Abstract: Both mindfulness and self-compassion have positive effects on mental health, and the relationship between them has always been the focus of researchers. And mindfulness and self-compassion are also closely related. Previous research has found that mindfulness can predict an individual's level of self-compassion, a component of which also includes mindfulness. However, different studies used the same two variables to measure, and the correlation coefficient between mindfulness and self-compassion was significantly different. Therefore, it is necessary to explore the reasons for the large correlation coefficient difference between mindfulness and self-compassion. In this study, meta-analysis method was adopted, combined with empirical research results from global research, to deeply explore the relationship between mindfulness-based interventions and self-compassion in different groups and different contrast ways, as well as the factors that affect the relationship between them.
Through a comprehensive search of Chinese and English literature up to November 2020, a total of 65 literature were finally included in this meta-analysis, including 62 English literature and 3 Chinese literature, and 99 independent effects were obtained, with a total sample size of 8103. In addition, 22 articles were included in the literature using horizontal comparison, 42 articles using longitudinal comparison; 16 articles using follow-up comparison. This paper mainly used Pearson correlation coefficient r as the effect value, and the effect value of the literature was coded as an effect value for each independent sample. If a paper reported multiple independent samples at the same time, the effect value would be coded separately to generate multiple independent effect sizes. After testing the publication bias through funnel plot, Rosenthals Classic Fail-safe N test, and Egger's test, it is found that there was no publication bias. After the heterogeneous test, the article also has high heterogeneity, so the analysis after the random effect model is selected. Sensitivity analysis showed that the degree of heterogeneity was effectively reduced after using One-study removed to gradually delete the research with higher heterogeneity, but regardless of the degree of heterogeneity, the intervention based on mindfulness was positively correlated with self-sympathy. This study discussed the moderating effects of different contrast methods (horizontal comparison, longitudinal comparison, follow-up comparison), sample groups (students, health care related workers, social related workers), and measurement tools (SCS, SCS-SF) on the relationship between mindfulness and self-compassion. The subgroup test showed that the relationship between mindfulness-based intervention and self-compassion was influenced by the moderating effect of contrast style, but not by the sample group or measurement tool.
The results showed that the mindfulness-based intervention had a significant positive correlation with self-compassion, and the mindfulness-based intervention had a positive impact on the level of self-compassion. The relationship between the mindfulness-based intervention and self-compassion was affected by the way of measurement comparison, but did not have a significant moderating relationship with the group and the measurement tool. The reason why there is such a big difference among many domestic and foreign studies is probably due to the difference of measurement methods and the difference caused by the in unity of measurement methods. Overall, in many experimental studies on the relationship between mindfulness and self-compassion, no matter which contrast method is used for measurement, the individual's level of self-compassion has increased after mindfulness-based intervention.

Key words: mindfulness, self-compassion, intervention outcome, meta-analysis