SWB comprises, on the one hand, the cognitive component of satisfaction with life as a whole, and on the other hand the presence of positive emotions (e.g., joy, interest, love, or pride) and the absence of negative emotions (e.g., loneliness, anger, anxiety) as affective components (Seligman, 2002; Fredrickson, 2009). PWB consists of autonomy, environmental mastery, personal growth, personal relationships, purpose in life, and self-acceptance (Keyes et al., 2002). SWB and PWB are related but empirically distinct (Keyes et al., 2002; Ring et al., 2007).

Su et al. (2014) recently reviewed, compared and integrated existing prominent well-being theories. They identified six aspects of PWB alongside SWB. These are key indicators based on established well-being theories, namely self-determination theory by Ryan and Deci (2000) and Ryff's (1995) theory of psychological well-being, and Seligman's (2011) PERMA well-being model. Building on these theories Su et al. (2014) developed a well-being questionnaire (Comprehensive Inventory of Thriving) measuring SWB and the following six theoretical based aspects of PWB:

Relationships: One of the three psychological basic needs postulated in the self-determination theory (Ryan and Deci, 2000) is the need for relatedness (besides competence and autonomy). Ryff (1995), Seligman (2011), and Diener et al. (2009) demonstrated the importance of positive relationships for different well-being concepts. Relationships include mutual support, community, trust, respect, belonging and the opposite of loneliness (Su et al., 2014).

Peterson and Seligman (2004) postulated the so-called “good character” to be essential for human functioning. Moreover, character strengths and their expression differ between individuals and make people unique. Peterson and Seligman (2004) identified 24 individual character strengths (e.g., bravery, humor, curiosity) which were assigned to six historically based human virtues, namely courage, justice, humanity, temperance, transcendence, and wisdom. The classification of character strengths are presented in the Values in Action Inventory of Strengths (VIA-IS) and 10 criteria were set which define character strengths in general (Peterson and Seligman, 2004). According to these criteria, character strengths are universally valid, contribute to the “good life” in several ways and are fulfilling, morally valued in themselves, do not evoke envy, have a negative counterpart, and constitute individual differences that are trait-like, measurable, unique and distinct from the other character strengths, have paragons and prodigies, can be absent in some people, and can be fostered by institutions and rituals. Peterson and Seligman (2004) argued that every person has three to seven character strengths which are especially important for the individual. These strengths are called “signature strengths” and are defined as central strengths for the individual which fulfill several criteria such as “a sense of ownership and authenticity,” “a feeling of excitement while displaying it (…),” or “intrinsic motivation to use the strength” (Peterson and Seligman, 2004). Short descriptions of the 24 character strengths are provided in online Supplementary Material.

Peterson and Seligman assumed that character strengths underpin well-being and lead to more positive emotions, engagement, meaning, accomplishment, and better relationships (Peterson and Seligman, 2004; Seligman, 2011). The empirical research also found character strengths in general to contribute to subjective well-being, mental and physical health and satisfaction with life as well as work satisfaction (e.g., Peterson and Park, 2011; Allan and Duffy, 2014; Douglass and Duffy, 2014). A positive relation of the “good character” with life satisfaction was identified in several studies (e.g., Ruch et al., 2007; Martínez-Martí and Ruch, 2014) in German speaking countries. Moreover, the 24 character strengths were analyzed in more detail in relation to SWB. Many studies have identified hope, zest, gratitude, curiosity, and love as the five character strengths most strongly related with life satisfaction (e.g., Park et al., 2004; Buschor et al., 2013). Littman-Ovadia et al. (2017) called these five strengths the “happiness strengths.” Park et al. (2004) further identified the five character strengths least related to life satisfaction: modesty, creativity, appreciation of beauty and excellence, judgment, and love of learning. Some studies found transcendence strengths in general (appreciation of beauty and excellence, gratitude, hope, humor, and spirituality) to be the strongest predictors of life satisfaction (Shoshani and Slone, 2013; Weber et al., 2013) and positive affect (Weber et al., 2013). In addition to replicating the highest correlations between life satisfaction and the happiness strengths, Proyer et al. (2011) found hope and spirituality to be the best predictors of future life satisfaction. Peterson et al. (2007) examined relationships between character strengths and the three routes to happiness: meaning, engagement and pleasure. They found that religiousness, gratitude, hope, zest, and curiosity were the character strengths associated with meaning, the character strengths of zest, curiosity, hope, perseverance, and perspective were strongly correlated with engagement, and humor, zest, hope, social intelligence, and love with the SWB aspect of pleasure. Some longitudinal studies have also focused on the link between character strengths and SWB, and shown the potential of fostering well-being through character strengths interventions (Proyer et al., 2013).

Overall most studies have focused on SWB (especially on life satisfaction) as a general well-being indicator, and less on PWB aspects (such as meaning or engagement) when exploring their relationships with character strengths. The results indicate that specific character strengths are more important than others (e.g., hope, zest, gratitude, curiosity, and love) and different strengths are important for different aspects of SWB (e.g., Weber et al., 2013). Results are not definite and vary (e.g., due to sampling issues) with respect to the importance of character strengths.

The purpose of the study was a more fine-grained analysis exploring the associations of the specific 24 character strengths with the distinct aspects of SWB and PWB in a sample of medical students. Fine-grained analyses are indicated due to the following reasons:

Firstly, more detailed understanding of relational aspects of character strengths and distinct well-being aspects is relevant for tailoring character strengths interventions. According to specific well-being aspects an individual may want to improve, an intervention can focus at a particular character strength which is most related to that particular well-being aspect. In the long term this can become a basis for guidelines for practitioners as well as for individuals.

Secondly, if there are different relationships between the 24 character strengths and well-being aspects it is important for future studies to take up a detailed strategy when exploring relationships between character strengths and well-being.

Thirdly, a more fine-grained analysis increases the comparability of research results. Future studies will need to take specific wording into account for a more precise, clear and sustainable research practice in relation to character strengths and different aspects of SWB and PWB.

Finally, we are not aware of any study analyzing relational aspects of character strengths and well-being in the context of medical students. This analysis will give a first basis for future character strengths intervention studies for the well-being of medical students.

In this study, we identified four main research questions examining data of medical students in their first year of study (t1) and one year later (t2).

The Board for Ethical Questions in Science of the University of Innsbruck provided approval for the study. Medical students participated in an online survey twice, with measuring times a year apart (t1: January/February 2015 with N = 178 and t2: January/February 2016 with N = 121). We offered direct feedback on participants' personal five signature strengths, medical education credits, and a raffle of medical books relevant to their current semester. We collected the data at a medical university in Austria as a part of a larger research project. Complete longitudinal data was available for N = 117 [66% female, mean age 20.3 ± 2.0 years (t1) and 21.4 ± 2.1 (t2) respectively]. Most were Austrian nationals (56%), German (23%), or Italian (20%); 2% had other nationalities. 74/50% (t1/t2) were single. 24/22% (t1/t2) were living alone, 11/12% (t1/t2) with a partner, 48/54% (t1/t2) in a flat share and 17/12% (t1/t2) with their parents or family of origin. Participants spent an average of 37.5 ± 15.4 (t1) and 35.6 ± 13.8 (t2) h per week studying. 17/26% (t1/t2) reported secondary employment with an average of 8.8 ± 4.0 (t1) and 9 ± 4.8 (t2) h per week.

We analyzed data at the level of specific character strengths. First, we computed descriptive statistics and internal consistencies (Cronbach's Alphas) of the measures. Second, we compared means of measurement at time 1 and 2 (paired sample t-test and Cohen's d) and analyzed intercorrelations of the means at t1 and t2. Pearson's correlations coefficients were interpreted as follows: r of < 0.10 = no effect, r = 0.10–0.29 = weak effect, r = 0.30–0.49 = moderate effect, r ≥ 50 = high effect. Third, we calculated correlations with demographics (age, sex, family status), as proposed in several studies (e.g., Harzer and Ruch, 2012; Wagner and Ruch, 2015). All subsequent analyses were controlled for demographics. To test hypotheses 1, 2, and 3, we computed partial correlations between character strengths and well-being aspects. We conducted Steiger tests (one-tailed) for comparing dependent correlations (Steiger, 1980).

In order to analyze the specific relationships between the well-being aspects and each of the 24 character strengths the data was analyzed as follows: (1) partial correlations between all character strengths and the six PWB aspects were computed; (2) hierarchical multiple regressions were calculated, including only significantly correlated character strengths identified by the partial correlations (first step: demographics, method: enter; second step: significantly correlated character strengths, method: stepwise).

We had hypothesized that hope, zest, gratitude, curiosity and love at t1 would be correlated most strongly with SWB at t1 or t2 (hypothesis 1). Results mostly confirmed hypothesis 1 at t1 and t2 for hope (r = 0.67/0.41), zest (r = 0.51/0.42), gratitude (r = 0.45/0.27), curiosity (r = 0.39/0.34), and love (r = 0.41/0.21). The character strengths of humor at t1 (r = 0.30) and spirituality at t2 (r = 0.31) were also moderately correlated with SWB. According to the Steiger tests the correlation of SWB and humor (t1) was only significantly different compared to the two highest correlations of SWB with zest (p = 0.003) and hope (p < 0.001). The correlation of SWB and spirituality (t2) was not statistically different to all of the happiness strengths correlations. All other character strengths had no significant or only small (r < 0.30) correlations with SWB (Table 2).

We had hypothesized that the happiness strengths at t1 would be correlated most strongly with PWB at t1 or t2 (hypothesis 2). Results confirmed hypothesis 2. High to moderate correlations were found for hope (r = 0.59/0.48), zest (r = 0.56/0.53), gratitude (r = 0.48/0.41), curiosity (r = 0.52/0.48) and love (r = 0.38/0.35). Compared to the correlations of character strengths and SWB, a greater number of character strengths was also moderately correlated with PWB at t1 and t2 (Table 2). No other character strength was significantly higher correlated with PWB than love (p > 0.05).

We had hypothesized that the “good character” (as measured by the VIA-120 total score) at t1 would show a significantly higher correlation with PWB than with SWB at t1 and t2 (hypothesis 3). This was confirmed by significantly higher correlation coefficients (p < 0.001) for PWB (r = 0.63/0.53) than SWB (r = 0.41/0.20). In particular, the PWB subcategories of mastery, meaning and engagement (large effect sizes), followed by relationships and optimism (moderate effect sizes), were significantly correlated with strengths possession in general, whereas autonomy showed a weak but still a significant relationship. This PWB aspect was by far the one with the lowest correlation with character strengths in general (Table 2).

Higher scores on measures of the 24 character strengths (t1) were associated with higher scores concerning the six aspects of PWB at t1 and t2 when controlling for demographics (strengths are given in an ascending order and have moderate partial correlations at least; Table 2). We performed stepwise multiple regression analyses for all character strengths that were previously identified as significantly correlated at each respective time. The aim was to identify the character strengths that would explain the most unique variance (Table 3). The following focuses on results at t2; all results can be found in Table 3.

Relationships were moderately related to a broad range of strengths (step 1, Table 2). We found zest, teamwork and spirituality to be the most relevant strengths for this aspect of PWB (step 2, Table 3).

Engagement was also moderately related to a broad spectrum of strengths. We found high effects for zest and curiosity. Regression analyses identified zest as the most important character strength, followed by persistence and spirituality.

For meaning, we found several moderate effects. Hope, zest and self-regulation were shown to be the most important character strengths by stepwise regression.

Mastery had the highest amount of significantly correlated character strengths (19 and 18 strengths at t1 and t2, respectively). Furthermore, compared with the other PWB aspects, mastery was most strongly influenced by the wisdom and knowledge strengths (Table 2). We found the highest standardized regression coefficients for persistence, followed by curiosity and love.

For autonomy, hope was the only strength with a moderate effect.

For optimism, the character strengths of hope and spirituality were the most important strengths.

We found some differences between t1 and t2 (Table 3): Besides the most important strengths of t2 (zest, teamwork and spirituality), the PWB dimension of relationships was additionally associated with hope and authenticity. Meaning at t1 was also significantly related to hope, zest and as well as to persistence, but not to self-regulation. Optimism was not only related to hope, but also to zest at t1. Relationships, engagement and meaning were associated with spirituality only at t2, but not in the cross-sectional design.

The study offers some theoretical implications. The happiness strengths were highly correlated with SWB and a broad range of PWB aspects. In general, these strong relationships mirror the assumption that tonic strengths fit in more situations than phasic strengths: Strengths that are relevant in a broader range of situations are called tonic (e.g., gratitude or zest), while phasic strengths are relevant for specific situations (e.g., bravery or leadership) (Peterson and Seligman, 2004). The VIA-120 measures the frequency of behavior patterns and tonic strengths could be generally used more often in a broader range of situations and may therefore be more central. Another reason for the high correlation between happiness strengths and especially SWB is that character strengths have something in common with positive emotions (Fredrickson, 2009). Güsewell and Ruch (2012) analyzed the relationship of character strengths and dispositional positive emotions in a German-speaking sample. They found low to moderate correlations between character strengths and emotions in general, but moderate to high correlations of a few strengths–emotions pairs. The highest correlations for the five happiness strengths in their study were: curiosity and hope with contentment (0.51^**; 0.67^**), zest and gratitude with joy (0.63^**; 0.54^**) and love (as a strength) with love (as an emotion) (0.53^**). The overlap between these pairings and results of longitudinal studies (e.g., Proyer et al., 2011) show the potential of positively influencing (subjective) well-being through character strengths interventions.

Our results also have several practical implications: The idea of improving the well-being of medical students by using character strengths interventions, in particular by fostering the character strengths of hope, zest, gratitude, curiosity and love based on our results warrant future research. Furthermore, if a specific well-being aspect should be targeted it may be helpful to choose a character strengths intervention that is aimed at a strength which is especially related to this well-being aspect. For example, Gander et al. (2016) adapted the well-known exercise “three good things”, where participants have to write down three positive things of the day and reflect what they have contributed to let them happen (Seligman et al., 2005), to different well-being aspects. The authors put the focus of the exercise in relation to one of the well-being aspects pleasure, engagement, meaning, positive relationships or accomplishment. To find a suitable and potentially effective intervention for the respective well-being aspect which should be addressed, knowing about the basic relationships between well-being aspects and character strengths is an important resource. This study is the first to report results of relationships between specific characters strengths and a broad range of well-being aspects in a sample of medical students of an Austrian university. The study shows that the PWB aspects mastery and relationships both were relevantly linked to a great amount of character strengths, indicating that these aspects of well-being may broadly benefit from different character strengths interventions. Across all results hope, zest and spirituality emerged as important key variables for a broad range of well-being aspects at t2. This provides sufficient rational to support future research in this area focusing on developing interventions that target these strengths in particular. For example, experimental research fostering the application of hope as a character strength could demonstrate whether well-being can be influenced directly or not. An example of targeting hope as a character strength is the exercise “best possible self,” where people write for 20 min about a positive possible self in the future, imagining the best that could happen in this period of time (King, 2001; Layous et al., 2013). Additionally, assuming a direct effect of curiosity and gratitude on hope and zest, it is plausible that fostering curiosity or gratitude may be an indirect way to foster hope and zest as well, because of the high intercorrelations between curiosity and both zest and hope, and gratitude and zest/hope. Causal links may be examined in future research.

To summarize, our study offers a fine-grained analysis of the relation between specific character strengths and a broad range of well-being aspects. Our analyses have shown that specific character strengths are linked in distinct ways to SWB and different aspects of PWB in medical students. The six subscales of PWB varied substantially in their associations with the 24 character strengths. These results underline the importance of more precise differentiation in research on character strengths and well-being. There is a need to investigate potential underlying causal relationships, which may help plan and design interventions to address the important research question of how to increase different well-being aspects by promoting specific character strengths. Future analyses may also inform practice with a view to targeting specific well-being aspects through interventions that focus on one or more corresponding character strengths. For example, in order to enhance the PWB aspects of autonomy (or meaning), a character strengths intervention targeting hope (such as the exercise “best possible self”) may be particularly beneficial. This further may allow for a more focused application of positive psychological strengths interventions in medical education.

The study was faced with several limitations:

Firstly, the sample is limited in its generalizability by the fact that all medical students came from one medical university in Austria, and by the relatively small sample size due to the study design (two times of measurement) and the length of the applied questionnaires. Moreover, all data was self-reported, which may have led to inflated correlations (common method bias). However, according to Podsakoff et al. (2003), separating measurements over time is one way of attenuating the risk of common method bias. In order to attain more generalizable results, future research should replicate these findings at different stages over the medical career (including more measurement points or longer intervals in between), in different countries (to take potential cultural differences into account; Joshanloo, 2014) and additionally, in the general population or in other specific samples over time. The impact of culture on specific links between character strengths and well-being has yet not been sufficiently empirically addressed in the scientific literature. Potential differences regarding the well-being of medical students living in western vs. non-western countries (e.g., with respect to the definition of the self, the “standing” of hedonism vs. euidaimonism, accepting vs. avoiding negative emotions, control/mastery vs. interpersonal harmony, life satisfaction vs. contentment, relevance of spirituality and religion; Joshanloo, 2014) and developmental aspects (students vs. physicians) should be addressed in future cross-cultural studies.

Nevertheless, in spite of the specificity of the sample, our study replicated the character strengths previously shown to be the most relevant to well-being (hope, zest, gratitude, curiosity, and love) in general populations (Park et al., 2004; Peterson et al., 2005, 2007). This shows that results from samples of the general population also have relevance for medical students in particular. Knowing which character strengths serve as protective factors and positive predictors of SWB and PWB may be helpful in shaping medical education with a view to fostering the well-being of students and mitigating risks such as burnout.

Secondly, the general question of causal vs. indicator variables is a common issue in well-being research (Fayers and Hand, 2002). The measures of well-being and character strengths can include variables that are indicators of the magnitude of the underlying construct, but they can also include variables that are part of the definition of the construct. That people with higher well-being also show higher values in specific character strengths (or higher means in general) does not automatically imply that these strengths lead to well-being in a causal sense. Perhaps well-being is not the consequence of character strengths, but reflects aspects of character strengths (or vice versa).

Thirdly, another limitation lies in the partly relatively low reliability coefficients of strengths scores. In contrast to previous validations of the VIA-IS and the VIA-120, some of the Cronbach's alphas in our study were lower than expected: Love of learning (t1/t2), authenticity (t1), love (t1), social intelligence (t1/t2), leadership (t1), teamwork (t2), forgiveness (t1/t2), prudence (t2), self-regulation (t1/t2) and appreciation of beauty (t2) had alphas ranging between 0.6 and 0.7, which indicates a questionable reliability in slight contrast to other studies (e.g., Ruch et al., 2010; Littman-Ovadia, 2015). In particular, modesty had an alpha of 0.57 at t1, which casts doubt on the internal consistency of that subscale. Means of modesty were relatively low and highly correlated between t1 and t2. Modesty showed no relevant relationships with any of the well-being aspects. Due to the low reliability, however, these particular results should be interpreted with caution.

The exclusion of the traditional clustering of character strengths into six “virtues” in our statistical analyses is an impulse for further thought rather than a limitation. Several studies have conducted factor analyses before but different factor structures were found: Three factors (McGrath, 2015), five (e.g., Peterson et al., 2008; Ruch et al., 2010; Harzer and Ruch, 2014) or six factors (Ruch and Proyer, 2015) were identified through statistical analyses. In addition to the lack of empirical evidence for six virtues based on the VIA-IS or VIA-120, all current interventions address character strengths, not virtues as an empirical construct. An analysis at the level of strengths allows for an interpretation of the effect of each character strength on well-being. Therefore we decided to use the theoretical construct of virtues only as conceptual clusters.

In this study we focused on distinct relationships between the different character strengths and several aspects of well-being. Another focus would have been an analysis based on participants' signature strengths, which have been shown to be especially relevant for the individual (e.g., for work satisfaction; Harzer and Ruch, 2012). Future research should address the level of character strengths (e.g., highest strength scores vs. lowest strengths scores) in a longitudinal study design in relation to well-being. Despite this issue, the fact that situational factors may allow or prohibit the application of strengths remains completely unaddressed in this analysis, and warrants further exploration.

Based on the results of this study, it is worth investigating the potential of character strengths interventions in medical education in order to understand potential underlying causal relationships. In practice, a character strengths approach may proactively improve students' well-being. Future research might address whether improvements in medical curricula can be achieved by (a) identifying character strengths that are generally helpful to medical students for specific outcomes (e.g., accomplishment, academic success, work engagement, or burnout), (b) investigating the effects of becoming aware of one's (signature) strengths, (c) assessing the impact of fostering these strengths through shaping studying conditions accordingly, and (d) evaluating the teaching and promotion of individual character strengths (e.g., through activities that build character strengths) in the educational context.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.