As women increasingly enter the labor market while maintaining family and household responsibilities, the incidence of work-family conflict may increase for female workers (40) and indeed for both male and female workers when partners both work and share responsibility for family care-giving (41). In Saudi Arabia, the Brookings Institution estimated an increase of 20 to 33% of women entering the workplace in 2021 (42). Balancing family life and work life has, therefore, been increasingly challenging in Saudi Arabia. This is especially true for nurses whose jobs entail psychological and physical demands associated with heavy workloads, attention to multiple patients, and the need to display positive emotions regardless of their mood (43).

The notion of work-family conflict originates in role theory (44) and the role strain hypothesis (45). Conflict can exist in both directions, work responsibilities interfering with family life (that is, work-family conflict, or WFC) and family responsibilities interfering with work life (family-work conflict, or FWC) (46). Among nurses, it was revealed that work-family conflict acted as a path through which toxic leadership impacted on distress of the nurses (47). To capture a fuller understanding of the interference between work and family responsibilities, researchers have suggested examining both WFC and FWC (33, 48).

As established earlier, WFC has been shown to be a critical factor impacting life satisfaction. Tang et al. (38) reported its negative impact on workers’ happiness, as well. FWC has also been negatively associated with life satisfaction (37) and happiness (49). Studies have highlighted the negative impacts of work-family conflict on nurses’ well-being in particular: WFC has been positively associated with nurses’ turnover intentions (50), job stress (51), and anxiety and depression (52). A negative relationship was also found among nurses between work-family conflict and both life satisfaction (53–55) and job satisfaction (56). Accordingly, the first hypothesis of this study is as follows:

The aforementioned research notwithstanding, there have been studies that found no relationship between work-family conflict and life satisfaction (57, 58), which suggests that there may be other pathways through which WFC impacts life satisfaction. Indeed, Taşdelen-Karçkay and Bakalım (54) found a mediating effect of work-life balance in this relationship. This relationship was also reported to be mediated by emotional exhaustion (60) and coping strategies (61). Others have also reported a mediation role of emotional exhaustion in the relationship between work-family conflict and job satisfaction, a construct that is related to life satisfaction among workers (62). However, there have been no studies that investigated the mediation role of self-compassion.

Self-compassion is a positive personality trait that involves treating oneself with a sense of care and kindness rather than self-criticism, especially in the face of adversity and failures (29, 63). Neff (64) adds that self-compassion includes cultivating a sense of connectedness with others rather than of isolation, and a mindful practice of not overly self-identifying with one’s painful thoughts and feelings (64). Self-compassion has been found to benefit individuals’ well-being (65). Self-compassion is linked to adaptive functioning, especially in times of loss, failure, and stress (66). (67) reported that self-compassion is correlated with positive mental health outcomes, such as less depression and anxiety and greater life satisfaction. Neff and Faso (68) reported a positive relationship between self-compassion and well-being in parents of children with autism. And (69) found a negative relationship between self-compassion and work-family interference.

It has been argued that self-compassion protects people when a negative situation is beyond their control or even when they are responsible for the negative circumstances (66); thus, nurses who are self-compassionate may be less affected by work-family conflict regardless of the situation. In previous research specifically with nurses, it was reported that self-compassion was positively associated with job satisfaction and negatively related to burnout (70). Further, self-compassion was an important predictor of relationship satisfaction and conflict resolution (71) and family functioning (72) among nurses. The study’s second set of hypotheses is as follows:

Gratitude is the emotion or state of being thankful. In one sense, it can be defined as the emotion that people feel when they receive something that they perceive as valuable, altruistic, or even costly (73), and more broadly can be defined as an attitude or disposition by which one notices and appreciates positive experiences and achievements (74, 75). The broaden-and-build theory (76) posits that positive emotions, including gratitude, help individuals’ build enduring personal physical, intellectual, and psychological resources. It has been postulated that people with high levels of gratitude experience more frequently and intensely positive emotions and use positive coping mechanisms, which enhance their mental well-being (31, 75, 77). Others have claimed that gratitude leads people to interpret other people’s behaviors in a positive way, perceive other people as supportive and available, which benefits both parties (74, 75, 78). Randolph (79) posited that expressing gratitude is essential for nurses to cope with the challenges of work and to enhance family relationships. Prior research has reported that gratitude increased happiness (80), resilience, job satisfaction, and a healthy work environment among nurses (81). Further, gratitude was claimed as beneficial at workplace. In a sample of 411 employees, Mahipalan and Garg (82) reported that gratitude moderated the relationship between workplace bullying and psychological capital of employees. A study involving 112 female employees in India found that gratitude was linked to family enrichment (83), and to spiritual well-being and distress (84). In a sample of healthcare employees in India, Garg et al. (85) concluded that gratitude moderated the relationship between workplace toxicity and turnover intentions. Gratitude was also related to vitality with a mediation of resilience (86). Gratitude was also reported as a moderator in the relationship between teasing and depression (87). Among nurses, gratitude was reported to mitigate the negative impacts of FWC on work engagement (88). Parallelly, it was claimed in previous research that the relationship between self-compassion and happiness was moderated by gratitude (89). Therefore, it is possible that the mediation effect of self-compassion may be different at different levels of gratitude. Accordingly, the third set of hypotheses for our study was as follows:

Nurses were recruited via social media using convenience sampling techniques. The design of this study is cross-sectional. The questionnaire was sent to respondents via email, Facebook, WhatsApp, and Twitter. Nurses were approached through their hospitals whose managers provided their emails; nurses were then asked to send the link of the questionnaire to other nurses they knew via different media platforms. The data were collected from December 10, 2022 to January 28, 2023 in the Kingdom of Saudi Arabia. A total number of 425 participants returned the completed survey, but only 368 surveys without missing data were used in the study. The mean age of the surveyed nurses was 35.18 with a standard deviation of 6.67. 70.65% were female, 28.2% were single, 65% were married, 5.7% divorced, and 1.1% widowed. About 34.2% had a diploma degree, 8.7% had a diploma after university degree, 52.7% had a university degree, 3.8% had a master’s degree, and 0.6% possessed a doctorate. Around 83.4% were working in the government sector, 3.6% in a semi-government sector, and 13% were working in the private sector. Around 90.2% were full-time workers, 2.7% part-time, and 7.1% were working irregular hours. About 5.2% had less than 1 year of experience, 5.7% had 1 to 3 years of experience, 16.3% 3 to 6 years, 16% between 6 and 9 years, 21.2% between 9 and 12 years, 13% between 12 and 15 years, and 22.6% had more than 15 years of experience. Around 8.7% had monthly income of less than 5,000 RS (Saudi Riyal), 31.2% had income between 5,000 and 9,000 SR, 31.5% between 9,000 and 13,000 SR, 17.4% between 13,000 and 17,000 SR, 5.4% between 17,000 and 21,000, 4% between 21,000 and 25,000, and 1.6% had monthly income greater than 25,000 SR. Finally, 62% were Saudi citizens, 22.8% were residents from non-Arab countries, and 15.2% were residents from other Arab countries.

Study participants completed questionnaires which included a set of demographic questions as well as the Satisfaction with Life Scale (90), the Subjective Happiness Scale (91), the Gratitude Questionnaire–6 (GQ-6) (74), the Self-Compassion Scale (92), and the Work-Family Conflict Scale (93).

All the data analyses were conducted using the RStudio software (94). For much of the analysis, an add-on software package called Process, developed by Hayes (39), was used. Descriptive statistics and Pearson correlations between the study variables were gathered first. Secondly, a mediation analysis was conducted using Model 4 of the Process software, and a moderated mediation analysis was performed using Process Model 8. We used 95% confidence intervals with 10,000 bootstrap samples. 95% confidence intervals that do not contain zero indicate statistically significant conditional indirect effects. The Process software also allows centering of variables that are part of the product terms, to set conditional effects for different levels of the moderator variables (mean-1SD, mean, and mean + 1SD), and to bootstrap results against violations of normality and homoscedasticity assumptions. It has been established by previous studies that age and gender were significantly associated with life satisfaction and happiness (95–97); therefore, we included these variables as covariates.

All procedures followed were in accordance with the ethical standards of the responsible committees on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. Approval for conducting this study was obtained from the ethics committee of Institutional Review Board of King Abdulaziz University, Jeddah in Saudi Arabia (No 343-253-1443). Informed consent was obtained from all participants.

The results of the descriptive statistics and the Pearson correlations are summarized in Table 1. The mean scores of participants on the scales were as follows: 24.15 for life satisfaction (SD = 6.48, range = 5–35), 31.7 for gratitude (SD = 5.75, range = 5–42), 20.09 for WFC (SD = 5.62, range = 5–35), 14.71 for FWC (SD = 7.14, range = 5–35), 41.73 for self-compassion (SD = 7.02, range = 12–60), and 19.31 for happiness (SD = 4.10, range = 4–28). As expected, life satisfaction was positively correlated with gratitude, self-compassion, and happiness, and negatively correlated with WFC and FWC. Happiness was positively correlated with life satisfaction, gratitude and self-compassion, and negatively correlated with WFC, and slightly with FWC. Finally, gratitude and self-compassion were positively correlated.

To answer Hypothesis 1, that is, WFC and FWC will be negatively associated with life satisfaction and happiness, we used the mediation analysis module of the Process software package with gender and age as covariates. The results show that WFC was negatively associated with life satisfaction (β = −0.15, p < 0.05, 95% CI [−0.21 to −0.09]) and with happiness (β = −0.07, p < 0.05, 95% CI [−0.12 to −0.03]) in the absence of a mediator. FWC was similarly negatively related to life satisfaction (β = −0.16, p < 0.01, 95% CI [−0.22 to −0.11]) and happiness (β = −0.03, p < 0.05, 95% CI [−0.11 to −0.03]) in the absence of a mediator. Therefore, Hypothesis 1 was supported.

When the mediator was included, WFC was negatively associated with life satisfaction (β = −0.15, p < 0.01, 95% CI [−0.29 to −0.06]), but not with happiness. Similarly, FWC was negatively associated with life satisfaction (β = −0.17, p < 0.01, 95% CI [−0.25 to −0.07]) but not with happiness. Self-compassion was related to life satisfaction (β = 0.29, p < 0.001, 95% CI [0.17 to 0.35]), happiness (β = 0.29, p < 0.001, 95% CI [0.23 to 0.34]), WFC (β = −0.30, p < 0.001, 95% CI [−0.50 to −0.26]), and FWC (β = −0.31, p < 0.001, 95% CI [−0.40 to −0.21]). As shown in Table 2, these relationships were established in the presence of covariates.

To test the mediation model, we used the bootstrap method and the results are summarized in Table 3. The indirect effect of self-compassion was −0.10 (95% CI = −0.13 to −0.04) on the relationship between WFC and life satisfaction, −0.08 (95% CI = −0.14 to −0.04) on the relationship between FWC and life satisfaction, −0.15 (95% CI = −0.21 to −0.09) on the relationship between WFC and happiness, and − 0.16 (95% CI = −0.22 to −0.10) on the relationship between FWC and happiness. None of these confidence intervals contained the value zero, which indicates that the indirect effects were statistically significant. Therefore, our hypotheses that self-compassion will mediate these relationships (Hypotheses 2a, 2b, 2c, and 2d) were supported.

To check the gratitude as moderator hypotheses (3a through 3f), we used the moderated mediation analysis module of the Process software package with gender and age as covariates. WFC was negatively associated with self-compassion (β = −0.28, p < 0.001), and the interaction with gratitude was statistically significant (β = −0.03, p < 0.001). Thus, Hypothesis 3a was supported. WFC was also negatively associated with life satisfaction (β = −0.13, p < 0.05), but the interaction with gratitude was not statistically significant. Thus, Hypothesis 3b was not supported. On the other hand, WFC was not related to happiness and the interaction with gratitude was also not statistically significant. Therefore, Hypothesis 3c was also not supported. FWC was negatively related to self-compassion (β = −0.21, p < 0.001) and the interaction with gratitude was not statistically significant. Therefore, Hypothesis 3d was not supported. FWC was negatively related to life satisfaction (β = −0.12, p < 0.01) and the interaction with gratitude was not statistically significant. Therefore, Hypothesis 3e was also not supported. Lastly, FWC was negatively associated with happiness (β = −0.06, p < 0.01) and the interaction with gratitude was statistically significant as well (β = −0.009, p < 0.01). Thus, Hypothesis 3 was supported (Table 4).

The indirect effects of self-compassion at different levels of gratitude are summarized in Table 5. The indirect effect of self-compassion on the relationship between WFC and life satisfaction was not statistically significant for participants with low levels of gratitude (β_ind = −0.009, 95% CI = −0.002 to 0.0007), but was statistically significant for participants with medium levels of gratitude (β_ind = −0.02, 95% CI = −0.06 to −0.001) and with higher levels of gratitude (β_ind = −0.05, 95% CI = −0.011 to −0.003), which indicates that the indirect effect of self-compassion on the relationship between WFC and life satisfaction was stronger for those with higher levels of gratitude. Similarly, the indirect effect of self-compassion on the relationship between WFC and happiness was not statistically significant for participants with low levels of gratitude (β_ind = −0.01, 95% CI = −0.04 to 0.01), but was statistically significant for participants with medium levels of gratitude (β_ind = −0.05, 95% CI = −0.08 to −0.03) and higher levels of gratitude (β_ind = −0.10, 95% CI = −0.015 to −0.06). Thus, the indirect effect of self-compassion on the relationship between WFC and happiness was also stronger for those with higher levels of gratitude.

The indirect effect of self-compassion on the relationship between FWC and life satisfaction was not statistically significant for participants with low levels of gratitude (β_ind = −0.01, 95% CI = −0.03 to 0.0008), medium levels of gratitude (β_ind = −0.02, 95% CI = −0.04 to 0.0006) or higher levels of gratitude (β_ind = −0.02, 95% CI = −0.06 to 0.0007). The indirect effect of self-compassion on the relationship between FWC and happiness was statistically significant for participants with low levels of gratitude (β_ind = −0.03, 95% CI = −0.06 to −0.003), medium levels of gratitude (β_ind = −0.04, 95% CI = −0.07 to −0.02) and higher levels of gratitude (β_ind = −0.06, 95% CI = −0.09 to −0.03). Thus, the indirect effect of self-compassion on the relationship between WFC and life satisfaction was beneficial for all levels of gratitude. These conditions are plotted in Figures 1, 2.

This study has theoretical and practical implications. For theoretical implications, this study contributed to the literature by establishing evidence of the role of self-compassion and gratitude in protecting nurses from work-family conflicts among a Saudi sample. These findings have practical implications for organizations in general and hospitals in particular. For hospital HR, it is important to plan programs to enhance self-compassion and gratitude of the nurses. These positive psychology interventions have proved their effectiveness, and it is crucial that Saudi hospitals plan effective interventions to promote nurses’ well-being, which would increase their performance. Gratitude and self-compassion are trainable and can therefore be practiced and enhanced, and they present affordable targets for interventions to enhance the well-being of workers in general and nurses in particular. Therefore, managers should consider bringing into the workplace positive psychological interventions focusing on growing gratitude and self-compassion. Further, workers themselves should learn about these positive interventions and attempt to integrate them into their psychological, work, and family lives.

Despite the contributions of this study, there are limitations that should be mentioned. First, the design was cross-sectional. Longitudinal designs are recommended for future research. Second, the study used convenience sampling and may have yielded sampling and selection bias, which hampers generalizability of findings; random sampling is recommended for future research. Third, the proportion of females was relatively high in the sample and it is advised that a more proportionate sample be tested. Fourth, the Subjective Happiness Scale has a relatively low Cronbach’s alpha and results should be interpreted bearing this in mind. Fifth, the data collection was made via online platforms, we cannot know in which conditions the respondents were when the completed the questionnaire. Further, the recruiting process via social media is beyond any control of who has participated and what the respondents are representing. Future research should use also other means of data collection to facilitate data cross-validation. Sixth, this study relied on subjective data, future research should include also objective measures.