Consensual qualitative research on free associations for compassion and self-compassion

Research goals

To the best of our knowledge, only in the two studies mentioned above (Pauley & McPherson, 2010; van der Cingel, 2011) has a qualitative analysis been used to explore compassion and self-compassion. Consequently, very little is known about the words people associate with these constructs, how the terms are conceptualised, and how they are interpreted. For that reason, our goal was to identify and categorise associations linked to the words compassion and self-compassion.

Given the ongoing debate over what compassion is, we were also interested in whether our participants would conceptualise these words more in terms of emotions, cognitions, or behaviours.

Sample

The participants were psychology students at Comenius University in Bratislava. They were selected on the assumption they have better skills of mentalisation, greater experience of psychological constructs and in their future work they will require an understanding of subjective human experiences. Of the 151 participants, 37 were men. The mean age was 22.2 years (SD = 4.4).

Testing materials

The participants were asked two open-ended questions “…without censoring, please write down the first three associations when you see the word compassion” and “…without censoring, please write down the first three associations when you see the word selfcompassion”.

Procedure

The data was collected as a part of a larger questionnaire battery on self-compassion and self-criticism. The questions on the free associations came at the beginning of the questionnaires immediately after the informed consent form and the socio-demographic information.

Data analysis

Emotional aspect

The most frequently represented domain for compassion and for self-compassion was Emotional Aspect, which is most fleshed out in terms of subdomains, categories, and characteristics for both stimulus words. The emotional states participants in our study associated with the words included both positive and negative feelings. This may contribute to the mixed nature of compassion, which is a mixture of negative emotions that are the result of suffering and positive feelings associated with the motivation to help a sufferer. The negative emotions to do with suffering associated with compassion were sadness and remorse, and to a lesser extent pain. Love and happiness were also mentioned by the participants and were included in the group of emotions with positive valence.

The emotional aspect of compassion and self-compassion was the most frequent and the richest domain, suggesting that people see compassion and self-compassion mainly as emotions. This domain consists of two subdomains—“Empathy” and “Emotions”. Participants connected compassion with the affective component in particular: either with specific emotions or with empathy—resonance with the emotions of others. Klimecki, Leiberg, Lamm, and Singer (2012) discuss the difference between empathy and compassion from a neuropsychological point of view. In their experiment they found greater negative affect and the activation of the brain areas associated with resonance with other peoples’ pain among participants who watched a video of suffering following empathy training. However, when the same video was watched after compassion training was conducted participants reported an increased positive affect and various brain areas were activated. They concluded that these socio-affective states are related, but also have significant differences that are indicative of their functions. The “Emotions” subdomain was made up of two categories: 1) “Synonyms for Emotions” and 2) “Specific Emotions”. The first category contained associations relating to the general concept of emotion, like “feeling” or “sentiment”. The second category was more specific. The “Specific Emotions” category includes specific emotions related to compassion. Negative emotions like “sadness”, or “remorse” were dominant. In their study Goetz et al. (2010) refer to compassion as a distinct emotion which differs from sadness, distress, love, and other emotions. A compassionate facial expression used in a study by Keltner and Haidt (1999) was most frequently identified as sadness. Quite a lot of positive characteristics were found in our data for compassion, like “love” and “happiness”, which partially supports the conclusion of Sprecher and Fehr (2005) that compassion could be a special kind of love. But the frequency of these associations was comparable with that of the negative characteristics, which could be interpreted as being a result of the fact that compassion is a mixed emotion. Compassion itself combines negative emotional aspects evoked by both situations of suffering and by emotions expressed by sufferers with whom we resonate emotionally. At the same time, compassion produces positive feelings brought about by helping behaviours. This positive feeling is our natural reward–physiological feedback in real time, a reaction to our behaviour towards a person in need of help (Ekman, 2014). The mixed nature of self-compassion can be seen in the mix of negative emotions brought about by suffering and the positive feelings associated with the relief of the suffering. Self-compassion was related to remorse, sadness, and unhappiness. On the other hand, self-compassion was connected to love and calmness. Interestingly, more emotional characteristics were found for self-compassion than for compassion. This could suggest that people are more emotionally involved with the self than with others. Not surprisingly, for both self-compassion and compassion, there were more free associations connected with negative emotions than with positive emotions. As compassion and self-compassion are felt in distressful situations or even situations involving suffering, it is no wonder that people emotionally resonate with the discomfort experienced by the people involved–either their own or that of others.

Behavioural aspect

The second frequency domain for compassion and self-compassion was the identical Behavioural Aspect domain. Behavioural Aspect has three main subdomains when related to compassion and two subdomains when related to self-compassion, and there is a difference only in one subdomain. The subdomains for which compassion and self-compassion are common are “display of help” and “display of favour”. In compassion, there is an extra category, “display of motivation”, which is relevant to Gilbert’s theory (Gilbert et al., 2017). Not surprisingly, display of motivation does not appear in relation to self-compassion because people are naturally driven to help themselves but need motivation to act compassionately towards others.

Biological aspect

The third most frequent domain for compassion was Biological Aspect, and for selfcompassion it was Evaluative Aspect. The biological aspect of compassion includes associations relating to living things. In addition to people, participants mentioned animals and in one case plants. In associations linked to human beings, relationships such as “friend” or “child” were mentioned, which is compatible with the evolutionary view that compassion has evolved to protect the weak (e.g., Goetz, Keltner, & Simon-Thomas, 2010). As the target of self-compassion is the self, there is no need to specify who it is associated with, but there is a need to describe the situations in which self-compassion occurs in more detail. A similar domain is living organisms in the case of compassion and types of situations in the case of self-compassion. The participants described compassionate situations as those in which a person suffers.

Evaluative aspect

The Evaluative Aspect domain was also identified in the data relating to compassion and self-compassion. Interestingly, the Evaluative Aspect of compassion was less frequently represented than was the case with self-compassion, where the Evaluative and Behavioural Aspects are second most frequent. The reason more people evaluated self-compassion than compassion could be that it creates more cognitive dissonance, which they feel the need to explain or at least comment on. As Leary et al. (2007) have suggested, self-compassion leads people to understand their role in situations of distress without feeling negative emotions towards the self. Negative evaluation, which is another subdomain found in the data, may therefore be connected to a lack of personal experience of self-compassion.

Cognitive aspect

The Cognitive Aspect domain included understanding. In relation to compassion, this domain mainly comprises general understanding and understanding for others and within self-compassion it concerns understanding of the self. These could also be referred to as cognitive empathy, the term used by Bloom (2016). Compassion involves understanding suffering, and was expressed by the participants as “understanding” or “understanding others”. In relation to self-compassion, the main category was “understanding self” and “general understanding”, while a further category was “balancing”, indicated by statements like “thinking about self” or “thinking about life”.

Definitions of compassion and self-compassion

The categories obtained from our data are similar to the definitions of compassion by Jazaieri et al. (2013) and Strauss et al. (2016), which state that compassion is an affective, cognitive, and behavioural process. We therefore consider compassion and self-compassion to consist of various aspects (Jazaieri et al., 2013; Strauss et al., 2016), as the data shows that the emotional aspect occurs most frequently and is possibly the most important domain for both constructs.

Both these definitions are relevant to self-compassion, with the only caveat being that it is directed towards the self and not other people. All five elements of the definition by Strauss et al. (2016) could be seen in our data. Recognition of suffering is present in the Biological Aspect of compassion and the Situational Aspect of self-compassion, which clearly indicate that people interpret it in terms of the suffering of the person. A feeling of empathy was frequently found in the data and so the whole subdomain is called empathy. The connection between empathy and living organisms as postulated in the theory by Strauss et al. (2016) is seen in the Biological Aspect domain. Tolerating uncomfortable feelings elicited in situations involving suffering were manifested in the data as negative emotional associations connected with compassion. The Behavioural Aspect was also evident in compassion, with participants reporting a general desire to help or concrete acts of helping like “hugging” or “supporting”. The participants’ answers on self-compassion mainly concerned helping and favouring one’s self. The second element in the definition by Strauss et al. (2016) was not so clear in our data because of the complexity of the meaning and the simplicity of our data. However, there were humanity associations for both compassion and self-compassion in our data.

Analysis of the most frequent associations

In order to summarise our findings we described the most frequent associations for both constructs. We decided to include only categories containing at least 10 associations. The analysis showed that the participants saw compassion as mainly consisting of empathy; the emotions of love, sadness and remorse; cognitive understanding; behavioural displays of help; and physical or mental closeness. They thought compassion was mainly directed towards people close to them, such as family and friends, and vulnerable people. Compassion occurs in situations involving loss or any other kind of suffering. The participants saw selfcompassion primarily in terms of the positive emotions of love and calmness; the negative emotions of unhappiness, sadness and remorse; cognitive understanding; behavioural displays of helping oneself by providing self-support and self-assurance. Self-compassion mainly appeared in relation to internal suffering in response to criticism or externally in relation to difficult situations. Self-compassion was evaluated in both negative (because of its misuse) and positive terms (because of its connection to virtues).

Limitations and future work

We have focused on free associations produced in response to the words compassion and self-compassion. We limited these to the first three associations. The associations have limited reporting value and, as we had no opportunity to ask the participants for details, the statements could have been misinterpreted. To some extent, we solved this by using the Consensual Qualitative Research method of data analysis. As very little is known yet about what compassion and self-compassion mean to people, it is a promising area for further qualitative research. In our future work, we would like to use open-ended questionnaires, in-depth interviews, or focus groups to explore people’s perceptions and experiences of compassion and self-compassion more deeply.

The fact that our sample comprised psychology students is another limitation. In the future it would be a good idea to analyse the free associations of a community sample and then compare the results. Also, it would be useful to use a quantitative methodology to test peoples’ constructions of compassion and self-compassion, as evidence from this kind of research is still viewed more seriously than that obtained using a qualitative methodology (Masaryk & Sokolová, 2012). In future, it would be useful to investigate what compassion and self-compassion are and whether there is a particular facial expression associated with these emotions as there are with other primary emotions.