When parents wish to reduce their work hours: Does sorting into occupations and work organizations explain gender differences in working-time adjustments?

Objective: This study investigates whether sorting into occupations and work organizations contributes to gender differences in parents’ likelihood to reduce their working hours. Background: While mothers reduce their work hours to reconcile their work and personal lives, fathers increasingly wish to reduce their hours but face obstacles to doing so. Mismatches between parents’ desired and actual work hours prompt the question of whether fathers’ realization of working-time reductions is constrained due to their sorting into more time-intensive occupations and/or work organizations. Method: Cross-classified multilevel models were applied to German linked employeremployee data analyzing gender differences in parents’ likelihood of reducing work hours. Including sorting indicators, the question of whether differences in full-time employed mothers’ and fathers’ working-time reductions were driven by sorting into different work contexts (occupations/work organizations) was explored. Results: The results confirmed that full-time employed mothers are more likely to reduce their work hours than full-time employed fathers. While occupations play almost no role in determining working-time adjustments, work context does at least partly contribute to parents’ variation in working-time reductions. However, neither gendered sorting into occupations nor gendered sorting into work organizations explained gender differences in parents’ likelihood of scaling back their work hours. Conclusion: It is concluded that gender differences in German parents’ reduction of working hours rather respond to traditional gender norms than being influenced by the different occupations or work organizations mothers and fathers sort themselves into.


Introduction
Gender inequalities in employment patterns (i.e. work hours, career interruptions) and employment outcomes (i.e. wages, authority) persist despite the alignment of male and female careers (Blau & Kahn 2017;Granato 2017;Schrenker & Zucco 2020). Especially, following the transition to parenthood mothers' and fathers' career paths diverge: it is mainly mothers who take on the main responsibility for childcare and who take advantage of working-time flexibility by reducing their work hours to better integrate work and family demands (e.g., Grunow, Schulz & Blossfeld 2012;Rosenfeld & Birkelund 1995;Van der Lippe 2001). Research provides evidence that mothers' part-time employment is a relevant contributor to the gender pay gap and likewise hampers mothers' access to management positions (Cha & Weeden 2014;Goldin 2014;Leuze & Strauß 2016;Schrenker & Zucco 2020). In contrast, fathers rarely work fewer hours and instead are more likely to work even longer hours after the birth of a child (e.g., Fagan 2004;Lewis, Campbell & Huerta 2008). More recently, however, the ideal of involved fatherhood has gained traction, and fathers' experiences of work-life conflict are becoming comparable to those of their female counterparts (Bünning 2015;Oberndorfer & Rost 2005;Zerle-Elsässer & Li 2017). Although fathers increasingly desire a reduction in work hours (Abendroth & Pausch 2018;Hobson & Fahlén 2009;Pollmann-Schult & Reynolds 2017), they continue to be less likely to realize reductions than mothers (Clarkberg & Moen 2001;Hobson & Fahlén 2009;Thornthwaite 2004). Hence, it seems that possibilities for work-hour reductions are gendered.
Previous research on parents' working-time adjustments mainly addressed household dynamics: While fathers' opportunities to reduce their working hours are constrained because of their financial responsibilities due to the gendered division of labor (Pollmann-Schult 2008;Pollmann-Schult & Reynolds 2017), mothers as primary caregivers are expected to reduce their working-time when returning to the labor market. However, less is known about how the work context shapes parents' opportunities to scale back work hours. This study explores whether gendered sorting into occupations and work organizations explains differences in parents' possibilities for work-hour reduction.
Gender segregation into occupations (Cha 2013;Huffman & Cohen 2004) and work organizations (Fuller 2017;Reskin 1993) has been used to explain gender differences in absolute work hours but has not yet received much attention as potentially explaining gendered working-time adjustments. Althaber & Leuze (2020) provided first evidence that transitions from full-time to part-time work are shaped by occupational working-time norms. However, men and women sort into different work organizations that vary in their work-hour norms, even when they have the same occupations (Bielby & Baron 1986;Bryan 2007). Hence, this article argues that it is essential to look at the impact of not only occupation but also work organization on working-time adjustments. Thus, it is asked: Are gender differences in parents '

possibilities of attaining a reduction in work hours driven by segregation of mothers and fathers into different occupations and/or work organizations?
This study adds to existing research in several ways. First, individual-and family-related explanations for the gender gap in the reduction of work hours are theoretically expanded by work-related explanations, such as gendered segregation into different occupations and work organizations. Second, this study expands previous research on consequences of labor-market segregation (Cha 2013;Leuze & Strauß 2016) by examining to what extent sex segregation into occupations/work organizations influences mothers' and fathers' possibilities of adjusting their work hours. Third, including individual preferences for reducing work hours allows exploring whether gender differences are influenced by mothers' and fathers' different wishes. On the one hand, mothers and fathers might develop different preferences for work hours throughout their careers, regardless of their work context, based on differences in family orientation (Correll, Bernard & Paik 2007). On the other hand, parents might sort into different occupations and work organizations where they are more or less likely to experience stress and high work pressures, resulting in different preferences for reducing their working-time.
Analyses are based on two waves of unique German linked employer-employee data (LEEP-B3) that combine information on employees' preferences for reducing their work hours with information on their actual realization of working-time adjustments within the work organization context. Therefore, a quasi-longitudinal approach by analyzing workhour differences between two observation periods and controlling for parents' sorting into work organizations and occupations can be used. This article is organized as follows: the second section sets the study into context of working-time reductions in Germany and current findings. The third section presents the theoretical framework and discusses the proposed hypotheses. The methodology and the results of the empirical analysis are presented in the fourth and fifth sections. Concluding remarks are drawn in the final section.

Gender differences in the reduction of work hours
Research provides extensive evidence that, regardless of modern couples' egalitarian gender beliefs before becoming parents, after the birth of their first child, German parents retraditionalize their division of labor, following the long tradition of the male-breadwinner ideal (Grunow 2013;Kühhirt 2012;Rosenfeld & Birkelund 1995;Schrenker & Zucco 2020). German parents' division of labor is particularly interesting because they receive contradictory signals from institutions: Since 2001, employees have the right to demand a part-time job; however, the employer can refuse if the firm is unable to change its work organization (Leitner, Ostner & Schratzenstaller 2004). Additionally, mothers and fathers are, since 2007, allowed to take parental leave simultaneously and two months of parental leave are explicitly reserved for fathers (Bünning 2015). Despite this support for both mothers' and fathers' reconciliation of work and private life, few childcare options (Kreyenfeld & Geisler 2006;Rosenfeld, Trappe & Gornick 2004) and joint taxation of German spouses (Hipp, Bernhardt & Allmendinger 2015) might counteract ambitions to foster gender equality in the labor market and might create incentives especially for women to reduce their hours. In 2018, approximately 70% of German mothers with young children worked part-time, but only 6% of fathers did (DESTATIS 2019). However, it becomes increasingly evident that mothers' and fathers' realized employment patterns are not in line with their preferences (Pollmann-Schult 2008;Van Echtelt, Gleebek & Lindberg 2006;Wanger 2011): While full-time working fathers are dissatisfied with their long work hours (Abendroth & Pausch 2018), women in part-time employment are wishing to increase their working-time (Harnisch, Müller & Neumann 2018).
Although some studies indicate that both men and women can realize reductions in working hours, when they prefer to do so, women's working-time reductions remain to be more extensive compared to men's (Seifert et al. 2016). Moreover, gender differences in the reduction of work hours are especially pronounced among parents of young children (Young & Schiemann 2018). However, there is evidence that institutional reconfigurations are effective in increasing fathers' likelihood of reducing their work hours, such as the introduction of so-called "daddy months" in parental leave regulations (for Germany: see Bünning 2015; for Sweden: see Duvander & Jans 2009). Although younger men scale back their work hours due to fatherhood, their adjustments still depend on their female partners' employment and are still substantially smaller than mothers' adjustments (Pollmann-Schult & Reynolds 2017). Moreover, research argues that adjustments in work arrangements for supporting the reconciliation of work and private life (i.e. paternal leave, working-time reductions) are more prevalent among highly educated fathers with highly educated partners (Geisler & Kreyenfeld 2011) and with more egalitarian gender attitudes (Pollmann-Schult & Reynolds 2017). When it comes to their previous involvement in carework and in paid employment, fathers who adjust their work arrangements for family reasons, by taking parental-leave, do not seem to deviate from others (Geisler & Kreyenfeld 2011).
Recent studies also took employees' work context into account for explaining workingtime adjustments and show that men working in occupations with lower working-time norms are indeed more likely to reduce their work hours, whereas they refrain from such reductions in occupations with pronounced long work-hour norms (Althaber & Leuze 2020). In contrast, women in occupations with long work hours are even more likely to change to part-time work. Focusing on work organizations Seifert and Colleagues (2016) found that realized reductions of work hours in line with employees' preferences are related to employer changes, indicating that work organizations also vary in expectations regarding employees' time investment in their work, which might influence their openness to reducing work hours.
One the one hand, empirical findings demonstrate the persistence of gender differences in reductions in work hours to the disadvantage of modern fathers, who increasingly wish to work reduced hours. On the other hand, research provides evidence that constraints in realizing desired work-hour reductions emanate not only from household dynamics but also from work contexts (occupations/work organizations). However, to the best of knowledge, there has not been a systematic analysis of occupations and work organizations and the specific roles they play in shaping fathers' and mothers' ability to realize work-hour reductions.

Gender differences in the reduction of work hours and gendered segregation into work contexts
This section provides theoretical reasoning for why gendered sorting into work contexts (occupations/work organizations) with different work-hour norms should contribute to gender differences in parents' ability to obtain reductions in work hours.

Occupational segregation
Research on occupational gender segregation provides extensive evidence that men and women are clustered in different jobs which is causing gender inequalities in employment outcomes (Card, Cardoso & Kline 2016;Leuze & Strauß 2016). Labor-market research attributes occupational gender segregation to supply-and demand-side factors derived from economic (Becker 1964;Polachek 1981) as well as social psychological and sociological theories (Padavic & Reskin 2002;Ridgeway & Correll 2004; for an overview: Achatz 2018).
On the supply side, the neoclassical economic perspective emphasizes women's sorting into occupations based on rational-choice considerations. As women anticipate lower employment continuity over their life-course compared to men, they sort into jobs in which career breaks due to family formation do not lead to the atrophy of skills and high wage losses (Polachek 1981). Additionally, the compensating differentials argument states that women, based on their role as homemaker, trade off working in occupations that allow for flexible working-time arrangements against lower economic rewards (Filer 1985). Sociological approaches emphasize the internalization of traditional gender roles during socialization that form men's and women's gender identity and cause gendered career goals to develop (Padavic & Reskin 2002). In alignment with gender identities, women are likely to work in occupations where female characteristics, such as being caring and communal, are valued (e.g., nurses and teachers), whereas men sort into occupations where the perception of being ambitious is valued (Gmür 2004). The doing gender approach (West & Zimmermann 1987) argues that gender is not ascribed as stated in socialization theories but actively produced in social interactions. Men and women choose how to act based on their knowledge of what is expected as appropriate behavior.
Demand-side arguments emphasize the relevance of employers' considerations when hiring. Employers might discriminate against women in hiring based on (conscious or unconscious) gender stereotypes, thereby restricting women's access to male-dominated jobs (Achatz 2018;Bernard & Correll 2010;Correll et al. 2007). Statistical discrimination states that employers lack information about workers' productivity when hiring, as a result of which they rely on categorical distinctions, such as gender or educational credentials, as indicators of employees' productivity (Bernard & Correll 2010). Based on employers' previous experiences, they are assumed to hire women for positions that are less timely demanding, whereas men are considered more suitable for time-intensive job.
Empirical evidence confirms that the occupational structure indeed entails expectations about workers' time investment and productivity (Carney & Junor 2014;Cha 2013;Cha & Weeden 2014;Leuze & Strauß 2016;Lott & Abendroth 2019) but mainly focuses on absolute work hours rather than flexibility in adjusting working-time. Research on absolute work hours indicates that male-dominated occupations (i.e., managerial, or professional occupations) are characterized by long work hours, expectations of overtime work and constant availability (Cha 2013;Epstein et al. 1999;Hipp & Stuth 2013). In these occupations, employees signal loyalty to employers or convey motivation (Cha 2013;Bryan 2007). Penalties for violating these norms are high (Blair-Loy 2003;Cha 2013). High work hours are also likely when employers want to keep productivity high at low labor costs, such as in production and operations occupations (Cha 2013;Ely & Meyerson 2010;Horrell & Rubery 1991). In contrast, women sort into more flexible jobs (Goldin 2014). Jobs that developed in the service sector were particularly well-matched to women who prefer to work shorter hours because of family obligations and employers' growing need for a flexible workforce (Horrell & Rubery 1991). Hence, occupations such as teachers, nurses, and social workers (Wrohlich 2017), have on average shorter work hours and allow for reconciliation of work and personal lives (Cha 2013;Damelang & Ebensperger 2020;Epstein et al. 1999;Leuze & Strauß 2016). Working in the same occupations, both men and women are more likely to work part-time when there is a higher number of women in an occupation (Malin 2020). The same was found to be true for the uptake of flexible measures (Minotte, Cook & Minotte 2010). To our knowledge, there was only one study systematically investigating the relevance of occupations' gender-composition to working-time adjustments: Althaber & Leuze (2020) showed that men's transitions from full-time to part-time employment are hampered by long occupational working-time norms.
Based on supply-and demand-side arguments, and previous findings, it is argued that mothers' and fathers' sorting into occupations with different working-time norms should not only explain different absolute working-times but also differential possibilities of scaling back work hours. First, when occupations with a greater share of female employees have lower working-time norms, the organization of work tasks may already allow flexible staff deployment and reduce structural barriers to work-hour reductions. Additionally, long-term monetary losses due to working-time reductions should be lower in female-dominated occupations and therefore not hinder employees' reducing work hours. In contrast, fathers employed in occupations with pronounced long work-hour norms may face constraints in reducing their work hours because the organization of tasks requires full-time work to fulfill job requirements. Additionally, fathers might refrain from scaling back working hours in male-typed occupations, because they fear being considered less productive. Reductions for personal reasons might be in line with female gender roles and therefore accepted in female-typed occupations but not in male-typed jobs. Lastly, based on demand-side arguments, mothers might be hired to less demanding positions that allow for flexible working-time arrangements, whereas fathers might be hired to time-intensive jobs that do not allow for any reductions to accommodate employees' preferences. Hence, it is hypothesized: H1: When considering the share of female employees on the occupation level, differences in the realization of work-hour reductions between mothers and fathers should decrease.

Gender segregation in work organizations
In addition to the long tradition of studying occupational stratification, organizational scholars emphasize the inequality-generating processes that are rooted in variations between work organizations (Baron & Bielby 1985; Tomaskovic-Devey & Avent-Holt 2019). Even if mothers and fathers work in the same occupation, they sort into different work organizations (Bielby & Baron 1984). Relational Inequality Theory (RIT) (Tomaskovic-Devey & Avent-Holt 2019) emphasizes the significance of work organizations' specific norms and rules that employees absorb and adhere to in generating inequality. Work organizations' expectations of employees' work hours potentially emphasize or mitigate occupational work-hour norms. Consequently, it is essential to consider fathers' and mothers' sorting into work organizations with different work-hour norms (Bryan 2007) because this is where working arrangements, such as individual working-time adjustments, are negotiated (Tomaskovic-Devey & Avent-Holt 2019).
Joan Acker's theory of Gendered Organizations (1990) claims that gender inequality is built into work organizations. Most organizations prefer male workers based on their preference for employees who are fully available to the work organization and face few distractions from their work as stated in the ideal worker norm (Williams 2000) and the work devotion schema (Blair-Loy 2003). Preferences are legitimized in work rules, such as standardized career ladders, job descriptions, or job evaluations (Acker 1990;Biernat, Tocci & Williams 2012). Following Acker (1990;2006), the engraining of such preferences into the logic of work organizations systematically disadvantages mothers who are the primary caregivers in households and therefore constrained in the hours they work. When work organizations are organized following the ideal worker norm that mainly fits male careers, gendered status differences are likely to develop there. In today's work organizations, it is still mainly men who hold high-status jobs that are also more time intensive, as opposed to women (Busch 2013;Huffman & Cohen 2004). However, work organizations adhere to different inequality regimes, that is, organizational rules and practices that disadvantage groups of employees with certain characteristics (Acker 2006; Tomaskovic-Devey & Avent-Holt 2019). Therefore, there is also variation in the extent to which work organizations follow traditional logics that advantage male employees who are better able to align with the ideal worker norm. This is, for example, the case for expected work hours that have been shown to vary between work organizations (Bryan 2007) as well as the provision of flexible work arrangements (Alemann, Beaufaӱs & Oechsle 2017;Bernhardt & Bünning 2018;Den Dulk 2001).
Despite increasing need for organizations to provide employees with flexible work arrangements due to societal changes and a changing workforce, some organizations face greater pressures than others to do so, causing workplace variation in possibilities for requesting working reduced hours. Furthermore, organizations that do make these changes might have different rationales for doing so (Abendroth & Diewald 2019;Alemann et al. 2017): On the one hand, they can be an investment in workers' employability. In that case, allowing more flexible work schedules is considered to increase organizations' economic returns because employees can remain employed and productive despite their family obligations (Abendroth & Diewald 2019). In organizations following this ideal, working with reduced hours should be supported by employers and colleagues, and, therefore, employees should be more likely to realize a reduction. On the other hand, following neoinstitutionalism, work organizations might opt for external legitimacy driven by demands from politics, media, or clients to support employees' reconciliation of work and personal lives (Beile & Jahnz 2007;Den Dulk 2001). The formal provision of flexible working-time arrangements, then, is a response to external pressures, while employers might still follow the idea of the ideal worker. When flexible arrangements are integrated due to external pressures, possibilities of realizing desired reductions should be reduced because they might still signal lower commitment and productivity and employees fear stigmatization and negative career outcomes in these contexts (Abendroth & Diewald 2019;Rudman & Mescher 2013;Williams, Blair-Loy & Berdahl 2013). This assumption is in line with Abendroth & Pausch (2018), who applied the Capabilities Approach (Sen 1992) in arguing that fathers' preference for a reduction in work hours is shaped by opportunities that seem available in the work context. Following this line of argument, scaling back might not be an accepted reconciliation strategy when the majority of employees works long hours. Hence, fathers might not feel entitled to act on their desire to reduce their working-time, as this would be a deviation from a full-time norm associated with the perception of lower productivity and less commitment than their colleagues in male-dominated work organizations. In contrast, working-time adjustments might be an accepted strategy in female-dominated work organizations and therefore realized more often. Hence, gender differences in mothers' and fathers' potential to reduce work hours should be shaped by not only differences in work organizations' provision of flexible work arrangements but also the different rationales that are followed when flexible working-time reductions are available.
Research on gendered sorting into work organizations argues, that mothers sort themselves into work contexts that allow them to adjust their work hours to their timely demands at home, even if they have to trade off increased possibilities for reconciliation against higher income and career prospects (Fuller 2017;Fuller & Hirsh 2018;Gangl & Ziefle 2009;Lott & Klenner 2018). Moreover, research shows that female-dominated work organizations are more likely to invest in family-friendly working arrangements, including possibilities of reducing work hours Den Dulk 2001). Especially women transitioning to motherhood are more likely to leave work organizations not offering flexibility in working-time arrangements . As a result, mothers are assumed to be more likely to be employed in work organizations offering more opportunities to reduce work hours. In contrast, based on fathers' role as male breadwinner, fathers might sort into work organizations with a pronounced norm of working full-time that restricts opportunities for reductions. Research provides evidence that fathers are more likely to be employed in high-wage and high-productivity firms (Bruns 2019), which might indicate increased timely demands and constrained ability to scale back hours.
Gendered sorting into work organizations might also be induced by employers: Based on employers' discriminatory hiring behavior, mothers whose available time for work is perceived to depend on their personal duties (Ridgeway & Correll 2004) might not be considered able to fulfill time expectations and therefore might not be granted access to demanding work organizations . In this case, fathers should be more likely to be hired into work organizations with pronounced long work-hour norms and restricted opportunities to scale back. Hence, it is assumed that gender differences in the likelihood of reducing work hours are driven by mothers' and fathers' sorting into different work organizations that support or hinder their adjusting their work hours: H2: When considering the share of female employees at the work-organization level, differences in the reduction of work hours between mothers and fathers should decrease.

Sample
The analyses are based on the first (2012-2013) and second wave (2014)(2015) of the German Linked Employer-Employee Panel Survey (LEEP-B3). The data were collected within the study "Interactions Between Capabilities in Work and Private Life: A Study of Employees in Different Work Organizations". The study was conducted in cooperation with the Institute for Employment Research (IAB) in Nuremberg ). The dataset combines information at the organization level (including administrative data) with information at the level of employees of these organizations and their partners or spouses (Jacobebbinghaus et al. 2015). First, work organizations were randomly selected from administrative records. Next, employees were randomly selected, and with their consent, their work histories were added to the survey data. The final data set is representative of workers who are subject to social security contributions and employed in large German work organizations with at least 500 employees (Diewald et al. 2014). The hierarchical data structure and the information on employees' occupations allows for clustering employees in occupations and work organizations.
Because working-time adjustments of mothers and fathers between waves 1 and 2 are of interest, only respondents who took part in the survey in both waves, with children living in their households, and who could be matched to work organizations were included. These are employees who remained with the same employer over the two observation periods. Because many mothers in the sample already realized a reduction of working hours, the sample was restricted to parents working full-time in wave 1 (>35 hours per week). Respondents also had to indicate whether they wanted to reduce their work hours in wave 1. After all restrictions were implemented, the final sample contained information on 1464 employees in 100 work organizations with 415 mothers and 1049 fathers (see Table 1). Work-hour adjustments were measured by subtracting contracted work hours in wave 2 from work hours in wave 1. The variable differentiates employees who reduced their working-time by at least 1 hour (1) from employees reducing their working-time by less than 1 hour or who increased their work hours (0). Approximately 17% of the employees in the sample reduced their work hours between the waves: 13% of fathers and 26% of mothers (see Table 2). Comparing these numbers to other studies, it becomes evident that the identified reductions vary, based on different measurements and observed time periods: Althaber and Leuze (2020: 331) described that 4% of men and 17% of women transitioned from working full-time to working part-time. Explicitly focusing on transitions from fulltime employment to part-time employment, the numbers are smaller compared to the current study that considers all reductions in actual working hours. Seifert and colleagues (2016: 305) found that within a period of three years about 40% of employees were able to realize reductions in actual working hours by at least three hours. Sopp and Wagner (2016: 56) found that between 1999 and 2013 about 19% to 25% of employees were able to realize their preferred reduction in working hours between two observation periods. These numbers are comparable to the results of the current study.

Sorting indicators
Mothers' and fathers' sorting into occupations is indicated using the 2-digit indicator of the 2008 version of the International Standard Classification of Occupations (ISCO). Occupations with less than 30 observations were combined with their closest neighbor (if the differences between occupational groups were marginal), leading to 30 occupational groups in the sample 1 . Sorting into work organizations was performed using the ID of each of the 100 1 To rule out that the different numbers of work organizations and occupational groups bias the results, additional sensitivity analyses, using the German Classification of Occupations (KLDB) on the three-digit work organizations. The models also include the share of female employees on the occupation and work organization levels, depicting the proportion of female employees in each occupation or work organization. The shares of female employees were generated based on aggregated information taken from the full-sample of the LEEP-B3 (including 8413 employees; see Table 1) and averaged over the two observation periods (on the level of occupations/work organizations).

Preference for reduction in work hours
The respondents were asked the following question to indicate their preferred work hours: "if you could choose your own number of work hours taking into account that your income would change according to the number of hours: how many hours would you want to work?". If the number of preferred work hours was lower than the actual number of hours employees worked, respondents are considered to prefer a reduction. The question is comparable to the SOEP questionnaire and does not include a prior filter question about level, were added (see Appendix, Table A1). Integrating this more detailed operationalization of occupations led to around 90 occupational groups in the sample. Due to low case numbers within each occupational category, the share of female employees was added based on information from the Microcensus. Results between different operationalizations are comparable. Hence, the smaller impact off occupations does not seem to be driven by the operationalization. Due to a rather small sample size in the all-parent sample, models were reproduced with a larger sample that included all full-time employed men and women (Appendix , Table  A1).
whether a reduction or increase of work hours was preferred. According to Rengers, Bringmann & Holst (2017), the indication of a preference for reducing hours is lower in studies that include a prior filter. Hence, the indication of preferences to reduce hours might be higher in the used data set (45%) compared to other sources (e.g., Microcensus). If parents preferred a reduction in work hours, they had to choose from a list of reasons why. Combining information from those two questions, a categorical variable was built, differentiating between employees with no preference for reduction, when employees did not wish for a reduction or wanted to increase their work hours, and those with a preference for reducing hours for family reasons, for leisure and self-care, or for decreasing stress and work-life conflict. 2 The variable was built based on a previous study by Abendroth & Pausch (2018) 3 , who focused on fathers' preferences for reducing their work hours.

Control variables
As work-hour adjustments are related to individual characteristics and the household situation, it was controlled for age, age squared, marital status, whether children younger than 6 years were living in the household, tenure, and whether the partner was in gainful employment (sample description in Table 2). For a robustness check, control variables on job characteristics, such as supervisory position and whether employees engaged in overwork, were also included. Because gendered sorting into work organizations and occupations as well as decisions about working-time reductions should be influenced by gender norms, individual gender ideology was considered as well.

Methods
To test assumptions on the relevance of gendered sorting into occupations and work organizations for parental gender differences to their likelihood of reducing work hours, linear probability models were estimated (Breen, Karlson & Holm 2018). The explanatory power of occupations and work organizations was investigated by exploiting the hierarchical structure of the data and estimating cross-classified multilevel models (Snijders & Bosker 2012). It is argued that individuals' working-time adjustments are shaped by occupational and work-organization contexts; therefore, it was accounted for occupations when quantifying the relative impact of work organizations on parents' work-hour reductions and vice versa. However, work organizations and occupations are not strictly hierarchically nested. Hence, in this type of models, employees are nested within groups of a two-way cross-classification of occupations by work organizations. The models were built stepwise: First, random intercept-only models accounting for employees' clustering in occupations and work organizations were estimated to disentangle the different variance components 2 Comparing the distribution of the preference-variable between panel respondents, work organization leavers and dropouts who only participated in survey wave 1 shows equal patterns between groups. Except that panel respondents are slightly less likely to wish for a reduction to reduce hours and mothers who wish to reduce their work hours for family reasons are more likely to change employers. Therefore, gender differences and the effect of work organizations on working-time adjustments might be underestimated due to employer changes. 3 For an overview of characteristics of fathers with/without a preference to reduce hours, also see Abendroth & Pausch (2018). in the reduction of work hours (variation due to occupational contexts/work organizations/individual differences). Using the variance components, the intraclass correlation coefficient (ICC) was calculated for each level. This indicator depicts the proportion of the overall variation in the dependent variable that was due to differences between occupations/work organizations/individuals. After gender was introduced in the second model (Model 2), individual-level control variables were added in Model 3. Model 4 includes the share of female employees within occupations, and the share of female employees within work organizations was added in Model 5. The models for all the employees (Models 1 to 5) allow for analyzing the correlation between gender differences in work-hour reductions and the context. Whether fathers' and mothers' working-time reductions were differently affected by work contexts was examined by applying the same procedure to all-fathers (Models 6 to 9) and all-mothers (Models 10 to13) samples.

Gender differences in opportunities to reduce work hours
The boxplots in Figure 1 show the variation in occupational/work organizational averages in mothers' and fathers' likelihood of reducing work hours. The figure indicates that in some work organizations, none of the employees reduced their work hours, while in others, more than 40% of the employees did so. Between occupations, the overall variation in workhour reductions was smaller, compared to work organizations, with a maximum of approximately 33%. Comparing the distributions between mothers and fathers shows a similar pattern for work organizations and occupations, mothers were in both contexts more likely to reduce their work hours than fathers. Additional t-tests indicate that gender differences in the mean values were significant on the work-organization and occupation levels (results not shown). While the median of mothers' reducing their working hours on the level of work organizations was at 16.7%, it was 14.3% for fathers. This indicates that fathers were on average less likely than mothers to reduce their work hours. The difference, however, was rather small. Looking at the median of mothers and fathers reducing their work hours on the occupation level, differences were more pronounced (20.5% mothers, 14% fathers). The boxplots show that the overall variation in working-time reductions was greater among work organizations than among occupations. Hence, the figure reveals that considering occupation alone does not paint the full picture of influences on parents' working-time reductions. Table 3 presents the results of the cross-classified multilevel models, including determinants on the occupation and work-organization levels. Firstly, the identified gender differences in working-time reductions for full-time employed mothers and fathers are examined. Looking at Model 2, fathers' likelihood to reduce their working hours was 12.9% lower, compared to mothers. Interestingly, the effect for fathers does not change significantly following the integration of individual preferences and small children (<6 years) in the household in Model 3. Hence, differences in the likelihood of reducing work hours do not seem to be driven by gender differences in work-hour preferences.

Gender differences in opportunities to reduce work hours and sex segregation
Turning to the relevance of gendered sorting into occupations and work organizations for differences in mothers' and fathers' possibilities to reduce their working hours, neither the share of women in occupations (Model 4) nor the share of women in work organizations (Model 5) seemed to influence employees' likelihood of reducing work hours in the sample. Hence, the findings do not support H1 and H2, which assumed that controlling for employees sorting into female-dominated occupations (H1) and work organizations (H2) would reduce the gender gap in parents' likelihood of reducing their contracted work hours. This findings were further supported when models were estimated for fathers (Models 8 and 9) and mothers (Models 12 and 13) separately: Neither mothers' nor fathers' likelihood of adjusting their work hours was influenced by women's representation in occupations or work organizations 4 . 4 There was a negative effect of women's share of employees in the work organization on fathers' working-time reductions which indicates that fathers were less likely to reduce their work hours when more women were in the work organization. However, the coefficient was only significant on the 10% level and was not robust to different model specifications with a) additional control variables, and b) when a 2-hours threshold for reductions is used (see Appendix, Table A2; A3). The current analyses do not suggest that gendered sorting into occupations/work organizations drives the gender gap in the reduction of working hours, however, looking at the explanatory power of the individual-, occupational and work organizational level shows that the work context shapes full-time employed workers' possibilities for working-time reductions, to some extent. Although, more than 90% of the variation in the likelihood of full-time employed workers reduction of work hours was related to individual-level differences (see Model 1; U1 Individual: 0.129***; ICC Individual: 93.16%), work organizations accounted for 5.33% (U3 WO: 0.007***; ICC WO: 5.33%) and occupations for 1.51% (U2 OCC: 0.002***; ICC OCC: 1.51%). These findings further suggest that work organizations were more relevant to employees' working-time adjustments than occupations. Adding individual-level covariates (Model 3) to the models, the between-level variance of occupations (U2 OCC: 0.001***) and work organizations (U3 WO: 0.008***) stayed almost unchanged, as did the individual-level variance (U1 Individual: 0.124***). Hence, it does not seem to be the case that the variation on each of the levels was driven by workforce composition.
Source: Author's calculations, LEEP 2012-2015N: 1464.     Interestingly, the residual variance terms also remained unchanged after the integration of the share of female employees on the occupation level (Model 3: 0.001***; Model 4: 001***) and work-organization level (Model 3: 0.008***; Model 4: 0.008***). Performing likelihood-ratio tests between Model 3 and Models 4 and 5 further indicate that neither the introduction of the share of female employees in occupations (Model 4) nor the share of female employees in work organizations (Model 5) increased the models' explanatory power. Again, highlighting that differences in the gender composition between occupations/work organizations does not seem to play a role for full-time employed workers' possibilities to reduce their working hours when staying with the same employer. Once again, results were supported, and even more pronounced, in the separate models for mothers and fathers: The clustering in work organizations accounted for approximately 7.27% of the variation in fathers' likelihood of reducing contracted work hours (Model 6). Within the group of mothers, the affiliation with work organizations even accounted for 7.37% (Model 10) and the integration of sorting indicators did not influence the explanatory power.
Lastly, the overall results remained stable when alternative model specifications with a) additional control variables, and b) a 2-hours threshold for reductions were applied (see Appendix, Table A2; A3).

Discussion
Despite the growing relevance of intensive fathering in the public debate and fathers' increased desire to reduce their work hours, it is still mainly mothers who use flexible work arrangements, such as reduced work hours, to manage conflicting demands from paid employment and their private responsibilities. This paper aimed to contribute to the understanding of persisting gender differences in the reduction of work hours, focusing on full-time employed parents' work context (occupations/work organizations) staying within the same work organization. Relying on economic and sociological explanations for occupational gender segregation as well as theories of gendered organizations and relational inequality, it was argued that working-time reductions are also influenced by normative barriers in occupations and work organizations. More specifically, it was assumed that mothers' sorting into work contexts with lower work-hour norms and fathers' sorting into contexts with a pronounced norm to work full-time contributes to gender differences in parents' potential to reduce their work hours.
The results show that full-time employed mothers staying with the same employer are more likely than fathers to reduce their contracted work hours. For these full-time employed parents, results further suggest that gender differences in working-time reductions are, in contrast to the derived hypotheses, not shaped by their sorting into different work contexts (occupations/work organizations). Moreover, findings indicate that parents' potential to reduce work hours is to a great extent shaped by individual differences, whereas affiliation with occupations or work organizations do play a smaller role. Work organizations are, however, more influential for individual working-time adjustments than occupations. This was the case for both, full-time working mothers, and fathers staying with the same employer in the sample.
The results of this study confirm previous research on gender differences in the reduction of working hours, indicating that working-time adjustments are a common strategy for mothers to reconcile work and family demands (Young & Schiemann 2014). Moreover, this article contributes to the literature by examining how parents' sorting into occupations and work organizations relate to gender differences in their ability to achieve reductions in work hours. Although previous research finds that gender segregation on the labor market explains gender differences in overall work hours (Damelang & Ebensperger 2020), the findings on working-time reductions are rather mixed (Althaber & Leuze 2020;Sopp &Wagner 2016). For the full-time employed parents who stayed with the same employer, the current study does not support the idea that their previous sorting into work contexts (occupations/work organizations) shapes differences in mothers' and fathers' potential to reduce their work hours. For this group of parents, the findings indicate that gender differences in work-hour reductions seem to be more responsive to broader societal gender norms. Fathers' lower likelihood of adjusting their work hours, regardless of their affiliation with occupations or work organizations, might be explained by societal beliefs about fathers as primary breadwinners. It may also reflect their need to comply with the ideal worker norm, characterized by full-time employment and undivided attention to their job. In contrast, mothers' reduction in work hours seems to be in line with their role as primary caregivers, an accepted strategy to reconcile work and family. Moreover, mothers may lack any options other than to reduce their work hours due to restricted opportunities to outsource personal duties that are still perceived as their responsibility.
The findings are stable, even when accounting for parents' preferences to reduce their work hours. Hence, results suggest that fathers' lower likelihood to reduce working hours is neither related to their lower preference for a reduction, nor does it seem to be the case that traditional fathers, following the ideal of the male breadwinner, sort into time intensive jobs without developing preferences for reducing work hours. This is in line with Abendroth's and Pausch's (2018) result that fathers' preference for reducing work hours rather develops when they are confronted with pronounced high work demands than in response to work organization norms. These results point to the prescriptive effects of traditional gender norms on parents' labor-market decisions and the path dependency following fathers' selection into working-time regimes, which leaves little room for adjustments when work pressures are high.
Although gendered sorting into work organizations did not explain gender differences in parents' work-hour reductions in the analyses as expected, the findings suggest that work organizations do have the power to shape parents' options for working-time reductions. Previous research points towards the role of regulated organizational practices for work-life balance (formal rules on flexible work / telework; flexible work arrangements available to all employees) as well as a weak ideal worker norm for shorter working hours among fathers (Bernhardt & Bünning 2018). Future research could further investigate whether organizational rules and practices regarding the reconciliation of work and private life affect employees' possibilities for working time adjustments, as well.
This study focusses on possibilities and constraints for working time adjustments within work organizations and occupations. Therefore, analyses do not account for alternative strategies to adjust work hours outside of the work organization or occupation. Some employees might be forced to change occupations or employer if they want to reduce their working hours but cannot realize this in their current work context (Seifert et al. 2016). Future research could investigate when these alternative strategies need to be utilized and whether they are specifically relevant for certain types of occupations/work organizations, or certain types of employees who are not given the possibility to realize working-time reductions in their current work context (e.g., employees in low-status jobs might sort out of their occupation/work organization for realizing a reduction in working hours). While employer changes have been shown effective for realizing reductions in working hours, the findings on occupational changes are mixed (Böheim & Taylor 2004;Seifert et al. 2016). Although the revolving-door argument (Jacobs 1989) suggests that especially mothers might leave male-typed occupations in order to realize a reduction, recent research indicates that selection processes likely happen when women enter male-typed occupations but not afterwards: women occupying leadership positions in male-typed occupations have been shown to not be more likely to leave those positions, compared to men (Malin & Wise 2018).
The analyses also had shortcomings. It was not possible to fully disentangle different mechanisms leading to actual reductions in work hours: Changes in work hours could result from negotiations between employees and supervisors, or they could be driven by employers' preferences for flexible staff deployment. In contrast, it is not clear whether employees not reducing their hours did not ask for a reduction or whether the claimed reduction was not granted. The sample only included large work organizations with at least 500 employees. Research indicates that preferred adjustments in work hours are more likely to be realized in small work organizations (with up to 20 employees; Seifert et al. 2016). This could cause overestimation of gender differences in working-time adjustments because mothers and fathers are both better able to align their work hours with their preferences in small work organizations. Despite these shortcomings, this study makes a relevant contribution to organizational research on working-time reductions by jointly examining two factors that have previously been shown to influence employees' absolute working-times: occupations and work organizations.