Job changes are significant life stressors, and losing a job takes several years to recover psychologically (9). According to the family stress model (10), economic instability (including unemployment, debt, and receipt of income transfers) places considerable strain on parents. The COVID-19 pandemic has forced parents to quickly transition to a new way of work (11). These transitions present all kinds of stressors, which might have repercussions on parents' capacities to be psychologically flexible, thus putting their QoL at risk (11).

Active opposition of various forms, including verbal, physical, sexual, financial, or psychological, among family members is often referred to as family conflict (12). According to adult attachment theory, high-quality close relationships, including marital relationships and parent–child relationships, have been consistently linked to better health, while conflicting relationships may influence parents' physiological responses to stress, risky health behaviors, susceptibility to physical illness, and poorer disease outcomes (13). When parents were asked to stay at home during the pandemic, they faced problems getting along with their family members and dealing with conflicts with their partner and children. There is evidence that a close relationship between family members could affect parents' QoL, and parents' health status can further affect children's development and well-being, creating a vicious circle in such a difficult period (14, 15).

Theories and empirical results based on role stress theory indicate that the combination of family and employment often creates more demands than one can handle, leading to role overload (16). Consequently, both job changes and family conflict can bring more stress and psychological problems for parents of young children.

The interaction effect of job changes and family conflicts on QoL may be different for mothers and fathers. COVID-19 has indeed changed parents' experiences with employment outside the household and the division of labor at home (17). Compared with fathers, mothers have reduced their working hours more and spent more time on childcare duties (18). Moreover, school and daycare closures increased caregiving responsibilities. Arguably, these changes increased the burden on women more than on men (19). Therefore, this work anticipates that mothers have experienced more job changes, perceived more family conflicts, and low QoL than fathers.

Additionally, the interaction role between job changes and family conflict on QoL may differentiate across the number of children. The traditional belief, “more children, more happiness,” is still entrenched in China (20). Evidence has shown that parents can obtain more emotional support from their children in a family with multiple children (21), and parents with more children perceive a higher level of life satisfaction. In addition, parents tend to rear more children if they can easily integrate work and family roles to reduce the double burden (22). Thus, the association between family conflict and QoL in a multichild family may not be enhanced by job changes.

The objective of this work was to examine the relationship between job changes, family conflict, and QoL among parents with a child aged 0–10, to assess the interaction effects between job changes and family conflict, and to explore the difference in gender and family size among them. We hypothesized that either job changes or family conflicts correlated with parents' poor QoL. We further hypothesized that pandemic-related job changes enhanced the relationship between family conflict and poor QoL. Moreover, we assumed that the abovementioned moderating effects varied across genders and families with different numbers of children.

This study was conducted based on an online survey (a web-based platform, https://www.wjx.cn/app/survey.aspx) from April 21 to April 28, 2020, during the COVID-19 lockdown in China. To nationally reflect the mental health status of parents affected by the pandemic and enlarge the sample size, Hubei province, the area most affected by COVID-19, and its neighboring province Henan and non-adjacent province Guangdong were selected for sample investigation.

We used various sampling techniques to reduce bias in situations where there were large populations in China. Random sampling is practically impossible because families have difficulty reaching during the pandemic. The advantages of convenience sampling approaches, such as collecting data quickly, were suitable for data collection during pandemics with school closures. In addition, we used cluster sampling because this method is usually adopted to recruit subjects in the school. Convenience sampling was used to select primary schools. The cluster sampling method was used to choose participants from selected schools. The selected schools were stratified by grade. Then, all classes were selected from each grade of the selected school, and all the students in these classes joined the survey. The head teachers helped us distribute the questionnaires to the subjects online.

Parents aged 18 years and above who had a child or children aged between 0 and 10 years old were invited to participate. The rationale for restricting the sample to parents with at least one 10-year-old or younger child was that younger children necessitate more direct supervision; therefore, it was expected that the COVID-19 pandemic may pose more specific challenges to them, such as more work–family conflict. These parents were instructed about how to answer the questionnaires by online guidance and to complete the questionnaires independently. In total, 1,286 parents participated in the survey. As 77 parents were excluded because they had no job, the final number of participants became 1,209. Each question of the online questionnaire was required to be answered; thus, no missing data were reported in our study. The study protocol was approved by the Institutional Review Board of Peking University, Beijing, China. All participants gave electronic consent after being informed of the aims of the survey and their rights to refuse to participate.

Quality of life was assessed with the global health items in the patient-reported outcomes measurement information system (PROMIS) (23). The checklist includes 6 items, which were answered on a five-point scale, Excellent(1)–Poor(5). See the attachment for details. The total scores were calculated by the sum of all items, with a higher score indicative of better QoL. Cronbach's alpha of QoL in this study was 0.916.

Job changes were assessed with a subset of items from the questionnaire of the COVID-19 and perinatal experiences study (COPE Study, https://osf.io/uqhcv/). The scale contained 15 questions about the changes in jobs related to the COVID-19 pandemic, such as telecommuting, reduced working hours, extended working hours, pay cuts, increased responsibility, increased supervision and reporting, and declining job security. For each question, 0 represented “No,” while 1 represented “Yes.” All items were summed to obtain a total score for job changes. A higher score represented more job changes. The internal consistency of job changes was 0.881.

Family conflict was measured with a 6-item self-made questionnaire about specific conflict behaviors, including marital conflict and parent–child conflict. Among them, three questions examined marital conflict, and three other items examined parent–child conflict. Each item was rated on a five-point scale ranging from never (1 score) to always (5 scores). For details, please page to attachment. Higher scores indicated more frequent marital and parent–child conflict. The Cronbach's α were 0.696 and 0.726, respectively. The questionnaire items were vetted by some sociological professors, and they agreed that there were no items that were culturally insensitive.

Confounding variables included demographic and socioeconomic characteristics as follows: exposure to COVID-19 (yes, no), gender (male, female), province (Hubei, Henan, Guangdong, else), number of children (1, 2, 3 or more), education (junior school and below, high school, college, undergraduate, and above), marital status (married, else), occupation (stable, unstable), and annual income ( ≤ 50,000, 50,000–1,00,000, ≥1,00,000). In the study, this term (stable job) refers to work that is continuous and safe employment, and there are no sudden layoffs or labor strife. By unstable jobs we mean work that is uncertain, insecure, irregular, and in which employees bear the risks of work and receive limited social benefits, relatively low income, and statutory protections (i.e., In China, employees work in state-owned enterprises, and public institutions are regarded as having a stable job. Unstable job employees refer to freelancer and gig workers). All covariates were selected according to previous related studies (24, 25).

Data in this study were analyzed with SPSS version 22.0. Descriptive statistics were calculated to describe the demographic characteristics of the subjects. Since QoL as the dependent variable is continuous, multiple linear regression analysis was performed, and each model controlled the same confounding variables, including exposure, gender, province, number of children, education, marital status, occupation, and annual income.

The main analyses consisted of multiple regressions on QoL in two steps. In model 0, each of the predictors was included separately to estimate their “raw” contribution to the prediction of QoL, and in model 1, the three predictors of job changes, marital conflict, and parent–child conflict were introduced to examine which component would be the most powerful predictor.

In the next multivariable linear regression analysis, job changes, marital conflict, and parent–child conflict and their interaction terms (job changes^*marital conflict and job changes^*parent–child conflict) were all standardized and entered to estimate the moderating effects of job changes on the relationships between marital conflict, parent–child conflict, and QoL.

In addition, the whole sample was divided into 2 groups by gender to examine gender differences in the moderating effects of job changes on the relationships between marital conflict, parent–child conflict, and QoL. In the final linear regression models, the whole sample was divided into 3 groups by the number of children in a family.

Several limitations of this study should be acknowledged. First, the selective bias resulting from convenient sampling methods used in our study might constrain the generalizability of our findings. Second, the proportion of fathers was <30% in our study, which may suggest that we may miss the data from fathers who are quiet busy with their job or other things. Thus, we should be cautious in generalizing our results. Third, our study only included limited job change forms without consulting vocational psychology, which may fail to fully reflect the relationship between job changes and young parents' QoL. Fifth, with the cross-sectional design of the current research, it is difficult to make a causal inference. Longitudinal designs are expected in future studies to help clarify the causal relationships.

Despite these limitations, the findings of this study have significant implications. First, the QoL among young parents experiencing job changes and family conflict should be given adequate attention during COVID-19 confinement, and corresponding proactive and applicable interventions can be proposed. For example, community organizations and social workers should pay more attention to the prevention of family conflict. In addition, fathers and one-child families need to implement psychological interventions (that mitigate marital conflict) and work insurance (that reduce financial pressure).