Introduction

Adolescence is a stage marked by deep physiological, cognitive, psychological, and behavioural changes [1]. These are often accompanied setbacks and difficulties [2] which may explain the emergence of emotional difficulties in this phase (e.g., increased emotional reactivity, lower ability for emotion regulation, symptoms of stress and anxiety) [3,4,5] and specific disorders such as eating disorders [6]. Most of the negative factors that contribute for the development of an eating disorder during adolescence are reported to be higher in girls, comparing to boys [7], and eating disorders are indeed known to mainly affect adolescent girls [8, 9]. Considering this, adolescence appears to be an optimal period to develop and maintain healthy social and emotional habits, and promote supportive environments in the family, at school, and in the wider community [10]. This effort is essential to protect individuals from hostile experiences and risk factors that may have a detrimental impact in their health and well-being [10].

Regarding the emergence of eating disorders, many influences exist during the teen years, including changes in one’s body shape, weight, and appearance [11]. Historically, research on eating disorders has been focusing on the influence of negative body image [12], with several studies pointing body image dissatisfaction, drive for thinness, perceived social pressure to be thin, and family history of eating concerns to be its soundest predictors [13, 14].

Recently, an effort to change the polarity of research on body image and eating-related disorders has produced insights on how holding a positive body image associates with several positive and protective outcomes in adolescence. These include valuing body image differences, self-confidence, self-attunement and empowerment [15, 16], body functionality, and sports participation [17]. Moreover, body appreciation, a facet of positive body image, is known to predict a decrease in dieting in adolescent girls [18], and to be negatively correlated with body mass index, body dissatisfaction, and disordered eating [17]. Body appreciation defines as “holding favourable opinions of the body regardless of actual physical appearance, acceptance of the body despite incongruences with media appearance ideals, respect toward the body by attending to its needs and engaging in healthy behaviours, and protection of the body by rejecting unrealistic appearance ideals” [19]. Although the factors that contribute for body appreciation are not thoroughly explored, especially in adolescence, some studies have pointed its association with protective factors, namely, unconditional and body-related acceptance by others [20,21,22], and self-compassion [23,24,25,26]. Considering that positive body image may constitute an effective way to prevent body image and eating-related disorders, further research on factors that contribute to body appreciation in adolescence seems to be of particular importance [12].

Compassion has been successfully incorporated in therapeutic approaches to body and eating-related disorders with adults [27] and has also been suggested to be applied in adolescents with such difficulties [28]. Indeed, compassion lies on the opposite side of attitudes associated with negative body image, such as harsh judgement, striving and competition [29, 30]. It defines as “a sensitivity to suffering with a commitment to try to alleviate and prevent it”, and may flow in three distinct directions: compassion for others, compassion received from others, and self-compassion [30].

Self-compassion has been associated with body appreciation in samples of adult females [25, 26], although studies using adolescent samples and exploring other directions of compassion are absent. Indeed, compassion associates with many positive outcomes, such as psychological and interpersonal processes [30]. One of the foundations of its adaptive character lies on the fact that the ability to access compassion from self and others has proved to be a central skill to prevent and cope with difficulties and severe psychosocial struggles [30].

Compassion activates the safeness-soothing affect regulation system, a system that is stimulated by signs of care, protection, and affection, associating with social safeness by promoting feelings of affiliation, connectedness, and closeness [31, 32]. Social safeness is defined as the experience of positive feelings of pleasure, connectedness, and safeness within one’s social world, and of feeling cared about and calmed by other people [33]. It is positively associated with a sense of warmth and security in social relationships and with the perception of compassion in other people [33]. Recent studies on positive psychology have reported that, in adults, it associates with body appreciation [25, 26]. Regarding negative correlations, social safeness associates with a lower tendency to perceive others as hypercritical and rejecting [33].

The present study aimed at testing a theoretical model exploring the association between self-compassion and receiving others’ compassion, social safeness, body appreciation and disordered eating, on female adolescents. Considering the need of studying protective factors against disordered eating in adolescence, this model explored two main hypotheses. First, it was hypothesized that having self-compassion competences and perceiving others’ compassion competences associate with higher social safeness and pleasure, which would associate with higher body appreciation. Second, it was hypothesized that higher body appreciative attitudes would associate with a weaker manifestation of disordered eating symptoms.

Methods

Participants

Participants in this study were 205 female adolescents, aged between 12 and 18 (M = 14.3; SD = 1.5). A total of 27.8% of participants (n = 57) reported that they lived in a rural area, while 72.2% in an urban area. These girls attended school from the 7th to the 11th grade, and the majority of participants were in the 10th grade (30%). Regarding BMI values, approximately 2% of the participants presented thinness (BMI z-scores under − 2SD), 76% presented normal weight (BMI z-scores between − 2SD and + 1SD), 18% presented with overweigh (BMI z-scores between + 1SD and + 2SD) and 5% with obesity (BMI z-scores above + 2SD) [34].

Measures

All measures used in the present study were administered in their Portuguese versions and were validated for the adolescent population. Cronbach’s alpha values for all measures are presented in Table 1.

Table 1 Cronbach’s alphas (α), means (M), and standard deviations (SD) for the total sample (n = 205), and intercorrelation scores on self-report measures
Full size table

Compassionate engagement and action scales (CEAS) [35, 36]

CEAS are a set of three scales that assesses compassion competences for others, from others to oneself, and self-compassion. Each one (with 13 items each) divides into two subsections assessing engagement (motivation and competencies to engage with, attentional sensitivity to suffering signs); and action (motivation to attain wisdom and skills to ease or prevent suffering). Respondents reflect on potentially difficult situations and rate their frequency (1 “Never” to 10 “Always”). In the present study, only the scales assessing self-compassion (CEAS_SelfComp) and compassion received from others (CEAS_CompFromOthers) were used. The Portuguese version for adolescents [36] revealed good validity (\(\alpha\) values from 0.71 to 0.92 in the subscales used in this study).

Social safeness and pleasure scale (SSPS) [33, 37]

SSPS evaluates current feelings of acceptance, belonging and safeness in social relationships. It is composed of 11 items to be rated in a scale from 0 (“Almost never”) to 4 (“Almost all the time”). SSPS has shown consistent psychometrics (α = 0.93) in the validation for Portuguese adolescents.

Body appreciation scale-2 (BAS) [38, 39]

BAS assesses participants’ acceptance, gratitude, care and respect to one’s own body features and rejection of harmful societal appearance ideals. Its 10 items are rated between 1 (“Never”) and 5 (“Always”). Good psychometric properties of the BAS have been reported (α = 0.94 for adolescent girls in the Portuguese validation study).

Eating disorder examination questionnaire (EDE-Q) [40, 41]

EDE-Q assesses the frequency and intensity of disordered eating attitudes and behaviours. It comprises 36 items divided into four subscales: restraint, weight concern, shape concern and eating concern. The global score results of the mean of all the subscales’ scores. Items are rated for the frequency of occurrence or severity. In the present study, only the global score was used. The Portuguese version showed good psychometric properties (α = 0.94).

Body mass index (BMI)

BMI was calculated by dividing participants’ self-reported weight by the square of their height (kg/m2). BMI values were analysed considering the World Health Organization’s growth reference data for 5 to 19 years [34]. BMI z-scores (“BMI for age”, age and sex sensitive; zBMI) were estimated according to WHO Child Growth Standards and used to determine BMI categories among participants [34].

Procedures

The present study is part of a wider research project on positive body image and well-being of Portuguese female adolescents. Ethics approval was obtained from independent ethics committees. Students from five middle and secondary schools in the centre of Portugal were invited and accepted to participate in the study. Criteria for inclusion were presenting informed consent form signed by a legal guardian, being aged between 12 and 18, and being able to autonomously complete self-report questionnaires. Principal, teachers at the schools, students and legal guardians were debriefed about the purposes of the study, the voluntary nature of the participation, the confidentiality of the data, and legal guardians and students were asked to sign a written informed consent form. Self-report instruments were counterbalanced and administered to participants at a class, with a researcher present to clarify doubts about the study.

Data analysis

Descriptive statistics (to explore the sample characteristics) and product-moment Pearson correlation analyses (analyse relationships between study variables) were computed using the software IBM SPSS (v.22; SPSS Inc., Chicago, IL).

A path analysis was conducted to examine the mediator roles of social safeness and body appreciation in the relationship between both self-compassion and compassion received from others and disordered eating. Age and zBMI’s effects were controlled for in this model, due to their influence on body image and eating related variables [6,7,8,9]. This analysis (inspection of assumed structural relations between the variables) was conducted using the software AMOS (v.22, SPSS Inc., Chicago, IL). The Maximum Likelihood estimation method was used to estimate model path coefficients and to compute fit statistics (with 95% confidence interval). Chi-square tests were used to test the significance of direct, indirect, and total effects. The Bootstrap resampling method was used to test the significance of the mediational paths, with 5000 Bootstrap samples and 95% confidence intervals around the standardized estimates of effects (total, direct and indirect). The mediational effect is considered to be statistically significant (p < 0.05) if zero is not included in the interval between the lower and the upper bound of the 95% bias-corrected confidence intervals [42]. Several goodness-of-fit measures were used to assess the adjustment of the tested model to empirical data: Chi-square (χ2) and Normed Chi-square (χ2/d.f.) [43]; Normed Fit Index (NFI) [43] Comparative Fit Index (CFI) and Tucker Lewis Index (TLI)[44]; and Root-Mean Square Error of Approximation (RMSEA) with 95% confidence interval.

Results

Descriptive and correlation analyses

Mean and standard deviation values of the studied variables are presented in Table 1.

Correlation analysis (Table 1) showed that self-compassion associated positively (with a magnitude between weak and moderate) with compassion received from others, feelings of social safeness and body appreciation, and negatively (with weak magnitude) with disordered eating and age. Compassion received from others associated positively and strongly with feelings of social safeness, and positively and moderately with body appreciation. This scale also showed negative and weak associations with disordered eating and age. Feelings of social safeness showed a positive and moderate association with body appreciation, and negative and weak ones with disordered eating and age. Body appreciation presented a strong negative association with disordered eating, and negative and weak associations with age and zBMI. Finally, age and zBMI correlated negatively and weakly.

Considering that age and zBMI presented significant correlations with variables in the study, its effects were controlled for in path analysis.

Path analysis

A fully saturated path analysis model (35 parameters) was first tested, accounting for 44%, 32%, and 45% of the variances of social safeness, body appreciation, and disordered eating, respectively. Some of its paths were not significant and were subsequently removed, and the overall model was readjusted. These paths were: the direct effect of social safeness on disordered eating (bSSPS = 0.00; SEb = 0.01; Z = − 0.23; p = 0.821), the direct effect of age on social safeness (bage = − 0.11; SEb = 0.32; Z = − 0.36; p = 0.723), the direct effect of self-compassion on disordered eating (bCEAS_SelfComp = 0.00; SEb = 0.01; Z = − 0.49; p = 0.627), the direct effect of compassion received from others on disordered eating (bCEAS_CompFromOthers = 0.00; SEb = 0.00; Z = 0.55; p = 0.582), the direct effect of zBMI on social safeness (bzBMI = 0.54; SEb = 0.45; Z = 1.21; p = 0.227), the direct effect of compassion received from others on body appreciation (bCEAS_CompFromOthers = 0.01; SEb = 0.00; Z = 1.45; p = 0.146), and the direct effect of age on disordered eating (bAge = − 0.09; SEb = 0.45; Z = 1.21; p = 0.227).

The readjusted and final path model is presented in Fig. 1. This model accounted for 43% of the variance of social safeness, 32% of the variance of body appreciation, and 44% of the variance of disordered eating. It presented a good fit to empirical data: CMIN/df = 1.119, χ2(7) = 7.836, p = 0.543 [43]; NFI = 0.978, CFI = 0.998, TLI = 0.993 [43, 44]; RMSEA = 0.024, p = 0.654, 95% CI 0.000–0.092 [44]. This final path model (with all paths significant at a significance level of 95%) allowed to notice that self-compassion presented positive direct effects of 0.12 on social safeness (bCEAS_SelfComp = 0.07; SEb = 0.03; Z = 2.23; p < 0.050), and of 0.20 on body appreciation (bCEAS_SelfComp = 0.01; SEb = 0.00; Z = 3.20; p < 0.010). It also presented indirect effects on body appreciation through social safeness (β = 0.05; 95% CI − 0.002 to 0.102), and on disordered eating through both social safeness and body appreciation (\(\beta\) = − 0.14; 95% CI − 0.225 to − 0.063). Compassion received from others presented a positive direct effect on social safeness of 0.61 (bCEAS_CompFromOthers = 0.30; SEb = 0.03; Z = 11.00; p < 0.001), and indirect effects on body appreciation through social safeness (β = 0.22; 95% CI 0.135–0.319) and on disordered eating through social safeness and body appreciation (β = − 0.13; 95% CI − 0.197 to − 0.077). Feelings of social safeness were shown to directly and positively impact on body appreciation (β = 0.36; bSSPS = 0.03; SEb = 0.01; Z = 5.89; p < 0.001), and indirectly and negatively on disordered eating through body appreciation (\(\beta\) = − 0.21; 95% CI − 0.300 to − 0.131). Body appreciation impacted directly on disordered eating (β= − 0.59; bBAS = − 0.97; SEb = 0.09; Z = − 11.06; p < 0.001). Effects of variables to be controlled for were also examined. Both age and zBMI associated directly with body appreciation (age: β = − 0.20; bAge = − 0.10; SEb = 0.03; Z = − 3.28; p < 0.010; and zBMI: β = − 0.15; bzBMI = − 0.12; SEb = 0.05; Z = − 2.58; p < 0.050), and zBMI associated directly with disordered eating (β = 0.23; bzBMI = 0.30; SEb = 0.07; Z = 4.41; p < 0.001. Age and zBMI also revealed an indirect effect on disordered eating via body appreciation (age: β = 0.12; 95% CI 0.034–0.20; and zBMI: \(\beta\)= 0.09; 95% CI 0.024–0.160).

Fig. 1
figure 1

Final path model. Standardized path coefficients among variables are presented. All path coefficients are significant at the 0.050 level. *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

Discussion

The present study aimed at proposing a model hypothesizing that being self-compassionate and receiving compassion from others associate social safeness, which associates with higher body appreciation and, in turn, with a weaker manifestation of disordered eating symptoms. This study may provide important insights on protective factors against disordered eating in adolescent girls.

Correlations between variables were explored. Similarly to results using adult samples [33], self-compassion and receiving others’ compassion associated positively with social safeness. To the extent of our knowledge, this finding was not yet reported in adolescents. One possible explanation is the link between compassion and the activation of the emotion regulation system associated to feelings of calmness, affiliation, and connectedness, that is, feelings of social safeness and pleasure [31, 32]. Self-compassion and receiving others’ compassion also associated positively with body appreciation. This correlation of self-compassion was already found in samples of adults [25, 26]. The finding that both self-compassion and receiving others’ compassion associate with body appreciation seems to be novel. This finding seems to underline that when adolescent girls have the competences to self-direct kindness and understanding, and perceive the same attitudes from close others, they seem more able to relate with their own bodies with kindness, acceptance, and protection. Moreover, compassion variables in this study associated negatively with disordered eating, an expected finding based on other studies’ findings [27, 28]. Social safeness and body appreciation correlated positively, and body appreciation correlated negatively with disordered eating, correlations already found in adult samples [25].

The model explored via path analysis explained 43%, 32% and 44% of the variances of social safeness, body appreciation and disordered eating symptoms. First, the model proposed a direct relationship between self-compassion competences and both social safeness and body appreciation. This finding is supported by previous studies in adults [25, 26] and suggests that, in female adolescents, self-compassion competences associate positively with an experience of the social world as safe and reassuring and an attitude of body appreciation. Self-compassion also presented an indirect association with body appreciation, mediated by social safeness, but with less noteworthy strength (β = 0.05) Regarding compassion received from others, it only associated directly with social safeness, an expected result based on Kelly and colleagues’ findings [45]. These authors reported that social safeness is strongly predicted by the amount of support one receives. Moreover, the association of compassion received from others with body appreciation was only indirect, but with significant strength (β = 0.22).

In sum, self-compassion competences and perceived compassionate competences from others presented an indirect association with disordered eating symptoms. This indirect association seems to be mediated by social safeness and by body appreciation. When accounting for age and zBMI’s effects, self-compassion and receiving compassion from others associated with higher social safeness, which associated with higher body appreciation, and in turn with a weaker manifestation of disordered eating symptoms. These results may mean that when adolescent girls perceive that others treat them in a kind and understanding manner in times of suffering, they may tend to experience positive feelings of safeness and pleasure in their social contexts. This may constitute a proper environment for them to relate with their body characteristics and idiosyncrasies in an understanding, accepting and kind manner. This finding is in line with the main finding of the present study: both competences for self-compassion and compassion received from others associated with a weaker display of disordered eating via higher feelings of social safeness and higher body appreciation.

One possible explanation for present results lies on Gilbert’s [29] input on social safeness, i.e., the lack of social safeness associates with the perception of a lower social rank position and thus with feelings of inferiority and shame. These are thought to provoke defensive and concealment strategies which make the individual more vulnerable to psychopathological symptoms [29]. Taking this into account, it may be possible that adolescent girls who have self-soothing competences and who perceive the same competences in others (directed to themselves) may feel safe in their social world. These feelings of social safeness may thus protect them from the body image dimension putting them at risk in their social rank (e.g., having body features that are not in line with social beauty standards does not diminish their perceived self-worth). Specifically, present findings may imply that social safeness in adolescents associates with a state of peace with their body image that may lower the probability of body-concealing behaviours, in which disordered eating is included.

Strengths and limits

Strengths of the current study include the use of a diverse sample of Portuguese girls, and the use of counterbalanced validated psychological measures. Nonetheless, the present study used a cross-sectional design, which prevents causal conclusions. Prospective studies are thus necessary to explore relations of causality. Another limitation is that disordered eating is a complex phenomenon; therefore, upcoming studies should explore the roles of other contributing variables. Finally, the use of self-report measures may imply bias, and it would be preferable to use assessment interviews.

What is already known on this subject?

Compassion-related and interpersonal factors are associated with a secure relationship with one’s body image. Positive body image associates with healthy eating attitudes and behaviours.

What does this study add?

In conclusion, the present study seems to underline the importance of establishing warm and caring relationships with oneself and others, and to provide novel information on female adolescents’ health and well-being. Although disordered eating is a multicausal phenomenon, results suggested the contribution of self-compassion and compassion received from others for the construction of a safe social world, as well as of a more peaceful relationship with body image in female teenagers, with a potential protective effect on disordered eating.

Taking into account that compassion is a skill that is susceptible to cultivation (i.e., practicing compassionate attention, thinking, feeling, and behavior in relation to one’s and others’ suffering; [46], and that it is believed that interventions to promote adolescents’ mental health should reinforce protective factors and heighten alternatives to risky behaviours [10], present findings may point to the development of programs for female adolescents (delivered in digital media, schools, or the community) that include the cultivation of compassionate and affiliative skills to promote positive body image and influence the manifestation of disordered eating. Regarding the clinical setting, these findings suggest a potential means to promote positive body image. This set of attitudes and behaviours is considered important for going beyond the elimination of negative symptoms and invest in relapse-prevention and maintenance of therapeutic gains in patients with eating disorders [47].