Self-compassion as a way to happiness

Self-compassion can be defined as the individuals’ ability to “being touched by and open to one’s own suffering, not avoiding or disconnecting from it, generating the desire to alleviate one’s suffering and to heal oneself with kindness. Self-compassion also involves offering nonjudgmental understanding to one’s pain, inadequacies and failures, so that one’s experience is seen as part of the larger human experience” (^{Neff, 2003}, p. 87). This concept emerges as an individual’s compromise to prevent or relieve suffer (^{Marques et al., 2018}). According to ^{Neff (2003}), self-compassion entails three basic components: self-kindness, common humanity, and mindfulness. Self-kindness refers to the individuals’ ability to extend kindness and understanding to oneself, avoiding harsh judgment and self-criticism; common humanity refers to individuals’ ability to seeing one’s experiences as part of the human experience avoiding feelings of failure; finally, mindfulness refers to individuals’ ability to hold one’s painful thoughts and feelings in balanced awareness and avoid over-identification.

Paying attention to the present moment has been seen as a source of happiness since it tends to promote intra and interpersonal well-being (^{Marques et al., 2018}). For instance, ^{Neff et al. (2007}) found that self-compassion is positively associated with happiness perception, optimism and positive affect, protecting individuals from anxiety and depressive symptoms. Also, ^{Raes (2011}) refers that self-compassion is an important protective factor against emotional problems since higher levels of self-compassion were associated with less depressive symptoms.

According to ^{Ryff (1989}), happiness results from the balance between several domains, namely self-acceptance, personal growth, purpose in life, positive relations with others, environmental mastery and autonomy. ^{Fordyce (1983}) beliefs that happiness is linked to a long-term general sensation of emotional well-being and life satisfaction and can be increased through the adoption of positive thinking and awareness of the present moment. This would contribute to avoid rumination and promote acceptance.

For all of these reasons, interventions targeting relaxation would be effective in reducing anxiety and stress by promoting positive emotions and the development of skills for dealing with physical and mental problems (^{Barnard & Curry, 2011}; ^{Marques et al., 2018}). Moreover, self-compassion, meditation and harmonious breathing should be trained to promote positive health behaviors (^{Sirois et al., 2014}).

Autogenic training

The autogenic training (AT) or “being yourself” (auto means self and genic means being) was originally developed by Johannes Schultz and was later refined with Wolfgang Luthe (^{Naylor & Marshall, 2007}). It emerged as a result of the findings related to the therapeutic hypnosis and some authors suggest that the AT can induce a meditative state (^{Yurdakul et al., 2009}). AT is seen as a way of intervention in the individuals’ development by focusing on primary prevention and skills, appearing as one of the main psychological treatment in group to improve individuals’ self-regulation skills (^{Krampen, 1996}). Thus, it is recognized by the American Psychological Association as an effective mind-body therapy (^{Atkins & Hayes, 2019}).

This technique promotes the process of self-knowledge with individuals adopting an active participation and responsibilization in their treatment (^{Carruthers, 1979}; ^{Houghton, 1996}; ^{Kanji & Ernst, 2000}). This means that the AT beliefs in individuals’ internal resources for own self-healing and allows to be aware of individuals internal manifestations (^{Kanji, 2000}).

The six-based exercises proposed by ^{Schultz (1967}) are still the core of the autogenic training and focuses attention on different parts of the body by using autosuggestion to repeat internal phrases that promote heaviness sensations, regulate cardiac activity, and control breathing. Relaxation is achieved using passive concentration that allows to contact mentally with the different parts of the body (^{Kanji, 2000}; ^{Kanji & Ernst, 2000}).

This allows achieving three important aspects of relaxation: reduce muscle tension in the body, change the style or the way people think, and change the content of what they are thinking (^{Naylor & Marshall, 2007}). Its regular practice allows adopting a new and well-succeed way of living and thinking (^{Kanji, 2000}). Some benefits of the AT include achieve peace of mind, develop skills to deal with life challenges, emotional balance, improved health and well-being, development of coping skills and control of body responses to stress; in short-term, it helps to neutralize acute stress and fatigue, as well as to promote mental and physical rest (^{Kanji, 2000}).

Guided imagery

Relaxation can benefit from include exercises focused on mental images that through imagery contribute to reduce body response to stress. Imagery is considered be therapeutic in many cultures, since it provides a deep journey through the inner life that will lead to self-knowledge - it is the mind thinking about our inner reality through images (^{Kestenberg et al., 2014}). ^{Hernández et al. (2014}) consider the breathing, the Schulz’s autogenic training (relaxation through mental control) and imagery (through mental sensations) are effective techniques in reducing anxiety.

According to Lynn and Rhue (1988) as cited in ^{Özü (2010}), the main aim of the guided imagery is to calm individual’s mind, promote individual’s skills to face negative emotions, and to improve individual’s sense of consciousness by using five senses. This will allow to teach individuals to induce inner power and eliminate negative emotions. These techniques are usually used in the initial phase of guided imagery since they facilitate the mind to be open to experiences, consciousness, and facilitate individuals’ processing skills of positive images. As highlighted by Morrison and Cometa (1977) as cited in ^{Özü (2010}), guided imagery can support other therapeutic techniques that aim to work individuals’ skills in different domains.

^{Apóstolo and Kolcaba (2009}) using a guided imagery intervention with psychiatric inpatients with depressive disorders found that it contributed to increase patients’ comfort and decrease patients’ depression, anxiety, and stress over time. This guided imagery intervention included progressive muscle relaxation, imagine relaxing natural scene, imagine meetings with significant figures, and create positive, comforting, and serene images of the hospital; all these stimuli would contribute to easiness, safety, positive images, and elimination of disturbing thoughts. Also, ^{Tugade and Fredrickson (2007}) suggest that guided mediation practices aiming at inducing positive experiences can contribute to prolong positive emotional experiences with benefits for physical and mental health.

Positive Psychology

Positive Psychology (PP) was defined by ^{Seligman (1999}) as a science that emphasizes the understanding and promotion of individuals’ positive qualities such as optimism, courage, ethics, meaning in life, interpersonal skills and social responsibility. It examines the conditions that allow the optimal functioning of individuals allowing to improving quality of life and preventing psychopathology. It aims to develop individuals’ skills and not to correct their weaknesses (^{Seligman & Csikszentmihalyi, 2000}).

This positive approach understands better the psychopathology, since it removes the focus from the symptoms treatment and directs their attention to the promotion of well-being (^{Marques et al., 2018}). It aims to make life happier and more significant (^{Seligman, 2004}). Thus, interventions based on this approach aim to promote positive psychological states such as gratitude, optimism, and mindfulness through systematic exercises (^{Nikrahan et al., 2016}). These exercises represent a complementary strategy to promote mental health being used as a self-help strategy (^{Bolier et al., 2013}). Relaxation is recognized as a way to promote positive emotions and experiences of calmness allowing the development of resources to improve health and achieve well-being (^{Estrela-Dias & Pais-Ribeiro, 2014}).

Music and its therapeutic purpose

Music is part of the relaxation interventions that aim to reduce stress, anxiety, depression and psychological tension (^{Klainin-Yobas et al., 2015}; ^{Lufiego et al., 2017}). This is because the human body has an innate predisposition to answer internally to sounds and rhythms from the external environment (^{Nunes-Silva et al., 2016}). Since the beginnings of humanity music is used for healing purposes but musicotherapy only emerged in the 20^th century, as a science that aims to understand the interaction between human being, music and their sound elements, as a way to promote quality of life (^{Nunes-Silva et al., 2012}).

Studies have shown that music is likely to reduce anxiety (^{Karagozoglu et al., 2013}) and stress, leading to happiness, tranquility and well-being. Thus, some types of music seem to keep mental health, having an important role in the emotional development of individuals (^{Nunes-Silva et al., 2012}). It facilitates the release of endorphins and cortisol reduction, promoting well-being and inducing relaxation (^{Chanda, & Levitin, 2013}; ^{Khalfa et al., 2003}; ^{Koelsch et al., 2011}). Thus, music is seen as a powerful tool to evoke emotions, especially positive ones.

Design

This randomized controlled trial followed CONSORT guidelines (2010). A parallel-group design with an equal randomization made by an independent observer by hand by ordering participants’ name and allocating them sequentially was conducted (allocation ratio of 1:1 patient for three groups: group 1 - intervention group, group 2 - active control group, and group 3 - passive control group). The allocation was performed twice because we recruited participants in two phases (one between 2019 May and July and the other between 2019 October and December). This was a no-blinded study (both participants and interventionists were not blinded to group assignment). The sample size was calculated based on the formula: n=[(Zα/2+Zβ)²×{2(ó)²}]/(μ1-μ2)² (^{Sakpal, 2010}). The stress subscale of DASS-21 was used considering the means and standard deviations of a similar intervention with the same population (^{Parcover et al., 2018}). A sample size of 123 participants, 41 in each arm, was sufficient to detect a clinically important difference of 4,3 points assuming a standard deviation of 7 using a two-tailed t-test of difference between means with 80% power and a 5% level of significance.

Participants were assessed in three moments: baseline, at the end of the intervention (6 weeks later), and at the follow-up (4 weeks later).

Participants

One hundred and sixty-one students met the eligibility criteria and agreed to participate. Participants’ age ranged from 17 to 46 years (M=20.96 years, SD=3.66). Participants were assigned to the experimental group (n=55), the active control group (n=51), or the passive control group (n=55). Participation followed the Declaration of Helsinki.

Participants were recruited during classes between May 2018 and December 2019. Inclusion criteria were: being a university student; agreeing to participate and complete the three assessment moments.

Interventions

Experimental group. Participants in the experimental group participated in 6 weeklies, 20 min. sessions of a relaxation protocol based on the PP approach (music was not included in this protocol). This number of sessions was chosen based on previous studies (e.g., ^{Atkins & Hayes, 2019}).

The relaxation protocol focuses on two essential aspects namely guided imagery and diaphragmatic breathing. Goals were to promote the experience of more positive emotions and reduce the prevalence of depression, anxiety, and stress. The relaxation protocol emerges as a facilitator for the management of thoughts and negative emotions while positive psychology emerges as a facilitator of positive emotions and, consequently, psychological wellbeing.

The intervention consisted of three phases: phase 1 - Beginning of Relaxation, phase 2 - Body Awareness, and phase 3 - Guided imagery.

In phase 1 - Beginning of Relaxation, participants were asked to sit comfortably in the chair with their palms facing down. Then, were invited to close their eyes gently to decrease attention to external stimuli. Here participants were asked to focus on diaphragmatic breathing (i.e., a breathing that accentuates movement in the lower part of the belly, requires less effort and is predominant in states of rest, relaxation and during sleep; ^{Elias, 2009}) and guided imagery, as essentials issues for leading to relaxation. Examples of specific instructions given to participants during the sessions: “Inhale deeply through your nose, filling your abdomen and feeling the weight of the air. Visualize air coming through your nostrils and being transported to your lungs, promoting relaxation, well-being, health, energy and vitality to every cell in your system”.

In phase 2 - Body Awareness - six established exercises or techniques proposed by ^{Schultz (1967}) were applied: (1) induce heaviness by using verbal cues that suggest heaviness in the body; (2) induce warmth by using verbal cues that call attention to warmth; (3) reduce heartbeat by using verbal clues that call attention to the heartbeat; (4) breathing practice by using verbal cues focus on breath; (5) abdominal practice by using verbal cues focus on abdominal sensations; and (6) head practice by using verbal cues on the coolness of the forehead. Examples of specific instructions given to participants during the sessions: “Repeat 5 times inwardly to yourself: “My arms are pleasantly heavy and relaxed”, “My heart beats peacefully”, “My breathing is calm and balanced”, “I am completely calm and relaxed”.

In phase 3 - Guided Imagery - the objective was to promote the development of personal resources that could be used as coping strategies to improve health and individuals’ well-being, by focusing on positive emotions. Examples of specific instructions given to participants during the sessions: “Now that you feel totally calm, relaxed and safe as you breathe in, imagine the air entering your system. By bringing more oxygen, energy, and vitality to each of your cells, your organs, and systems, you also bring love, peace, well-being, joy, harmony, strength, health, and healing. As you exhale, imagine that you expel everything that imprisons you, tires you and wears you out, what you no longer want to keep in your life. Feel untied, freer, happier.”

Finally, a last phase was included to leave the relaxation state. Participants were guided to experience feelings of gratitude toward themselves for engaging in self-care; then, they were asked to return from the relaxation state, awakening gently the five senses.

To ensure intervention fidelity, an explicit treatment protocol (with the autogenic training and instruction to diaphragmatic breathing and guided imagery) was developed, and the intervention was carefully supervised by senior researchers, one of them with training in the therapeutic model.

Active control group. Participants assigned to the active control group participated in 6 weekly, 20 min. sessions of relaxing music. In the beginning of each session participants were asked to close their eyes and to do abdominal breathing so participants could make the most of that time. The music was chosen based on some characteristics that are essential to boost a relaxation state, namely: slow tempo and slow-moving melody, harmonious combination of instruments and sounds of nature (^{Khalfa et al., 2003}; ^{Nunes-Silva et al., 2012}, 2016).

Passive control group. Participants assigned to the passive control group did not participate in any intervention, but they had the opportunity to receive the relaxation protocol training after completion of the study (i.e., in the beginning of the following school semester).

Measures

Sociodemographic questionnaire. Data on sex, age, scientific area of study and curricular year were collected. Participants were asked about the importance given to mental health in a seven points Likert-type scale. Participants from the intervention group, the active control group and the passive control group filled out a questionnaire at three times: baseline, 6 weeks post-intervention and 4 weeks follow-up.

Depression, anxiety and stress. The Depression, Anxiety, and Stress Scale (DASS-21; ^{Lovibond & Lovibond, 1995}; Portuguese version: ^{Pais-Ribeiro et al., 2004}) was used to measure depression, anxiety, and stress. The DASS-21 is composed by 21 items divided into three subscales: depression (7 items), anxiety (7 items), and stress (7 items). Items are rated on a Likert-type scale from 0 (did not apply to me at all) to 3 (applied to me very much or most of the time). Depression, anxiety, and stress scores are calculated using the sum of the items, with higher scores indicating higher depression, anxiety, and stress symptoms. Cronbach’s α in our study was .84 for depression, .81 for anxiety, and .88 for stress.

Self-compassion. The Self-Compassion Scale (SCS; ^{Neff, 2003}; Portuguese version: ^{Castilho & Gouveia, 2011}) was used to measure self-compassion and its three basic components: self-kindness, common humanity, and mindfulness. The SCS is composed by 26 items and comprises six subscales: self-kindness (5 items; item example: “I try to be loving towards myself when I’m feeling emotional pain”), self-judgment (5 items; item example: “When times are really difficult, I tend to be tough on myself” - reversed), common humanity (4 items; item example: “When things are going badly for me, I see the difficulties as part of life that everyone goes through”), isolation (4 items; item example: “When I’m feeling down, I tend to feel like most other people are probably happier than I am” - reversed), mindfulness (4 items; item example “When something upsets me I try to keep my emotions in balance”), and over-identification (4 items; item example “When I’m feeling down I tend to obsess and fixate on everything that’s wrong” - reversed). Items are rated on a Likert-type scale from 1 (almost never) to 5 (almost always). Subscale scores result from the mean of subscale items. Cronbach’s α in our study was .93.

Subjective happiness. The Subjective Happiness Scale (SHS; ^{Lyubomirsky & Lepper, 1999}; Portuguese version: ^{Pais-Ribeiro, 2012}) was used to measure subjective happiness. It is composed by four items, two ask participants to characterize themselves using rating relatives to peers, and two offer participants descriptions of happy and unhappy individuals asking them to rate the extent to which each item describes them. Cronbach’s α in our study was .79.

Data analysis

Data were analyzed using SPSS version 26. Means and standard deviations were used to characterize the sample as well as to examine the study variables for the total sample and for each group (intervention, active control group and passive control group). α was set at 0.05 (two-tailed).

To examine the efficacy of the intervention group and compare differences among groups over time (first aim), Multilevel Modeling (MLM) was used. MLM is a useful method to examine data with repeated measurements and organized in nested levels. Two-level multilevel models were used to examine the effect of the relaxation protocol on depression, anxiety, and stress: measurement occasions (level 1) nested within individuals (level 2).

To examine the relationship between study variables (second aim), Pearson correlations and regression analyses were used. To examine the moderating role of self-compassion in the association between subjective happiness and depression, anxiety, and stress, the PROCESS macro (Model 1 - simple moderation) (^{Hayes, 2018}) was used. In this model, we included subjective happiness as independent variable, depression, anxiety, and stress as dependent variables, and self-compassion as moderator.

Recruitment

Participants were assessed for eligibility during May 2018 and December 2019. A total of 161 participants consented to participate and were randomized with 55 participants being allocated to the intervention group, 51 to the active control group and 55 to the passive control group. During the study, eight participants from the intervention group, ten from the active control group and four from the passive control group withdrew from the study (e.g., lack of time and difficulty in combining the AT with school activities). The final number of participants included in the analysis was 139; 47 in the experimental group, 41 in the active control group and 51 in the passive control group (see Figure 1). The overall inclusion rate was 85% for the intervention group; 80% for the active control group and 93% for the passive control group. Figure 1 shows the flow diagram that is available for reporting an RCT (^{Moher et al., 2012}).

Note. Adapted from the Consolidated Standards of Reporting Trials (2010).

Figure 1 Flow diagram of the progress 

Descriptive results

Baseline demographic characteristics for each group are presented in Table 1. Most of the participants were women (90.6%), with a mean age of 20.96 (SD=3.65; range 17-46). In terms of curricular year, 12.9% (n=18) were in the 1^st year, 8.6% (n=12) were in the 2^nd year, 63.3% (n=88) were in the 3^rd year, 14.4% (n=20) were in the 4^th year, and 0.7% (n=1) were in the 5^th year. Most of the participants were attending a degree in human and social sciences (80.6%; n=112), while the remaining were attending a degree in life and environment sciences (19.4%; n=27). Means and standard deviations for each dependent variable follows:

Table 1 Sample characteristics 

Note. ^aA higher score indicates more importance attributed to mental health (range 1-7); ^bHigher scores indicate higher levels of depression, anxiety, and stress (range 0-3), higher levels of self-compassion (range 1-5) and higher levels of perceived happiness (range 1-7).

Depression. Paired-sample t tests indicated significant differences between depression at baseline and depression after intervention [t(138)=3.12, p<.01] and between depression at baseline and depression at follow-up [t(138)=2.33, p<.05]. Participants presented higher depression at baseline (M=.55; SD=.51) in comparison to after intervention (M=.43; SD=.47) and to follow-up (M=.46; SD=.46).

Anxiety. Paired-sample t tests indicated no significant differences between anxiety at baseline and anxiety after intervention [t(138)=1.46, p=.146] neither between anxiety at baseline and anxiety at follow-up [t(138)=1.53, p=.127].

Stress. Paired-sample t tests indicated significant differences between stress at baseline and stress after intervention [t(138)=2.19, p<.05] and between stress at baseline and stress at follow-up [t(138)=2.15, p<.05]. Participants presented higher stress at baseline (M=1.05; SD=.67) in comparison to after intervention (M=.93; SD=.64) and to follow-up (M=.93; SD=.63).

Levels of self-compassion (M=3.12; SD=.63) and levels of subjective happiness (M=5.03; SD=.98) were below the mean - considering the mean of the Likert scale. The importance given to mental health was high (M=6.81; SD=.46).

Depression

Base model. In the first step a base model was estimated with no predictors. In this model, depression was estimated by its mean (M=.48), a specific term representing within-subjects variability (=.13) and an error term associated with each observation (=.10). Intra-class coefficient was .551 which justifies the exploration of predictors of change in individuals’ depression.

Unconditional model. This model adds the time as predictor of changes in individuals’ depression. Results revealed a significant change in depression over time (-.05, p<.05) (see Table 2). Depression reduced 0.05 units and their global mean increased 0.57. The analyses over time reduced the estimate of the error variance of 9.7% (from .103 to .093) showing that the time explained some of the observed variance in the depression. As this variance is statistically different from 0 (p<.001), other predictors should be examined.

Table 2 Results from MLM analyses regarding Psychological Morbidity 

Note. Condition 1 - intervention group; Condition 2 - active control group; Passive control group was used as reference.

Conditional model. This model adds the condition (i.e., intervention group, active control group or passive control group) as predictor of change in individuals’ depression. When the condition was added as predictor, the time was no longer a significant predictor of changes in individuals (=-.04; p=.25) and the global mean of depression reduced to .53. None of the conditions showed a significant interaction with time (p=.58 e p=.98). It is important to note that the variance error did not change (.093).

Anxiety

Base model. In the first step a base model was estimated with no predictors. In this model, anxiety was estimated by its mean (M=.48), a specific term representing within-subjects variability (=.15) and an error term associated with each observation (=.10). Intra-class coefficient was .594 which justifies the exploration of predictors of change in individuals’ anxiety.

Unconditional model. This model adds the time as predictor of changes in individuals’ anxiety. Results revealed a non-significant change in anxiety over time (-.03, p>.05) (see Table 2). The analyses over time reduced the estimate of the error variance of 32% (from .103 to .070) showing that the time explained some of the observed variance in the anxiety. As this variance is statistically different from 0 (p<.001), other predictors should be examined.

Conditional model. This model adds the condition (i.e., intervention group, active control group or passive control group) as predictor of change in individuals’ anxiety. When the condition was added as predictor, the global mean of anxiety reduced to .45. None of the conditions showed a significant interaction with time (p=.50 e p=.28) and only one significant interaction between the intervention group and the passive control group was found (=.27; p=.05) (see Table 2). This means that anxiety increased 0.27 units in the intervention group when compared to the passive control group. It is important to note that the variance error did not change (.070).

Stress

Base model. In the first step a base model was estimated with no predictors. In this model, stress was estimated by its mean (M=.97), a specific term representing within-subjects variability (=.14) and an error term associated with each observation (=.14). Intra-class coefficient was .649 which justifies the exploration of predictors of change in individuals’ stress.

Unconditional model. This model adds the time as predictor of changes in individuals’ stress. Results revealed a significant change in depression over time (-.06, p<.05) (see Table 2). Stress reduced 0.06 units and their global mean increased 1.09. The analyses over time reduced the estimate of the error variance of 50% (from .144 to .072) showing that the time explained some of the observed variance in the stress. As this variance is statistically different from 0 (p<.001), other predictors should be examined.

Conditional model. This model adds the condition (i.e., intervention group, active control group or passive control group) as predictor of change in individuals’ stress. When the condition was added as predictor, the time was no longer a significant predictor of changes in individuals (=-.02; p=.62) and the global mean of stress reduced to .87. None of the conditions showed a significant interaction with time (p=.21 e p=.66) and only one significant interaction between the intervention group and the passive control group was found (=.41; p=.01) (see Table 2). This means that stress increased 0.41 units in the intervention group when compared to the passive control group. It is important to note that the variance error did not change (.072).

Limitations and future directions

It is important to acknowledge the limitations of the present study. First, the protocol employed was brief since it included only six sessions. While some studies have shown that very brief interventions can be effective (e.g., see ^{Kanji & Ernst, 2000} for more details), other authors suggest that at least eight to 12 sessions are needed to allow participants to learn the technique and lead to positive benefits (e.g., ^{Linden, 1994}). It is possible that for these specific participants and context, this number of sessions can be insufficient to produce benefits - but this is only one possible explanation. For this reason, future studies with university students should test the efficacy of longer interventions. Also, in order to be effective, it is recommended that participants train the technique on a regular and daily basis (^{Abuín, 2016}; ^{Krampen, 1996}). In this study, no control was made regarding participants’ training of the technique outside the session (e.g., at home). Additionally, the autogenic training was offered in a group format. What may makes it difficult to adequate the intervention to the individuals’ needs and characteristics.

In terms of methodology, the efficacy of the study was assessed only using self-report measures, which can lead to participants’ biases. Future studies should assess its efficacy also using a qualitative approach (e.g., interviews with participants to explore perceived benefits or (un)helpful aspects of the therapy). Additionally, other variables should be included; while this protocol seems ineffective in improving depression, anxiety, and stress, it could contribute to improve other areas of psychological functioning in university students (e.g., emotion regulation, resilience, among others). For this reason, future studies should explore how this protocol impacts other variables.

Finally, most of the participants were women which limits conclusions with the minimum required of 41 participants in one arm of the RCT. Therefore, future studies should include larger and more heterogeneous samples (including more men) in each arm and, eventually, explore the moderating role of sex in the efficacy of relaxation training men and women may differ in their reaction to the relaxation protocol.