Non-expert views of compassion: consensual qualitative research using focus groups

Sample

There were 56 participants in the focus groups. They were Slovak and Czech men (48%) and women (52%) aged 18 to 75 years old. The participants included 12 Czechs. As Czechs and Slovaks have a shared history of living together in one state and understand each other’s languages, we were able to do the focus groups together with everybody speaking their own native language. This enabled us to enrich the data on understanding compassion with data from two cultures. The mean age of the sample was 38.69 with a standard deviation of 18.01. We conducted 10 focus groups in total: three groups of men only (5 members in each), three with women only (6 or 5 members in each), and four focus groups with men and women together (6 members in each). We selected lay participants with no experience of the helping professions (based on sociodemographic data—education and work), as data on professional helpers had been collected in previous research (Baránková & Halamová, 2018; Crowther et al., 2013; Halamová et al., 2018; Kneafsey et al., 2016; Van der Cingel, 2011).

Consensual Qualitative Research

To reduce the risk of distortion we decided to use Consensual Qualitative Research (CQR) analysis, developed by Hill, Thompson, and Williams (1997). In this kind of analysis, the data are investigated by a research team, which means that multiple perspectives and views can be taken into account before a consensus is reached. The whole process is watched over by an independent auditor. Hill et al. (1997) stressed that objectivity is improved throughout the process of data interpretation as all researchers in the CQR team evaluate the possible expectations prior to getting the raw data for analysis. First, the data categorization is done separately, then the team discuss the best ways of eliminating subjectivity in the process (Hill et al., 1997).

In our study there were two assessors in the core research team. The assessors discussed their expectations beforehand. Afterwards, each assessor separately received the raw transcripted data from all the focus groups. The assessors tried to match the statements with the categorizations created in the previous research (Halamová, Baránková, Strnádelová, & Koróniová, 2018). This was because the findings reflected the multidimensional definition of compassion by Strauss et al. (2016). The statements were categorized by each assessor individually and then discussed. After reaching a consensus on the domains, subdomains, categories, subcategories, and characteristics, the assessors submitted their categorization to the auditor who checked it and gave feedback. The auditor’s comments were considered and implemented in the final version of categorization. No inter-coder reliability was computed, the results were reached by consensus between the assessors and the auditor.

Emotional aspect

The Emotional aspect domain occurred most frequently in the categorization. It contained definitions of compassion in terms of emotional states. The most frequent subdomain was Emotions. Some participants perceived compassion as a Synonym for emotion (“For me it is an emotion.”) or a Specifically named emotion. Specific positive emotions that were mentioned were Joy (“It can also be positive compassion, something like joy…”) and Love (“There are many forms of compassion - for example love.”). Surprisingly, Love was mentioned only once, contrary to what Gilbert’s (2017) theory. Unlike in our previous research on free associations for compassion (Baránková & Halamová, 2018; Halamová et al., 2018), in this data, there was no support for compassion as a form of love (Sprecher & Fehr, 2005). It is probable that the participants mentioned love when they wanted to be concise by naming three free associations but articulated it differently in the time and opportunity provided by the focus groups. Love could have been present in compassion towards close people but not mentioned by name. The descriptions and categorization of negative emotions were more diversely described and categorized compared to positive emotions. Many participants discussed the similarity versus the dissimilarity of compassion versus sadness and remorse. A few participants thought compassion identical to Sadness (“It evokes sadness in people…”). But more participants thought that compassion differed from sadness: Sadness is not compassion (“A person can be sad but they don’t have to feel compassion in that situation.”); Sadness is followed by compassion (“First there is compassion and then you can feel sadness…”); Compassion is sadness plus help (“... well, sadness… we can be sad about something but I don’t think help has to be there like in compassion.”); and Compassion is a broader term than sadness (“Compassion is on a higher level than sadness.”).

Remorse was also a controversial topic for the interlocutors. The number of participants who thought compassion the same as Remorse (“I feel remorse for someone who is suffering.”) was approximately equal to those who thought it was Not the same as remorse. Differences between compassion and remorse were discussed using these subcharacteristics: Compassion is remorse plus help (“... remorse needn’t involve help, that’s the difference.”); Compassion is not remorse (“I can’t feel remorse for someone who I share compassion with.”); Remorse signalizes superiority (“Remorse is only for poor you...”); Remorse is negative and compassion positive (“Remorse is more of a negative emotion.”). In our data compassion was sometimes confused with remorse or pity. Some of the participants thought that remorse is related to superiority, which is similar to findings by Schantz (2007) and Frazer (2006).

Our findings support previous quantitative research that showed that people are not used to compassion and find compassion towards oneself and receiving compassion from others difficult and that they can be fearful of compassion (Gilbert et al., 2011). In the data, Fear of compassion was found in various forms: Fear of compassion for others (“...I don’t know how to express emotions like this… it feels weird...”); Fear of compassion from others (“For me, it is unbearable, uncomfortable, I run away from these situations.”) and Fear of compassion towards oneself (“I feel angry rather than sad, but I never feel compassion towards myself at all.”).

Participants thought compassion was like Pain (“In most cases when you see that someone is suffering, you can feel the pain of the sufferer.”) and for some participants, compassion was seen as Both a positive and a negative emotion (“Compassion relates to happy feelings and also to sad feelings, depending on the circumstances.”). Also, some participants mentioned Nonspecific positive emotions (“The very essence of compassion is good, compassion is positive.”) and Nonspecific negative emotions (“I feel very uncomfortable in situations of compassion.”).

The subdomain Related to empathy shows that for many people compassion is The same as empathy (“For me, it is the same; compassion and empathy are the same.”). A smaller number of participants saw compassion as Not the same as empathy. This category had five subcategories. Participants thought that Compassion is more emotional than empathy (“Empathy is maybe more rational and compassion more emotional.”); Compassion is empathy plus help (“I can see the similarities, but one fundamental difference is help – compassion is empathy plus helping someone.”); Compassion is related to suffering (“... compassion is about the bad thing that happens.”); Empathy is not compassion (“...you can see the same problem through his/her eyes when you feel empathy and through your own eyes, when you feel compassion.”) and most of the participants thought that Empathy is a broader term than compassion (“Compassion is more concrete.”). As in research by Kneafsey et al. (2016), participants described compassion as being very close to or the same as empathy with small differences, for example, that empathy is a part of compassion. As Singer and Klimecki (2014) stated, the brain regions involved when compassion is felt differ from those when empathy is felt. Bloom (2016) is another who distinguished between the two constructs (empathy defined as feeling what a person believes the other person feels and compassion as understanding the other person’s situation, based on a similar past experience, together with the desire to end suffering, accompanied by helping behaviour). On top of that, he recommends using rational compassion rather than empathy. This is supported by the research findings of Klimecki, Leiberg, Ricard, and Singer (2014) that people who had received training to feel compassion or empathy showed increased helping behaviours while empathy training increased negative affect in response to other people suffering. It is important to differentiate between these two similar constructs if we are to better cultivate and exploit compassion.

Behavioural aspect

Participants also perceived compassion through the behavioural part of this construct, as doing something to alleviate suffering. This domain consists of 3 subdomains: 1) Display of expression, 2) Display of help, and 3) Display of favour. Participants described Displays of expressions mainly in connection with facial expressions. Specific facial expression was the most numerous category and had the following subcategories: Gaze (“...you can see it through his/her eyes.”); Neutral lip corners (“Lip corners are never pulled down.”); Sad eyes (“I have sad eyes…”); Sad eyebrows (“Outer eyebrows are up and inner down.”); Raised eyelid (“...and you open your eyes wide, because you want to show that you can understand that.”); Dropped lip corners (“The lips go downwards and there is no smile on the face…”); compared to Raised lip corners (“Someone is just smiling…”). Kneafsey et al. (2016) found behaviour connected to compassion was significant too.

Haidt and Keltner (1999) suggested that the facial expression of compassion is most confused with sadness; however, in our data, participants thought compassion and sadness differed. Most of the participants thought sadness was different from compassion, even though many of them felt urged to comment on sadness in relation to compassion. Statements related to sad facial expressions showed that participants saw sad eyes, a sad mouth, and sad eyebrows as partly expressing compassion for example.

Participants thought that there was no specific facial expression for compassion Nonspecific facial expressions (“Each person is unique, and we cannot react in the same way in the same situation.”) and some believed that there was no facial expression for compassion or did not know if there was.

There were also the categories of General expression Verbal (“He/she can read our words…”), Paraverbal (“You can feel it from the tone of voice.”), and Nonspecific nonverbal expressions which consisted of Posture (“...or overall body posture.”); Gestures (“...it´s also about gestures.”) and Mirroring (“I try to look like that person.”). Mirroring appeared in the data in the same way as suggested in the theories, as the part of empathy that includes compassion ( Goetz et al., 2010).

The subdomain Display of help contains the categories Help and Support. Help consisted of 1) Searching for help (“Try to find an opportunity to help.”) 2) Desire to help (“Compassion and the desire to help… to relieve someone’s suffering.”) 3) A concrete act of help (“You can help materially.”), and 4) Help should be provided (“...you have to help in that case.”). Similarly, Kneafsey et al. (2016) saw compassion as the motivation and responsibility to help someone in need. Compared to health care professionals, non-experts mentioned the need to help less often. The category Support comprises: 1) General support (“There is support.”) 2) Behavioural support (“Verbal support or gestural support.”), and 3) Mental support (“Something mentally supportive.”).

Display of favour was divided into three categories: Behavioural closeness [Touch (“Just a hug, that’s the greatest help.”) or Physical closeness (“Sometimes you don’t have to say anything, just be with me.”)], Mental closeness [Interest (“...just bigger interest.”), Companionship (“I’m here with you, you´re not alone.”), Care (“I will show that I care.”), Humanity (“It is a deeply human feature…”), Being present (“We can give some time and space to share things and just to be there.”), and Kindness (“The person is kinder.”)], and Display of motivation (“...sometimes it gives you the energy to do something about it.”). For Gilbert (2017), compassion is related to constructs such as altruism, benevolence, heartedness, prosociality, kindness, and love. However, in our data, these constructs were mentioned rarely.

Biological aspect

There were three subdomains in the Biological aspect of compassion: People was the most frequent, followed by Animals and Plants, which were represented far less. Compassion for people was discussed as Types of compassionate relationships involving Vulnerable people (“For example elderly people in public transport.”) and people close to us, so according to the participants closeness in a relationship is important in compassion Related to closeness in a relationship (“It depends if they are close to me or a stranger.”) or not important Unrelated to closeness in relationship (“Compassion is still the same - whether it is my brother who breaks his arm or my neighbour…”). Therefore, our findings support the importance of reciprocity in compassion and its relation to kin or non-kin relationships (DeSteno, 2015).

Types of compassionate situations mentioned in the discussions were Loss (“Loss of someone close.”), Physical suffering (“Cancer or the look of a sick child.”) or Homelessness (“For example helping someone who has no home.”), and General suffering (“Compassion is suitable in difficult situations…”). While physical suffering, like diseases or injuries, was mentioned frequently, participants completely omitted the possibility of mental suffering, which may be related to the taboo around mental health issues in Slovakia.

Evaluative aspect

People evaluated compassion as an important virtue in life in the category Evaluation of importance (“Compassion is necessary in society; if there ´s no compassion, there will be great chaos.”) and only one person labelled compassion as unimportant – Evaluation as unimportant (“Compassion is not always beneficial.”). In the category Evaluation of compassion, Evaluation of adequacy of situation (“It depends on what caused the situation and why it occurred.”) was important, as was Evaluation of deservedness (“I don’t feel compassion if someone is doing bad things and something terrible happens to him/her.”). According to our data and those of Gilbert (2010), the situational and personal aspects of compassion are evaluated before motivation and helping behaviour are provided to the person suffering. Participants pointed out in Compassion limits (“If you had to share compassion with everyone, you would go mad.”).

Altruism in compassionate behaviour was also evaluated and categorized as Misuse of compassion (for someone’s benefit) (“it is very often exploited or artificially created.”) and as Reciprocity (“I want to help someone who has helped me in a bad situation.”), which is similar to DeSteno’s (2015) statements that we help on a reciprocal basis. From the discussions in the focus groups, it emerged that compassion is perceived as Innate with further development (“Everyone has compassion and if it is not developed, it is the individual’s fault.”) which is consistent with findings of Kneafsey et al. (2016). Only a few participants said that there were some people who have no capacity at all for compassion.

Cognitive aspect

The Cognitive aspect domain was the least frequent one and had a subdomain Understanding which was divided into three categories: General understanding (“...the person starts thinking about it, processing it.”), Understanding others (“To understand another.”), and Understanding a situation (“I try to figure out what happened and why he/she is suffering.”). The Cognitive aspect or understanding is an important part of compassion according to many scholars (e.g. Kneafsey et al., 2016; Strauss et al., 2016; Van der Cingel, 2011).