Abstract
Keywords
Introduction
The transition to parenthood is a significant life event that introduces new responsibilities and roles (Flaherty & Sadler, 2022), often leading to stress due to the demands of child rearing. As parental responsibilities increase, along with societal expectations, the pressures of parenting can become overwhelming, potentially undermining the quality of parenting. Parenting stress is defined as a negative psychological and physical response to the challenges of parenting (Deater-Deckard, 2008), and it tends to intensify when there is a discrepancy between the expected and available resources. Parents may feel overwhelmed when they perceive themselves as underinformed, less competent, or lacking emotional and practical support from partners, family, or friends (Deater-Deckard, 1998). Furthermore, modern social factors such as evolving values, increased workforce participation by women, and diverse family structures further exacerbate this stress, compounding the challenges new parents face.
Maternal parenting stress is particularly common, and failure to manage effectively can negatively affect child development (Rajgariah et al., 2021). Elevated maternal parenting stress is linked to negative outcomes, such as increased internalizing and externalizing problem behaviors in children (de Maat et al., 2021; Dey et al., 2021; Jiang et al., 2023; Neece et al., 2012; Owen et al., 2006) and impairments in executive functioning and problem-solving skills (Park et al., 2023; Tan et al., 2017). Additionally, maternal parenting stress can adversely affect children’s social–emotional development (Anthony et al., 2005; Carapito et al., 2020; Colletti et al., 2008; Ostberg & Hagekull, 2013) and their adjustment to school (Conger et al., 1995; Rogers et al., 2009). These findings underscore the importance of implementing effective strategies to reduce maternal parenting stress in order to support children’s adaptive behavior and supporting their cognitive, social–emotional, and academic development during critical growth stages.
While it is well established that maternal parenting stress plays a significant role in shaping child development, much of the existing research has relied on cross-sectional data, providing a static view rather than a developmental perspective over time. Although some studies have explored long-term patterns of maternal parenting stress, their findings have been inconsistent (Chang & Fine, 2007; Park et al., 2023; Stone et al., 2016; Williford et al., 2007), pointing to meaningful variation in how mothers experience and manage parenting stress. These inconsistencies suggest that maternal parenting stress may not follow a single, uniform pattern across individuals. To address this gap, the current study adopts a longitudinal approach, using latent class analysis to identify distinct patterns of maternal parenting stress. Moreover, considering the dynamic nature of the mother–child relationship in the context of maternal employment, this study explored how these stress trajectories differ by maternal employment status.
Maternal Parenting Stress of Employed and Unemployed Mothers
Since maternal parenting stress is heavily shaped by the conditions and contexts surrounding mothers, maternal employment plays a critical role in influencing it. Like many industrialized countries, South Korea has experienced a significant rise in maternal employment in recent decades, influenced by both growing female labor force participation and economic demands that necessitate continued employment after childbirth. However, several South Korean mothers lack access to extended family support, such as assistance from grandparents or other relatives, which exacerbates the challenges of child rearing. This absence of support has contributed to South Korea’s notably low fertility rate, which has decreased from 4.53 children per woman in 1970 to a projected 0.72 in 2023, marking the sharpest decline among Organization for Economic Cooperation and Development (OECD) countries (Statistics Korea, 2024b).
Despite a significant drop in fertility, the proportion of working mothers living with children aged under 18 in South Korea reached 62.4% in 2024, showing a gradual upward trend (Statistics Korea, 2024a). However, a national gender equality survey revealed that 68.9% of wives primarily assume responsibility for household chores, and 32.2% of mothers reported difficulties in balancing work and childcare obligations (Ministry of Gender Equality and Family, 2021). These data suggest that maternal employment significantly influences maternal parenting stress, indicating a need for national policies and support systems that can better address the unique challenges faced by working mothers.
Employed mothers often struggle to balance multiple responsibilities, such as household duties, childcare, and professional obligations (Yeo & Teo, 2013). Their employment typically leaves them with limited time for their children, which increases the strain of managing both their work and family lives (Kim et al., 2013). The need to divide psychological and physical energy between these competing roles can lead to exhaustion (Forgays & Forgays, 1993; Kim et al., 2013). Moreover, in South Korea, the Confucian ideal that mothers should primarily act as caregivers intensifies feelings of guilt among working mothers who feel that they are not fully engaging with their children (Hong & Lee, 2019).
Contrary to the common assumption that stay-at-home mothers enjoy a more relaxed lifestyle than their employed counterparts, their daily routines are often equally demanding. Unemployed mothers frequently set higher expectations for their children, leading to a unique set of stressors that differ from those experienced by working mothers (Crnic & Greenberg, 1990). By dedicating themselves fully to child rearing, unemployed mothers may struggle to obtain adequate support because domestic responsibilities are often undervalued compared with external employment (Aslam & Adams, 2022; Dillaway & Paré, 2008). Additionally, the continuous presence of unemployed mothers with their children can limit mothers’ opportunities for adult interaction, thereby intensifying the stresses associated with parenting (Crowley, 2014).
Research on maternal parenting stress in relation to maternal employment status presents a complex and varied picture. Some studies indicate that working mothers face higher levels of parenting stress than their stay-at-home counterparts, suggesting that the combined demands of professional and domestic responsibilities contribute to this increase in stress (Dey et al., 2021; Rajgariah et al., 2021). Conversely, other research shows that stay-at-home mothers may experience greater parenting stress owing to the monotonous nature of household tasks, the absence of external rewards, and social isolation from friends and family (Crnic & Greenberg, 1990; Kahn & Cuthbertson, 1998; Song et al., 2014; Yoon et al., 2009). Additionally, some studies have found no significant differences in parenting stress levels between employed and stay-at-home mothers (Forgays & Forgays, 1993; McCarten, 2003; Yeo & Teo, 2013). These varied findings suggest that employed and stay-at-home mothers may experience parenting stress in fundamentally different ways, indicating that these groups should not be considered a single, homogeneous category in studies of maternal parenting stress. Employed and stay-at-home mothers are distinct groups with unique stress-related characteristics. Therefore, this study examined the potential differences between these groups by conducting a measurement invariance test to assess whether such distinctions exist within the latent structure.
Purpose of This Study
Several previous studies have shown that maternal parenting stress generally decreases over time following childbirth (Berryhill & Durtschi, 2017; Chang & Fine, 2007; Park et al., 2023; Pope et al., 2005; Williford et al., 2007). However, these studies also demonstrated significant variance, suggesting that the longitudinal patterns of maternal parenting stress can vary widely among individuals (Chang & Fine, 2007; Park et al., 2023; Stone et al., 2016; Williford et al., 2007). Therefore, this study identified distinct trajectories of maternal parenting stress among different subgroups using a latent class analysis.
Furthermore, this study used a multi-group analysis to investigate whether the latent classes differ by employment status. Although previous research has examined differences in maternal parenting stress patterns based on employment status, it has not explored these differences within the context of latent classes. For example, Park and Park (2016) showed that stay-at-home mothers experienced higher levels of parenting stress than employed mothers immediately after childbirth, with significant increases in stress until the child reached the age of three. Meanwhile, employed mothers’ stress levels showed no significant changes over the same period.
Our study aimed to address this gap in the literature by identifying latent classes in the longitudinal patterns of maternal parenting stress and determining whether these classes and their characteristics vary based on maternal employment status. In the presence of such differences, the study explored their possible underlying causes. Our research questions were as follows:
Method
Data
In this study, we used data from the Panel Study on Korean Children, conducted by the Korea Institute of Child Care and Education. We compared the classification of the latent classes for maternal parenting stress among groups of mothers with young children, based on maternal employment status. The nationally representative panel data set collects annual surveys from a cohort of children born in 2008 and tracks their personal characteristics and the environmental factors related to their parents, households, and schools as they grow. The longitudinal panel data are ideal for identifying latent classes of maternal parenting stress at multiple time points and assessing the consistency of these classifications between employed and unemployed mothers. This is because the data include repeated annual surveys of the same children and households.
To classify maternal parenting stress based on changes over time, more than three time points of longitudinal data are required (Curran et al., 2010). Accordingly, this study utilized data from the third to the eighth waves of the Panel Study on Korean Children (PSKC), tracking children from age three to eight. The analysis began at the third wave, corresponding to when the children were 3 years old—a key developmental and social transition point in South Korea. At this stage, many children begin attending daycare centers or preschools, typically coinciding with the conclusion of parental leave. Therefore, we began collecting data from the third wave, when the children were 3 years old—a common age for mothers to return to work after parental leave and for children to start preschool or daycare programs (Blum et al., 2023).
According to the Ministry of Education (2023), as of 2020, the enrollment rate in early childhood education—including kindergartens and childcare centers—for children aged 3 to 5 was 94.0%. This widespread participation reflects the norm of institutional early education beginning around age three, which in turn facilitates maternal return to the workforce by providing structured childcare. Entry into elementary school usually occurs at age six or seven (first grade), bringing additional changes in family routines and educational expectations. These shifts in the caregiving and educational landscape may influence the trajectory of maternal parenting stress throughout early childhood.
The final sample consisted of 1,796 mothers, including only those with verifiable maternal employment status at the third wave. At that time, the average age of mothers was 32.8 years (
Measures
Response Variable: Maternal Parenting Stress
Maternal parenting stress was assessed using a scale designed to measure the burden and distress experienced by mothers in their parental role. We calculated the mean score of the 11 items on the scale and used it as an indicator variable across 6 time points. Each item was rated on a 5-point Likert scale, ranging from 1 (Not at all) to 5 (Very much), with higher scores indicating a greater level of burden and distress related to the parental role. The contents of each item used to measure maternal parenting stress, along with the descriptive statistics for each time point categorized by maternal employment status, are presented in Table 1. To ensure the reliability of the scale, Cronbach’s alpha was calculated separately for each group.
Details of Items Related to Maternal Parenting Stress.
The Third to Eighth waves correspond to child ages 3 to 8.
Cronbach’s alpha.
Data Analysis
This study used a multiple-group latent class analysis (MG-LCA) to explore whether classifications of latent classes for parenting stress types remain consistent across groups of mothers differentiated by employment status. As an extension of latent class analysis, MG-LCA enables comparisons of latent class structures across different groups. Following the method proposed by Kankaraš et al. (2010), we analyzed the measurement invariance of parenting stress between employed and unemployed mothers at the latent and indicator levels via four key steps.
First, we identified the optimal number of latent classes for each group based on employment status. We used various information criteria (ICs) to determine the best-fitting model, including the Akaike information criterion (AIC; Akaike, 1974), Bayesian information criterion (BIC; Schwarz, 1978), consistent AIC (CAIC; Bozdogan, 1987), and sample-size adjusted BIC (SABIC; Sclove, 1987), with lower values indicating a better model fit. Additionally, we employed model comparison tests, such as the Lo–Mendell–Rubin adjusted likelihood ratio test (LMR-LRT; Lo et al., 2001) and bootstrapped likelihood ratio test (BLRT; Peel & McLachlan, 2000), to verify the significance by comparing models with
Second, once we confirmed that the number of latent classes was consistent across groups, we applied MG-LCA to test measurement invariance at the latent level using the combined data set. During this process, we included employment status as a categorical covariate in the latent class model to represent the groups being compared. This method is considered the most straightforward for incorporating group effects in latent class models (Kankaraš et al., 2010). The group variable can have both direct and interaction effects on each indicator and latent variable (McCutcheon & Hagenaars, 1997). Direct effects are the group variable’s immediate influence on the indicator variable, whereas interaction effects arise when the group variable moderates the latent variable’s effect on the indicator variable. By progressively constraining these direct and interaction effects, we created three types of nested models. First, the non-invariance model allows both direct and interaction effects. Second, the partial invariance model allows direct effects but constrains interaction effects across groups. Third, the structural invariance model constrains both direct and interaction effects. We sequentially compared these models to determine whether the latent class structure of maternal parenting stress was equivalent across the two groups at the latent level based on ICs and model comparison tests, as in the first step. In this hierarchical model comparison, if a more constrained model was supported, we considered it to be a more parsimonious and statistically sound model theoretically and statistically (Eid et al., 2003; Sohn, 2016).
Third, when the non-invariance or partial invariance models were supported, we tested measurement invariance at the indicator level to identify which indicators contributed to the lack of invariance. As in the first step, we evaluated model fit using ICs and model comparison tests. We also examined the values of bivariate residuals, calculated based on the Pearson chi-square statistic, to compare the chi-square statistic corresponding to the difference in degrees of freedom between the baseline model and the comparison model (Vermunt & Magidson, 2021). Bivariate residuals help assess model fit by comparing models that include or exclude specific variables and are particularly useful in deciding whether to include direct effects (Moors, 2004). Although bivariate residuals do not directly assess interaction effects, excessively large residuals may suggest the presence of unmodeled interaction effects, thus offering an indirect means of evaluation (Brown, 2015). We used Latent Gold 6.0 software (Statistical Innovations, 2021) for these analyses because it can analyze complex latent class models with hierarchical structures and multiple groups. The research model for this study is shown in Figure 1, in which the red arrows represent the direct effects of the group variable on the indicator variables, yellow arrows indicate the interaction effects, and blue arrows show the effects of the group variable on the latent variable.
Research model.
Results
Determining the Optimal Number of Latent Classes for Maternal Parenting Stress by Group
To determine the optimal number of latent classes by group for maternal parenting stress, differentiated by maternal employment status, we systematically increased the class count from two to five. For each model, we analyzed the ICs, model comparison tests, entropy, classification error, and proportions of each latent class. The findings from these analyses appear in Table 2.
Model Comparison Results by Number of Latent Classes for Each Group.
Bolded values represent the final selected latent class solution for each group.
In terms of ICs (AIC, BIC, CAIC, and SABIC), all groups exhibited decreasing values with the addition of more latent classes. When the values decrease continuously, the optimal number of latent classes is typically found at the “elbow point,” where the rate of decrease is the sharpest (Nylund et al., 2007). In our study, the largest reduction in ICs occurred when the number of latent classes rose from two to three, as shown in Figure 2. For the classification quality indicators, the highest entropy values for both groups were observed with two latent classes. However, the model comparison tests (BLRT and LMR-LRT) indicated significance for all class solutions, demonstrating that these tests alone were not adequate to determine the optimal number of classes. Therefore, we needed additional criteria to make an informed decision (Nylund-Gibson et al., 2014). Moreover, in the three-class model, the smallest latent class accounted for 21.2% of employed mothers and 25.0% of unemployed mothers, satisfying the minimum 5% threshold for latent class proportions (Berlin et al., 2014).
Changes in information criteria with increasing number of latent classes by group.
After considering both statistical indicators and interpretability, we selected a three-class model as the most appropriate for both groups. This model demonstrated substantial decreases in ICs and high classification quality beyond the two-class solution.
Before testing measurement invariance for the identified latent classes across groups, we examined the characteristics of each class separately for employed and unemployed mothers based on the coefficients for maternal parenting stress. Figure 3 shows the latent class profiles for each group. Both groups exhibited a similar classification pattern across three classes: high, moderate, and low levels of parenting stress. Specifically, the first latent class, which represents the highest level of parenting stress, accounted for the largest proportion in both groups (employed mothers: 47.8%, unemployed mothers: 40.4%). The second class showed a moderate level of stress (employed mothers: 30.9%, unemployed mothers: 36.6%), and the third class exhibited the lowest level of parenting stress (employed mothers: 21.2%, unemployed mothers: 25.0%).
Latent class profiles by group.
Quantitative differences were observed between the groups in terms of their initial stress levels and rate of change over time. Initial stress levels for unemployed mothers were slightly higher across all classes compared with those of employed mothers. Additionally, although both groups showed an overall decrease in stress levels over time, the rate of decrease varied. Employed mothers in the high-stress class experienced a steeper reduction in stress levels after the fifth wave, but the decline for unemployed mothers was more gradual. This suggests that stress levels among high-stress employed mothers tended to stabilize more quickly over time, whereas stress reduction for unemployed mothers occurred at a slower pace.
Results of the Multi-Group Latent Class Analysis
All groups were confirmed to have three latent classes in the classification of maternal parenting stress. Based on this finding, we conducted an MG-LCA using the combined data set, with employment status as a grouping variable. We applied the method for testing measurement invariance proposed by Kankaraš et al. (2010), and the results at the latent level are presented in Table 3. By selecting the model that best fits the data, we can determine the degree of measurement invariance. This model selection helps clarify whether the similarity between groups is due to the types of latent classes themselves or the conditional response probabilities within each class.
Model Comparison Results by Number of Latent Classes for Each Group.
Bolded values represent the final determined measurement invariance model.
A comparison between the non-invariance and partial invariance models revealed that all ICs were the smallest for the partial invariance model. Although the result of the BLRT was significant, that of the LMR-LRT was not. Based on these results, the partial invariance model, which constrains the interaction effects between the group and latent variables to be equal across groups, was deemed more suitable than the non-invariance model. Subsequently, a comparison between the partial and structural invariance models showed that, except for the AIC and LMR-LRT, all ICs were lower in the structural invariance model, and the result of the BLRT was significant. This indicates that the direct effects of the group variable on indicator variables were also absent across groups. Accordingly, the structural invariance model, which constrains both direct and interaction effects to be equal across groups, was selected as the final model. This final model showed that the latent class types and response patterns for indicator variables were equally consistent across groups.
Although we selected the structural invariance model, we also conducted an indicator-level measurement invariance test to identify potential sources of non-invariance among the indicator variables. To accomplish this, we compared the structural invariance model (Model 3) with a series of alternative models (Models 3-1 to 3-6), allowing the direct effects of the group variable on each indicator variable to vary across groups.
Table 4 shows that allowing the direct effect of the group variable on the third-wave indicator variable to differ across groups resulted in lower AIC and SABIC values compared with those of the structural invariance model. Additionally, the BLRT
Results of the Indicator-Level Measurement Invariance Analysis for Maternal Parenting Stress Classification Across Groups.
Bolded values represent the final selected indicator variables causing latent non-invariance.
In summary, these results demonstrated that the direct effect of the group variable on the third-wave indicator variable was not consistent across groups. This finding suggests that the group variable, which represents employment status, may affect this specific indicator independently of the latent variable.
Based on the results that support the invariance of latent classes for maternal parenting stress across groups, we analyzed the differences in the response categories for each indicator variable, as shown in Figure 4. Greater discrepancies in the graph’s shape suggest potential differences in the perceptions of indicator variables between groups. One indicator variable (third wave), identified in the model comparisons as having a non-invariant direct effect across groups, exhibited larger differences in response categories—particularly in categories 3 and 4—compared with the invariant indicators (fourth to eighth waves). This finding suggests that differences in response categories for the single non-invariant indicator cannot be fully explained by differences in latent class membership alone, implying the likelihood of specific group differences in the response patterns for this particular indicator variable.
Group differences in responses to levels of maternal parenting stress.
Conclusion and Discussion
This study examined the latent class patterns of maternal parenting stress over a 6-year period among mothers with 3-year-old children, focusing on classification differences based on maternal employment status and testing for measurement invariance. We conducted an MG-LCA to assess measurement invariance at both the latent and indicator levels and to identify indicator variables that contribute to non-invariance. Below, we summarize the detailed findings, discuss implications, and outline limitations and directions for future research.
Summary of Findings
First, to verify measurement invariance in the latent class classification of maternal parenting stress across groups, we conducted a latent class analysis separately for each group to identify the optimal number of latent classes. The analysis showed that both groups exhibited the same optimal solution of three latent classes (high, moderate, and low), suggesting that the latent class structure for assessing maternal parenting stress was consistent across groups. Although the classification patterns of the latent classes were similar, quantitative differences were observed between the groups, particularly in how stress levels changed over time.
Across all classes, unemployed mothers showed slightly higher initial stress levels, which may be attributed to their limited access to external support and a lack of structured daily routines—factors known to contribute to prolonged parenting stress (Gallie & Russell, 2009; Nomaguchi & Milkie, 2003). In contrast, employed mothers showed a more gradual reduction in stress levels over time, especially in the high-stress class, suggesting that employment may provide access to resources, social networks, and structured schedules that help manage stress more effectively (Nomaguchi & Milkie, 2020; Parkes et al., 2015). Additionally, the slower decline in stress levels observed among unemployed mothers, particularly in the high-stress class, may reflect the restricted access to structured supports and social resources, which are more accessible to employed mothers (Rizzo et al., 2013).
Given the higher initial stress levels observed among unemployed mothers, these findings emphasized the need for early, targeted interventions, such as community-based support programs, mental health resources, and structured social networks, to effectively alleviate long-term stress from the outset. For employed mothers, policies that promote work–life balance, including flexible hours and accessible childcare, may further aid in reducing parenting stress by reinforcing the stability provided by employment. Future research should examine these dynamics in greater detail, focusing on how specific support mechanisms differentially affect parenting stress trajectories in employed versus unemployed mothers.
Second, because both groups were confirmed to have the same number of latent classes, a measurement invariance analysis conducted at the latent level to determine whether the latent class structures for maternal parenting stress were consistent across employed and unemployed mothers. The results supported the structural invariance model as the final model. This finding indicates that the interaction effects between the latent variable (maternal parenting stress) and the group variable (employment status), as well as the direct effects of employment status on indicator variables, were equivalent across groups. In other words, regardless of employment status, mothers showed similar response patterns and comparable stress trajectories over time (Kankaraš et al., 2010; Morin et al., 2016).
The observed structural consistency suggests that maternal parenting stress is shaped by a common psychological framework that persists across employment contexts. This suggests that although employment status may shape how intensely stress is experienced, it does not appear to change the fundamental structure of maternal parenting stress. Therefore, interventions that address core elements—such as emotional strain and the burden of multiple roles—may be broadly effective across mothers in varied socioeconomic contexts. Furthermore, the consistency in direct effects across groups suggests that mothers interpreted and responded to stress indicators at a similar baseline level, reinforcing that the foundational aspects of stress perception and response are unaffected by employment status (Hair et al., 2006; Steenkamp & Baumgartner, 1998). This invariance model showed that the maternal parenting stress structure was inherently robust across employment groups, which may simplify the development of standardized support programs. However, further exploration could help identify any subtle, context-specific differences that may require targeted adaptations for each group, ensuring that the support strategies address individuals’ unique needs effectively.
Finally, although structural invariance was supported overall, a further analysis of measurement invariance at the indicator level showed a direct effect of the group variable at the third time point, when the child was 3 years old. This finding suggests that the likelihood of mothers’ responses at this specific age might differ between groups owing to additional factors specific to each group that are unrelated to the identified latent class structure. While the overall classification of maternal parenting stress levels remained consistent across groups, unique characteristics associated with the child being 3 years old might affect how employed and unemployed mothers perceived and reported stress.
These findings underscore the complex interplay between maternal roles and external demands. The consistent latent class structures across groups suggest that foundational aspects of maternal parenting stress are not merely situational, but instead reflect broader psychological processes common to mothers regardless of employment status. This extends previous research (e.g., Crnic & Greenberg, 1990; Forgays & Forgays, 1993), which often emphasized dichotomous group comparisons, by revealing nuanced patterns within groups. Moreover, the non-invariance detected at the child’s age of three may point to a developmental inflection point, where external obligations (for employed mothers) and intensive caregiving demands (for unemployed mothers) diverge most starkly. Thus, this study contributes to the literature by showing not only the structural similarity in maternal parenting stress but also identifying periods of heightened divergence that warrant targeted policy and psychological interventions.
The age of three represents a significant developmental milestone, characterized by increased demands related to physical activity, early socialization, and growing communication skills, all of which can add to parental stress (Deater-Deckard, 2008). At this stage, children develop greater physical mobility and curiosity, necessitating closer supervision, and they begin to interact more with peers, testing social boundaries (Siegler et al., 2010). Additionally, their emerging communication skills lead to more frequent expressions of independence and assertiveness, both verbally and physically, which can further contribute to elevated maternal parenting stress (Cole et al., 2004; Keane & Calkins, 2004). For employed mothers, balancing these demands with work responsibilities may heighten stress owing to time constraints and fewer opportunities for hands-on caregiving, potentially amplifying work–family conflict during this stage (Chung & van der Horst, 2018; Cooklin et al., 2015). In contrast, unemployed mothers may experience different stressors tied to full-time caregiving, potentially feeling the effects of limited structured routines or external support, both of which are known to contribute to sustained stress (Berry & Jones, 1995; Nomaguchi & Milkie, 2020).
Implications
These findings indicate that the structural invariance of maternal parenting stress remained consistent across groups during the developmental period spanning ages three to eight. The absence of significant differences in parenting stress levels between employed and non-employed mothers challenges existing theoretical expectations. On one hand, theories such as the multiple stressor theory (McBride, 1990) or role strain theory (Goode, 1960) suggest that employed mothers would experience higher stress due to occupying multiple roles. On the other hand, the role accumulation theory (Sieber, 1974) posits that employment may buffer stress by providing additional rewards. Instead, the results in this study align more closely with theoretical perspectives such as the role context approach (Moen et al., 1989). This framework suggests that maternal parenting stress is influenced not merely by employment status or the number of roles occupied, but by qualitative aspects of these roles—for instance, individual differences in mothers’ adaptation and engagement with their work or the overall quality of their job (Forgays & Forgays, 1993; Gómez, 2006; Haggag et al., 2011).
Taken together, these findings suggest that while the sources of maternal parenting stress may differ qualitatively, the overall magnitude of stress may not vary significantly between groups. This highlights the importance of developing tailored interventions that account for individual contextual factors, rather than relying on overly simplified or uniform approaches. Supporting both employed and unemployed mothers may require approaches tailored to the distinct stressors each group encounters. Future studies should consider contextual factors—particularly those arising at key stages of child development—to better identify when and how interventions can be most effective in reducing maternal parenting stress. Programs such as parenting stress management workshops, structured peer support groups, or evidence-based cognitive–behavioral interventions for parents have the potential to help maintain balanced stress levels, enhance adaptive parenting practices, and ultimately improve child outcomes. Additionally, exploring how maternal parenting stress interacts with child development and contextual factors, such as socioeconomic status and social support networks, could provide valuable insights for refining these interventions. Such research would be instrumental in developing optimized support strategies tailored to the distinct needs of employed and unemployed mothers, leading to healthier and more resilient family dynamics.
Limitations and Future Directions
While this study significantly advances our understanding of the relationship between maternal parenting stress and employment status through the application of MG-LCA, several limitations warrant attention. First, although mothers were categorized dichotomously as employed or unemployed, this classification does not capture the variation in employment conditions, such as part-time versus full-time work, job flexibility, or the mothers’ preferences and motivations for employment. Prior research suggests that the subjective fit between mothers’ work preferences and actual employment status plays a critical role in shaping parenting stress (Chung & van der Horst, 2018; Nomaguchi & Milkie, 2003). Future studies should incorporate these nuanced employment-related factors to more comprehensively assess how work contexts influence maternal well-being. Second, while grouping mothers as either employed or unemployed offered a useful starting point, it inevitably glosses over the wide variation in lived experiences within each category. A mother balancing a part-time job with strong family support likely navigates stress very differently from one in a full-time role with limited help. Factors such as income, education, and access to support systems can deeply influence how stress unfolds over time. Future research that takes these differences into account will offer a fuller, more accurate picture of maternal parenting stress. Third, this study focused exclusively on Korean mothers raising children aged three to eight. While the findings offer meaningful contributions within the Korean context and during early childhood, they may not generalize to other cultural settings or developmental stages. In Korea, it’s still common for mothers to carry most of the responsibility for raising children, even when they’re working full-time. Social expectations often push them toward a style of parenting that values constant involvement and personal sacrifice. These pressures—combined with limited support from fathers and strong societal views about how mothers should behave—can shape how parenting stress is experienced and expressed. Because of this, the findings from this study may reflect uniquely Korean patterns that don’t necessarily apply elsewhere. To better understand how maternal parenting stress changes over time and across different cultures, future research should include cross-national comparisons and cover a wider range of developmental periods.
Footnotes
ORCID iDs
