Today, many children, adolescents, and adults are confronted with increasing expectations at home, school, and work and often feel overwhelmed and stressed. Thus, research has started to shift from being focused on risk and vulnerability to potentially protective factors. One of these protective variables is self-compassion. This construct originates from Buddhist philosophy and was defined by Kristin Neff (
2003) as an open and understanding attitude toward oneself when confronted with one’s own weaknesses, inadequacies, and suffering. Self-compassion consists of three components: self-kindness, common humanity, and mindfulness. Neff contrasted those main components with an antagonist, each receiving three bipolar components: 1) self-kindness versus self-judgment, 2) common humanity vs. isolation, and 3) mindfulness vs. overidentification. Self-kindness means with kindness and respect when confronted with personal failure instead of being harsh on oneself. Common humanity refers to being able to see failures as part of human nature and thus feeling a connection to other people rather than feeling alone and isolated. The third component, mindfulness, is a mental state of openness and acceptance when confronted with negative events and suffering instead of suppressing or overthinking (Gruber et al.,
2020). In 2016, Neff updated her definition of self-compassion and added emotional, cognitive, and attentional processes. According to her, people respond emotionally to failure or suffering (with self-kindness or self-judgment), cognitively understand their dilemma (feeling connected or isolated), and focus their attention on this condition (being mindful or overidentifying), thus always balancing between being compassionate and uncompassionate (Neff,
2016).
Self-compassion and Its Effects in Adults
Since the implementation of self-compassion in 2003, many studies have investigated potential positive effects on mental health and overall protective functions of self-compassion in adults. Barnard and Curry (
2011) found positive correlations with well-being. A meta-analysis by MacBeth and Gumley (
2012) examined healthy adults and reported a robust, significant negative association between self-compassion and overall psychopathology. More positive associations have been reported for quality of life (Van Dam et al.,
2011), interpersonal conflict resolution (Yarnell & Neff,
2013), body appreciation (Pullmer et al.,
2021), happiness (Neff et al.,
2007), overall psychological well-being (Krieger et al.,
2015; Yarnell & Neff,
2013; Zessin et al.,
2015), emotional intelligence (Heffernan et al.,
2010), positive affect (Krieger et al.,
2015; Neff et al.,
2007), self-improvement motivation (Breines & Chen,
2012), and wisdom (Neff et al.,
2007) in clinical and non-clinical samples. Negative associations for adults have been found with psychological distress (MacBeth & Gumley,
2012), eating pathology (Pullmer et al.,
2021), depression (Castilho et al.,
2015; Gilbert & Procter,
2006; Krieger et al.,
2016a; Körner et al.,
2015; MacBeth & Gumley,
2012; Raes,
2010,
2011), anxiety (Gilbert & Procter,
2006; MacBeth & Gumley,
2012; Raes,
2010), discrimination (Pullmer et al.,
2021), negative affect (Krieger et al.,
2015; Neff et al.,
2007), rumination (Raes,
2010), shame (Gilbert & Procter,
2006), and academic failure (Neff et al.,
2005).
In general, mindfulness-based compassion practices (MBCPs) are of high scientific interest, and many studies investigate their implementation in routine therapy (Blanck et al.,
2018; Kuyken et al.,
2016; Mander et al.,
2019). There are different approaches to improving MBCP, namely “Acceptance and Commitment Therapy” (ACT; Hayes et al.,
2012), “Compassion-Focused Therapy” (CFT; Gilbert,
2009), “Mindfulness Self-Compassion” (MSC; Neff & Germer,
2013), and “Making Friends With Yourself” (MFY; Bluth et al.,
2016). ACT mentions self-compassion as one of its main targets, while others invented special treatment forms to foster self-compassion in adults (CFT, MSC) and in children and adolescents (MFY).
Self-compassion and Its Effects in Children and Adolescents
Adolescence is often a difficult and vulnerable phase in life, as young people undergo massive changes. Adolescents are confronted with biological and cognitive changes, e.g., body growth or changes in brain structures and functioning (Blakemore et al.,
2009; Susman & Dorn,
2009). Moreover, as they try to develop self-efficacy and their own identity, they often feel stress, which is associated with depression and anxiety (Bryne et al.,
2007; Grant,
2013). Furthermore, gender differences seem to play an important role during this time. Male adolescents seem to be more resilient to stress (Parker & Brotchie,
2010). Female adolescents were reported to be twice as likely to experience depression and two to three times more likely to experience anxiety than males (Beesdo et al.,
2009; Thapar et al.,
2012). Given these aspects, many researchers argue that self-compassion may play an important role as a protective factor, especially during adolescence (Cunha et al.,
2016; Neff & McGehee,
2010).
Although this field of research seems to be promising, the number of published studies is lagging behind research in adults (Muris et al.,
2016b). A first study found strong negative correlations between self-compassion, trait anxiety (
r = −0.73), and depressive symptoms (
r = −0.60) (Neff & McGehee,
2010). Subsequent studies found negative correlations between self-compassion, perceived stress, negative affect, anxiety, depression, anger, vulnerable narcissism, panic complaints, suicidality, and post-traumatic stress symptoms, and positive correlations with mental health; effect sizes were medium to large (
r between.|55| and.|70|) (Barry et al.,
2015; Bluth & Blanton,
2014; Marshall et al.,
2015; Zeller et al.,
2015). A meta-analysis confirmed a strong negative link between self-compassion and psychological distress, with large effect sizes, while a literature review found strong evidence for self-compassion playing an important role in the prevention of depression (Marsh et al.,
2018; Pullmer et al.,
2019) and confirmed gender and age-associated differences (Bluth & Blanton,
2015; Castilho et al.,
2017; Sun et al.,
2016). These findings emphasize the importance of self-compassion as a resilience factor in adolescents.
Relevant Factors in Developing Self-compassion
For a better understanding of the relevance of self-compassion in mental health, it is important to closely examine factors that help or hinder its development. There are first indicators that self-compassion interventions help improve mental health across societies (Finlay-Jones et al.,
2018; Lou et al.,
2022). One important etiological model goes back to the compassion research of Gilbert and colleagues (Gilbert,
2009; Gilbert et al.,
2014; Gilbert et al.,
2011; Gilbert & Procter,
2006). They found that fear of positive emotions and fear of self-compassion can be relevant risk factors that come along with various mental health problems. People high in shame and self-criticism were most fearful of positive emotions due to a hyperactive threat system, insecure adult attachment, and, most often, experiential avoidance. These characteristics mostly stem from abusive, invalidating, or neglectful backgrounds, certain parenting practices, and classical conditioning (Gilbert et al.,
2014). Thus, it can be assumed that the development of self-compassion is closely related to a wider social context of adolescents: socio-economic factors, marginalization, and social exclusion increase the probability to experience abuse, invalidation, and neglect.
With regard to parental practices, cultural differences were found between Western and Eastern cultures. In collectivistic cultures that use shame, “loss of face,” or self-criticism as a method to control/regulate one’s behavior (e.g., Taiwan or Japan), people tend to be less compassionate and self-compassionate than in cultures where parents are more forgiving and warm (e.g., Thailand, USA) (Finlay-Jones et al.,
2018; Lou et al.,
2022; Montero-Marin et al.,
2018; Neff et al.,
2008). Gilbert and Procter (
2006) point out that warmth and reassurance are important concepts in developing any form of compassion. Furthermore, Neff and colleagues (2008) found a gender gap in self-compassion. American men are far more self-compassionate than women, a fact that could not be replicated for Eastern cultures.
The Assessment of Self-compassion
The Self-Compassion Scale (SCS) by Neff (
2003) is the most common assessment tool for self-compassion. It consists of 26 items on six subscales representing the three bipolar components of self-compassion. Neff (
2003) originally reported a six-factor and a higher-order factor solution. The results of subsequent studies were inconclusive. Some replicated Neff’s results (Benda & Reichová,
2016; Castilho et al.,
2015; Cunha et al.,
2016; Dundas et al.,
2016); others did not (Costa et al.,
2016; Hupfeld & Ruffieux,
2011; Lopéz et al.,
2015; Petrocchi et al.,
2013; Williams et al.,
2014). By now, there are numerous suggestions for the ideal factorial structure: a single-factor model (Deniz et al.,
2008), a two-factor model representing self-compassion (SCS-POS) and self-coldness (SCS-NEG) (Costa et al.,
2016; Lopéz et al.,
2015; Stolow et al.,
2016), a three-factor model representing the basic components of self-compassion (Hupfeld & Ruffieux,
2011), a four-factor model (Zeng et al.,
2016), and a six-factor model (Benda & Reichová,
2016; Castilho et al.,
2015; Costa et al.,
2016; Cunha et al.,
2016; Dundas et al.,
2016; Hupfeld & Ruffieux,
2011; Petrocchi et al.,
2013). In 2019, Neff and colleagues reexamined the factorial structure in various samples and found a six-factor model and a single-bifactor model to best fit the data. A bifactor model allows covariances between factors that rise from a general factor (here: self-compassion) and also tolerates the individual factors to contribute to variance in their own item subset (Reise et al.,
2010). These findings were supported by further research (Neff et al.,
2017; Tóth-Király et al.,
2016).
In recent years, there has been debate about the use of the total score. Muris and Petrocchi (
2017) uttered concerns because of the total score containing negative components. They argued that this might lead to an overestimation of the negative relationship between self-compassion and psychopathology. In 2019, Neff emphasized the validity of working with a total score, underlining that the SCS is used to examine self-compassion with its positive and negative aspects, as she defined it in 2003 (Neff,
2016; Neff et al.,
2019).
Translation and Cultural Adaptation of the SCS-CA
So far, the SCS for adults has been translated into many different languages, e.g., Czech (Benda & Reichová,
2016), Dutch (Lopéz et al.,
2015), German (Hupfeld & Ruffieux,
2011), Greek (Mantzios et al.,
2013), Italian (Petrocchi et al.
2013), Iranian (Azizi et al.,
2013), Portuguese (Castilho et al.,
2015), Spanish (Garcia-Campayo et al.,
2014), and Turkish (Deniz et al.,
2008). Some of the problems in replicating the factor structure may stem from translating the original English version to other languages, as the items have to suit the culture and therefore are sometimes adapted, which is a common problem in translating questionnaires (Auer et al.,
2000; Behling & Law,
2000). The German version SCS-D (Hupfeld & Ruffieux,
2011) replicates Neff’s six-factorial solution and was merely translated and not culturally adapted (see also: Heim et al.,
2021). Stolow and colleagues (
2016) adapted the English version of the SCS into a version for adolescents, so that they could better understand what the items mean (e.g., Item 2: “When I’m feeling down, I tend to obsess and fixate on everything that’s wrong” was adapted to Item 2: “When I feel sad or down, it seems like I’m the only one who feels that way”). The SCS for adolescents has not been as well distributed as the original SCS. There is only a 17-item English version SCS Youth for children from 10 to 14 years old (Neff,
2021) and a Portuguese version (Cunha et al.,
2016). In order to further spread the SCS for adolescents, we translated Stolow et al.’s (
2016) questionnaire for adolescents into German, creating the SCS-CA.
Discussion
Self-compassion is widely seen as an important buffer and resilience factor against stress, depression, and anxiety in adolescents (Marsh et al.,
2018), though the vast majority of research has been conducted in adult samples. The aims of the present study were to (1) translate the English version of the SCS for adolescents (Stolow et al.,
2016) into German and (2) examine its fit in a community sample of German children and adolescents aged 10 to 19 years old. In detail, we explored the factorial structure of the SCS-CA, comparing two, three, and six-factorial solutions using CFAs. Furthermore, we evaluated different forms of reliability and validity of the SCS-CA and examined potential gender differences in our sample.
The overall total score for SCS-CA was very similar to the results of other research groups, with a total of M = 3.1 (Bluth & Blanton,
2014; Cunha et al.
2016). However, the subscales with the highest and lowest mean scores differed from the findings of Cunha and colleagues (
2016). In our sample, self-judgment showed the highest mean score (
M = 4.0), and overidentification and common humanity (both
M = 3.0) had the lowest mean scores. This might be an effect of the translation or cultural understanding and should be examined closely in future research (Auer et al.
2000; Behling and Law
2000). The CFA showed the six-factorial solution to best fit our data, with good fits for
RMSEA (0.059) and
SRMR (0.066) but suboptimal fits for
CFI = 0.880 and
TLI = 0.863. This replicates the findings of Hupfeld and Ruffieux (
2011) for the SCS-D. They explained their findings by explaining that structural equation modelings show discrepancies from a perfect simple structure. Therefore, they recommended not overrating the indices with poor fit. Although the
TLI is sensitive to weak correlations between different factors and thus often stays below the cut-off (Sharma et al.
2005), the present results should be interpreted with caution. Future studies should further investigate this in other samples. The two-factorial solution showed a less good fit to our data than the six-factorial solution, with two fit indices showing adequate fits (
RMSEA = 0.073,
SRMR = 0.077) and two fit indices showing poor fits (
CFI = 0.804,
TLI = 0.787). Perhaps, self-compassion is a more complex construct and cannot be covered by one general self-compassion factor or a two-factorial solution focusing on self-kindness and self-coldness. All six factors and their interactions might be necessary to properly represent peoples’ experience of self-compassion, as defined by (Neff
2003; Neff et al.
2019). The three-factorial solution showed poor fit, replicating the findings of Hupfeld and Ruffieux (
2011). Thus, the assumption that the SCS might represent the three basic components of self-compassion instead of six different subscales might not be appropriate.
In our sample, all six subscales correlated significantly, replicating the findings of Hupfeld and Ruffieux (
2011) and Cleare et al. (
2018) but opposing the findings of Stolow et al.(
2016). Overall internal consistency of the SCS-CA was good, with Cronbach’s α = 0.89 for the total score, subscales ranging from α = 0.48 to α = 0.71. The reliability for the subscale mindfulness was very poor (α = 0.48), replicating the findings of Hupfeld and Ruffieux (
2011). More research is needed to explain the very poor reliability of the mindfulness subscale. We suspect it to either be affected by cultural or translation effects or because adolescents may not have developed a static understanding of mindfulness as a concept yet.
The SCS-CA had good content, criterion, and construct validity. We found high positive correlations between five of six SCS-CA subscales and the German version of the CAMM. The German version of the CAMM mainly contains items that map states of being not mindful and is, therefore, reverse coded. Being self-compassionate is closely related to mindfulness, as the latter is one of six subscales of self-compassion. Therefore, our results of high levels in SCS-CA and CAMM German versions fit the theoretical assumptions (Greco et al.
2011). Furthermore, the SCS-CA showed high correlations with ILK, a measure of life satisfaction for children and adolescents (Achenbach
1991). This replicates the results of other research teams supposing self-compassion to be a potential resilience factor in adults and youths (Bluth and Blanton
2015; Marsh et al.
2018; Neff and McGehee
2010). We found negative correlations between SCS-CA total and YSR-SF total (
r = −0.68), indicating that higher levels of self-compassion come with lower levels of psychopathology. This is in line with the results of other research groups investigating adult samples and children and adolescents (Bluth and Blanton
2014; Krieger et al.
2016a; Muris and Petrocchi
2017; Petrocchi et al.
2013; Raes
2010; Terry et al.
2013; Van Dam et al.
2011). The inverse correlations of the negative subscales of the SCS-CA (self-judgment, isolation, and overidentification) with YSR-SF total were bigger than the inverse correlations of the positive subscales (self-kindness, common humanity, and mindfulness). They ranged from
r = −0.54 to
r = −0.67 (negative subscales) and from
r = −0.26 to
r = −0.50 (positive subscales). This is contrary to the findings of a meta-analysis conducted by Muris and Petrocchi (
2017) that found positive correlations between the negative subscale of SCS and psychopathology, but similar to the findings of Muris (
2016a). More research is needed to further clarify these associations.
With a view to the subscales of SCS-CA and YSR-SF, we found interesting patterns, resembling the findings of Muris (
2016a). All SCS-CA subscales showed significant negative correlations with all YSR-SF subscales. Interestingly, the internalization subscales showed higher correlations with all SCS-CA subscales. Furthermore, negative subscales of the SCS-CA showed higher negative correlations with internalizing symptoms than with externalizing symptoms. This can be seen as additional evidence that differences exist between SCS subscales and their influence on internalizing and externalizing symptoms. More research is needed to further examine the relation of self-compassion and its subscales with psychopathology in children and adolescents. Being aware of the current discussion about the appropriateness of using the total score of the SCS and the reverse-coded negative items (Muris
2016a; Muris et al.
2021; Muris and Petrocchi
2017; Neff et al.
2017), we decided to report both, a total score and scores for all subscales.
With a view on potential gender effects on self-compassion, our findings are in line with those of other research groups for both adult and adolescent samples (Bluth and Blanton
2014,
2015; Cunha et al.
2016; Neff et al.
2019; Petrocchi et al.
2013; Yarnell et al.
2015). We found significant gender differences for the SCS-CA total, the overidentification and isolation subscales, and a trend for self-judgment. The effect sizes ranged from small to medium. Overall, males were significantly more self-compassionate than females. However, the gender differences were limited to the negative subscales of the SCS-CA, which is not in line with the findings of other research groups (Cunha et al.
2016; Petrocchi et al.
2013). The fact that the biggest difference was found for the subscale isolation, indicating that young females feel far more isolated (having lower scores on isolation) than young men, is of great importance. Given that the rate of depression in young women is about twice as high as in young men (Thapar et al.
2012), feeling isolated and alone with one’s problems may be an important factor. There is growing evidence that the subscale isolation plays an important role in developing and maintaining depression (Van Dam et al.
2011). Following Bluth and Blanton (
2015), we cautiously assume that girls are more self-deprecating when they are confronted with their failures. They might ruminate more, feel more isolated, and might, in succession, develop depressive symptoms.
Some limitations of this study should be acknowledged. First, the study design was cross-sectional and therefore can only provide information about the strength of correlations between self-compassion, mindfulness, quality of life, and psychopathology, but not about the direction of causality. The generalizability of the results is limited, as our sample was not representative of the general population and only consisted of healthy participants. Future research should focus on longitudinal and experimental research to closely examine the relationship between self-compassion, well-being, and psychopathology in adolescents. Some researchers have demonstrated that self-compassion can predict depression in later life (Krieger et al.
2016a; Raes
2011); these studies were conducted in adults. We recommend examining these associations with adolescents, too, in order to use self-compassion as early as possible and potentially hinder the development and maintenance of depression in early life. In a randomized controlled study (MARS-CA; (Kalmar et al.
2022)), we will assess the longitudinal and predictive validity of SCS-CA in a clinical sample. This study is currently being carried out at an outpatient center for cognitive therapy with children and adolescents in Germany. Second, our study design did not allow us to examine test-retest reliability, as the design was cross-sectional. Third, we conducted CFAs investigating two, three, and six-factorial solutions. Future research is advised to examine these results using developing statistical techniques, e.g., exploratory structural equation models (ESEM) and, ideally, including the test of a bifactorial model as proposed by Neff et al. (
2019).
Fourth, information about construct validity is limited, as we did not include measures to assess self-esteem, extraversion, neuroticism, and emotional intelligence. We did not want our set of questionnaires to take more than 30 min to complete; thus, we focused on the assessment of self-compassion, mindfulness, quality of life, and psychopathology. Future research should take discriminant validity into account. Fifth, we investigated self-compassion in a community sample. Future research should examine adolescent clinical populations to better understand the link between self-compassion and psychopathology and to deduce potential interventions and implications for therapy. Sixth, the high age range of the current study is a limitation, as from the viewpoint of internal validity, it remains unclear whether the validity is specific to certain age groups. However, it is a strength at the same time, as our sample reflects a broad range of children and adolescents and hence represents the community under investigation in a broad range, reflecting high external validity. Seventh, as in most studies investigating self-compassion, the major part of the sample consists of white females/males and children (Bluth et al.
2016; Neff et al.
2019); it is of very high importance that future studies investigate samples of other cultural backgrounds and ethnic minorities. We could not address this important research question in our sample, but it has to be highlighted that future studies addressing this issue are of utmost importance.
Implications for practice
The present study showed that the SCS-CA is a reliable and valid instrument for assessing self-compassion in children and adolescents. Given the fact that adolescence is a critical period full of challenges and massive developmental steps such as forming an own identity, finding one’s way in society, feeling alone and isolated with problems, and changes in body and brain (Blakemore et al.
2009; Susman and Dorn
2009), self-compassion seems to be a potential beneficial construct for helping young people cope with life (Marshall et al.
2015; Neff and McGehee
2010). As the probability of suffering from a mental disorder rises over the course of puberty, it is important to further investigate potential resilience factors. One of those seems to be self-compassion, as it was proven to buffer psychological distress, e.g., depression, anxiety, and stress (Barry et al.
2015; Marsh et al.
2018; Pullmer et al.
2019; Zeller et al.
2015). Further fostering interventions to improve self-compassion in children and adolescents is of particular importance. Along with manualized group settings such as MFY (Bluth et al.
2016) or therapy approaches like ACT (Hayes et al.
2012), we strongly recommend conducting randomized controlled trials in naturalistic settings, as done in adult psychotherapy research. The development of online training to facilitate access to psychotherapists might be an additional option (Hunt et al.
2021; Krieger et al.
2016b; Mander et al.
2019).
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.