Compassion has received increased attention in clinical psychology over recent decades (e.g. Galante et al., 2014; Gilbert, 2010). It has been conceptualized as a felt response to suffering that involves caring as well as an authentic desire to ease distress (Goetz et al., 2010). Compassion is furthermore thought to be an evolved motivational response referring to “the sensitivity to suffering in self and others, with a commitment to try to alleviate and prevent it” (Gilbert, 2017). The academic discourse on compassion and mindfulness has been informed largely by Buddhist philosophy and has been incorporated into clinical psychology in order to understand and alleviate mental distress, as well as to facilitate growth and promote well-being (e.g. Craig et al., 2020; de Vibe et al., 2012; Klimecki et al., 2013; Mongrain et al., 2011).
With regard to the conceptualization of compassion, compassion can be understood both as a state and as a trait (Goetz et al., 2010). Compassion as a state is triggered by distinct situational antecedents and consists of brief and context-specific expressions of compassion. Compassion as a trait involves the tendency to experience and express compassion in different situations and is understood as a general style of responding that is stable over time and contexts (Goetz et al., 2010). The scientific literature on compassion has also focused on its relation to other constructs, in particular empathy. Although compassion and empathy are related concepts (Jeffrey, 2016; Neff & Pommier, 2013), there is a consensus that they are different concepts (Bivins et al., 2017). In empathy, a person can empathize with someone else’s sadness or distress, as well as their joys, gratitude, or admiration. Compassion on the other hand usually occurs as a response to someone else’s pain. Moreover, compassion, unlike empathy, contains a motivational component aimed at acting in a way that relieves the suffering of others (Caouette & Price, 2018).
Both compassion for others and self-compassion are associated with individuals’ psychological well-being. For example, a compassionate action towards others performed daily for 1 week has been shown to lead to significant increases in self-reported happiness, compared to a control condition (Mongrain et al., 2011). Furthermore, Klimecki et al. (2013) found support for higher experienced positive affectivity in participants after brief compassion training compared to a control condition. In general, recently published meta-analyses from different fields suggest that the deliberate cultivation of compassion has beneficial effects on prosocial behaviors and can be perceived as a coping strategy that fosters positive effects when confronted with the distress of others (Kirby et al., 2017; Quaglia et al., 2021; Sinclair et al., 2021).
In a recent article, Pommier et al. (2020) presented a new measure of compassion for others called the Compassion Scale. The construction of the Compassion Scale was based on the work of Neff (2003) on self-compassion resulting in a scale called the Self-Compassion Scale, which is the most widely used measure of self-compassion (Quaglia et al., 2021). In Neff’s conceptualization, self-compassion is understood as comprising three sets of opposing components along the dimensions: emotional response (self-kindness vs. self-judgements), cognitive understanding (common humanity vs. isolation), and paying attention to personal suffering (mindfulness vs. overidentification). In a similar way, in the Compassion Scale, compassion for others is operationalized as a trait involving cognitive, emotional, and attentional aspects and encompassing the dimensions of kindness vs. indifference, common humanity vs. separation, and mindfulness vs. disengagement in response to the suffering of others. In this context, kindness is conceptualized as being concerned for and caring towards others who are in pain, as well as having the desire to support those in need. Common humanity is understood as recognizing that it is a common human experience to experience hardship as well as a sense of connection to those who are suffering. Lastly, mindfulness is conceptualized as a form of balanced awareness which neither avoids nor becomes lost in the pain of others; rather, it refers to being willing to pay attention and listen to others when they are suffering.
During the development process of the Compassion Scale (Pommier et al., 2020), the originators settled upon a 16-item measure of compassion with adequate internal consistency in all studied samples (Cronbach α coefficients ranged from 0.77 to 0.90). The final scale had a 4-factor structure, with 3 factors positively related to compassion according to the intended dimensions (“Kindness”, “Common humanity”, and “Mindfulness”) and 1 factor negatively related to compassion (“Indifference”). With regard to discriminative and convergent validity, the Compassion Scale showed positive correlations with other measures of different forms of compassion, empathy, agreeableness, and altruism, whereas the correlations were negative when concerning unfavorable attitudes towards others and neuroticism (Pommier et al., 2020).
As previously described, recent years have seen an increased theoretical and empirical interest in compassion as a psychological construct, and—as always—adequate measures of central concepts are a necessary prerequisite for further research. In a review from 2016, Strauss et al. (2016) identified important shortcomings in the existing self-report measures of compassion, and the Compassion Scale was developed partly to address these shortcomings (Pommier et al., 2020). It is also increasingly used in research and translated to different languages (Lucarini et al., 2023).
On this background, the purpose of the present study was to translate the Compassion Scale (Pommier et al., 2020) to Swedish and to examine whether the factor structure of the Swedish version of the Compassion Scale was the same as that of the original. Furthermore, we also wanted to test the construct validity of the Swedish version of the Compassion Scale by investigating its relationship to measures of personality and empathy. In addition, the purpose of the study was to examine how scores on the Swedish version of the Compassion Scale related to sociodemographic variables (gender, age, and income).
Method
Participants
Sample 1
In total, there were 4507 responses collected from a sample that poorly represented the Swedish population with regard to gender and age (67.8% were women and the average age was 32.2 years, compared to 49.6% women and 41.7 years in the general population (Ekonomifakta, 2023; Statistics Sweden, 2023). To increase the representativeness, a series of random selections from the general sample were made to achieve a final sample that matched the composition of the adult population in Sweden with regard to gender and relative size of different age groups. This final sample consisted of 1712 individuals. Sample 1 was used for investigating the factor structure of the Swedish version of the Compassion Scale as well as its relationship to personality variables, age, gender, and income. Descriptive statistics on demographic variables for this final sample are presented in Table 1. The sample size was not based on any specific a priori power calculations. Instead, we used convenience sampling and aimed for a sufficiently large sample size to achieve both general representativeness and the possibility to make reliable between group comparisons.
Table 1
Descriptive statistics on demographic variables in Sample 1
Variables | Frequency (n = 1712) | Relative frequency |
|---|---|---|
Gender | ||
Man | 854 | 49.8% |
Woman | 838 | 48.9% |
Other/do not want to say | 20 | 1.2% |
Age | ||
20–39 | 548 | 32.0% |
40–59 | 592 | 34.6% |
60– | 572 | 33.4% |
Country of birth | ||
Sweden | 1452 | 84.8% |
Other | 256 | 14.9% |
Main occupation | ||
Working | 744 | 43.5% |
Student | 556 | 32.5% |
Retired | 372 | 21.7% |
Other | 40 | 2.3% |
Sample 2
Sample 2 consisted of 353 participants (270 women, 82 men, and 1 that stated “Other/do not want to say”). With regard to occupation, 44% stated “Student”, 47% “Employed”, and 9% “other” as their main occupation. In Sample 2, 60% of the participants were between 18 and 29 years old, while 14% were between 30 and 39, 15% between 40 and 49, and 11% over 50 years old. Sample 2 was included to investigate the relationship between the Swedish version of the Compassion Scale with a measure of empathy. The sample was recruited as a part of another project of the same research group and its size was not based on specific a priori power calculations in relation to the present study. Rather, the size was determined by the number of people which assented to participate, but we made sure that it exceeded the number of participants required for correlation estimates to stabilize (Schönbrodt & Perugini, 2013).
Measures
Compassion Scale
The 16-item version of the Compassion Scale (Pommier et al., 2020) was translated into Swedish using a translation—back translation procedure (Wang et al., 2006). Two researchers—fluent in both English and Swedish and with extensive experience of research in relation to compassion, empathy, mindfulness, and related constructs—performed the translations. One translated from English to Swedish, and one back-translated to English. Discrepancies were discussed by the translators until a final Swedish translation of all items were agreed upon. Four of the 16 items are negatively worded and 12 positively. The participants are asked to rate “how often do you feel or act in the way described by the statements” on a 5-point scale ranging from Almost never to Almost always. In Table 3, the items are listed in both English and Swedish.
Big Five Inventory (BFI)
Personality traits according to the 5-factor model were measured using The Big Five Inventory (BFI; John & Srivastava, 1999). The BFI is a 44-item questionnaire measuring the factors Extraversion, Agreeableness, Conscientiousness, Neuroticism, and Openness. Extraversion includes traits related to activity, sociability, and assertiveness; Agreeableness concerns traits such as altruism, friendliness, trust, and modesty; Conscientiousness consists of traits such as orderliness, perseverance, and industriousness; and Neuroticism refers to emotional instability and experiencing negative emotions, such as anxiety, sadness, and irritability. Finally, Openness incorporates intellect, fantasy, aesthetic interest, and curiosity.
The BFI has been well validated by previous research and has been shown to have adequate psychometric properties and good convergent validity with more comprehensive measures of the 5-factor model (John & Srivastava, 1999). The Swedish version of the scale has previously been validated (Zakrisson, 2010), and in the present sample, it had adequate levels of internal consistency across all factors (0.75 < α < 0.84).
Narcissistic Personality Inventory (NPI) 16-item version
The 16-item version of the NPI developed by Ames et al. (2006) measures subclinical narcissistic personality traits. Narcissism concerns a grandiose sense of self and a pre-occupation with one’s own success (Ames et al., 2006). The NPI has been shown to have adequate psychometric properties, as well as meaningful face, discriminant, internal, and predicative validity (Ames et al., 2006). The Swedish version of the scale used in the present study showed adequate internal consistency (Cronbach’s α = 0.71).
Interpersonal Reactivity Index (IRI)
The Interpersonal Reactivity Index (IRI) is a 28-item questionnaire measuring individual differences in different aspects of empathy (Davis, 1983). The measure has four subscales: Perspective Taking (the tendency to spontaneously adopt the psychological point of view of others), Fantasy (the tendency of the respondents to transpose themselves imaginatively into the feelings and actions of fictitious characters in books, movies, and plays), Empathic Concern (“other-oriented” feelings of sympathy and concern for unfortunate others), and Personal Distress (“self-oriented” feelings of personal anxiety and unease in tense interpersonal settings). In previous research, the IRI has been shown to have adequate psychometric properties regarding reliability and validity (Davis, 1983; Pulos et al., 2004). The Swedish version of the scale used in the present study showed adequate internal consistency for all subscales (0.72 < α < 0.81).
Sociodemographic Variables
With regard to the sociodemographic variables under investigation, participants were asked to state their gender (female/male/other), age (in whole years), and average monthly income (in Swedish kronor). Descriptive statistics are presented in Table 1.
Data Analyses
The factor structure (including standardized factor loadings) of the Swedish version of the Compassion Scale was tested with confirmatory factor analysis (CFA) using Amos for SPSS. CFA was used as the data analytic method since it allows for the assessment of fit between observed data and an a priori conceptualized model (Mueller & Hancock, 2015).
The factor structure tested was the four-factor model, where all latent factors were correlated, suggested by the originators (Pommier et al., 2020). Maximum likelihood estimation was used to assess the parameters of the models. None of the error terms was allowed to correlate. To estimate model fit, in addition to the χ2 value, we selected three fit indices: the root mean square error of approximation (RMSEA), the comparative fit index (CFI), and the Tucker-Lewis Index (TLI). On these indices, Hu and Bentler (1998) suggest the values below 0.06 (RMSEA) and above 0.95 (TLI and CFI) as indicative of good model fit. In addition to model fit statistics and item factor loadings, the internal consistency (Cronbach’s alpha and McDonald’s omega) was calculated both for the entire scale and the different subscales.
Prior to performing the analyses, the data were screened for violations of the assumptions of confirmatory factor analyses. Here, the assumptions of univariate normality (skewness < ± 2 and kurtosis < ± 7; Byrne, 2010), multivariate normality (using Q-Q and P-P plots), and sufficiently large sample size (n > 200) were met. Since the assumptions of both univariate and multivariate normality were met, we retained the more robust maximum likelihood estimation method, even though the response scale of the individual items can be considered ordinal (Harris, 2001; Robitzsch, 2020; Tabachnick & Fidell, 2007).
Evidence for convergent validity was assumed to be found primarily in relation to Agreeableness among the 5-factor model dimensions (John & Srivastava, 1999) and Empathic concern (Davis, 1983) among the empathy dimensions measured with the IRI. Agreeableness, which is about being friendly and altruistic, is constructed as functionally related to compassion and here a positive association was found by Pommier et al. (2020). Empathic concern refers to sympathetic feelings towards others and has obvious potential in relation to compassion (also shown in Pommier et al., 2020). With narcissism, we expected a negative association since the narcissistic pre-occupation with oneself can be regarded as in some way opposite to caring for others, which is an essential aspect of the concept of compassion. The other personality factors and empathy dimensions will be explored but no specific assumptions were formulated.
Finally, based on previous research with regard to sociodemographic variables, our assumption was that total scores on the Compassion Scale would be negatively related to age (e.g. Grühn et al., 2008) and income (e.g. Piff et al., 2012). With regard to gender, we expected females to—on average—have higher compassion scores than men (e.g. Baez et al., 2017).
Results
Confirmatory Factor Analysis
The factor structure of the Swedish version of the Compassion Scale was examined using confirmatory factor analysis in Sample 1 in accordance with what has been described in the “Method” section. The analysis indicated that the suggested factor solution provided acceptable fit with the data (CFI = 0.95; RMSEA = 0.067; TLI = 0.94), where 0.06, 0.95, and 0.95 were considered indicative of good model fit (Hu & Bentler, 1998), and an RMSEA value below 0.08 is considered acceptable (Browne & Cudeck, 1993). In the next step of the analysis, the factor loading for each item was calculated, and the results are presented in Table 2. The analysis shows that every item except one had a factor loading that exceeded the 0.50 criteria suggested by Hair et al. (2006). The only item that did not reach this threshold was the item “I notice when people are upset, even if they don’t say anything”, which had a factor loading of 0.30 on the Mindfulness factor. Despite the fact that the item did not exceed the 0.50 criteria suggested by Hair et al. (2006), we decided to keep the item in order to retain the structure of the original scale and since it exceeded the minimum factor loading suggested by others in the field (e.g. Tabachnick & Fidell, 2007).
Table 2
Factor loadings for the Compassion Scale
Items | Factors | |||
|---|---|---|---|---|
K | CH | M | I | |
If I see someone going through a difficult time, I try to be caring toward that person (SWE: “Om jag ser att någon går igenom en svår tid, försöker jag vara omtänksam mot den personen”) | 0.78 | |||
I like to be there for others in times of difficulty (SWE: “Jag tycker om att finnas där för andra när de har det svårt”) | 0.77 | |||
My heart goes out to people who are unhappy (SWE “Jag känner starkt med de som är olyckliga”) | 0.60 | |||
When others feel sadness, I try to comfort them (SWE: “När andra är ledsna försöker jag trösta dem”) | 0.70 | |||
I realize everyone feels down sometimes, it is part of being human (SWE: “Jag inser att alla känner sig ledsna ibland, det är en del av att vara människa”) | 0.65 | |||
I feel it’s important to recognize that all people have weaknesses and no one’s perfect (SWE: “Jag tycker att det är viktigt att vara medveten om att alla människor har brister och att ingen är perfekt”) | 0.76 | |||
Despite my differences with others, I know that everyone feels pain just like me (SWE: “Trots våra olikheter vet jag att andra känner lidande precis som jag”) | 0.57 | |||
I feel that suffering is just a part of the common human experience (SWE: “Jag anser att lidande är en naturlig del av vad det innebär att vara människa”) | 0.56 | |||
I pay careful attention when other people talk to me about their troubles (SWE: “Jag är mycket uppmärksam när andra pratar med mig om sina bekymmer”) | 0.80 | |||
I notice when people are upset, even if they don’t say anything (SWE: “Jag lägger märke till när andra är upprörda, även om de inte säger något”) | 0.30 | |||
I listen patiently when people tell me their problems (SWE: “Jag lyssnar tålmodigt på andra när de berättar för mig om sina probem”) | 0.79 | |||
When people tell me about their problems, I try to keep a balanced perspective on the situation (SWE: “När andra berättar för mig om sina problem, försöker jag behålla ett balanserat perspektiv på situationen”) | 0.55 | |||
I am unconcerned with other people’s problems (SWE: “Jag är likgiltig inför andras problem”) | 0.58 | |||
I think little about the concerns of others (SWE: “Jag bryr mig inte så mycket om andras bekymmer”) | 0.67 | |||
I try to avoid people who are experiencing a lot of pain (SWE: “Jag försöker undvika människor som upplever mycket lidande”) | 0.70 | |||
I can’t really connect with other people when they’re suffering (SWE: “Jag kan inte riktigt knyta an till människor som lider”) | 0.53 | |||
In the next step, the bivariate correlations between the 4 subscales were calculated and are presented in Table 3. As can be seen, the correlations between the subscales were large (according to the criteria suggested by Gignac & Szodorai, 2016) between all subscales.
Table 3
Bivariate correlations (Pearson’s) between the four subscales of the Compassion scale (n = 1712)
Kindness | Common humanity | Mindfulness | Indifference | |
|---|---|---|---|---|
Kindness | - | 0.32 | 0.61 | − 0.56 |
Common humanity | - | 0.33 | − 0.24 | |
Mindfulness | - | − 0.35 | ||
Indifference | - |
In addition, to determine whether the Swedish version of the Compassion Scale can be reliably treated as a summated scale, a hierarchical model was tested in a CFA where the 4 factors loaded on a general latent variable. The results indicate an acceptable model fit for this hierarchical model (CFI = 0.94; RMSEA = 0.075; TLI = 0.93). Finally, the internal consistency for both the entire scale and the different subscales was calculated. The entire scale had an internal consistency of 0.83 when measured as Cronbach’s α, 0.80 when measured as McDonald’s ω. For the subscales, the internal consistencies were α = 0.75 and ω = 0.73 for Kindness, α = 0.71 and ω = 0.70 for Common Humanity, α = 0.66 and ω = 0.65 for Mindfulness, and α = 0.63 and ω = 0.61 for Indifference. Hence, the performed analyses indicate that the Swedish version of the Compassion Scale has the same factor structure as the original scale and that the scale is reliable with regard to the entire scale. With regard to the different subscales, the internal consistency was satisfactory (α and ω > 0.70; McNeish, 2018; Nunnally, 1978) for Kindness and Common Humanity, but a low for Mindfulness and Indifference. When a factor only includes a few items, however, α-coefficients between 0.60 and 0.70 are considered acceptable (Hair et al., 2014).
Convergent and Discriminant Validity
To test the convergent and discriminant validity of the Swedish version of the Compassion Scale, we investigated the bivariate correlations between the Compassion Scale scores and scores from the BFI, the NPI (based on Sample 1), and the IRI (based on Sample 2). The results are presented in Table 4.
Table 4
Correlations between the Swedish version of the Compassion Scale and 5-factor personality (BFI), narcissism (NPI), and empathy (IRI)
Measure | Sample | n | r with | ||||
|---|---|---|---|---|---|---|---|
CS total | K | CH | M | I | |||
BFI Agreeableness | 1 | 1686 | 0.47** | 0.47** | 0.15** | 0.42** | − 0.36** |
BFI Neuroticism | 1 | 1682 | − 0.04 | − 0.01 | 0.01 | − 0.03 | 0.09** |
BFI Openness | 1 | 1686 | 0.24** | 0.17** | 0.22** | 0.13** | − 0.13** |
BFI Conscientiousness | 1 | 1684 | 0.24** | 0.19** | 0.01 | 0.31** | − 0.22** |
BFI Extraversion | 1 | 1686 | 0.24** | 0.31** | − 0.04 | 0.23** | − 0.21** |
NPI | 1 | 1678 | − 0.06* | − 0.02 | − 0.04 | − 0.06* | 0.07** |
IRI Perspective | 2 | 353 | 0.42** | 0.23** | 0.29** | 0.40** | − 0.28** |
IRI Fantasy | 2 | 353 | 0.25** | 0.29** | 0.03 | 0.21** | − 0.19* |
IRI Empathic concern | 2 | 353 | 0.59** | 0.60** | 0.15* | 0.36** | − 0.54** |
IRI Personal distress | 2 | 353 | − 0.08 | 0.10 | − 0.18* | − 0.13 | 0.02 |
As expected, in Sample 1, there was a positive correlation between Compassion Scale and BFI Agreeableness. Also, in line with previous findings (e.g. Melchers et al., 2016), we found moderate to large positive correlations with Openness, Conscientiousness, and Extraversion, whereas the relationship to Neuroticism was weaker. The negative correlation with NPI was also in alignment with what was expected.
With regard to the different aspects of empathy measured with the IRI in Sample 2, the correlations between the Compassion Scale and the scores on all subscales, except “Personal Distress”, were statistically significant. As predicted, the correlations were strongest between the Compassion Scale and the subscale “Emphatic Concern”, which measures feelings of sympathy and concern for unfortunate others.
Finally, the predictions regarding the associations between compassion, income, and age were examined in Sample 1 using Pearson’s product moment correlations, whereas the prediction regarding gender differences was tested using independent samples t-test. The results indicate that women (M = 4.14; SD = 0.41) generally scored higher than men (M = 3.88; SD = 0.42) on the Compassion Scale with a moderate effect size (t (1686) = 13.0, p < 0.001; Cohen’s d = 0.63), and that compassion had weak negative correlations with both age (r = − 0.07, p < 0.001; n = 1688) and average monthly income (r = − 0.09; p < 0.001; n = 1688). Hence, the predictions regarding the associations between scores on the Compassion Scale and the sociodemographic variables were all supported by the results from the performed analyses.
Discussion
Given the increased interest in compassion in both clinical and non-clinical research (Gilbert, 2020), the purpose of the present study was to validate a Swedish translation of the Compassion Scale (Pommier et al., 2020). The results of the study show that the Swedish version of the Compassion Scale had the same factor structure as the original version, with 4 distinct but correlated factors relating to the aspects of “Kindness”, “Common humanity”, “Mindfulness”, and “Indifference”. It should be noted however, that one of the items (“I notice when people are upset, even if they don’t say anything”) had a low loading (0.30) on its latent factor. Notably, this item had a lower factor loading on its latent factor also in the original version of the scale (Pommier et al., 2020). A possible interpretation of this is that the item does not only measure the mindfulness aspect of compassion but also contains an element of interpersonal sensitivity or receptiveness. In line with this interpretation, in the original version of the scale, the item showed an almost equally strong association with the factor “Common humanity” as with “Mindfulness” (Pommier et al., 2020).
In addition, the results indicate that all the factors are correlated with each other and are reliably related to a single common factor, i.e. Compassion. Compared to the original English version, the correlation between the subscales was slightly lower in our study, where they ranged from 0.24 to 0.61, compared to 0.25 to 0.89 in the data reported for the community samples for the original version (Pommier et al., 2020). The interscale correlations found in the present study, however, are quite similar to the ones reported in a recently published article on an Italian version of the Compassion Scale (Lucarini et al., 2023). Just as in the original version of the scale, the strongest correlations are observed between subscales “Mindfulness” and “Kindness”, and the weakest correlations are observed between “Common Humanity” and “Indifference”.
Furthermore, the results indicate that the developed measure provided reliable scores, when it comes to scores on the total scale, and—though to a lesser extent—with regard to the different subscales. When compared to the original version of the scale, the Swedish version of the scale developed in the present study showed similar internal consistency when it comes to total score (0.83 in the present study and 0.88 in the original article). At the subscale level, the internal consistencies were a bit lower in the present study compared to what was reported in the community samples for the original version of the scale (0.63–0.75 compared to 0.74–0.85). In this context, it should be noted that the Italian version of the Compassion Scale (Lucarini et al., 2023) also showed somewhat lower levels of internal consistency (0.69 to 0.83) and that the internal consistencies of the subscales were significantly lower in other samples reported in the article by Pommier et al. (2020), where Cronbach’s α coefficients varied between 0.47 and 0.68 in one of the samples. In the present study, the original article, and in the article by Lucarini et al. (2023), the lowest internal consistency is generally reported for the Mindfulness and Indifference subscales. This might indicate that the items of these subscales are in need of revision in order to more reliably capture the intended constructs, or that the operationalizations of these concepts are particularly sensitive to sampling issues and to variability in interpretation. It should be noted, however, that the subscales only consist of 4 items, and that for such short scales, Cronbach’s α coefficients between 0.60 and 0.70 can be considered acceptable (Hair et al., 2014).
Regarding convergent and discriminant validity, the performed analyses indicate that the Swedish version of the Compassion Scale generally shows the expected relationship to related measures of personality and individual differences in empathic responding. The identified associations position the construct in a personality framework with the strongest link to Agreeableness, which refers to friendliness and therefore has the most obvious connection to compassion, hence indicating convergent validity. The small to moderate relations to the other positively valenced personality traits, namely Extraversion, Openness, and Conscientiousness, also corroborated the findings by Pommier et al. (2020). The positive correlation with the dimension of empathic concern furthermore supported the convergent validity of the scale since this aspect of empathy relates to feelings of warmth and concern for others (Davis, 1983) and hence demonstrates a clear resemblance to compassion (particularly the compassion subscales kindness and negatively to indifference). The negative association with narcissism was small, but nevertheless conveyed the expected discriminant validity. Finally, when it came to the analyses of scores on the Swedish version of the Compassion Scale and sociodemographic variables, the results supported the predictions that women would generally score higher than men on the Compassion Scale, and that compassion would have negative correlations with both age and average monthly income. Hence, in sum, the results of the study supported the discriminant and convergent validity of the Swedish version of the Compassion Scale, though it should be noted that some of the associations (for example with income, age, and narcissism) were quite weak.
Limitations and Future Research
There are important limitations that should be taken into account when interpreting the results of the study. One such limitation is that the measurement validation process rests entirely upon self-report measures. This is a weakness that the present study shares with other studies aiming to develop measures of compassion (e.g. Pommier et al., 2020), and in future research it would be valuable to test the convergent and discriminant validity of the scales using for example behavior samplings, to investigate whether scores on the Compassion Scale are related to differences in everyday behaviors, or ratings by others. Another important limitation, which the present study also shares with the other articles on the same subject, is that the study is cross-sectional and hence does not allow for the analyses of how compassion measured with the Compassion Scale and its associations with other variables unfold over time. In future research, it would be valuable with prospective designs to capture such temporal dynamics and processes. Given the variable and sometimes a bit low, future research should also further investigate if particularly the aspects of compassion related to Mindfulness and Indifference can be more reliably measured. Despite the limitations discussed above, the present study contributes to the scientific literature by describing the psychometric evaluation of a Swedish version of the Compassion Scale (Pommier et al., 2020) and providing support for the adequacy of its psychometric properties as well as for its convergent and discriminant validity.
Declarations
Ethics Approval
The study has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The project has also been performed in full accordance with the Swedish Ethical Review Act and approved by the ethics review board at the Department of Psychology, Lund University.
Informed Consent
All participants provided their informed consent prior to participating in the study.
Use of Artificial Intelligence
Artificial intelligence was not used in any part of the project.
Conflict of Interest
The authors declare no competing interests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.