Self-compassion as a healthy attitude toward the self: Factorial and construct validity in an Italian sample

https://doi.org/10.1016/j.paid.2017.06.028Get rights and content

Highlights

  • A complete Italian version of the Self-Compassion Scale is proposed.

  • Factor analyses sustain the employment of both six subscales and a total score.

  • Self-compassion confirms to be a healthy self-attitude, alternative to self-esteem.

  • Self-compassion is independent of narcissism and self-enhancement.

  • Self-compassion is associated with well-being controlling for self-esteem.

Abstract

This paper aims to propose the Italian version of the Self-Compassion Scale (SCS, Neff, 2003b) and to provide new evidence concerning its relationships with various forms of self-view and well-being. In the first study, we analysed whether the scale preserved its original psychometric features. Among the tested models, both a six-factor and a bifactor model showed adequate fit indexes, sustaining the employment of both the six subscales and a total self-compassion score. In the second study, through Confirmatory Factor Analysis and partial correlations, we explored convergent, divergent, and predictive validity of the scale. As expected, self-compassion was associated with, yet distinguishable from, self-esteem and low labile self-esteem scores, and it was unrelated to narcissism and self-enhancement. -Moreover, self-compassion maintained its link with well-being variables also controlling for self-esteem, labile self-esteem, narcissism, and self-enhancement. Findings suggest that self-compassion may be conceived as a healthy self-attitude, alternative to self-esteem, as it is related to self-esteem benefits (low labile self-esteem and well-being), but not with its potential downsides (narcissism and self-enhancement). Therefore, self-compassion appears as a self-caring disposition that does not lead to overly positive self-evaluations and self-image enhancement.

Introduction

Psychological research has recently directed an increasing attention toward the construct of self-compassion. The theoretical definition of this construct is anchored to the broader concept of compassion, which has been defined as an affective state elicited by a non-judgmental awareness of the others' pain, leading to the desire to alleviate the others' sufferings (Neff, 2003a, Wispé, 1991). Likewise, self-compassion has been conceived as a positive attitude that arises in front of one's own experiences of troubles and suffering (Neff, 2003a). In particular, self-compassion consists of three bipolar components. The first dimension is self-kindness, which describes attitudes of kindness and understanding toward the self despite one's limits and failures, in contrast to the opposite pole, self-judgment, characterized by harsh self-criticism. The second component is common humanity, which is the awareness that one's negative experiences are part of the human nature and are shared with all the other humans. Its opposite pole, isolation, involves a sense of separation from others, as if one's own troubles and suffering were abnormal and unique. The third dimension is named mindfulness, as it describes a balanced awareness of personal negative experiences; its opposite pole, over-identification, represents a process of identification with one's own difficulties, promoting ruminant thoughts. Although these components have been theorized as conceptually distinct, they are all equally important in the definition of self-compassion. Therefore, self-compassion can be conceived as a dynamic system, resulting from the interaction of its three bipolar components (Neff, Whittaker, & Karl, 2017).

Self-compassion is most often measured with the Self-Compassion Scale (Neff, 2003b), which consists of 26 items, designed to capture both the positive and negative poles of the three components, i.e. self-kindness vs. self-judgment, common humanity vs. isolation, and mindfulness vs. over-identification. In the original validation article, involving three samples, Neff (2003b) reported acceptable indexes of fit both for the six-factor solution and a higher-order model including the six first-order factors and a total self-compassion score as second-order latent variable. These preliminary analyses supported a twofold use of the scale, i.e. calculating the six subscales' scores and an overall score of self-compassion (Neff, 2003b). Actually, most of the self-compassion research focused on the global self-compassion score.

However, in the validation process of translated versions of the SCS, some inconsistent findings emerged about its factorial structure. First, several translations have confirmed the six-factor structure (e.g. Azizi et al., 2013, Castilho et al., 2015, Chen et al., 2011, Lee and Lee, 2010), while in other studies the six-factor solution emerged only by dropping out some items (Neff and Vonk, 2009, Petrocchi et al., 2014). Secondly, the higher-order model found support in Chinese undergraduates (Chen et al., 2011) and Portuguese clinical and community samples (Castilho et al., 2015), but not in other analyses (e.g. Costa et al., 2016, Hupfeld and Ruffieux, 2011, Petrocchi et al., 2014, Williams et al., 2014). The lack of a clear second-order factorial structure may question the computation and interpretation of a SCS global score (e.g. Muris et al., 2016, Muris and Petrocchi, 2016, Neff, 2016a).

Based on a critical examination of the scale, an alternative two-factor model has also been proposed, in which a self-compassion factor is computed by collapsing the three positive subscales scores, while the items of the negative poles saturate a single factor, named self-criticism (Costa et al., 2016, López et al., 2015, Muris, 2015) or uncompassionate behavior (Neff, 2016a). Notably, this two-factor solution does not reflect the original theorization of the scale (Neff, 2016a). Indeed, the SCS was conceived as a tripartite instrument, able to disentangle three types of individual differences when dealing with personal failures and sufferings. First, differences in the way the self is treated, i.e. in a kind or denigrating way. Second, differences in the appraisal of difficulties, which may be seen as part of the human nature or as a source of isolation. Third, differences in the attitude toward personal sufferings, which may elicit a balanced awareness or an identification with one's own problems.

Given the complexity and heterogeneity of these findings, Neff, 2016a, Neff, 2016b proposed a different approach to test the dimensionality of the Self-Compassion Scale: the employment of a bifactor solution (Reise, Bonifay, & Haviland, 2013). In a bifactor model, a target factor directly influences item responses and, in turn, the items are measures of different group factors (Reise, Moore, & Haviland, 2010). The target factor consists in a single trait that accounts for some proportion of common item variance for all items, while group factors explain additional common variance for each subscale. Moreover, it is possible to compute, through the omega index and the omega hierarchical (McDonald, 1999), the percentage of total variance score explained by the general factor, the group factors, and error. The computation of a global scale score will be sustained if the large majority of the observed variance is accounted for by the target factor (Reise et al., 2010). The bifactor model may be an accurate way to represent self-compassion (Neff et al., 2017), as the target factor consists in the overall self-compassion score, thus explaining some proportion of common item variance for all items, and the six subscales represent the group factors, which account for additional common variance of the subscales. The computation of a total self-compassion score would be justified if the large majority of the observed variance was accounted for by the general self-compassion factor.

Since the work of Neff, 2016a, Neff, 2016b, the bifactor model has received support from several studies on the SCS. Tóth-Király, Bőthe, and Orosz (2016) performed bifactor CFA and bifactor Exploratory Structural Equation Modelling (ESEM) on the Hungarian version of the SCS and found that the latter provided the best fit to the data, also compared to six-factor first-order CFA and ESEM. In the French validation of the SCS (Kotsou & Leys, 2016), the six-factor solution showed the best fit, but the omega index of the still acceptable bifactor model highlighted the relevance of the total score accounting for self-compassion. Also in the Brazilian validation of the SCS, the six-factor solution showed the best fit, but the bifactor approach led to acceptable fit indexes (de Souza & Hutz, 2016).

Neff et al. (2017) have recently compared the fit of one-factor, two-factor correlated, six-factor correlated, higher-order, and bifactor solutions. All these factorial structures have been explored in samples of undergraduates, community adults, meditators, and individuals with a history of recurrent depression. The higher-order model, the one-factor, and the two-factor correlated solutions demonstrated poor fit across all the samples, thus their use did not seem justifiable. On the contrary, the fit indexes supported the six-factor solution across all the samples and the bifactor model within three samples (undergraduates, community adults, and meditators). Notably, although the six-factor solution showed the best fit to the data, in the bifactor model the large majority of variance was accounted for by the general self-compassion factor. This supported the computation of a total scale score, despite the presence of multidimensionality, but also of the six subscales scores (Neff et al., 2017, Reise et al., 2013a). Concluding, validations of translated versions of the SCS should also test a bifactor model, in addition to the other solutions already investigated (Neff, 2016a, Neff, 2016b, Neff et al., 2017).

Several research findings suggest that self-esteem and self-compassion are strongly related but distinguishable constructs (Neff, 2011, Neff and Vonk, 2009). Self-esteem consists in a positive self-evaluation, related to attitudes of self-respect and worth (Klein, 1992), and to low self-perceptions of unworthiness and inadequacy (Rosenberg, 1989). As also self-compassion involves positive self-attitudes and feelings of self-worth (Neff, 2011), convergent validity could be assessed investigating the association between the SCS and the Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965), the most used instruments for the assessment of global self-esteem. Consistently, previous studies showed correlations coefficients between SCS and RSES ranging from 0.57 to 0.59 (Leary, Tate, Adams, Batts Allen, & Hancock, 2007; Neff, 2003a, Neff, 2011, Neff and Vonk, 2009).

Notably, self-compassion has been theorized as a healthy attitude toward oneself, somehow alternative to self-esteem, as it may offer similar benefits without involving its potential downsides (Neff, 2003a, Neff, 2011, Neff and Vonk, 2009). Indeed, self-esteem is an adaptive self-attitude and one of the most important aspect of mental health (James, 1983), as suggested also by its associations with high levels of happiness and optimism and low scores of anxiety and depression (Pyszczynski, Greenberg, Solomon, Arndt, & Schimel, 2004). Nevertheless, research findings suggested that what people do to get and maintain high levels of self-esteem may lead to some problematic correlates, such as narcissism and self-enhancement (Baumeister et al., 2003, Crocker and Park, 2004).

Narcissism is an inflated, unrealistically positive self-view (Campbell & Foster, 2007), generated by a limitless and dysfunctional pursuit of self-esteem. A narcissistic self-view may represent a “social trap”, characterized by short term benefits, but long term aversive consequences in terms of underperformance, poor interpersonal relations, and even aggression (Campbell & Buffardi, 2008). While past research often found a positive association between self-esteem and narcissism (e.g. Crocker & Park, 2004; see also Stronge, Cichocka, & Sibley, 2016), no significant association emerged between self-compassion and narcissism (e.g. Leary et al., 2007, Neff, 2003b, Neff et al., 2007). The last finding is not surprising, since narcissism is characterized by pervasive grandiosity and feelings of entitlement (e.g. Emmons, 1987), perceptions that should be unrelated to self-compassion, as the latter does not involve judging attitudes toward the self or the others (Neff & Vonk, 2009). Consistent with these features and previous findings, a statistically non-significant relation between SCS and narcissism should be conceived as a proof of divergent validity.

The second potential downside of self-esteem is self-enhancement, which consists in positive illusions about the self that push individuals to perceive themselves as better than the others (e.g. Sedikides and Gregg, 2008, Taylor and Brown, 1988). Although self-enhancement is usually associated with well-being (e.g. Taylor & Brown, 1988), research also suggests that it may inhibit processes of learning and growth (e.g. Baumeister et al., 2003, Kim et al., 2010). While it's well established that high levels of self-esteem are combined with self-enhancing tendencies (e.g. Baumeister et al., 1989, Rosenberg et al., 1995), no results are available concerning the association of self-compassion with self-enhancement. This fact is quite surprising, as several authors have theorized that self-compassion and self-enhancement should be unrelated (e.g. Breines and Chen, 2012, Leary et al., 2007, Neff, 2011, Neff and McGehee, 2010, Neff et al., 2007). Indeed, although self-compassion is a source of positive self-regard, it does not imply an overly positive evaluation of the self or positive illusions about the self-image (Breines and Chen, 2012, Leary et al., 2007, Neff, 2011). A statistically non-significant relationship between self-compassion and self-enhancement would represent a test of divergent validity, useful to confirm that self-compassion consists in a positive, but realistic, self-view, without feelings of superiority over the others or inflated self-illusions.

Occasionally, self-esteem may fluctuate based on specific outcomes and circumstances (Crocker, Luhtanen, Cooper, & Bouvrette, 2003). Such an instability may turn into a labile self-esteem (e.g. Butler et al., 1994, Dykman, 1998), characterized by daily oscillations between positive and negative self-evaluations, depending on personal experiences. People with this unstable kind of self-esteem are particularly sensitive to negative events and vulnerable to depression symptoms (Dykman, 1998). Previous research demonstrated that a high self-esteem is able to protect individuals from these fluctuations, as shown by the negative association between self-esteem levels and self-esteem instability (e.g. Zeigler-Hill, Chadha, & Osterman, 2008). A self-compassionate disposition is likely to offer a similar benefit. As self-compassion promotes attitudes of kindness and acceptance toward the self when facing failures (Neff & Vonk, 2009), it should also mitigate the shifts between positive and negative self-appraisals driven by favorable or adverse circumstances, fostering stable self-worth feelings. Notably, the relationship between self-esteem and labile self-esteem has yet to be investigated. A negative association between the two constructs would represent a further test of convergent validity of the SCS.

A final important positive correlate of self-esteem is well-being (Diener, Diener, & Diener, 1995). Self-compassion seems to share with self-esteem also this important benefit, as it showed (for a review, see Barnard & Curry, 2011) associations with positive affectivity (Leary et al., 2007, Neff and Vonk, 2009, Neff et al., 2007), life satisfaction, self-reported happiness, optimism (Neff et al., 2007), and general psychological health (Marshall et al., 2015). Importantly, a study corroborated the predictive validity of the SCS by showing that self-compassion preserved its associations with well-being indicators also controlling for self-esteem (Neff & Vonk, 2009). Nevertheless, no evidence is present in the literature about the effects of self-compassion on well-being controlling for narcissism, self-enhancement, and labile self-esteem. Such a proof would be important to further sustain the predictive validity of the scale.

Section snippets

Aims and studies overview

The aim of the present paper is twofold. First, in Study 1 we aimed at proposing an Italian version of the Self-Compassion Scale that reproduces, as far as possible, the psychometric characteristics of the original English version by Neff, 2003b, Neff, 2016a, Neff, 2016b. Second, in Study 2 we aimed at investigating issues of convergent and divergent validity of the scale, assessing its relations with self-esteem and self-views scales, and of predictive validity, analyzing its association with

Study 1

The aim of the first study was to investigate the psychometric properties and the factorial structure of the Italian version of the SCS. In the original validation of the scale (Neff, 2003b), both the total score of the SCS and each subscale were characterized by a good reliability, and men presented higher levels of global self-compassion. We expected to find these features in the Italian version too. Moreover, we tested several factorial structures of the scale, trying to replicate and extend

Study 2

In the second study, we tested validity issues of the SCS in its Italian version in the framework of self-related perceptions and views. Based on previous research findings and theoretical reasons (Neff, 2003a, Neff, 2011, Neff, 2016a, Neff and Vonk, 2009), we tested convergent validity investigating the relations that self-compassion holds with self-esteem levels and labile self-esteem, expecting a positive link with the former and a negative association with the latter. Regarding divergent

General discussion

The aim of the present research was to propose an Italian version of the Self-Compassion Scale (Neff, 2003b) and to investigate issues of convergent, divergent, and predictive validity of the scale, extending previous research findings on its relations with self-related perceptions and views.

In the first study, we presented the Italian version of the scale, developed with the aim to preserve the original psychometric characteristics (Neff, 2003b, Neff et al., 2017). In particular, we focused on

Conclusions

This paper tested – to our knowledge for the first time – the relationships between self-compassion, self-related perceptions and views through confirmatory factor analysis, a more stringent test compared to simple correlations (Bagozzi & Burnkrant, 1979). All the relationships emerged in the CFA supported the convergent and divergent validity of the Italian version of the SCS and, in a broader sense, of the construct in general. Likewise, partial correlations provided new evidence on the

References (74)

  • K.A. Bollen

    Structural equations with latent variables

    (1989)
  • J.G. Breines et al.

    Self-compassion increases self-improvement motivation

    Personality and Social Psychology Bulletin

    (2012)
  • A.C. Butler et al.

    A comparison of self-esteem lability and low trait self-esteem as vulnerability factors for depression

    Journal of Personality and Social Psychology

    (1994)
  • W.K. Campbell et al.

    The lure of the noisy ego: Narcissism as a social trap

  • W.K. Campbell et al.

    The narcissistic self: background, an extended agency model, and ongoing controversies

  • P. Castilho et al.

    Evaluating the multifactor structure of the long and short versions of the self-compassion scale in a clinical sample

    Journal of Clinical Psychology

    (2015)
  • J. Chen et al.

    Reliability and validity of Chinese version of self-compassion scale

    Chinese Journal of Clinical Psychology

    (2011)
  • J. Costa et al.

    Validation of the psychometric properties of the self-compassion scale. Testing the factorial validity and factorial invariance of the measure among borderline personality disorder, anxiety disorder, eating disorder and general populations

    Clinical Psychology & Psychotherapy

    (2016)
  • J. Crocker et al.

    Contingencies of self-worth in college students: Theory and measurement

    Journal of Personality and Social Psychology

    (2003)
  • J. Crocker et al.

    The costly pursuit of self-esteem

    Psychological Bulletin

    (2004)
  • P.J. Curran et al.

    The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis

    Psychological Methods

    (1996)
  • E. Diener et al.

    Factors predicting the subjective well-being of nations

    Journal of Personality and Social Psychology

    (1995)
  • E. Diener et al.

    The satisfaction with life scale

    Journal of Personality Assessment

    (1985)
  • B.M. Dykman

    Integrating cognitive and motivational factors in depression: Initial tests of a goal-orientation approach

    Journal of Personality and Social Psychology

    (1998)
  • R.A. Emmons

    Narcissism: Theory and measurement

    Journal of Personality and Social Psychology

    (1987)
  • A. Fossati et al.

    Proprietà psicometriche della versione Italiana del narcissistic personality inventory [psychometric properties of the Italian version of the narcissistic personality inventory]

    Rivista di Psicologia Clinica

    (2008)
  • D.P. Goldberg et al.

    User's guide to the GHQ

    (1988)
  • G.R. Hancock et al.

    Rethinking construct reliability within latent variable systems

  • L. Hu et al.

    Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives

    Structural Equation Modeling

    (1999)
  • J. Hupfeld et al.

    Validierung einer deutschen version der self-compassion scale (SCS-D)

    Zeitschrift für Klinische Psychologie und Psychotherapie

    (2011)
  • W. James

    The principles of psychology

    (1983)
  • Y. Kim et al.

    Know thyself: Misperceptions of actual performance undermine achievement motivation, future performance, and subjective well-being

    Journal of Personality and Social Psychology

    (2010)
  • H.A. Klein

    Temperament and self-esteem in late adolescence

    Adolescence

    (1992)
  • I. Kotsou et al.

    Self-compassion scale (SCS): Psychometric properties of the French translation and its relations with psychological well-being, affect and depression

    PloS One

    (2016)
  • M.R. Leary et al.

    Self-compassion and reactions to unpleasant self-relevant events: The implications of treating oneself kindly

    Journal of Personality and Social Psychology

    (2007)
  • W.K. Lee et al.

    The validation study of the Korean version of self-compassion scale with adult women in community

    Journal of Korean Neuropsychiatric Association

    (2010)
  • T.D. Little et al.

    To parcel or not to parcel: Exploring the question, weighing the merits

    Structural Equation Modeling

    (2002)
  • Cited by (0)

    View full text