Introduction

The sense of wellbeing in one’s life has many positive effects on personal and societal levels. Therefore, subjective wellbeing (SWB) has been a highly investigated research area over many years. SWB is defined as “people’s overall evaluations of their lives and emotional experiences” (Diener et al., 2017, p. 87). It describes the quality of life from the perspective of a subject and takes into account various aspects of wellbeing, such as overall life satisfaction, positive affect, negative affect, and happiness (Diener et al., 1999). Research on predictors of SWB has demonstrated that its components are predicted by a number of sociodemographic factors (such as age, sex, marital status, or employment) and contextual factors (e.g., life events; Hentschel et al., 2017), and, above all, by personality traits (Diener et al., 2017). Findings consistently showed significant relationships between SWB and personality. These studies included both single features and general personality models (the big three or big five; DeNeve & Cooper, 1998). The latter may be particularly useful for SWB researchers, as they establish the basis for assessing the importance of other predictors of wellbeing (see Ozer & Benet-Martinez, 2006).

However, in recent years, a growing number of studies confirm the utility of the HEXACO model (Ashton & Lee, 2007) competing with the B5/FFM model (John & Srivastava, 1999). The HEXACO proposes an additional personality factor (honesty-humility) which is responsible for moral behavior. According to the research results, the HEXACO model outperformed the B5 model in predicting essential criterion variables, such as manipulativeness, delinquency, or materialism (see Ashton et al., 2014). Despite growing popularity of a new model, attempts to implement it in SWB research are sparse. The vast majority of studies on wellbeing from the last years were conducted with the use of the B5 model (DeNeve & Cooper, 1998; Steel et al., 2008), and this may lead to a situation of uncertainty about the predictive validity of the HEXACO in SWB research. This study aims at contributing to the clarification of this issue.

The purpose of the current study is to compare the predictive validity of the competing general personality models (i.e., the B5 and the HEXACO) in relation to the emotional dimensions of SWB, treated both as characteristics that changes over time (i.e., a state) and as a stable feature (a trait). Despite the fact that SWB is often considered a relatively stable trait, a considerable number of researchers find it important to measure fluctuations of SWB dimensions in everyday life (Diener, 1996; Reis et al., 2000). Because SWB treated as a trait or as a state should be considered conceptually and mathematically independent (Nezlek, 2001), it would be beneficial to compare the predictive validity of both personality models on both levels of SWB measurement (i.e., interpersonal and intrapersonal). To our knowledge, only Aghababaei and Arji (2014) and Romero et al. (2015) have directly compared the predictive validity of the B5 and the HEXACO models in predicting SWB (as a trait) using cross-sectional design. They did not find evidence that the HEXACO outperformed the B5 in predicting stable happiness and stable life satisfaction.

Most of the research concerning SWB had cross-sectional nature, which stems from considering wellbeing a stable, trait-like individual difference. However, a different research strategy which postulates the multiple measurements of SWB in a specified period of time is also possible. This approach enables identification of associations of situational variables and/or internal states of an individual with fluctuations of wellbeing. According to Scollon et al., (2003, p. 27) “(…) experience sampling is most useful in assessing the affective components of SWB, particularly because these components are vulnerable to distortions in memory.” Measuring SWB as a state can be of great significance when the goal of the research is the investigation of associations between emotional wellbeing and daily activities, daily life experiences, and specific daily events, or between therapy results and changes in SWB.

Current Study

Searching for the universal taxonomy of personality traits, which began in the second half of the twentieth century, has led to the conclusion that five superordinate constructs are the most accurate in describing personality structure (see Digman, 1990). At present, the B5 has become the most widely used model of personality and researchers can choose between different well-established and widely used instruments which are intended to measure its domains and facets (e.g., Costa & McCrae, 1992; Gosling et al., 2003). The B5 model consists of the following bipolar factors: neuroticism, extraversion, openness to experience, agreeableness, and conscientiousness. According to John and Srivastava (1999, p. 121), neuroticism includes “negative emotionality such as feeling anxious, nervous, sad, and tense,” extraversion involves “an energetic approach to the social and material word,” openness illustrates “the breadth, depth, originality, and complexity of an individual’s mental and experiential life,” agreeableness describes “a prosocial and communal orientation towards others,” and conscientiousness concerns “socially prescribed impulse control that facilitates task- and goal-directed behavior.”

The six-factor model of personality (HEXACO) was introduced by Ashton and Lee (2001) at the beginning of the twenty-first century and since then has received significant empirical validation. The HEXACO broadens personality structure of the B5 with a sixth factor known as honesty-humility that is related to moral aspects of human behavior. Also, two other factors, known from the B5 model, were modified. The HEXACO emotionality contains some elements of neuroticism, but it excludes anger and includes sentimentality (an element of the B5 agreeableness). In turn, the HEXACO agreeableness incorporates components of anger, hostility, and impatience (Lee & Ashton, 2004). The correlations between the corresponding traits of both models range from moderate (r > 0.5) for neuroticism/emotionality and agreeableness to high (r ≥ 0.7) for extraversion, conscientiousness, and openness (Ashton et al., 2014).

Many studies have compared the predictive validity of both models. It has been shown that compared to the B5 traits, the HEXACO traits demonstrate better predictive validity for outcomes related to honesty-humility factor, such as workplace delinquency (Lee et al., 2005), the “dark triad” of personality traits (Machiavellianism, narcissism, and psychopathy; Lee & Ashton, 2005), unethical decision making (Ashton & Lee, 2008), materialism (Górnik-Durose & Pilch, 2016), and also for such constructs as student life outcomes (Thalmayer et al., 2011) and many others (for a review see Ashton et al., 2014).

Empirical studies focused on associations between personality and wellbeing have been conducted since the 1960s (e.g., Wilson, 1967). In their well-known study, Costa and McCrae (1980) proved the negative correlation of neuroticism and the positive correlation of extraversion with emotional components of SWB. With the popularization of the five-factor taxonomy of personality traits, two other predictors of SWB (agreeableness and conscientiousness) have been identified. They both showed positive relationships with SWB (Ozer & Benet-Martinez, 2006; Steel et al., 2008).

So far, only a few studies on SWB have used the HEXACO instead of the B5. Pollock et al. (2016) obtained significant correlations between the HEXACO extraversion, agreeableness, and conscientiousness factors and the three SWB components (positive for life satisfaction and positive affect, and negative for negative affect), but honesty-humility and openness were unrelated to SWB. In turn, MacInnis et al. (2013) reported a positive relationship between the HEXACO openness and positive affect and a negative relationship between honesty-humility and negative affect. Aghababaei and colleagues investigated the associations of the HEXACO and other variables with happiness and life satisfaction on several samples with different cultural backgrounds (e.g., Aghababaei, 2014; Aghababaei & Błachnio, 2015). In all these samples HEXACO extraversion was strongly correlated with both happiness and life satisfaction, but the correlations of the other HEXACO traits with the SWB indicators were weak and inconclusive.

The above literature review shows that the associations of general personality traits with SWB seem well established. Nevertheless, personality still remains important in SWB research. Currently, a number of researchers aim at identifying the potential moderators of relationships between personality traits and SWB, such as cognitive mechanisms (Tkach & Lyubomirsky, 2006), self-efficacy (Strobel et al., 2011), or orientations to happiness (Pollock et al., 2016). Additionally, controlling for differences in personality enables investigating the incremental validity of different constructs, such as alexithymia (Páez et al., 2013) or emotional intelligence (Koydemir & Schütz, 2012) above and beyond general personality traits in predicting SWB. At present, researchers can choose between the two well-established personality models. In our study we try to assess which model constitutes a better taxonomic framework for SWB research. Examining the unique contributions of each domain of the B5 and the HEXACO and also an overall contribution of the two personality models to the prediction of SWB may help researchers to choose a better option for their study.

In the current study we investigated an affective component of SWB, to which we will refer as emotional wellbeing (EWB). Past research showed that personality (especially extraversion and neuroticism) was more strongly related to the affective than to the cognitive dimension of SWB (Schimmack et al., 2002a, 2002b). The association between the cognitive component of SWB and personality was often mediated by the affective component (Schimmack, 2008). Moreover, investigating momentary life satisfaction with the experience sampling method (i.e., several times a day) may be relatively more problematic. The measurement of life satisfaction requires a judgment of one’s life and it does not necessarily change considerably within a short time (Schimmack et al., 2002a, 2002b).

The affective element of SWB in the present study was considered both a state and a trait. Firstly, EWB was measured as a state in everyday life situations with the use of experience sampling methodology, which allowed a more precise assessment of its dimensions and helped to reduce recall bias. Secondly, indicators of trait-like EWB were calculated by averaging scores for each person for every momentary variable. This approach is in line with the recommendations of Diener (2012, p. 595) and Fleeson and Gallagher (2009).

In summary, this study aimed to examine the predictive validity of the B5/FFM and the HEXACO domains on an emotional aspect of SWB. As we presented in previous paragraphs, only relationships between B5 and EWB are well established. Using both personality models allows to compare these models in terms of their predictive validity, i.e., their ability to predict momentary and stable EWB. On the basis of previous research results, we expect that extraversion (positively) and neuroticism/emotionality (negatively) would be significant predictors of EWB. We also predict significant albeit relatively weaker associations of EWB with agreeableness, conscientiousness, and openness.

To date, relationships between honesty-humility and EWB have been inconclusive. For example, honesty correlated negatively with negative affect; however, it was not associated with happiness (MacInnis et al., 2013; Romero et al., 2015). On the other hand, in their meta-analysis Lyubomirsky et al. (2005) quoted many studies which showed the relationships between SWB and such constructs as altruism, moral goodness, helpfulness, unselfishness, cooperativeness, or organizational citizenship behavior, which are in turn related to the honesty-humility factor. Based on this reasoning, we predict that honesty-humility would be a positive predictor of EWB.

In the present study brief measures of personality were used. This kind of personality measures seems to be useful for research purposes when the intention is to control for general personality traits, and a more detailed analysis is not necessary. Despite the fact that brief measures sometimes may lack internal consistency, they may still demonstrate good predictive validity, good test-criterion correlations, and their results may be stable over time (Gosling et al., 2003; Ziegler et al., 2014). Brief and easily administered personality measures may be a good choice for assessing personality factors when other variables are the main interest, when repeated measurements are made or in cross-cultural studies (Sorokowska et al., 2014).

Method

Participants and Procedure

A sample from the Polish general population was recruited via the snow-ball sampling technique. It consisted of 669 individuals (155 males and 514 females; age M = 24, SD = 7.3) who participated as volunteers. The participants were university students (49%), had jobs in a wide range of areas (49%), or were neither studying nor working at the time of data collection (2%). The sample size for regression analyses was not determined a priori. The results of a post-study power analysis showed the power of 0.99 for this study.

During the initial part of the study, the subjects completed online self-report questionnaires (the TIPI and the HEXACO-60) and provided socio-demographic information. Two weeks later they started the participation in the experience sampling procedure. The procedure enables repeated sampling of participants’ current feelings and/or behaviors in real time and natural environment (see Shiffman et al., 2008; Conner et al., 2009). The data in the current study concerned momentary affective states experienced in daily life and were collected using participants’ smartphones. A unique code was assigned to each participant to ensure anonymity. Personal data were not collected, except for a phone number. Six times a day for eight consecutive days the participants received notifications (via text messages sent to the mobile phones) with a link to an online survey. They were asked to log in and answer the survey questions on their mobile device as quickly as possible. However, the data were analyzed using multilevel modeling for longitudinal data (level 1, momentary affect; level 2, personality traits), which indicates that missing observations are allowed (because this analytic approach is robust to missing observations; see Nezlek, 2001, 2007). During the experience sampling period, each participant received 48 notifications (8 days × 6 notifications = 48), and they responded, on average, to 30 notifications (Me = 33, SD = 9.2, range 6–42), which gave a sum of 20,267 survey responses for analyses.

Measures

Emotional Wellbeing (Momentary-Level Measures)

Momentary mood was assessed with the use of the single-item question “How do you feel right now?” with the answers on a bipolar scale from 0 (very bad mood) to 10 (very good mood). A similar measure of emotional wellbeing was previously used in many studies (e.g., Knabe et al., 2010; Kross et al., 2013). Momentary mood served as the first indicator of momentary EWB.

Momentary positive affect (PA) was measured with the single item “Do you feel positive emotions (e.g., happy, joyful, satisfied, enthusiastic, etc.).” Momentary negative affect (NA) was measured with the single item “Do you feel negative emotions (e.g., sorrowful, depressed, dissatisfied, anxious, etc.).” The subjects answered these questions using a 5-point scale (1, not at all; 5, very much). These simple measures do not include ratings of any specific positive or negative emotions, but instead, they allow participants to assess the intensity of these emotions (assigned to the two categories—positive and negative) that they currently feel. Thus, the aggregation of separate ratings of very different emotional states (which is considered problematic) was not needed (see Knabe et al., 2010). The indicators of PA and NA were used to calculate momentary affect balance (a-balance), which was obtained by subtracting NA score from PA score (see Bradburn, 1969). Momentary a-balance served as the second indicator of momentary EWB. Momentary mood and a-balance were used as outcome variables in multilevel analyses.

Emotional Wellbeing (Person-Level Measures)

For each participant we calculated the mean scores of momentary mood, PA, NA, and a-balance. The scores aggregated within persons served as indicators of stable EWB. Stable mood and stable a-balance were used as outcome variables in regression analyses (Csikszentmihalyi & Hunter, 2003, p. 187).

Personality Traits

The Polish version of the Ten Item Personality Inventory (TIPI; Gosling et al., 2003; Sorokowska et al., 2014) was used to assess personality in the B5 model. It consists of 10 items with responses on a 7-point scale (1, strongly disagree; 7, strongly agree). The TIPI converged with widely used B5 measures, correlated as expected with many constructs and had good test–retest reliability (Gosling et al., 2003; Sorokowska et al., 2014). We calculated the reliability of the subscales (Cronbach’s alpha; emotional stability/neuroticism α = 0.6; extraversion α = 0.57; openness α = 0.47; agreeableness α = 0.37; conscientiousness α = 0.73) which (as in the original version) were the lowest for agreeableness and openness. However, it is important to note that the TIPI was not formed with the intention to maximize internal consistency, but the priority was to reflect different facets of the measured constructs (Gosling et al., 2003). Thus, test–retest reliability (which is good for the Polish version of the TIPI) is a more appropriate method of testing for reliability. The TIPI is probably the most widely used brief measure of the B5 and is available in more than 20 language versions (gosling.psy.utexas.edu).

The Polish version of the HEXACO-60 (an adaptation by Piotr Szarota; hexaco.org) was used to assess the six dimensions of personality (Ashton & Lee, 2009). It consists of 60 items with responses on a 5-point scale (1, strongly disagree; 5, strongly agree). The HEXACO-60 is a shorter version of the HEXACO-PI-R and can be used when the analysis at the facet level is not a priority. The reliability of the subscales was good in our sample (honesty-humility α = 0.78; emotionality α = 0.8; extraversion α = 0.73; agreeableness α = 0.76; conscientiousness α = 0.81; α = openness 0.73).The HEXACO-60 is a popular measure of a six-factor model of personality and is available in many language versions (hexaco.org).

Statistical Analyses

To determine the unique contribution of the B5 and the HEXACO traits in predicting stable EWB, a series of multiple regression analyses was performed with stable mood and stable a-balance as outcome variables. First, all the B5 traits were regressed on the indicators of stable EWB. After that, all the HEXACO traits were regressed on the indicators of stable EWB. The coefficients of determination (R2) of the models were then compared. To verify whether the B5 traits have incremental validity over and above the HEXACO (and vice versa) in predicting stable EWB, a series of hierarchical regression analyses was performed. First, the B5 traits were entered in step one and the HEXACO traits were added in step two. After that, the HEXACO traits were entered first into the regression and the B5 traits were added in step two. Then the squared semi-partial correlation coefficients (Δ R2) of the models were compared.

At the beginning of the multilevel analysis, the unconditional random intercept models for momentary mood and momentary a-balance were estimated in order to check whether there was a substantial within-person variability in momentary EWB. The models have no predictors at any level and they produce estimates for variance components at both levels of the analysis (sigma-square [σ2] for level 1 and tau [τ] for level 2). These estimates were used to compute how much of the total variance of momentary EWB can be accounted for by observations and participants and check whether multilevel analysis is appropriate for the data. In the next step a series of multilevel analyses was conducted to estimate the contribution of the B5 and the HEXACO in predicting momentary EWB. Time of measurement was included as level-1 control variable (within-person: β0j + β1j*(TIMEij) + rij) because this factor can influence momentary EWB (Egloff et al., 1995). At level 2 of the analysis associations between the B5/HEXACO traits and individual differences in ratings of momentary EWB were estimated. Each personality model was analyzed separately (between-person (β0j = γ00 + γ01(trait 1) + γ02(trait 2) + γ03(trait 3) + …. + u0j; β1j = γ10 + u1i)).Footnote 1 The strength of relationships between the B5/HEXACO models and momentary EWB (pseudo R-square) was assessed by comparing u0j (the residual error of the intercept) from unconditional models to u0j from conditional models that included the predictors (see Nezlek, 2007, p. 796).

To determine whether the B5 has incremental validity over and above the HEXACO (and vice versa) in predicting momentary EWB, hierarchical multilevel analyses were performed with momentary mood and momentary a-balance as criterion variables and the B5 and the HEXACO traits as predictors. First, the B5 traits were entered in step one into the regression and the HEXACO traits were added in step two. After that, the HEXACO traits were entered first and the B5 traits were added in step two. The incremental validity of these models was assessed using two different methods: by comparing pseudo R-square coefficients (Nezlek, 2007) and deviance statistics (Raudenbush et al., 2011). When two multilevel models are nested within one another, it is possible to compare their deviances using the chi-square difference test. Deviance may be treated as a measure of a model misfit and lower deviance indicates that a model fits the data better (Raudenbush et al., 2011). First, pseudo-R-square coefficients for the full models (with momentary mood and momentary a-balance as outcome variables and the B5 and the HEXACO as predictors) were compared with pseudo-R-square coefficients for the corresponding models with the B5 traits or the HEXACO traits and delta-R-squares were calculated.Footnote 2 Second, the deviance statistics of the corresponding models were compared.

Results

Big Five and HEXACO Traits as Predictors of Stable EWB

The results of regression analyses with personality traits as predictors and stable EWB as an outcome are displayed in Table 1 (models A). When all the B5 traits were regressed on the indicators of stable EWB, the models explained 22% of the variance in stable a-balance (F(5, 663) = 37.3, p < 0.001) and 15% of the variance in stable mood (F(5, 663) = 24.4, p = 0.001). Neuroticism turned out to be the strongest (negative) predictor of stable a-balance. The remaining B5 traits showed positive and significant, albeit weaker, relationships with stable a-balance. The similar pattern of relationships emerged when stable mood was a criterion variable: all the B5 traits were significant predictors of mood, and the negative association between neuroticism and mood was relatively the strongest one.

Table 1 Stable EWB (mood and affect balance) predicted by the B5 and the HEXACO
Full size table

When all the HEXACO traits were regressed on the indicators of stable EWB, the models explained 17% of the variance in stable a-balance (F(6, 662) = 14.0, p < 0.001) and 12% of the variance in stable mood (F(6, 662) = 16.7, p < 0.001). Extraversion was identified as the strongest predictor of stable a-balance and stable mood. Stable a-balance was also associated with emotionality (negatively), agreeableness, and conscientiousness, whereas stable mood was related to emotionality (negatively) and conscientiousness. Honesty-humility and openness turned out not to be associated with stable EWB.

Incremental Validity of the Big Five and HEXACO Models in Predicting Stable EWB

The results of hierarchical regression analyses are displayed in Table 1 (models B). When the B5 traits were entered in step one and the HEXACO traits were added in step two (a-balance: F(11, 657) = 14.37, p < 0.001; mood: F(11, 657) = 13.71, p < 0.001), the HEXACO added some incremental validity over and above the B5 in predicting stable a-balance (ΔR2 = 0.03, p < 0.001) and stable mood (ΔR2 = 0.03, p < 0.001). When the HEXACO traits were entered first into the regression and the B5 traits were added in step two, the B5 also added some incremental validity over and above the HEXACO in predicting stable a-balance (ΔR2 = 0.08, p < 0.001) and stable mood (ΔR2 = 0.05, p < 0.001). When the B5 and the HEXACO traits were entered together in step two, the B5 neuroticism (negatively), agreeableness, and openness were associated with both EWB indicators. Among the HEXACO traits, only extraversion was still a positive predictor of EWB. Agreeableness, which previously had no relationship with stable mood, became its negative predictor when the B5 was controlled. In summary, both the HEXACO and the B5 had some incremental validity over and above the other one.

The Big Five and HEXACO Traits as Predictors of Momentary EWB

The unconditional random intercept models showed that for momentary mood (σ2 = 4.52, τ = 1.2) and momentary a-balance (σ2 = 3.64, τ = 0.78) a substantial part of their variance was at the within-person level (79% and 82%, respectively). Thus, multilevel analysis is appropriate to analyze these variables. The results of the analysis are displayed in Table 2 (models A). According to the results, all the B5 traits were significant predictors of momentary mood (pseudo R2 = 0.18) and momentary a-balance (pseudo R2 = 0.27) and neuroticism emerged as their strongest predictor (βs = 0.3 and 0.32, respectively). The remaining B5 traits revealed weaker associations with momentary EWB indicators (βs = 0.1–0.16). In turn, three out of six HEXACO traits (emotionality, extraversion, and conscientiousness) turned out to be significant predictors of both momentary mood (pseudo R2 = 0.15) and momentary a-balance (βs = 0.1–0.16; pseudo R2 = 0.22), and additionally agreeableness predicted momentary a-balance. The HEXACO extraversion presented the strongest associations with momentary EWB (βs = 0.3 and 0.32).

Table 2 Multilevel models predicting momentary EWB (affect balance and mood) from the B5 traits and the HEXACO traits
Full size table

Incremental Validity of the B5 and HEXACO Models in Predicting Momentary EWB

The results of hierarchical multilevel analyses with momentary mood and momentary a-balance as outcomes and all the B5 and the HEXACO traits as predictors are displayed in Table 2 (models B). Some of the relationships lost significance due to the covariance between predictors (which measure similar constructs). When momentary mood was an outcome variable, the B5 explained an additional 5% of its variance over and above the HEXACO (an increase in pseudo R2 = 0.05), and the HEXACO explained an additional 2% of its variance over and above the B5 (ΔR2 = 0.02). In turn, when momentary a-balance was an outcome, the B5 explained an additional 8% of its variance over and above the HEXACO (ΔR2 = 0.08), and the HEXACO explained an additional 3% of its variance over and above the B5 (ΔR2 = 0.03). Thus, both personality models explained between 2 and 8% of additional variance in EWB; however, the increase in R-square was higher for the B5.

Deviances for models B were lower than these for models A. When model B for mood was compared with the HEXACO model, the reduction in deviance was significant (χ2 = 42.38, df = 5, p < 0.001), and the comparison for the B5 also showed significant differences in deviance (χ2 = 25.09, df = 6, p < 0.001). For a-balance the results were similar: the reduction in deviance was significant when model B was compared with both the HEXACO model (χ2 = 65,56, df = 5, p < 0.001) and the B5 model (χ2 = 30.23, df = 6, p < 0.001). Thus, model B fitted the data better than both the B5 and the HEXACO models separately, which means that all the personality models added incrementally to the prediction of EWB when the other model was controlled. However, both the values of deviance coefficients from model A and the above analysis indicate that B5 seems to have a relatively larger contribution to explaining the variance in momentary EWB.

Discussion

All the B5 traits were uniquely linked to both momentary and stable EWB and neuroticism was its strongest (negative) predictor. In turn, the pattern of relationships between the HEXACO traits and EWB was less consistent. Three of the HEXACO factors (emotionality, extraversion, and conscientiousness) demonstrated the unique associations with momentary and stable mood. The same factors and agreeableness were uniquely associated with momentary and stable a-balance. Among the HEXACO traits, extraversion was the strongest predictor of EWB.

The general pattern of relationships between personality and EWB was rather consistent for the two EWB indicators; however, both personality models predicted a larger portion of the variance in a-balance than in mood. The B5 outperformed the HEXACO predicting 27% of the variance in momentary a-balance (the HEXACO, 22%) and 18% of the variance in momentary mood (the HEXACO, 15%). The B5 also accounted for greater variance in stable a-balance (22%) and stable mood (15%) than the HEXACO did (17% and 12%, respectively). The B5 showed incremental validity (5–7%) over and above the HEXACO in predicting momentary and stable EWB. The HEXACO also added incrementally to the prediction of EWB over and above the B5; however, an increase in explained variance was lower in this case (2–3%).

The current study results are generally in line with previous findings. In many studies all the B5 factors were related to SWB (DeNeve & Cooper, 1998; Steel et al., 2008). However, in our study only neuroticism (one out of two B5 domains associated with temperamental dispositions to experience emotions; DeNeve & Cooper, 1998) showed, as expected, stronger relationships with EWB than the remaining three B5 traits. The relationships between extraversion and EWB were in our sample somewhat lower than was expected. The B5 openness, which usually has relatively the weakest associations with overall SWB, was consequently associated with EWB in our sample. However, this result is not surprising as openness tends to be connected with emotional aspects of wellbeing (Steel et al., 2008).

The obtained associations between the HEXACO and EWB are also congruent with previous research. In past studies the HEXACO extraversion was consequently found to be the strongest predictor of stable SWB (e.g., MacInnis et al., 2013; Pollock et al., 2016), whereas the relationships between stable SWB and emotionality, agreeableness, conscientiousness, and openness were weaker and inconsistent across studies. Our findings were not consistent with the expectations for honesty-humility, which in previous research had the weakest associations with SWB. Honesty-humility (i.e., being fair and genuine in dealing with others) characterizes the variation in personality which is not explained enough by five-factor conceptualizations (Ashton & Lee, 2007; Ashton et al., 2014). This factor is to a large extent responsible for the advantages of the HEXACO over the B5 in predicting numerous outcomes, such as interpersonal or moral behaviors (e.g., Ashton & Lee, 2008). However, our data show that the relationship between momentary EWB and honesty-humility seems to be rather minimal, which may be partially the reason why the HEXACO less well predicted SWB compared to the B5.

Generally, the pattern of our findings seems to indicate that the B5 predicted momentary and stable EWB better than the HEXACO did. Thus, our study replicated the results of Aghababaei and Arji (2014) and Romero et al. (2015) who did not find any advantage of the HEXACO over the B5 in predicting stable SWB in samples from Iran and Spain. At the same time, we extended their findings by exploring the relations between personality and momentary affect measured repeatedly in natural settings. Moreover, our results showed that some of the corresponding personality dimensions from the B5 and the HEXACO (which correlated moderately in the current study) exhibited different relationships with EWB indicators. It is visible in the case of openness—only the B5 openness predicted EWB, even though these dimensions of both models are characterized as very similar (Ashton et al., 2014). The HEXACO extraversion was consequently the strongest predictor of EWB while the B5 extraversion was not, although the differences between these two measures of extraversion are usually not reported in the literature (Ashton et al., 2014). In turn, emotionality factor from the HEXACO cannot be treated as an equivalent of neuroticism because of a few dissimilarities in their components (see Sect. 1); therefore, different strengths of relationships between these factors and EWB obtained in the current study may be attributed to the conceptual differences between the two personality models.

The personality spectrum represented by the B5 dimensions seems better fit the area of research on SWB. However, the advantages of the B5 over the HEXACO in predicting EWB might be particularly visible when one wants to use a wide spectrum of “normal” personality traits whose relationships with SWB are well established. On the other hand, when extraversion is of interest, the HEXACO model might be more useful because of a high predictive power of extraversion measured by the HEXACO questionnaires.

Conclusions and Limitations

The B5 and the HEXACO personality models were compared in order to determine the value of each model in predicting momentary and stable EWB. The impact of the B5 model of personality in predicting EWB seems to be greater. However, both personality models added the unique variance to the prediction of EWB. Thus, each model was able to explain a portion of EWB variance which was not accounted for the other one. Researchers who consider including general personality traits in their research could also take into account the fact that the spectrum of personality traits represented by the B5 might be more useful because all the B5 factors emerged as significant predictors of momentary and stable EWB. When a short measure of general personality traits is needed and personality is not a primary subject of interest, the TIPI may be recommended as a good choice in EWB research.

The current study has some limitations connected with a sample used that was large enough but was composed primarily of relatively young and well-adjusted volunteers, which limits the generalizability of the results. The study depends on one-item self-report measures of mood, NA, and PA; thus, future research should use longer and more sophisticated measures of EWB to confirm our results. It could be also useful to compare directly other measures of the HEXACO (e.g., the IPIP HEXACO) and the B5 (e.g., the NEO-FFI) in relation to EWB, which is important because of the possible differences in results depending on the type of questionnaire used.

We suppose that the predictive advantages of the B5 over the HEXACO in the current study cannot be attributable to the differences in the lengths of the questionnaires used. Other studies have shown that the predictive advantage of one personality model as compared to another was not a result of any differences in the length of the scales (Ashton & Lee, 2007). On the other hand, we assessed the B5 traits using a well-validated but short measure, which earlier correlated with other variables less strongly than multi-item measures of the B5 (Gosling et al., 2003, p. 523). Thus, it is possible that a longer measure could provide even more meaningful arguments in favor of the B5 model (Credé et al., 2012).