Who Are Working from Home Parents in China?: Comparing Working from Home Mothers and Fathers

Abstract

Working from home (WFH) has increasingly been adopted globally as a family-friendly arrangement that enables employees with greater family responsibilities to reconcile work and family demands. However, little is known about which parents take up WFH in the Chinese context. Using nationally representative pre-pandemic data from the China General Social Survey 2015, this paper examines the sociodemographic determinants of parents working from home in China, with particular attention to gendered patterns among mothers and fathers. The results show no statistically significant gender difference in the overall likelihood of WFH among parents. However, the sociodemographic determinants of WFH show different patterns among mothers and fathers. Specifically, Chinese mothers are more likely to WFH when they are engaged in non-standard employment arrangements and do not live with grandparents, whereas Chinese fathers are more likely to WFH when they are employed in the agricultural sector and hold more traditional gender role attitudes. A positive but weak association is observed between the number of children and mothers’ likelihood of WFH, while no such association is found among fathers. In addition, unlike in Western contexts, WFH uptake among Chinese parents is not found to be concentrated among more advantaged groups.

1. Introduction

With the rise of dual-earner households, working from home (WFH) and other flexible working arrangements (FWAs) have slowly but steadily increased in Western contexts as family-friendly practices aimed at attracting and retaining employees while supporting them in balancing work and family demands before the pandemic (T. D. Allen et al. 2015; Mele et al. 2023). In China, however, the adoption of WFH was relatively limited before the pandemic. Workplace cultures in China, characterised by heightened supervision and face-to-face interaction (Long et al. 2013; Himawan et al. 2022) and embedded in a high-power distance workplace environment, can constrain the adoption of working from home arrangements (Raghuram and Fang 2014; Himawan et al. 2022). As a result, homeworkers accounted for only about 0.6% (around 8 million workers) of the total workforce in China in 2018 (CIKD 2020; MOHRSS 2018). During the COVID-19 pandemic, WFH experienced a rapid and widespread expansion and has more recently been promoted by governments as a strategy to support employees with substantial family responsibilities (General Office of the State Council 2024). However, despite this recent expansion, the structural conditions, such as distinctive workplace culture and gendered role norms, continue to shape the WFH uptake in China. Exploring who was able to WFH before the pandemic can therefore provide a baseline for understanding how the uptake of WFH is shaped in the Chinese context.

Although WFH provides employees with greater perceived control over their workplace and schedules, potentially helping them accommodate work and family responsibilities, the intersection of WFH and gender remains deeply complex, particularly for parents who attempt to navigate competing norms around caregiving and work commitments. WFH has the potential to promote work–family balance and contribute to greater gender equality in the division of labour, as the perceived control over their workplace and schedule can enable fathers to increase their presence and involvement in family responsibilities (Madsen 2003; Johnson et al. 2007; Carlson et al. 2021), and reduce the likelihood that mothers with young children exit the labour market due to childcare responsibilities (Chung and Van Der Horst 2020; Lyttelton et al. 2022). However, under the traditional gender role norms, WFH may generate a gendered paradox, as it can intensify gendered asymmetric devotion to work and family responsibilities for parents, particularly with intensified mothers’ devotion to childcare and housework (Chung 2020; Chung and Van Der Horst 2020; Chauhan 2022; Clar-Novak 2025). WFH can also generate flexibility discrimination toward those who use this arrangement, with mothers being more likely to experience motherhood penalties, while fathers, under traditional masculinity norms, tend to experience flexibility stigma, which can act as a barrier to their use of FWAs (Chung 2020; Cook et al. 2021; Yucel and Chung 2023).

Existing research shows that employees with higher levels of education and those working in highly skilled occupations are more likely to access WFH or other FWAs in Western contexts before and after the pandemic (Brescoll et al. 2013; Chung 2019; Minkus et al. 2022; Barrero et al. 2023; Rubino et al. 2024). In addition, mothers are more likely to choose WFH than fathers (Golden 2008; Mas and Pallais 2017), especially after parenthood (S. Wang and Cheng 2025). However, women, including mothers, are generally less likely than men to be granted FWAs by their organisations, partly due to managerial biases that women are more likely to use FWAs for family-related reasons (Brescoll et al. 2013; Munsch 2016). Moreover, other studies report no significant gender differences in access to schedule control (Chung 2019) or the use of WFH among parents (Chung 2019; Cook et al. 2021).

Yet little is known about the sociodemographic profiles of WFH parents, let alone the potentially different sociodemographic determinants of mothers and fathers who work from home in China, where WFH is a relatively new practice and FWAs in supporting parental responsibilities have not been broadly recognised by most Chinese organisations (UNICEF and CPA 2022). Chinese working parents commonly rely on family networks to meet childcare needs, particularly support from grandparents (Sun et al. 2022; UNICEF and CPA 2022). Given its unique gender ideologies and workplace culture characterised by the ideal worker norm (Acker 1990) and presenteeism, China offers a distinctive context for exploring the dynamics of WFH and gender among parents with dependent children. The unique characteristics of WFH arrangements in China may result in distinctive differences in the sociodemographic determinants of mothers and fathers who WFH compared to those in Western cultural contexts. It is thus crucial to move beyond general assumptions and examine the sociodemographic profiles of Chinese parents who work from home. This study, therefore, examines whether the proportion of mothers and fathers working from home differs, as well as the sociodemographic characteristics associated with WFH for mothers and fathers separately. Specifically, the study addresses the following research questions:

• Are there gender differences in the likelihood of working from home among Chinese parents?
• What are the sociodemographic characteristics of mothers and fathers working from home in China, respectively?

This study makes a specific contribution by portraying the sociodemographic profile of parents working from home in China and examining gendered patterns in these associations. Importantly, by doing so, it provides empirical evidence on whether patterns widely documented in Western research—such as the concentration of WFH among more advantaged workers and gender differences in its uptake—also emerge in the Chinese context. By focusing on pre-pandemic evidence, the study further provides a baseline for understanding how access to WFH among parents may have evolved during and after the pandemic. It also contributes to broader debates on the ideal worker norm (Acker 1990) and gendered role norms (Pleck 1977) by situating WFH within a distinct institutional and cultural context.

2. Literature Review

In this study, WFH focuses on workers who perform their jobs at or from home during their regular working hours (Peters and Van Der Lippe 2007; Van Der Lippe and Lippényi 2020). Although WFH is often considered a form of telework that involves the use of information and communication technologies to perform job tasks at home (T. D. Allen et al. 2015; ILO 2020), this study focuses primarily on the change in work location itself, from the office or other workplaces outside the home to the home. This means that the definition encompasses a wide range of home-based work arrangements, including but not limited to digital technology-based telework. It also includes workers in lower-skilled occupations who may not rely on digital technologies, but whose work is spatially embedded in or near the home. In addition, this study encompasses employees who WFH on a part-time basis (i.e., hybrid arrangements), rather than restricting WFH to full-time homeworking. In practice, most homeworkers adopt hybrid working arrangements, in which employees split their time between the workplace and remote work (Hopkins and Bardoel 2023).

Existing research on the sociodemographic characteristics associated with WFH mothers and fathers has primarily examined two dimensions: WFH provision and the perceived availability of such arrangements, with the latter often shaped by flexibility stigma (Chung 2019; Van Der Lippe and Lippényi 2020; Cook et al. 2021). Meanwhile, the evidence indicates that WFH among parents is disproportionately concentrated among more advantaged groups (Chung 2020; Barrero et al. 2023), and those patterns of uptake are gendered (Golden 2008; Ewald and Hogg 2022).

Previous studies from Europe and the United States have shown that access to WFH is more prevalent among more advantaged employees, specifically highly educated and highly skilled workers, before and after the pandemic (Chung 2020; Barrero et al. 2023). Although policy support and improved accessibility in the post-pandemic period have enabled many jobs to be performed from home (Bick et al. 2023), research both before and after the pandemic continues to show that WFH remains concentrated among higher-skilled workers and varies across sectors. Drawing on data from 30 European countries, Chung (2018) demonstrates that access to flexible working arrangements is stratified by occupational status, with higher-skilled and more secure workers significantly more likely to benefit than lower-skilled or insecure workers. The use of WFH is inclined towards the more privileged group, probably because even when flexibility policies are designed to be universal, they often disproportionately benefit more privileged groups or workers with a vested interest, such as highly educated or higher-income families (Chung 2018; Cook et al. 2021). Another explanation is that jobs which can be performed remotely are often those with higher levels of skills (Paul 2022).

Most previous research in the U.S. and European contexts has shown that mothers with childcare responsibilities are more likely than their male counterparts to take up WFH arrangements (Golden 2008; Ewald and Hogg 2022; Rahman Fatmi et al. 2022). Mothers, who are mostly regarded as the main caregivers, may be more likely to adopt flexible working arrangements to accommodate caregiving and other family responsibilities. While several studies also indicate that mothers with childcare responsibilities do not necessarily have greater access to WFH than fathers (Chung 2019), this may be because mothers are less likely than fathers to be granted FWAs by their organisations (Brescoll et al. 2013; Munsch 2016). Pre-pandemic UK data show that while mothers overall had higher uptake of FWAs than fathers, WFH was the only FWAs more common among fathers than mothers (Cook et al. 2021).

These seemingly inconsistent findings can be attributed to variations in the uptake of WFH according to the type of FWAs, cross-national cultural norms, and structural differences in employment between mothers and fathers. Compared with reduced hours, schedule and location flexibility is less likely to provoke gendered senses of entitlement or flexibility stigma (Cook et al. 2021). The higher likelihood of income loss from reduced hours can be understood as reinforcing the perception that such arrangements violate traditional masculine norms (Cook et al. 2021). Furthermore, countries with stronger work–life balance cultures and more egalitarian gender norms tend to exhibit lower levels of flexibility stigma (Chung and Seo 2024), suggesting that gendered patterns in WFH uptake vary across cultural contexts. In addition, differences in employment status between mothers and fathers can shape these patterns. For instance, a quantitative study conducted in Spain during the COVID-19 pandemic found that women’s lower representation in occupations with opportunities for WFH was partly due to their overrepresentation in part-time jobs, which are less likely to be compatible with WFH (Rubino et al. 2024).

However, previous studies on the sociodemographic characteristics of employees who are WFH in China are limited (e.g., Xiong et al. 2023; Zhang et al. 2024). Research specifically examining parents from a gender perspective remains scarce, according to the databases of Scopus and Web of Science from 2005 to 2025. A quantitative study in the Chinese context using panel survey data collected during the COVID-19 pandemic examined the teleworkability of employees and found that individuals with higher education, urban residence, coastal access, and women were more likely to have jobs with greater teleworkability (Zhang et al. 2024). Another pre-pandemic quantitative study in China investigated the preferences in WFH of employees and found that women expressed a stronger preference for WFH compared to men (Xiong et al. 2023). However, this study further suggests that although women were more likely to prefer WFH, this preference was not necessarily driven by childcare demands. Although it has been suggested that the intensive involvement of grandparents in childcare may reduce the demand for mothers to WFH in the Chinese context (Xiong et al. 2023), the role of grandparental support in shaping parents’ WFH uptake has not been empirically examined in existing research. Hence, this paper also examines the role of grandparents in shaping the gendered patterns of parents’ WFH uptake in China.

3. Theories and Hypotheses Development

3.1. Ideal Worker Norm, Presenteeism and Chinese Workplace Culture

The WFH uptake can be shaped by the unique workplace culture in China, which emphasises establishing connections through daily interaction in the workplace (Long et al. 2013) and is characterised by high work intensity (J. J. Wang 2020).

The hierarchies in organisations are gendered, as reflected in the legitimisation of those fully committed to paid work and the devaluation of those whose commitments are divided between work and family (Acker 1990). Drawing on this assumption, subsequent scholars highlighted the notion of the ideal worker norm, which prescribes that the ideal worker should be dedicated to paid employment, work long hours, and arrange other duties around paid work commitments (Williams 2000; Kelly et al. 2010). Moreover, prolonged workplace presence is often constructed as a sign of high performance and fulfilling the ideal worker norm (Lott and Abendroth 2020). Long working hours are closely tied to physical presence, since the performance of extended work is usually made legible to organisations through employees’ ongoing visibility at work. “Face time” signals employees’ job commitment to their organisation: those who are physically present tend to be perceived as more committed and hard-working than those who are not (Cristea and Leonardi 2019; Williamson et al. 2023). Within workplace cultures structured by the ideal worker norm, the absence of physical presence associated with WFH arrangements can be interpreted as a sign of diminished work dedication and commitment.

Long hours of presence in the workplace are closely associated with the demonstration of work commitment in the Chinese context. The prevalence of high-intensity schedules and an “always-on” culture can be understood as the outcome of the intersection between global capitalist pressures and traditional Confucian values of hierarchy and obedience (J. J. Wang 2020). The average weekly working hours of urban employed men and women in China were 49.4 and 46.8 h, respectively (National Bureau of Statistics of China 2024a), exceeding those reported in neighbouring East Asian countries that are also known for long working hours, such as South Korea and Japan (Tsai et al. 2016). A prominent example is the “996” work schedule—working from 9 a.m. to 9 p.m., six days a week—which has become widespread in China’s technology sector (J. J. Wang 2020). At the same time, Chinese managerial practices are embedded in a context of high-power distance, characterised by heightened supervision and the centralisation of authority in leaders and managers (Raghuram and Fang 2014; Himawan et al. 2022). This hierarchical orientation can shape the implementation of WFH arrangements, as irregular physical presence in the workplace may be perceived by managers as a loss of control over employees. Moreover, daily interactions in the workplace are highly valued in Chinese organisational culture because they help cultivate connections with colleagues and company leaders, which are closely tied to career advancement (Long et al. 2013). In this context, the ideal worker norm is closely associated with sustained physical presence and visibility in the workplace. Consequently, WFH in China may be perceived as limiting opportunities to demonstrate work commitment, maintain interpersonal connections, and even access career advancement.

As a result, WFH in China may be less likely to be concentrated among more advantaged groups, as is often observed in Western contexts (Chung 2020; Barrero et al. 2023). Moreover, most parents in China still have limited access to family-supportive work arrangements (UNICEF and CPA 2022). Beyond the implication of WFH as a family-friendly policy for employees with standard employment relationships, another group of workers engaging in WFH are those in nonstandard employment, such as freelancers, gig workers, and the self-employed (Spreitzer et al. 2017). Around 60% of workers in China were engaged in nonstandard employment in the mid-2000s (Y. Zhou 2013). In urban labour markets, about 35% of workers were in nonstandard employment in 2016, and these workers typically experience greater precarity, including limited labour protections, poorer worker conditions, and longer hours (Y. Wang 2019). In recent years, the number of nonstandard workers in China has further increased with the emergence and expansion of new economic forms. This informal labour market has been increasingly shaped by the expansion of the gig economy, sharing economy, and digital platform-based work in China (Cai et al. 2021), which comes with high accessibility to working out of the office.

3.2. Parents, Gender and Role Theory

Parents with greater caregiving demands have been found to be more likely to engage in WFH arrangements (Golden 2008; Chung 2019). The increased perceived control over time and workspace associated with WFH can provide working parents with greater flexibility in managing competing family responsibilities. From a role theory perspective, individuals simultaneously occupy multiple social roles, each of which entails distinct normative expectations (Biddle 1979). Working parents, therefore, often face heightened role strain as they attempt to meet the demands of both the work and family domains. Within this context, WFH can be understood as a strategy through which parents seek to negotiate and accommodate these conflicting role expectations. Accordingly, this study expects that Chinese parents experiencing greater family demands are more likely to take up WFH.

The uptake of WFH among parents is often gendered. A role can be understood as “the behaviour related to normative expectations” (V. L. Allen and Van De Vliert 1984, p. 3), and role expectations between men and women tends to be gendered within both the work and family domains (Pleck 1977). Specifically, under traditional gender role norms, mothers are often expected to specialise in the family domain, particularly to childcare, while men with children may still prioritise their roles in the work domain. Meanwhile, flexibility in the workplace and work schedules is often regarded as enabling employees to accommodate greater family responsibilities. In this sense, mothers are more likely than fathers to adopt WFH arrangements to accommodate family obligations in Western contexts (Golden 2008; Ewald and Hogg 2022).

Moreover, the ideal worker norm reinforces these gendered expectations by privileging continuous devotion to paid work, an ideal that is, in practice, embodied as male (Acker 1990). In such contexts, women are often required to act like men in the workplace as these are perceived as manifestations of work devotion and commitment (Brumley 2014). Mothers, who are normally expected to be the primary caregivers, may therefore face greater pressure to reconcile family demands.

Meanwhile, in the Chinese context, mothers are expected to embody both the “ideal mother” and the “ideal worker,” whereas Chinese fathers are primarily considered in relation to the “ideal worker” norm (Ji 2025). Although fathers have increased their presence in families in recent years, partly due to growing recognition of the benefits of paternal involvement for children’s development (Xu and O’Brien 2014; Li 2020), the gendered division of labour in the family domain remains largely unchallenged (Ji et al. 2017). According to recent national statistics, Chinese men spend on average 50 min more per day on paid work than women, whereas women devote approximately 100 min more per day than men to unpaid work (National Bureau of Statistics of China 2024b). Faced with more competing demands between work and family roles, mothers may be more likely than fathers to consider flexible working arrangements as a way to remain competitive in the labour market while managing caregiving responsibilities and accommodating their dual roles.

Overall, the ideal worker norm, characterised by long working hours and presenteeism workplace culture, alongside gender role expectations in the Chinese context, may shape gendered patterns in parents’ uptake of WFH. In work environments where long working hours and physical presence signal commitment and productivity, employees who seek to maintain an “ideal worker” image may be less likely to adopt WFH because this work arrangement can reduce their physical visibility in the workplace. At the same time, gender role expectations can further intersect with the ideal worker norm to shape how mothers and fathers respond to the uptake of WFH. Chinese fathers may be more likely to conform to the ideal worker norm by maintaining visible workplace presence and prioritising paid work commitments, rather than adopting WFH arrangements. Chinese mothers, however, often face intensified expectations to fulfil both the roles of the “ideal worker” and the “ideal mother.” Within this context of dual role expectations, WFH may be viewed by mothers as a strategy to reconcile work and family responsibilities. Therefore, compared with fathers, Chinese mothers may be more likely to take up WFH as a way to accommodate competing demands in the work and family domains.

3.3. Hypotheses

Based on the literature review and theories, this study proposes the following hypotheses.

Drawing on the ideal worker norm (Acker 1990) and gendered role expectations in the Chinese context, mothers may be more likely than fathers to adopt flexible working arrangements to reconcile competing work and family responsibilities, while fathers may be more likely to conform to the ideal worker image that emphasises long physical presence at the workplace. Previous empirical evidence also suggests that women in China may have higher teleworkability than men (Zhang et al. 2024). Hence, this study expects a gender difference in the likelihood of working from home, with mothers being more likely than fathers.

Furthermore, drawing on role theory, which suggests that individuals experiencing incompatible roles across different domains are more likely to experience role strain (Biddle 1986), prior empirical studies also show that parents with greater family responsibilities are more inclined to work from home (Golden 2008; Chung 2019). Therefore, this study expects that the age of the youngest child and the number of children are positively associated with parents’ likelihood of WFH. Considering traditional gender role norms in the Chinese context, where mothers are more likely to be expected to be the primary caregivers within households (Ji et al. 2017), this study further examines whether this association differs between mothers and fathers.

In China, grandparental support plays an important role in alleviating parents’ childcare responsibilities, which may in turn reduce parents’ need to rely on flexible working arrangements, particularly for mothers, who are more likely to assume primary caregiving responsibilities (Sun et al. 2022; Xiong et al. 2023). Therefore, this study expects that not co-residing with grandparents is positively associated with mothers’ likelihood of working from home.

Based on the ideal worker norm embedded in a culture of long-hours presenteeism and the characteristics of China’s less regulated labour market (Tsai et al. 2016), and the prevalence of informal work arrangements (Cai et al. 2021), parents who work from home in the Chinese context may be less likely to belong to more privileged groups and more likely to be engaged in nonstandard employment. Nonstandard employment is characterised by multiparty employment relationships, temporary work, and part-time work (ILO 2016; Y. Wang 2019). Hence, this study expects that not having a formal work contract is positively associated with parents’ likelihood of WFH; working in the private sector is positively associated with parents’ likelihood of WFH; and holding multiple jobs is positively associated with parents’ likelihood of WFH.

Additionally, based on the previous research that men with more traditional gender role attitudes are more likely to have a sense of stigma in using FWAs (Blair-Loy 2010; Yucel and Chung 2023), this study also expects that more egalitarian gender role attitudes are positively associated with fathers’ likelihood of WFH.

4. Methods

4.1. Data and Sample

This study presents findings from an analysis of WFH status using data from 2015 China General Social Survey (CGSS 2015), a nationally representative cross-sectional survey. The China General Social Survey is the earliest national, comprehensive and continuous academic survey project in China, jointly launched by Renmin University of China and the Hong Kong University of Science and Technology in 2003, employing a multistage stratified sampling method to generate nationally representative samples in mainland China. CGSS 2015 surveyed Chinese individuals living in 478 regions across 28 provinces, autonomous regions, and municipalities in 2015, obtaining a final sample of 10,968 participants. The survey consists of six modules, and only one-sixth of the total participants were randomly selected to answer the East Asian social survey module and the International social survey program module, using a multistage stratified probability proportionate sampling method. The question on WFH status in CGSS 2015 was part of the International Social Survey Program module, meaning that approximately 1828 participants were randomly selected to provide information on their WFH status. See http://www.cnsda.org/index.php (accessed on 21 May 2025) for more on this survey.

4.2. Sampling

In the CGSS 2015 survey, a total sample of 302 observations, accounting for 3% of the total survey sample, was obtained after excluding cases with missing values and non-parents. After sequentially applying the selection criteria—including restricting to employed respondents, those with a co-resident partner, and those living with at least one dependent child aged under 18—a final analytical sample of 302 parents was obtained (see Table S1 in Supplementary Materials for the detailed sample construction procedure). Among these, 33.74% reported that they WFH at least sometimes (n = 102); the rest of the observations reported that they rarely or never WFH (n = 200).

In this study, parents are defined as employed mothers or fathers who live with their partners and at least one dependent child in the household, as co-resident parents constitute the group most directly exposed to daily childcare and household responsibilities, and this represents the predominant family composition in China (OECD and OECD Korea Policy Center 2023). Therefore, parents who are not co-residing with a partner, as well as widowed or single parents, are excluded. In this sample, 302 parents live with a co-resident partner, while parents report having no partner or not co-residing with their partner and are therefore excluded from the analysis (see Table S2 in Supplementary Materials). The preliminary analysis showed no statistically significant association between partnership status and WFH status (p = 0.69 > 0.05). Nevertheless, it should be acknowledged that parents without a co-resident partner may lack partner support in sharing domestic responsibilities, which may lead to a greater demand for WFH to manage family responsibilities.

Dependent children are defined as those under the age of 18, in accordance with the official age definition of dependent children in China, and living with their parents.

4.3. Measures

Working from home. In CGSS 2015, WFH was measured using a single item: “Do you work from home frequently during your normal working hours?” This question is intended to capture the frequency with which respondents perform paid work from home during their regular working hours. It should be noted that this measure encompasses different forms of work conducted at or from home, rather than referring only to high-skilled work that relies on digital technology.

Respondents can choose one from the following options: “always = 1, very often = 2, sometimes = 3, rarely = 4, never = 5” in this survey. A dummy variable was created for WFH to distinguish parents who are working from home from those who are not. Importantly, in this study, engaging in WFH includes not only frequent homeworking but also occasional homeworking (i.e., “sometimes”), which aligns with the most common form of WFH in practice, where most homeworkers work from home only occasionally, such as two days per week (Madsen 2003; Hopkins and Bardoel 2023). Hence, parents who responded with “always, very often and sometimes” are classified as working from home. Respondents who answered ‘rarely’ and ‘never’ are categorised as not working from home in this study.

Total number of children. This variable measures the total number of children in the household and was treated as a continuous variable.

Age of the youngest child. This variable captures the age of the youngest child in the household and was reported in years.

Gender. Gender was coded 0 = female and 1 = male.

Co-residence with grandparents. This variable was measured using the survey question asking whether the family member lives and eats together with the respondent. A dummy variable was constructed and coded 0 = not co-residing with grandparents and 1 = co-residing with grandparents.

Gender role attitudes. This variable was firstly assessed with five items (e.g., “To what extent do you agree that men should prioritise their careers, while women should prioritise their families?”). The five items were recoded so that higher values indicate more egalitarian gender attitudes. Principal component analysis (PCA) is employed to determine the principal component and calculate a comprehensive score (see Tables S3 and S4 in Supplementary Materials). The PCA reveals that the first four questions in the gender role attitudes section were strongly correlated and thus used to compute the comprehensive score. The fifth question (i.e., “To what extent do you agree that husbands and wives should share household chores equally”?), due to its high uniqueness (>0.90), is excluded. Cronbach’s alpha for the resulting scale is 0.66, suggesting acceptable internal consistency (Cortina 1993). Higher scores indicate more egalitarian gender role attitudes. The KMO value of 0.71 and a significant Bartlett’s test of sphericity (p < 0.001) indicate that the data are suitable for factor analysis.

Education level and partner’s education level. These variables capture the highest level of educational attainment of the respondent and their partner, respectively, and were treated as categorical variables with three categories: 1 = lower education, 2 = secondary education, and 3 = higher education.

Work contract. This variable was measured based on whether the respondent had signed a written employment contract with their employer and was coded 0 = no formal employment contract and 1 = fixed-term or open-ended employment contract.

The sector and agriculture work. The sector was coded 0 = non-public sector and 1 = public sector. In addition, drawing on prior research in the Chinese context that agricultural work often offers greater flexibility compared to other sectors (H. M. Wang et al. 2020), the sector was also classified as agricultural or non-agricultural (0 = non-agricultural sector and 1 = agricultural sector).

Employed partner. This variable indicates whether the respondents’ partner is currently employed and was coded 0 = not employed and 1 = employed.

Multiple jobs. This variable indicates whether the respondent holds more than one job. It was coded 0 = one job and 1 = having more than one job.

4.4. Analytical Scheme

Binary logistic regression models are adopted to analyse the association between WFH status and a set of explanatory variables in the CGSS 2015 dataset. To systematically assess the contribution of different sets of predictors and to disentangle their effects, explanatory variables were entered sequentially across four stages, including sociodemographic, family-related, socioeconomic status (SES), and work-related variables. This stepwise approach allows for a clearer understanding of how each group of variables influences the likelihood of WFH. Subsequently, two separate binary logistic regression models were conducted—one model for mothers and one model for fathers—to compare potential gendered patterns in their sociodemographic characteristics associated with WFH status. Given the exploratory nature of this study, coefficients with p-values < 0.10 are flagged with a plus (+) sign in the results tables to indicate potentially emerging trends.

5. Results

5.1. Descriptive Analysis

This section firstly presents a descriptive analysis comparing respondents who work from home and those who do not.

Table 1. Gender differences in working from home status.

Working from Home Status	Working from Home	Not Working from Home	Total
Mother (N)	57	106	163
(%)	34.97%	65.03%	100.00%
Father (N)	45	94	139
(%)	32.37%	67.63%	100.00%
Total (N)	102	200	302
(%)	33.77%	66.23%	100.00%

Data: China General Social Survey (2015).

illustrates the distribution of mothers and fathers based on their WFH status. According to the CGSS 2015 data, a relatively small proportion of Chinese parents (33.77%) report WFH at least occasionally. Among mothers, 34.97% report WFH at least occasionally, while 65.03% do not. Similarly, 32.37% of fathers report WFH at least occasionally, compared to 67.63% who do not. Although the proportion of mothers working from home is slightly higher than that of fathers, there is no significant association between gender and working from home status among parents (p = 0.635 > 0.05).

The original survey item measured the frequency of working from home using five response categories: “always,” “very often,” “sometimes,” “rarely,” and “never.”

Table 2. Distribution of working from home frequency among mothers and fathers.

	Working from Home			Not Working from Home		Total
	Always	Very Often	Sometimes	Rarely	Never
Mother (N)	11	16	30	51	55	163
(%)	6.75%	9.82%	18.40%	31.29%	33.74%	100.00%
Father (N)	6	14	25	52	42	139
(%)	4.32%	10.07%	17.99%	37.41%	30.22%	100.00%
Total (N)	17	30	55	103	97	302
(%)	5.63%	9.93%	18.21%	34.11%	32.12%	100.00%

Data: China General Social Survey (2015).

reports the full distribution across these categories for mothers and fathers. Overall, 5.63% of parents reported always working from home, 9.93% very often, and 18.21% sometimes, while 34.11% reported rarely and 32.12% never working from home. This distribution indicates that most parents who are categorised as working from home do so only occasionally, particularly in the “sometimes” category, whereas relatively few parents report always working from home.

Table 3. Key demographic characteristics among parents by workplace arrangements.

	Working from	Not Working	Total
	Home	from Home
Age	37.06	37.08	37.07
Total children
One	69.61%	79.50%	76.16%
More than one	30.39%	20.50%	23.84%
Youngest child age	8.42	8.62	8.55
Living with grandparents	25.49%	36.00%	32.45%
Education level
Low education	46.08%	37.00%	40.07%
Secondary education	37.25%	36.50%	36.75%
Higher education	16.67%	26.50%	23.18%
Partner Education level
Low education	44.12%	36.50%	39.07%
Secondary education	33.33%	39.50%	37.42%
Higher education	22.55%	24.00%	23.51%
Employed partner	87.25%	82.00%	83.77%
With formal contract	29.41%	52.00%	44.37%
Sector
Public sector	16.67%	34.00%	28.15%
Agricultural sector	15.69%	5.00%	8.61%
Another private sector	67.64%	61.00%	63.24%
Multiple jobs	4.90%	2.00%	2.98%
Observations	102 −33.74%	200 −66.23%	302 −100%

Data: China General Social Survey (2015).

presents the differences in the demographic, family-related, SES, and work-related characteristics between parents who WFH and those who do not. Households where parents WFH tend to have a higher number of children on average compared to those where parents do not WFH. The youngest child in households where parents work from home is only slightly younger than in households where parents do not work from home. The proportion of parents living with grandparents is substantially lower among those who WFH (25.49%) than among those who do not (36.0%). This result indicates that parents who WFH tend to be less likely to live with grandparents than those who do not work from home. Regarding SES-related variables, parents who WFH, as well as their partners, have a lower proportion of higher education attainment compared to those who do not WFH. The proportion of parents with an employed partner is higher among those who WFH than among those who do not. In terms of work-related characteristics, around 30% of parents who WFH have a formal employment contract, whereas over 50% of parents who do not WFH hold such contracts, suggesting that parents who WFH are considerably less likely to be in standard employment. Parents who WFH are more likely to be employed in the private or agricultural sectors compared to their counterparts who do not WFH. Additionally, 5% of parents who WFH have multiple jobs, compared to 2% of those who do not.

5.2. Regression Analysis on Sociodemographic Determinants of Parents’ Working from Home

Table 4. Binary logistic regressions on parents’ working from home, odds ratio.

	(1)	(2)	(3)	(4)
	Model 1	Model 2	Model 3	Model 4
Gender	0.898	0.826	0.949	0.917
	[0.554, 1.454]	[0.500, 1.364]	[0.555, 1.623]	[0.520, 1.619]
Age	1.001	0.996	0.994	0.999
	[0.966, 1.038]	[0.943, 1.051]	[0.941, 1.050]	[0.941, 1.059]
Total child	1.413	1.289	1.178	1.129
	[0.933, 2.140]	[0.833, 1.994]	[0.727, 1.911]	[0.681, 1.871]
Youngest child		0.986	0.977	0.968
		[0.915, 1.064]	[0.904, 1.056]	[0.891, 1.053]
Living with grandparents		0.582 +	0.518 *	0.468 *
		[0.332, 1.020]	[0.290, 0.926]	[0.250, 0.874]
Gender role attitudes		0.658 **	0.675 *	0.667 *
		[0.492, 0.881]	[0.497, 0.917]	[0.482, 0.923]
Education level
Ref: Lower education
Secondary education			0.960	1.560
			[0.443, 2.081]	[0.683, 3.562]
Higher education			0.482	0.966
			[0.171, 1.364]	[0.313, 2.982]
Partner education level
Ref: Lower education
Secondary education			0.856	1.117
			[0.389, 1.885]	[0.484, 2.577]
Higher education			1.345	2.725 +
			[0.491, 3.689]	[0.908, 8.177]
Employed partner			1.867	2.059 +
			[0.862, 4.044]	[0.906, 4.678]
With formal contract				0.410 **
				[0.212, 0.793]
Public sector				0.421 *
				[0.194, 0.915]
Agricultural sector				3.189 *
				[1.231, 8.259]
Multiple jobs				6.313 *
				[1.378, 28.914]
N	302	302	302	302
pseudo R2	0.007	0.037	0.054	0.124

Data: China General Social Survey (2015). Odds ratios are reported; 95% confidence intervals in brackets. Pseudo R² refers to McFadden’s pseudo R². + p < 0.1, * p < 0.05, ** p < 0.01.

presents the binary logistic regression results for parents’ overall likelihood of WFH. Model 1 shows that gender, age, and the number of children are not significant predictors of parents’ likelihood of WFH. Different from the proposed hypothesis, this result indicates that Chinese mothers do not show a higher likelihood of taking up WFH than fathers. Family-related variables, including co-residence with grandparents, the age of the youngest child, and gender role attitudes, are introduced in Model 2. The results show that more egalitarian gender role attitudes are significantly negatively associated with the likelihood of working from home among parents (OR: 0.66). This finding is contrary to the proposed hypothesis. Meanwhile, co-residence with grandparents is marginally negatively associated with the likelihood of WFH (p < 0.10), although this association is not statistically significant at the 95% confidence level in this model.

After controlling for educational level, partner educational level and partner employment status in model 3, co-residence with grandparents becomes statistically significantly negatively associated with WFH. The results show that parents who live with grandparents have odds of WFH that are about 48% lower than those who do not live with grandparents (OR: 0.52). However, SES-related variables are not statistically significantly associated with parents’ likelihood of WFH. Meanwhile, the association between the number of children and the likelihood of WFH remains not statistically significant after co-residence with grandparents and other SES-variables were added in model 2 and model 3. Taken together, these results may suggest that grandparental support in childcare could reduce the need for parents to rely on WFH as a strategy to accommodate childcare demands. This result aligns with previous studies suggesting that the widespread availability of grandparental childcare support in China may reduce parents’ reliance on WFH to accommodate childcare demands (Sun et al. 2022; Xiong et al. 2023).

Model 4 introduces work-related variables. Holding a formal work contract is significantly negatively associated with parents’ likelihood of WFH (OR: 0.41), suggesting that Chinese parents without a formal contract tend to be more likely to WFH. Parents employed in the public sector have 58% lower odds of WFH compared to those in the non-public sector (OR: 0.42). Whereas employment in the agricultural sector is significantly positively associated with parents’ likelihood of WFH (OR: 3.19). Having multiple jobs is significantly positively associated with parents’ likelihood of WFH, with odds more than five times higher than those with a single job (OR: 6.31), although the estimate should be interpreted with caution given the wide confidence interval. Moreover, after controlling for work-related variables in this model, having an employed partner and having a partner with higher education are marginally positively associated with the likelihood of WFH (p < 0.10), although these associations do not reach the 95% confidence level.

Overall, the coefficient for gender remains stable and statistically insignificant across all models, indicating that the likelihood of WFH does not differ significantly between mothers and fathers even after controlling for family, socioeconomic, and work-related characteristics. Parents who have a higher likelihood of WFH tend to lack a formal work contract, be employed outside the public sector, hold multiple jobs, not be co-resident with grandparents, and hold less egalitarian gender role attitudes. SES-related variables are not significantly associated with parents’ likelihood of WFH in China. Child-related variables (number of children and age of the youngest child) are not statistically significant predictors of WFH. The lack of statistical significance for child-related variables may be influenced by the average age of the child being older (with a mean age of over eight years old) and limited statistical power due to the overall sample size.

5.3. Regression Analysis on Sociodemographic Determinants of Parents’ Working from Home by Gender

Table 5. Binary logistic regressions on mothers’ working from home, odds ratio.

	(1)	(2)	(3)	(4)
	Model 1	Model 2	Model 3	Model 4
Age	0.988	0.972	0.971	0.965
	[0.940, 1.039]	[0.901, 1.049]	[0.898, 1.050]	[0.881, 1.056]
Total child	1.594 +	1.489	1.406	1.292
	[0.923, 2.753]	[0.836, 2.653]	[0.744, 2.656]	[0.661, 2.524]
Youngest child		0.994	0.988	0.984
		[0.898, 1.100]	[0.888, 1.100]	[0.871, 1.112]
Living with grandparents		0.398 *	0.336 *	0.271 **
		[0.180, 0.881]	[0.146, 0.773]	[0.109, 0.675]
Gender role attitudes		0.775	0.752	0.696
		[0.541, 1.109]	[0.497, 1.139]	[0.442, 1.097]
Education level
Ref: Lower education
Secondary education			1.389	2.443
			[0.427, 4.515]	[0.665, 8.979]
Higher education			0.727	1.875
			[0.163, 3.238]	[0.340, 10.332]
Partner education level
Ref: Lower education
Secondary education			0.485	0.610
			[0.148, 1.593]	[0.168, 2.215]
Higher education			1.313	4.527 +
			[0.303, 5.689]	[0.810, 25.287]
Employed partner			3.195	4.286
			[0.312, 32.677]	[0.350, 52.405]
With formal contract				0.177 ***
				[0.066, 0.473]
Public sector				0.240 *
				[0.073, 0.791]
Agricultural sector				1.247
				[0.281, 5.531]
Multiple jobs				17.006 *
				[1.896, 152.520]
N	163	163	163	163
pseudo R2	0.014	0.049	0.081	0.206

Data: China General Social Survey (2015). Odds ratios are reported; 95% confidence intervals in brackets. Pseudo R² refers to McFadden’s pseudo R². + p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001.

and

Table 6. Binary logistic regressions on fathers’ working from home, odds ratio.

	(1)	(2)	(3)	(4)
	Model 1	Model 2	Model 3	Model 4
Age	1.012	1.014	1.014	1.014
	[0.960, 1.067]	[0.935, 1.100]	[0.933, 1.102]	[0.927, 1.108]
Total child	1.242	1.079	0.807	0.828
	[0.643, 2.398]	[0.532, 2.187]	[0.349, 1.863]	[0.333, 2.060]
Youngest child		0.976	0.946	0.953
		[0.870, 1.096]	[0.838, 1.069]	[0.836, 1.085]
Living with grandparents		0.839	0.667	0.673
		[0.371, 1.896]	[0.281, 1.582]	[0.262, 1.726]
Gender role attitudes		0.483 **	0.495 *	0.529 *
		[0.287, 0.814]	[0.289, 0.847]	[0.299, 0.939]
Education level
Ref: Lower education
Secondary education			0.523	0.815
			[0.166, 1.652]	[0.234, 2.837]
Higher education			0.263	0.427
			[0.051, 1.351]	[0.078, 2.345]
Partner education level
Ref: Lower education
Secondary education			1.546	1.989
			[0.509, 4.699]	[0.594, 6.656]
Higher education			1.088	1.658
			[0.235, 5.030]	[0.335, 8.199]
Partner employment			1.940	2.199
			[0.789, 4.769]	[0.847, 5.706]
With formal contract				0.928
				[0.328, 2.621]
Public sector				0.425
				[0.129, 1.403]
Agricultural sector				5.153 *
				[1.374, 19.330]
Multiple jobs				1.556
				[0.097, 24.901]
N	139	139	139	139
pseudo R2	0.003	0.050	0.081	0.136

Data: China General Social Survey (2015). Odds ratios are reported; 95% confidence intervals in brackets. Pseudo R² refers to McFadden’s pseudo R². * p < 0.05, ** p < 0.01.

present the binary logistic regressions on the likelihood of mothers and fathers working from home separately. Hierarchical logistic regression models were conducted for mothers and fathers separately, entering demographic, family-related, socioeconomic, and work-related variables step by step.

Table 5 presents the sociodemographic factors associated with WFH uptake among mothers. In model 1, neither age nor the number of children is statistically associated with mothers’ likelihood of WFH, but the association between the number of children and mothers’ WFH is marginally positively significant at the 10 per cent level. However, model 2 shows that this association disappears and coefficients become weaker once additional family-related variables are included, including the number of total children, age of the youngest child, living with grandparents, as well as gender role attitudes. This pattern may reflect that the effect of the number of children is partially absorbed by other family-related variables, particularly given the limited sample size. By contrast, co-residence with grandparents is significantly negatively associated with mothers’ likelihood of WFH. The odds of WFH for mothers who live with grandparents are 60% lower than for mothers who do not live with grandparents (OR: 0.40). After controlling for SES-related variables in Model 3, the negative association between the likelihood of WFH and co-residence with grandparents becomes slightly stronger (OR: 0.34). This finding may suggest the importance of grandparental support in shaping mothers’ work arrangements in the Chinese context.

Model 4 shows that holding a formal employment contract is significantly negatively associated with mothers’ likelihood of WFH, with odds of WFH about 82% lower than those without a formal contract (OR: 0.18). Similarly, the finding also suggests that mothers employed in the public sector are significantly less likely to WFH, with odds about 76% lower than those employed outside the public sector (OR: 0.24). By contrast, holding multiple jobs is significantly positively associated with mothers’ likelihood of WFH (OR: 17.01), although this estimate should be interpreted with caution given the wide confidence interval, which likely reflects the relatively small number of respondents in this category. These findings may suggest that mothers without formal employment contracts, those working in the private sector, and those holding multiple jobs have a higher likelihood of working from home, whereas those in more stable or formal employment arrangements are less likely to do so.

Table 6 presents the sociodemographic factors of fathers’ uptake of WFH. Contrary to the proposed hypothesis, egalitarian gender role attitudes are significantly negatively associated with fathers’ likelihood of WFH, suggesting that fathers with more traditional gender role attitudes are more likely to WFH (OR: 0.48 in Model 2). In contrast, other family-related variables, including the number of children, the age of the youngest child, and co-residence with grandparents, are not statistically significantly associated with fathers’ likelihood of WFH across the four models. Taken together, these findings suggest that fathers’ uptake of WFH is less likely to be driven by the need to accommodate childcare or other family obligations. Similarly, Model 3 introduces SES-related variables, yet none of these factors—fathers’ educational attainment, their partners’ educational attainment, or partner employment status is significantly associated with fathers’ likelihood of WFH.

In Model 4, additional work-related variables are included. The results show that employment in the agricultural sector is significantly positively associated with the likelihood of WFH among fathers; specifically, those in agriculture have higher odds of WFH compared to those in non-agricultural sectors (OR: 5.15). Other work-related variables, including holding a formal work contract, employment in the public sector, and holding multiple jobs, have not shown a statistically significant association. Moreover, the coefficient for gender role attitudes increases slightly from Model 2 (OR: 0.48) to Model 4 (OR: 0.53), suggesting that this negative association becomes slightly weaker after controlling for SES and work-related variables.

Given that fathers employed in the agricultural sector are more likely to hold traditional gender role attitudes, this raises the possibility that the observed association between gender role attitudes and fathers’ likelihood of WFH may be driven by the concentration of fathers with more traditional gender role attitudes in agricultural employment. The following table, therefore, restricts the sample to fathers working in non-agricultural sectors to examine whether the observed association remains consistent when those in agricultural employment are excluded.

An additional analysis was conducted, excluding fathers employed in the agricultural sector, to examine the association between gender role attitudes and the likelihood of WFH among fathers in the non-agricultural sector.

Table 7. Binary logistic regression of working from home among non-agricultural fathers, odds ratio.

	Non-Agricultural Father
Age	1.029
	[0.933, 1.134]
Total child	1.090
	[0.409, 2.903]
Youngest child	0.923
	[0.800, 1.064]
Living with grandparents	0.491
	[0.174, 1.390]
Gender role attitudes	0.537 *
	[0.291, 0.992]
Education level
Ref: Lower education
Secondary education	0.918
	[0.252, 3.341]
Higher education	0.455
	[0.080, 2.598]
Partner education
Ref: Lower education
Secondary education	1.842
	[0.518, 6.550]
Higher education	1.436
	[0.279, 7.390]
Employed partner	2.984 *
	[1.005, 8.860]
With formal contract	0.964
	[0.334, 2.778]
Public sector	0.414
	[0.123, 1.393]
Multiple jobs	2.016
	[0.118, 34.442]
N	123
pseudo R2	0.105

Data: China General Social Survey (2015). Odds ratios are reported; 95% confidence intervals in brackets. Pseudo R² refers to McFadden’s pseudo R². * p < 0.05.

shows that egalitarian gender role attitudes remain significantly negatively associated with fathers’ likelihood of WFH among those employed in non-agricultural sectors. This result is consistent with the findings from the full sample of fathers, indicating that among fathers who are not engaged in agricultural work, those with more traditional gender role attitudes tend to be more likely to WFH. This result further suggests that the association between fathers’ likelihood of WFH and gender role attitudes does not appear to be mainly explained by the possibility that fathers employed in the agricultural sector are more likely to hold traditional gender role attitudes.

Moreover, after excluding fathers employed in the agricultural sector, having an employed partner becomes statistically significantly positively associated with the likelihood of working from home. This result suggests that fathers with employed partners have a higher likelihood of WFH compared to those whose partners are not employed. However, the child-related variables remain statistically insignificant in relation to these fathers’ likelihood of WFH. These results collectively tend to imply that fathers whose partners are less likely to stay at home to take on family responsibilities are more likely to take up WFH. In other words, the uptake of WFH among fathers does not appear to be directly explained by childcare demands within the households but may rather be shaped by their partners’ availability to accommodate childcare responsibilities.

5.4. Robustness Check

Several checks were run to confirm the robustness of the findings (see Supplementary Materials). First, to address potential bias arising from sparse data and the possibility of separation in the logistic regression models, a robustness check using Firth logistic regression was conducted. Specifically, two Firth logistic regression models were estimated separately for mothers and fathers to examine the sociodemographic factors of their likelihood of WFH. As the robustness check results show (see Tables S5 and S6 in Supplementary Materials), the overall patterns of association remain highly consistent with the main logistic regression results. The key variables identified in the baseline models remain statistically significant and retain similar directions and magnitudes of association. This suggests that the main findings are unlikely to be driven by sparse-data bias or separation problems in the original models. Although the confidence intervals for the multiple jobs’ variable among mothers in Table S6 remain relatively wide, this likely reflects the limited sample size of mothers working from home.

Second, because the main analysis treats WFH as a binary outcome, I evaluated whether the results would differ if WFH was measured on a three-point scale capturing the frequency of parents’ WFH (see Table S7 in Supplementary Materials). I therefore estimated ordered logistic regression models. Compared with the previous models, the significance of the gender role attitudes and sector weakens under the ordered logistic model. Both coefficients become marginally significant, failing to reach the 95% confidence level. However, most conclusions remained unchanged, and the directions of the effects were consistent, indicating that treating WFH as a dummy variable may not be problematic when testing the sociodemographic characteristics of WFH parents. The proportional odds assumption was tested using the Brant test after estimating the final ordered logistic regression models. The Brant test results (p = 0.054 > 0.05) indicate that the proportional odds assumption is not violated at the conventional 5% significance level, although the result is near the threshold. Subsequently, I checked the robustness of the findings by comparing potential gendered differences in mothers’ and fathers’ sociodemographic characteristics associated with WFH status (see Table S8 in Supplementary Materials). The ordered logistic models produced results that are highly consistent with those from the binary logistic regressions. For mothers, only the effect of the sector becomes insignificant at the 95% level. For fathers, employment in the agricultural sector and gender role attitudes continue to be statistically significant predictors of WFH. Overall, the robustness checks confirm the stability of the main findings.

6. Discussion and Conclusions

Using a nationally representative dataset, this article portrays the sociodemographic characteristics associated with parents’ likelihood of WFH in China before the COVID-19 pandemic in 2015. These findings offer a baseline for understanding how the uptake of WFH may have transformed during and after the pandemic.

In summary, the results based on the CGSS 2015 dataset provide partial support for the proposed hypotheses. Several proposed hypotheses are not supported by the data. Importantly, the findings show no statistically significant gender difference in the likelihood of WFH between mothers and fathers, which does not support the study’s main hypothesis. While we expected more egalitarian attitudes to be positively associated with fathers’ likelihood of working from home, the findings indicate that fathers with more traditional gender-role attitudes have a higher likelihood of WFH. Furthermore, some expectations receive only limited or marginal support. The number of children is marginally positively associated with mothers’ likelihood of WFH (p < 0.10), while no such association is observed among fathers. Several expectations receive clear support. Mothers’ likelihood of WFH is positively associated with not having a formal work contract, working in the private sector, and holding multiple jobs. Fathers’ likelihood of WFH, by contrast, is positively associated with employment in the agricultural sector. The results also support the hypothesis that co-residence with grandparents is negatively associated with parents’ likelihood of WFH, although this association is observed only among mothers. The following paragraphs discuss these findings further.

First, the results do not provide evidence of a gender difference in the likelihood of working from home among Chinese parents in the CGSS 2015 dataset. Although mothers’ likelihood of WFH is marginally positively associated with the number of children (p < 0.10), fathers’ uptake of WFH does not appear to be directly driven by childcare demands. This pattern is inconsistent with prior evidence from Western contexts, where mothers with greater caregiving responsibilities are generally more likely than fathers to adopt WFH arrangements to balance work and family demands (Golden 2008; Mas and Pallais 2017; Ewald and Hogg 2022). Given that mothers are typically expected to assume primary caregiving responsibilities in the Chinese family context, one might expect them to be more likely than fathers to rely on WFH arrangements to accommodate childcare needs. However, the findings do not support this expectation. One possible explanation is that WFH has not become a primary strategy for Chinese parents to accommodate childcare and other family responsibilities, particularly in the pre-pandemic period covered by the dataset.

In the Chinese context, under asymmetrically gendered role expectations (Pleck 1977) and the ideal worker norm (Acker 1990), mothers continue to adjust their work arrangements to accommodate dual roles and meet expectations of being both an “ideal mother” and an “ideal worker.” However, fulfilling the role of ideal worker is often regarded as fulfilling the role of “ideal fathers” (Ji 2025), and reduced physical presence in the workplace may jeopardise these Chinese fathers’ image as ideal workers. In line with this pattern, this paper further shows that fathers who hold more traditional gender role attitudes tend to have a higher likelihood of WFH, even in non-agricultural sectors. Nevertheless, those fathers with more traditional gender role attitudes tend to be less likely to expand their involvement in family responsibilities when they can WFH (Kurowska 2020). As a result, it may be that these Chinese WFH fathers are not primarily motivated by an intention to increase their devotion to unpaid work within the family. These results can further explain why childcare demands are not the primary driver of fathers’ likelihood of WFH.

Second, WFH arrangements in China are less likely to be linked to privileged groups among both mothers and fathers in the pre-pandemic period, compared with the Western context. Prior research in Western contexts shows that access to FWAs has typically been higher among those with higher levels of education and strong job performance (Brescoll et al. 2013; Barrero et al. 2023; H. Wang et al. 2023; Rubino et al. 2024). However, this study finds that education level is not statistically significantly associated with the likelihood of WFH for either mothers or fathers in China. Instead, the results show that Chinese mothers who have a higher likelihood of WFH tend to be those without formal employment contracts, those employed in the private sector, and those holding multiple jobs. Chinese fathers in the agricultural sector also have a higher likelihood of WFH than those who are not.

Considering the expansion of the informal employment market generated by the gig economy and digital platform economy in China (Cai et al. 2021), the results in this paper indicate that WFH parents in China tend to be concentrated within non-standard work employment, particularly before the pandemic. This work employment in China is typically characterised by lower job stability, intensified workloads, and a lack of employment protection and bargaining power (Kalleberg and Hewison 2013; Nawakitphaitoon and Tang 2021). In line with previous research, FWAs in China are often utilised to supply household income by undertaking additional work (H. M. Wang et al. 2020). In addition, the deviation of WFH parents in China from the occupationally privileged groups further indicates that Chinese parents are less likely to adopt WFH when it does not align with the ideal worker image (Acker 1990), which emphasises physical presence in the workplace and long working hours and as visible indicators of devotion to work (Lu and Cooper 2022). This pattern further suggests that WFH in China may partly reflect home-based or informal work arrangements rather than flexible work arrangements as commonly understood in Western contexts, particularly before the pandemic, when WFH was less widespread and less legitimised in China.

Third, mothers and fathers show different patterns in the work-related determinants of WFH. Chinese mothers who take up WFH are more likely to be situated in informal employment. Specifically, the results suggest that mothers without a formal employment contract, those working in the private sector, and those holding multiple jobs have a higher likelihood of working from home. Informal employment is closely related to weak bargaining power in labour markets and increasingly concentrated among women in China (J. Wang et al. 2016). However, fathers work in the agricultural sector have a higher likelihood of WFH than those do not. For fathers, engagement in WFH appears to be more strongly shaped by the nature of agricultural work. A high proportion of agricultural workers are self-employed, and farming activities can often be carried out from or near the home (Garrote Sanchez et al. 2021). The spatial embeddedness of agricultural work within or in proximity to the home thus affords these fathers greater flexibility in arranging their work schedules.

Fourth, this study suggests that mothers who live with grandparents have a lower likelihood of WFH than those who do not, while this association is not statistically significant among fathers. This pattern may suggest that mothers with greater intergenerational support are less likely to rely on WFH to accommodate work and family demands. Previous research on WFH in the Chinese context has emphasised the importance of grandparents in sharing childcare responsibilities, which can shape parents’ WFH experiences, particularly those of mothers (Sun et al. 2022; Xiong et al. 2023). Intergenerational support may substitute for flexible working arrangements in balancing work and childcare responsibilities for Chinese mothers. However, prior studies also argue that intergenerational support mainly alleviates housework burdens rather than reducing mothers’ childcare time. Under the norm of intensive mothering (Hays 1996) in China, mothers increase their interactions with children when they have substantial support from grandparents, as grandparents greatly share housework (Craig and Powell 2012; M. Zhou et al. 2022). These previous studies indicate that the availability of intergenerational support tends to mainly alleviate mothers’ burden on housework rather than reducing their childcaring involvement. This pattern may further help explain why the effects of the number of children do not become statistically significant even after controlling for co-residence with grandparents. Although these findings reflect patterns observed in the pre-pandemic period, they also suggest that intergenerational support in China may have broader implications in the post-pandemic context for how parents manage their work arrangements and childcare responsibilities.

This study contributes to filling the research gap by exploring parents’ uptake of WFH arrangements, portraying the sociodemographic determinants of parents working from home in China, with particular attention to gendered patterns among mothers and fathers. Furthermore, this study is also an exploratory and preliminary study to capture the sociodemographic profile of WFH parents in the Chinese context, focusing on work patterns and family-related characteristics, using a nationally representative dataset. This exploration extends the understanding of parents’ WFH arrangements in a range of Western cultural contexts with relatively well-developed flexible working policies to an underexplored East Asian setting.

The findings further contribute to the theoretical understanding of the ideal worker norm in the context of flexible work arrangements. Existing research often highlights the stigma attached to fathers’ use of flexible work arrangements, as traditional masculinity norms expect men to devote themselves more to the workplace than to family demands (Chung 2020; Cook et al. 2021; Yucel and Chung 2023). However, the findings of this study suggest that the tension between WFH and the ideal worker norm in China also reflects a broader institutional context that strongly emphasises visible physical presence and long working hours as indicators of work commitment. In such a context, WFH may be less likely to be legitimised as a common work arrangement for parents. Second, traditional gendered role norms would suggest that mothers are primary caregivers (Pleck 1977) and, therefore, more likely to adopt WFH to accommodate childcare responsibilities. However, the findings imply that the gender differences in the uptake of WFH can also be shaped strongly by cultural and institutional factors. Specifically, in the Chinese context, flexible working arrangements have not yet been widely institutionalised, particularly before the pandemic. Instead, childcare responsibilities are more commonly shared through intergenerational support from grandparents. These findings also have broader policy implications. Promoting WFH needs to be accompanied by stronger institutional legitimacy for flexible working arrangements in China. This measure could help reduce the stigma associated with adopting such arrangements, provide parents with additional options for reconciling paid work and care responsibilities, and may also contribute to improving fertility rates.

There are several limitations to this research that should be considered when interpreting the findings. First, the use of cross-sectional data does not allow for causal inference. For example, the observed positive association between holding multiple jobs and the likelihood of WFH could also be interpreted inversely: parents who already work from home may be more likely to take on additional jobs due to the flexibility of remote work arrangements. Longitudinal data would be better suited to examining how changes in family structure and employment conditions shape parents’ adoption of WFH over time. Second, the CGSS 2015 data reflect the profile of WFH in China more than a decade ago, prior to the COVID-19 pandemic. Some structural and institutional factors identified in this study may continue to shape WFH adoption beyond the pandemic. However, since the pandemic has substantially transformed the prevalence and social acceptance of WFH globally, the findings in this study may be less able to capture contemporary post-pandemic practices. While the discussion considers the broader implications of these findings, the results should be cautiously used to make direct claims about post-pandemic WFH patterns. Nevertheless, the CGSS 2015 remains the most recent large-scale publicly available dataset that enables a national-level analysis of WFH among parents in China and therefore provides an important baseline for future comparative research using post-pandemic data. Third, although the CGSS dataset is nationally representative, the number of parents working from home remains relatively small after sample selection. A limited sample size may reduce statistical power, making it more difficult to detect statistically significant associations and potentially obscuring meaningful relationships (Button et al. 2013). In addition, the subgroup analyses conducted separately for mothers and fathers further reduce the sample size, which may lead to instability in coefficient estimates. Relatedly, the inclusion of multiple explanatory variables in models estimated with relatively small subgroup samples may increase the risk of model overfitting. For this reason, the results are interpreted with caution, particularly those with relatively wide confidence intervals. Despite these limitations, the dataset still enables meaningful statistical comparisons of key differences in WFH patterns between Chinese mothers and fathers. Future research could build on this study in several ways. First, using longitudinal datasets collected in the COVID-19 pandemic would allow researchers to examine how the expansion of remote work across sectors has reshaped the sociodemographic characteristics and experiences of parents working from home in China. Second, future research could adopt a life-course perspective by examining how parents’ likelihood of WFH varies across different stages of family life, particularly as mothers’ and fathers’ involvement in childcare often changes with children’s age. Such analyses would provide a more nuanced understanding of how family structure and institutional contexts jointly shape parents’ engagement with WFH.