

The initial validation of the realness construct included nine studies (Hopwood et al., 2021). The first study demonstrated that realness is a core feature of existing multidimensional authenticity measures, but this core feature is also contaminated by positive valence, as demonstrated by strong positive correlations between common authenticity measures and agreeableness. In the second and third studies, a self- and informant-report Realness Scale was developed with a unidimensional structure. This measure correlated positively with existing authenticity measures as well as measures of mental health and well-being but was uncorrelated with agreeableness. Studies four and five confirmed positive correlations with adaptive traits (low neuroticism, high extraversion, high conscientiousness, and high openness) but negligible correlations with agreeableness. Study six used a peer-nomination strategy to show that, among friends that are equally liked, real friends are seen as less agreeable than polite friends. Study seven showed a retest correlation of 0.74 across 5 months, and no change in level over that span. In study eight, a reliable observer-coded version of the Realness Scale was validated. Finally, in study nine, the Realness Scale was translated to German and validity evidence from questionnaires and a situational-judgement measure showing that realness is related to social competence, but again unrelated to agreeableness. A follow-up study confirmed convergent validity with other measures of authenticity (McCutcheon et al., 2022).

Individual differences in realness

Correlations between realness and validating variables in these initial validation studies provide a relatively detailed picture of the construct’s nomological net, including how realness differs from other measures of authenticity, in cross-sectional data. These results suggested that people who are more real tend to be more emotionally mature, socially competent, and adaptive in general. This pattern can be summarized relatively well using big five traits. People who are more real have lower levels of neuroticism, implying that they are more likely to be able to tolerate the discomfort that comes with being real despite social pressure to do otherwise. Their greater social skills and positive affect, as indicated by positive associations with extraversion, might enable them to better navigate such situations. Higher conscientiousness suggests that people who are more real are also more committed to principle but also social tact in their interactions with others in a way that could facilitate realness. Finally, higher levels of openness to experience could lead people who are more real to have less normative views about the world, and thus more opportunities to differ from others in a way that potentiate opportunities for realness.

These adaptive correlates have also been observed in most other measures of authenticity (e.g., Hicks et al., 2019; Sheldon et al., 1997; Wood et al., 2008). The main difference between realness and common authenticity measure involves associations with agreeableness, as discussed above. The lack of association between realness and agreeableness suggests that people who are real can overcome the social pressure to be polite or follow social norms during times when there is a choice between either avoiding social awkwardness or expressing how one really feels. Put the other way around, this null association suggests that people who are truly agreeable in terms of their motivations to connect with others are likely to be agreeable and real at the same time, whereas people who are truly disagreeable motivationally – in situations in which those motivations may create social discomfort for self or others – will with either tend to be real and disagreeable or fake and agreeable. Notably, informant report data suggested that others see more real vs. more polite friends as more communal and kinder despite the lack of an association between realness and agreeableness in self-report data (Hopwood et al., 2021). This indicates that others tend to appreciate realness and even perhaps interpret a communal motivation to real behavior, perhaps even when it seems disagreeable in the moment. In other words, people seem to like others who are real, even when that means they are not always agreeable.

Social factors in realness

Being real – saying how you feel, think, or what you want even when you expect that others may not like it - requires a certain level of confidence, comfort, and self-assurance. This confidence can be bolstered by contextual resources that support autonomous self-expression, whereas it can be inhibited by contextual stressors that constrain self-expression. Horney (1950) emphasized that stress will tend to decrease the likelihood realness, whereas support from close attachment figures tends to enable it. Several authors have emphasized that a perceived lack of social support encourages people to fit in rather than be themselves (May, 1953; Rogers, 1961; Maslow, 1968). Psychotherapeutic approaches that regard realness as an outcome of effective treatment assume that succumbing to social pressure in a way that suppresses self-expression is an understandable response to stressful environments, whereas realness is promoted via the social support provided by the therapeutic relationship (Winnicott, 1958; Rogers, 1961). In general, cross-sectional findings in the literature have suggested that people report being more authentic if they experience greater social support (Didonato & Krueger, 2010; Sorrell et al., 2023; Wang et al., 2019) and less stress (Ryan et al., 2005; Sun & Liang, 2022). It follows that a valid measure of realness should be positively correlated with social support and negatively correlated with social stress. This hypothesis has not been tested in previous research. In addition to replicating previously identified correlations between realness and the big five traits, this study aims to examine whether realness is positively associated with social support and negatively associated with social stress.

Stability

Most research on authenticity and realness conducted thus far has focused on correlates using cross-sectional designs or changes over the short-term using ambulatory designs (Fleeson & Wilt, 2010; Heppner et al., 2008; Lenton et al., 2013, 2016; Lutz et al., 2023; Ryan & Deci, 2006; Ryan & Ryan, 2019; Sedikides et al., 2017;, 2019). Studies conducted over longer-term intervals suggest that authenticity measures function similarly to traits (Bleidorn et al., 2022), with relatively little mean-level change, retest correlations over extended intervals around 0.70, and decreasing stability estimates in younger adults (Bleidorn, 2015; Roberts & Davis, 2016) and longer retest intervals (Boyraz et al., 2014; Choi et al., 2022; Hill et al., 2013; Impett et al., 2008; Leroy et al., 2013). In the only previous realness study, the retest stability was 0.74 in college students over 9 months, and there was no mean-level change (d = 0; Hopwood et al., 2021). In this study, we aim to build upon this research by examining changes in realness across 4 waves over 18 months. We also hoped to use these data to examine longitudinal correlates between realness and criterion measures (Bleidorn et al., 2019) given the possibility that environmental factors during young adulthood could affect realness trajectories (Bleidorn et al., 2009; Hopwood et al., 2011; Kokko et al., 2013). However, as described below, realness was too stable to examine co-developmental effects.

Current study

The goals of this study were to (a) replicate cross-sectional associations between the big five traits and realness indicating negative associations with neuroticism and positive associations with extraversion, openness to experience, and conscientious, (b) examine whether realness is associated with greater social support and lesser social stress, and (c) estimate the stability of realness over time. We initially planned to examine longitudinal correlations between changes in realness and changes in personality traits, stress, and social support, but individual differences in change in realness were not sufficient for the appropriate analyses.

Method

Sample and Procedure

This study represents a secondary analysis of an existing data set. The original study was approved by a local Institutional Review Board, and all participants provided informed consent. Participants were recruited through different channels, including the university’s study abroad office, flyer postings, and advertisements. Assessments were accessible with a personalized survey link through Qualtrics, an online survey platform, and participants were paid for participating.

Surveys were administered at baseline, 4 months after baseline, 8 months after baseline, and 18 months after baseline. A total of 412 participants provided data on relevant variables for this study; retention was 86% at Time 2, 79% at Time 3, and 72% at Time 4. Age ranged from 18 to 31 years (M = 19.53, SD = 1.85) and participants were 82% women, 16% men, and 2% identifying as another gender. The sample was 28% White, 29% Asian, 27% mixed, 1% Black, and 15% indicated another ethnicity; 22% indicated that they were of LatinX/Hispanic origin.

This study was not preregistered thus all hypotheses were exploratory. We interpreted correlations as significant at p < .05 and focused interpretation primarily on the magnitude of effect sizes. Analyses were conducted in R.

Study variables

Table 1 provides the means and standard deviations for all variables at each wave. We also computed both Cronbach’s alpha and McDonald’s hierarchical omega for all variables at each wave as internal consistency values (see Table 1). Table 2 contains the correlations between our study variables at Time 1.

Realness was measured using the Realness Scale (Hopwood et al., 2021), and items were rated from 1 (false) to 4 (very true). Cronbach’s alpha and McDonald’s hierarchical omega ranged from 0.78 to 0.85 across all waves.

Big Five personality traits (neuroticism, extraversion, openness, agreeableness, and conscientiousness) were measured at baseline with the Big Five Inventory − 2 (Soto & John, 2017), including 60 brief characteristics that were rated from 1 (strongly disagree) to 5 (strongly agree). Both internal consistency values ranged from 0.80 to 0.89 across all variables and waves.

Perceived stress was assessed with the 10-item Perceived Stress Scale (Cohen, 1988), and items were rated from 0 (never) to 4 (very often). Alphas and omegas ranged from 0.82 to 0.85.

Perceived social support was measured with the Perceived Social Support Scale (Zimet et al., 1988), and items were rated from 1 (very strongly disagree) to 7 (very strongly agree). Alphas and omegas ranged from 0.90 to 0.94.

Table 1 Study descriptive statistics and internal consistency values at each wave
Full size table
Table 2 Study variable intercorrelations at Time 1
Full size table

Transparency and open Science

No data were excluded in this study. Although none of the analyses for this paper had been conducted before its conception, this is a secondary analysis of an existing dataset, and the research team are familiar enough with the data that it would have been unrealistic to preregister hypotheses. Thus, this should be considered an exploratory study, albeit with predictions based on previous literature. Data and R script are available at https://osf.io/juqan/?view_only=db1131f8cd904924ae739e0004b9c037.

The main study includes two cohorts of undergraduate students. The first cohort contains a group of students who studied abroad for 4 weeks over the summer and a control group of students who did not. The second cohort contains a group of students who did not study abroad (although there were a group of students who were set to go abroad) and were assessed from the onset of the COVID-19 pandemic. Both cohorts were assessed monthly over 8 waves followed by a follow-up 18 months from the first assessment. We used the second cohort from the main study because of the availability of longitudinal assessments across four waves for all study variables. Four other studies used data from the main study (Bühler et al., 2022; Hopwood et al., 2021; Nissen et al., 2022; Weber et al., 2023). Hopwood et al. (2021) examined retest correlations and mean-level change in realness between the first and second assessment in cohort 1 (non-overlapping sample). Nissen et al. (2022) examined selection and socialization effects of study abroad in a range of variables including realness, and found that realness, along with most other variables neither predicted entry into study abroad nor was affected by having studied abroad. Nissen et al. (2022) did not use data from cohort 2 and did not examine associations between realness and other study variables. However, it did confirm that there were very modest changes in realness over the course of the year, and that there were significant individual differences in change. Bühler et al. (2022) examined cohort differences in the criterion variables used in this study but did not examine realness. Finally, Weber et al. (2023) examined connections between self-esteem and self-concept clarity but did not examine realness.

Results

We used the dplyr (Wickham et al., 2022), lavaan (Rosseel, 2012), and psych (Revelle, 2022) packages of R (R Core Team, 2022) for analyses. We first fit univariate growth curve models with linear slopes for each of the study variables to confirm adequate model fit. All univariate models fit the data reasonably well (West et al., 2012; Table 3). Table 4 provides the unstandardized intercept and slope means and variances from the univariate growth models. Notably, neither the slope nor the slope variance was significant for realness. This suggests that realness was highly stable over time in terms of both average changes (Table 1) in the cohort and variation in those changes across individuals within the cohort (Tables 4 and 5). Interestingly, realness showed fewer individual differences in change than social support and the personality traits neuroticism, extraversion, and agreeableness. Non-significant individual differences in change precluded analyses of longitudinal covariation between realness and other variables.

Finally, we computed correlations between a latent variable composed of all four estimates of realness, and a corresponding latent variable composed of all four estimates of the other variables (Table 6). All cross-sectional associations were consistent with estimates using raw variables (Table 2) and our predictions: higher realness scores were related to lower neuroticism and stress and higher extraversion, openness, conscientiousness, and social support.

Table 3 Model fit indices for univariate growth curve models
Full size table
Table 4 Unstandardized parameter estimates from univariate growth curve models
Full size table
Table 5 Realness retest correlations
Full size table
Table 6 Realness standardized correlations with criterion variables
Full size table

Discussion

There were three major findings from this study. First, we replicated cross-sectional associations between realness and lower neuroticism and higher extraversion, openness, and conscientiousness, as well as the absence of a correlation between realness and agreeableness. Second, we found that lower levels of stress and higher levels of social support are associated with a higher likelihood of being real. Third, we found that realness was highly stable over the course of 18 months both in terms of average level of change in the sample overall and variation in changes across individuals within the sample.

Realness and the big five

Correlations between realness and big five personality traits have now been replicated across six separate samples with different big five instruments and in two different languages. Overall, the emerging pattern is very robust. People who are real are more likely to be less neurotic and more extraverted, open to experience, and conscientious, but they are not more likely to be agreeable.

As described in the introduction, the general pattern of adaptive personality correlates is consistent with previous theories indicating that realness is a marker of general psychological health. Moreover, specific mechanisms can be inferred about the role of each of these traits in supporting authentic behavior. Lower neuroticism might predict realness because negative feelings such as anxiety might lead people to bend to social pressure rather than expressing themselves. People who are extraverted might be more real because their better social skills may help them find solutions to complicated interpersonal situations and because positive affect can buffer the potential negative consequences of doing so. Conscientious people might be more real because of their commitment to doing the right thing, given that being direct with others about how one feels is generally considered the right thing to do in most situations. Open people might be more real because they have more opportunities to do so, departing as they tend to do from social convention.

Perhaps most important is that unlike other measures of authenticity, realness is not correlated with agreeableness. Being real means being able to always express your true self, but the real test of realness comes when there are social costs. Both real and “fake” people can be agreeable in moments when doing so is advantageous for the self and for others. It is during moments when there are potential costs – such as when standing up to power to speak the truth risks censure or loss of freedom, or when telling someone how you really feel may hurt their feelings, or when saying what you really want may set you apart from others in a way that risks social status – that reveal whether people are truly real. These situations require the elevation of being true to the self above other concerns, and in particular above maintaining social harmony. The potential for being real to be disagreeable has been strongly emphasized in foundational authenticity scholarship (Horney, 1950; Jongman-Sereno & Leary, 2019; Maslow, 1968; Rogers, 1961; Rosenblum et al., 2020) but has been lost in contemporary empirical research on the construct (Kernis & Goldman, 2006; Knoll et al., 2015; Sheldon et al., 1997; Wood et al., 2008). In this way, realness corresponds more closely to these foundational theories than existing measures of authenticity.

This is not to say that it is always disagreeable to be real, or that it is advantageous to always be real. It seems most likely that the most adaptive people will be able to figure out when they should choose to reveal themselves and when they should not, and thus realness across all situations would be very unlikely to be adaptive. However, the positive correlations in this study and previous research with adaptive features indicates that people who are real are more likely to be psychologically mature and healthy. This suggests an adaptive balance between being able to express oneself truly while minimizing social and personal costs.

Realness and social sontext

Given its general advantages for social behavior and well-being it is important to also understand the environmental conditions that promote realness. Two candidate features explored in this study are stress and social support. It is reasonable to infer that people who are under greater levels of stress are less likely to be real, particularly under social pressure not to. Stress tends to motivate conformity (Lewis et al., 2008) and diminish social creativity and problem-solving (Kassymova et al., 2019). When social contexts are chaotic or threatening, it is often advantageous or strategic to go with the flow, rather than to take the risk of standing out. Indeed, in this study we found that people who were under greater stress also tended to be less real, supporting this line of thinking.

It also stands to reason that realness is encouraged by the sense of support from those around you. Being real involves a risk of being ostracized, criticized, or out-grouped. When you feel supported, you may be more likely to take that risk, both to those who are close and those who are not. With those to whom you are close, being real may deepen relationships, particularly when there is a strong foundation of support that implies that being genuine will not lead to a relationship rupture. With those to whom you are not close, there is less of a personal risk of being real even if this does cause personal or social costs, because of the security provided by the support from others. Consistent with these ideas, in this study we found that people who felt stronger social support were also likely to be more real. Taken together, these findings may provide some initial evidence that the promotion of authentic behavior, which is commonly given as a developmental goal (Schachter & Ventura, 2008), can be achieved in part by reducing stress and increasing social support (Hoeeg et al., 2018). Future longitudinal and experimental studies are needed to test the causal role of social factors in promoting realness.

Stability of realness in young adults

Similar to previous research on personality traits (Bleidorn et al., 2022), longitudinal research with other authenticity measures (Boyraz et al., 2014; Choi et al., 2022; Hill et al., 2013; Impett et al., 2008; Leroy et al., 2013) and a previous two-wave study on realness over 4 months, we found that realness operates like a personality trait in terms of stability. It is highly stable in terms of both average scores in the sample (18 month d = − 0.08, nonsignificant slope) and the rank ordering of individuals over time (18 month r = .64, nonsignificant individual differences in change). In fact, whereas both the slope term and individual differences in change in a univariate growth model for realness were not significant, several other variables in the study had significant slopes and individual differences in change. Specifically, there were slight but significant decreases in openness, conscientiousness, stress, and social support over the course of the study, and variation in change was significant for neuroticism, extraversion, agreeableness, and social support. These findings suggest that realness was even more stable than some personality traits, although differences were slight and confidence intervals for these values overlapped. The general point is that individual differences in realness are highly stable over time, to the degree that it functions similarly to other personality traits (Schwaba & Bleidorn, 2018).

Limitations

The results of this study and these kinds of speculations are tempered by several study limitations. First, all data were self-reported. Although replication across studies and across cross-sectional and longitudinal data in this study helps limit method effects, they may still partially explain some findings. Moreover, it is possible that a different pattern of findings would be obtained with different measurement approaches. Future work should incorporate multimethod assessments to solidify study findings.

There were several limitations related to our sample. Participants were largely privileged college students from a WEIRD culture. Although they were not psychology students participating for course credit, and the university from which they were sampled is relatively diverse in its ethnic profile, these results may not generalize to individuals in other contexts, countries, or age groups. Moreover, women were overrepresented in this study, and this could have influenced results. Future work in other settings would strengthen confidence in these findings. Peculiar to this study was that we sampled individuals during the COVID-19 pandemic (March 2020 – July 2021), which was also a time of other stressors for Californians due to summer wildfires, a contentious national election, and significant racial tensions (e.g., murder of George Floyd and subsequent importance of the Black Lives Matter movement) in which many students were actively involved. It is not clear how these contextual factors may have affected our findings.

Finally, several methodical factors are important to consider when interpreting these results. Findings may be due in part to the timescale at which we sampled our participants, and different patterns could result if they were assessed more frequently or across a longer duration. We only focused on between-person individual differences, but it would also be useful to know about within-person patterns, such as whether people are more real during times they experience less stress and more social support relative to their own personal average. Finally, different designs should be used in future work to examine whether changes in personality and social context explain changes in realness, if changes in realness promote changes in context, or if there are other connections between these variables.

Conclusion

Realness was recently introduced to capture a core and specific feature of authenticity, namely, the tendency to be oneself in times when there is situational pressure not to. Previous cross-sectional research established that realness is generally adaptive and specifically associated with low neuroticism and high extraversion, openness to experience, and conscientiousness, but is unrelated to agreeableness. This study replicated those findings. It further showed associations with high social support and low stress in young adults. Results also confirm that realness is a highly stable construct, similar to other personality traits. The results of this study strengthen emerging evidence for the value of realness to help clarify some ambiguities in the authenticity literature and provide a firmer basis for future work on its developmental, personological, and contextual mechanisms.