Housework Reallocation between Genders and Generations during China’s COVID-19 Lockdowns: Patterns & Reasons

Round 1

Reviewer 1 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

The authors have addressed most of my previous comments, and the paper is improved. One minor comment on Table 3 (regression): it may not overturn the results, but it is more conventional to treat residence (1 = rural, 2 = suburban, 3 = urban) as categorical rather than continuous.

Author Response

Thank you immensely for your valuable and insightful feedback, as well as your unwavering support in enhancing the quality of the paper! Following your suggestions, I have revised the table, replacing the previous variable with a categorical one, where 1 indicates urban and 0 non-urban settings. This modification has yielded consistent results.

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The authors have thoroughly addressed every concern and query raised by the reviewers. I appreciate their efforts in revising the manuscript in accordance with these comments. Overall, the manuscript displays strong writing, a sophisticated research design, and provides valuable insights.

Author Response

Thank you immensely for your valuable and insightful feedback, as well as your unwavering support in enhancing the quality of the paper!

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Based on a collection of 1,669 surveys and 100 interviews conducted via internet platforms such as WeChat and Weibo, this research paper examines the reallocation of household chores during the strict lockdowns imposed in China due to the pandemic. The study findings indicate a significant increase in men's involvement in grocery shopping but a decrease in their participation in cooking, cleaning, and laundry during the lockdown. While younger generations also experienced an increase in household labor, this change was predominantly observed among women. The reasons provided for these shifts primarily revolved around the concept of 'doing gender.' The increased availability of time allowed women, particularly those employed full-time, to enhance their engagement in household chores. Men, on the other hand, focused mainly on grocery shopping, as the lockdown context temporarily associated this task with traditional masculine attributes. The authors of the study highlight the unique opportunity presented by the stringent lockdowns in China as an invaluable experimental setting to observe the potential reallocation of housework between genders. Overall, the manuscript is well-written, the research design is sophisticated, and the findings offer valuable insights. I have some minor suggestions for improvement, and I hope these comments are constructive in enhancing the manuscript. 

The authors acknowledge the unique characteristics of the Chinese social context, including the enduring influence of Confucianism, the persistent gendered division of labor within households, and the prevalence of intergenerational households. To provide a comprehensive understanding of the findings in this manuscript, it would be advantageous for the authors to provide a description of the distribution of the four specific tasks (grocery shopping, cooking, cleaning, and laundry) based on previous studies that utilize national representative data from sources other than internet surveys such as WeChat and Weibo. This comparison would allow readers to better interpret the findings of this study by highlighting any differences in gender divisions of these tasks observed in internet-based data compared to other sources. By considering a broader range of studies, the authors can provide a more nuanced understanding of the gendered division of household chores within the Chinese social context. 

Data

The survey and interview data were obtained through paid services on WeChat and Weibo. It would be helpful for the authors to provide more detailed information about the data collection process. Specifically, readers may be interested in knowing whether any sampling procedures were applied during data collection. Additionally, it would be valuable to know if the authors applied any adjustments or weighting factors to account for potential biases in the data, particularly if highly educated individuals were overrepresented in the sample. 

Analysis

The authors have indicated that the samples used in the study lean towards respondents with higher education. It would indeed be valuable to explore how education could influence the division of housework between genders during the lockdown. Therefore, it would be beneficial if the authors could provide additional analyses that examine the distribution of housework between genders based on educational attainment. This analysis would allow for a more comprehensive understanding of how education may impact the gendered division of labor during the lockdown period. 

Findings and discussion

The clarity of the results presented in Figure 1 is somewhat lacking. For instance, it is unclear what the eight patterns of housework reallocation refer to. It would be beneficial for the authors to provide examples that illustrate the findings depicted in Figure 1. Similarly, the information and numbers presented in Table 5 and Table 6 are not sufficiently clear. It is expected that the authors provide a clearer explanation of the findings listed in the tables. The authors could enhance clarity by incorporating examples that illustrate the data presented in Tables 5 and 6. 

As the authors have suggested, the strict lockdown measures implemented during the pandemic in China provide a valuable and unique experimental setting to observe the re-arrangement of housework when economic factors and time availability are held constant. It would indeed be beneficial for the authors to provide a more elaborate discussion comparing the unique findings in China with the patterns observed in other societies during lockdowns. 

Minor errors 

In Figure 1 on page 9, it appears that the bottom-left code (or reason) is obscured by the label why. Please revise this to ensure the clarity of the figure.

Reviewer 2 Report

Comments and Suggestions for Authors

This paper aims to contribute to the literature on the division of unpaid labor during the Covid-19 pandemic by adding research on China, which also expands the focus of work on the division of labor to include a household perspective rather than the typical couple-based perspective. It finds that changes in the division of housework revolved around men taking on increased shopping tasks and women taking on more housework across types, particularly in intergenerational households where younger members took on more unpaid labor. The paper argues that the “doing gender” perspective explains (changes in) the division of labor within households in China during the pandemic.

The topic of this paper on the division of labor within households in China during the pandemic is under-studied and this paper provides a welcome addition to the literature on Covid-19. I enjoyed reading the paper and I think there are some interesting findings in the paper (e.g., the finding that “primary houseworkers attempted to ‘balance’ the load after other family members carried more than before” (line 629) provides insight into the salience of time in how respondents think about the meaning of “balance” in the context of the division of labor).

At the same time, I think there are weaknesses in the analysis and methods/data that I outline below.  

One of the primary premises of the argument seems to be that the lockdown period in China was equivalent to a period of no market work such that the division of labor within households could not stem from inequalities in paid work (time) between men and women and thus were instead attributed primarily to “doing gender”. I don’t think that premise is plausible. According to a citation in this online article about work in China during the pandemic, https://www.publicbooks.org/working-in-china-in-the-covid-19-era/#fn-37819-2: “nationwide, 10 percent of employees were still working at the end of January during the Chinese New Year; 36 percent returned to work in February and 28 percent in March.” So by the time lockdown officially ended in most provinces according to Table 1, 46 percent of the workforce had returned to work. Even though China had the most stringent lockdown regulations, lockdown didn't necessarily exempt people from market work especially as working from home infrastructure was put into place. The paper may likely need a reframed argument if no evidence to support the plausibility of the no-market-work premise can be provided. Relatedly, the paper often misleadingly presents the research in "experimental" terms (e.g., line 74: “experimental opportunity”; line 116: "test which determinant(s)"), when the research design is not an experiment. In addition to the unsupported no market work premise, the data was collected at one (post-"external shock") time period and depends on respondents' own recollection of a pre-Covid time period. In general, Covid-19 is a poor instrument for any quasi-experimental research design because it completely violates exclusion assumptions.

I think Figure 1 is interesting and I think the analysis would benefit from thinking more precisely about how the gender distribution in the sample informs the findings and how that also contextualizes the types of conclusions that can be drawn. In particular, the validity of the argument that the division of household labor in China during the pandemic is primarily the result of “doing gender” is undermined by the interview sample being so skewed female (78%). The paper does not discuss the limitations of having a sample biased in this way. One limitation is that the argument seems to be based mostly on the accounts of women and how they view actions of men in their households through a gendered lens rather than on the accounts of men redefining how they “do” gender as they take on greater responsibilities in specific categories of housework. For example, the paper argues that “the explanation of time availability was not explicitly used to justify men’s increased participation in household labor” (line 557) but quotes from women respondents are used so this is from the viewpoint of women in the sample.

I think coding shared housework activities as +1 to “to each gender” (Line 380) does not make sense analytically. By adding one unit to each, those cases maintain the status quo. An arrangement of equality/sharing should be its own category.

I also think a discussion of the gender and age composition of the households of respondents is missing. Is it more likely than women comprise a bigger proportion of the average multiperson household-- in which case, would we consider an "equal division of labor" within the household to mean more women technically are responsible for household activities? The composition of households also seems relevant for the discussion of the gendered division of labor among younger versus not and unmarried versus not respondents—are there systematic differences in the composition of households for these subgroups that provide an alternative explanation for the findings of gendered differences? I think organization of the findings on subgroup differences could also be improved by presenting them separately and with clearer justification for why these subgroups matter. E.g., for generational differences, this could be done by pulling it out into its own subsection rather than under "Narrated reasons for both genders' increased participation". In addition, re: line 717/Table 6: If there are only 22 cases of family reunions in the data, what is the justification for highlighting it for discussion in the analysis?

I think the paper could also benefit from a discussion of its limitations stemming from data collection and specifically with regard to allowing interview respondents to respond in a written format to the interview guide questions. A first step is to provide the percentage breakdown of participants’ chosen response format. Social-desirability bias is relevant for both verbal and written responses but especially so when responses aren’t extemporaneous and the paper should discuss how this bias was mitigated or at least how it informs the interpretation of the findings. I also think clarifying the specific time period that respondents were instructed to recall in their responses versus the data collection time period (May-August 2020) would be helpful for clarifying how respondents disentangled (or not) different timelines with regard to Covid-19 lockdown phases in China and the factors and outcomes of interest. 

Survey error based on the questionnaire presented in Table 2 also seems to be an undiscussed limitation. The “Activity-specific change” questions and choices presume that cooking, cleaning, shopping, and laundry tasks were unequally distributed and do not provide the choice that a task was done equally/shared by men and women in the household. The question “During the COVID-19 lockdowns in early 2020, was the housework division more equal between genders?” and options “More equal than before; less equal than before; same as before” provides information about the direction of change in division of labor but coupled with the lack of “shared/equal” category in the questionnaire, it’s not extracting precise/accurate information about gender differences in the division of labor. Overall, the survey data doesn’t seem to be adding very much to the analysis, especially in comparison to the richer interview data. Also, sometimes the findings aren't very clearly presented as following from the survey data or the interview data (or both) so erring on the side of specificity in the presentation of the results would be helpful. 

I think quote selection could also be improved insofar as there seems to be a mismatch between the argument being made and the quote used for evidence (if a quote is used). Many of the quotes provide only descriptions from the respondents about what the division of labor was rather than insight into how respondents are thinking about/doing gender in changing their housework behaviors (e.g., lines 484-492; 501-507; 508-513; 690-694). I’d suggest tamping down on breadth of the respondents described in favor of a number of very good quotes to support the argument.

Minor issues:

-          Line 32: Why is (Carlen, 1995) cited when the sentence is about evidence about the Covid-19 pandemic?

-          Line 387 (and other places where interview data is discussed): It doesn’t make sense to use “statistically significant” (or "significant" in other places in the paper) if there’s no significance test run for the interview data (which I am not recommending).

Comments on the Quality of English Language

There are a few typos and confusing turns of phrases where I wasn't quite sure what the meaning intended was, e.g.:

- Line 49: "closed-off management"

- Line 450: "(over)parity"

- Line 578: "derived resources"

- Line 711-713: "primary labor supply" and "secondary labor supply" evoke "primary and secondary labor market" which seems unintended

Language around gender in the paper is quite binary, frequently using phrases like "both genders."

Reviewer 3 Report

Comments and Suggestions for Authors

This study employs a mixed methods approach to examine the patterns and explanations for the gender division of household labor during the initial covid-19 pandemic lockdown in China. While I find the topic interesting and the context important, I have several theoretical and methodological concerns. Below are my suggestions.

 

1. The authors claim that the initial covid-19 pandemic lockdown provides an experimental setting to address the endogeneity between money and time. However, it is questionable whether the relative employment and occupation status between spouses is completely out of the equation during lockdown when it comes to how couples decide the allocation of domestic labor. Pre-lockdown employment status and occupation can still be important predictors of household power dynamics. Furthermore, if men are more likely to be in those occupations in which job security and wages are less affected by the lockdown, then the observed changes between and during lockdown is not entirely due to the external shock and the increased time availability. Instead, it is still largely a result of the within-couple difference in monetary resources such as employment and occupation. The authors may consider conducting regression analysis to account for these characteristics. At a minimum, it is important to clarify whether the sample consists of mainly dual-income couples or breadwinner-homemaker-type couples and to what extent the composition is representative of the country.

 

2. The authors discussed how the division of labor has changed among multi-generational households, which is interesting and challenges the conventional focus on nuclear households. The data on these analyses seem to come exclusively from the interviews, but not the surveys. Furthermore, living arrangements also changed among the interview respondents between and during lockdown. Therefore, the results on the division of labor before and during lockdown are not directly comparable for these respondents. The observed changes in the division of labor before and during lockdown may be a result of changes in household structure, as opposed to changes in time availability. Those who experienced a change in living arrangements and those who did not may also be systematically different. The same issues apply to the authors’ analyses of households with and without paid house helpers. Some respondents may have experienced changes in the presence of paid house helpers before and during lockdown, and those who experienced changes and those who did not may differ in important ways (such as economic resources) that affect the division of domestic labor.

 

3. In Table 3.II, the authors claim that they removed “others” from the analysis (line 377), which means the sum of the % of households where the gender of the primary houseworker is male and the % where the gender of the primary housework is female should be 100%. But this does not appear to be the case. Furthermore, it is unclear how “Diff. Std. Dev.*” is calculated. The t-test should be testing the difference before and after lockdown (e.g., for shopping .61-.33 = .28, yet the statistic shown is .54). Tables 4 and 5 also have no statistical significance tests. It is unclear why the authors choose to only show the statistical significance test in Table 3 but not in others. 

 

4. Table 5 is very confusing. 1) The authors should label the table more clearly (or at least clarify in the table’s notes) as to what the “male” and “female” columns mean. 2) It is unclear what we can conclude from the subsample analysis by respondent’s age when the numbers of observations are so small (five or fewer). The overall changes have already shown evidence of “youngerization”). What additional information does the analysis by age offer? 3) The author’s definition of “rate of change” appears to change from table to table. In Table 4, the rate of change is (during – before)/before. However, in Table 5 and Table 6, the rate of change is (before – during)/before.

 

Comments on the Quality of English Language

The discussion of the quantitative results in section 7.1 can be more streamlined. It is clearer if you present the main arguments first and then for each main argument, present and discuss the supporting evidence.