Validation Study Of The Self-compassion Scale

Running head: Self-compassion in adolescence

1

Marina Cunha, Ana Xavier, & Paula Castilho (2016). Understanding self-compassion in

adolescents: Validation study of the Self-Compassion Scale. Personality and Individual

Differences, 93, 56–62. doi:10.1016/j.paid.2015.09.023

Understanding self-compassion in adolescents: Validation study of the Self–

Compassion Scale

Marina Cunha, Ana Xavier, & Paula Castilho

Abstract

Self-compassion is an adaptive self-attitude when considering personal

inadequacies or difficult life situations and seems to be crucial to adolescent’s

experience. However, self-compassion remains less investigated in adolescence. This

paper aims to analyse the psychometric properties of Self-Compassion Scale (SCS;

Neff, 2003) and test its six-factor structure through a Confirmatory Factor Analysis in a

representative sample of adolescents. The sample consists of 3165 adolescents, aged

between 12 and 19 years old (Mage = 15.49) from Portuguese schools. Results confirm

the six-factor and second-order structures of the SCS and the measurement invariance

across gender. The SCS and subscales also revealed good internal reliability and

convergent validity with measures of positive emotional memories, depressive, anxiety

and stress symptoms. Overall, our findings suggest that the SCS is a valid and reliable

measure to assess self-compassion among adolescents.

Keywords: Self-compassion; SCS; Adolescence; Confirmatory factor analysis;

Measurement invariance

Running head: Self-compassion in adolescence

2

1. Introduction

In the last years, there has been an increasing interest and a growing body of

research about the nature and implications of self-compassion in various fields from

health and education to business and sociological domains. Self-compassion is rooted in

Asian philosophy and according to Neff (2003a) entails three main interacting

components: Self-Kindness (i.e., the tendency to be kind and understanding toward

oneself in instances of pain or failure rather than being harshly critical); Common

Humanity (i.e., the ability to perceive one’s experiences as part of the larger humanity

experience rather than seeing them as separating and isolating); and Mindfulness (i.e.,

being aware of present moment experience in a balanced manner rather than over-

identifying with one’s negative emotions).

In face of negative external events (e.g., experiences of failure, shame, bullying),

painful or distressing feelings are not avoided, supressed or perpetuated by self-critical

evaluations but instead are seen as part of a shared human experience through an

attitude of acceptance, kindness, compassionate and non-judgmental, allowing proactive

and effective behaviours (Neff, 2003a). Indeed, several studies have shown that self-

compassion is negatively associated with anxiety, stress, depression, rumination

(Castilho, Pinto-Gouveia, & Duarte, 2015; Neff, Rude & Kirkpatrick, 2007; Raes,

2010), shame (Gilbert & Procter, 2006), academic failure (Neff, Hseih, & Dejitthirat,

2005). On the contrary, self-compassion is strongly and positively linked to

psychological well-being, happiness, life satisfaction, optimism, emotional intelligence,

interpersonal connectedness (Neff et al., 2007).

Although research on self-compassion in adolescent samples is scarce, some

studies have recently emerged. Overall, these findings point out that self-compassion is

associated with positive psychological indicators (e.g., early memories of warmth and

Running head: Self-compassion in adolescence

3

safeness, secure attachment, mindfulness, mental health, life satisfaction, emotional

well-being; Cunha, Martinho, Xavier, & Espírito-Santo, 2013; Bluth & Blanton, 2014;

Marshall, Parker, Ciarrochi, Sahdra, Jackson, & Heaven, 2015; Neff & McGehee,

2010). Inversely association patterns between self-compassion and maladaptive

outcomes were found (e.g., negative affect, aggression, trauma-related symptoms,

depression, anxiety; Barry, Loflin, & Douvette, 2015; Bluth & Blanton, 2015; Tanaka,

Wekerle, Shmuck, & Paglia-Boak, 2011; Vettese, Dyer, Li, & Wekerle, 2011; Zeller,

Yuval, Nitzan-Assayag, & Bernstein, 2014).

The major developmental tasks during the transition from childhood to

adulthood, including identity formation, playing different social roles, autonomy from

parents, necessity of belonging and acceptance of peers group, make adolescence a time

of heightened vulnerability to suffering (Steinberg & Morris, 2001). Thus, self-

compassion could be beneficial for this age group by providing a way for adolescents to

perceive their failures or mistakes proportionally and in a balanced perspective, to

experience supportive and warmth feelings towards themselves without engaging in the

problematic process of self-criticism, evaluation and social unfavourable comparisons

(Neff & McGehee, 2010). Therefore, the cultivation of self-compassion may function as

a protective strategy and may be a preventive and intervention target for adolescents in

order to foster their resilience and well-being in various contexts of their lives (e.g.,

family, school, friendships and community).

The majority of research conducted on self-compassion has widely used the

Self-Compassion Scale (SCS; Neff, 2003a). Neff (2003a) conducted several

confirmatory factor analyses (CFA) and results determined that a single higher-order

factor of self-compassion could explain the inter-correlations between the six subscales

(NNFI=.90; CFI=.91), indicating that this scale may be analysed through its six

Running head: Self-compassion in adolescence

4

subscales separately or as an overall score. Neff (2003a) also found good internal

reliability for the total score and for their subscales and good convergent and divergent

validities. The SCS has been adapted and validated in other countries. The Portuguese

version in non-clinical and clinical adult samples (Castilho et al., 2015) confirmed the

six-factor and higher-order structures of the SCS and found good psychometric

properties, high internal consistency for the total score (α = .94 for non-clinical and α =

.92 for clinical samples) and subscales (ranging between .70 and .88).

An exploratory study was conducted to adapt the Portuguese version of the SCS

for adults to adolescents. This adaptation took into account the cultural and linguistic

issues, but did not alter the items’ content of the original version, by maintaining the

semantic equivalence across languages. For instance, in some items an example was

added in order to make them more comprehensible for adolescents (e.g., item 5 “I try to

be loving towards myself when I’m feeling emotional pain (e.g., do or say something

kind towards myself)”; item 6 “When I fail at something important to me I become

consumed by feelings of inadequacy (e.g., feelings of failure)”; item 22 “When I’m

feeling down I try to approach my feelings with curiosity and openness (e.g., without

judgements or trying to avoid them)”. Overall, these results showed promising

psychometric properties for the SCS, despite its construct validity was not analysed

(Cunha, Xavier, & Vitória, 2013). Therefore, this study aims to examine the factorial

structure of the SCS in a large sample of adolescents from community and test the

measurement invariance across gender. Finally, convergent validity is also explored

with other related measures.

2. Method

2.1. Participants

Running head: Self-compassion in adolescence

5

The sample consists of 3165 adolescents, 1461 males (46.2%) and 1704 females

(53.8%), with a mean age of 15.49 (SD = 1.59) ranging between 12 and 19 years old.

This sample is from 7th to 12th grade (M = 9.70, SD = 1.43). No gender differences were

found for age, t(3007,314) = 0.572, p = .567, and years of education, t(3053,235) = -1.648, p =

.099.

2.2. Measures

Self-Compassion Scale (SCS; Neff, 2003a; Portuguese version for adolescents:

Cunha et al., 2013) comprises 26 items and six subscales: Self-Kindness; Self-

Judgment; Common Humanity; Isolation; Mindfulness; Over-Identification.

Respondents were instructed with the sentence “how I typically act towards myself in

difficult times” and were asked to answer each item according to a 5-point scale (1=

Almost Never; 5= Almost Always). Subscale scores are computed by calculating the

mean of subscale item responses. To compute the total score of SCS, the Self-Kindness,

Common Humanity, and Mindfulness are summed with reverse scores of the Self-

judgment, Isolation, and Over-identification subscales. Higher scores indicate greater

self-compassion. In the original version, the total score showed an excellent internal

consistency (α = .92) and the six subscales revealed adequate coefficients of internal

consistency, ranging between .75 and .81.

Early Memories of Warmth and Safeness Scale (EMWSS; Richter, Gilbert,

& McEwan, 2009; Portuguese version for adolescents: Cunha, Xavier, Martinho, &

Matos, 2014) measures recall of feeling warm, safe and cared for in childhood (e.g., “I

felt that I was a cherished member of my family.”). This is a 21-item scale rated on a 5-

point scale (0 = No, never; 4 = Yes, most of the time). Richter and colleagues (2009)

found a high Cronbach’s alpha of .97. The Portuguese version for adolescents also

Running head: Self-compassion in adolescence

6

revealed an excellent internal reliability (α = .95). In the current study, EMWSS

Cronbach’s alpha was .95.

Depression, Anxiety and Stress Scales (DASS-21; Lovibond & Lovibond,

1995; Portuguese version: Pais-Ribeiro, Honrado, & Leal, 2004) depression, anxiety

and stress symptoms. The 21-items indicate negative emotional symptoms rated on a 4-

point scale (0-3). Lovibond and Lovibond (1995) found high internal consistency (α =

.91 for depression, α = .84 for anxiety and α = .90 for stress). In the Portuguese version,

the internal consistency was good (α = .85, .74 and .81, respectively). In the present

study, the Cronbach’s alpha for subscales were .88, .83 and .86 respectively.

2.3. Procedures

This adolescents’ sample was collected from public schools in the north and

centre regions of Portugal. Ethical approvals were obtained by the Portuguese Ministry

of Education and the National Commission for Data Protection. The head teacher of the

school and parents were informed about the goals of the research and gave their written

consent. Adolescents assented to participate and were informed about the purpose of the

study and aspects of confidentiality. They voluntarily participated and filled out the

instruments in the classroom. The teacher and research assistant were present to provide

clarification if necessary and to ensure confidential and independent responding.

2.4. Data Analysis

Statistical software IBM SPSS (v.20) and AMOS (v. 18) was used. Descriptive

statistics, independent sample t tests, Cohen’s d and the effect size correlation were

calculated. Pearson correlations were computed to assess the convergent and divergent

validities.

Running head: Self-compassion in adolescence

A CFA using a maximum likelihood estimator (ML) was performed for factor

validity and model invariance across gender. The chi-square and simultaneously the

following goodness-of-fit indices were analysed: GFI, CFI, TLI ≥ .90, acceptable, ≥ .95,

good; RMSEA ≤ .06, good, ≤ .08, acceptable. The improvement of model fit was based

on Modification Indexes (MI > 11; p ≤ .001) by adding sequentially correlational

measurement errors for the residuals with higher MI values. The best fitting model as

determined by chi-square difference test and comparison indices, such as AIC and

ECVI, with smaller values indicating superior models and more stable model for

population under study (Kline, 2005).

Regarding local adjustment of the model, all standardized factor loadings should

be significant (p < .05) and equal or greater than .50. The average variance extracted

(AVE) was analysed to assess the discriminant validity of the measure (Hair, Anderson,