Abstract
Can women overcome motherhood penalties by quickly returning to their jobs after childbirth? Do employers discriminate against fathers who take extended family leave? To answer these questions, I exploit some unique features of Germany's parental leave and job application system. My field experiment shows that mothers who only took the mandatory leave of two months are less likely to be invited to a job interview than mothers who stayed home for a year. There is, however, no difference between fathers who took short versus long periods of leave. The results of the supplementary laboratory experiment support my theoretical claim that women who “lean in” and violate the norm of being “a good mother” are judged more negatively than norm-violating men, who benefit from their culturally ascribed higher status in professional settings. My study hence underscores that women are required to enact traditional family roles to “fit in” but men are not. Fathers have more leeway in their behaviors and are evaluated according to a more flexible range of criteria than mothers.
Sociological research has repeatedly shown that parenthood is a key driver of the persistent labor market inequalities between men and women. All else being equal, mothers across a large number of occupations and countries are less likely to be considered for a job interview than childless women (e.g., Baert et al., 2018 for Flanders; Bygren et al., 2017 for Sweden; Correll et al., 2007 for the U.S.; González et al., 2019 for Spain; Hipp, 2020 for Germany). In most countries, mothers, moreover, tend to earn less than comparable fathers and less than childless men and women (e.g., Budig et al., 2012, 2016). For men, by contrast, having children is not detrimental. Fathers do not experience discrimination when applying for jobs (e.g., Bygren et al., 2017 for Sweden; Correll et al., 2007 for the U.S.; González et al., 2019 for Spain; Hipp, 2020 for Germany) and even tend to out-earn comparable childless men (e.g., Mari, 2019 for Germany and the UK).
The gendered treatment of parents by employers may, however, be changing with mothers’ increasing labor force participation and the widespread desire of younger people to divide paid and unpaid work more equally (e.g., de Laat, 2023; Pedulla & Thébaud, 2015 for the U.S.). When mothers are not the primary caregivers, they should no longer experience labor market disadvantages—or should at least experience these disadvantages to a lesser extent. Yet, as fathers take on an increased role in childcare, they may expose themselves to the same disadvantages that mothers experience and may be penalized for violating the male breadwinner norm.
This paper investigates how nontraditional distributions of childcare responsibilities affect mothers’ and fathers’ likelihood of getting a job. Does a quick return to paid work after childbirth allow women to overcome motherhood penalties? Do fathers experience employer discrimination when they move away from the traditional breadwinner role by taking a longer than average period of family leave? To answer these questions, I conducted a field and a laboratory experiment in which I used the length of family leave to indicate either traditional or nontraditional parental engagement in childcare.
Both experiments were conducted in Germany, which—for both theoretical and practical reasons—is an ideal setting for examining how changes in the traditional division of labor affect mothers’ and fathers’ likelihood of getting a job. First, despite some changes in recent decades, the division of paid and unpaid work has remained very gendered (Schober & Zoch, 2019), and women incur comparatively high motherhood wage penalties (e.g., Mari & Cutuli, 2021). Second, Germany's relatively generous family leave policy encourages both parents to take time off after childbirth, and fathers’ leave take-up has increased considerably over the past decade. Third, job applications in Germany commonly include a wide range of private information, including marital status, number of children, and information on both the duration of and reasons for career interruptions, which allows me to unobtrusively manipulate the duration of family leave in the application materials.
To test how employers react to applicants who signal nontraditional childcare roles, I submitted applications from a pair of comparable male or female candidates for real job openings and kept a record of all interview invitations each applicant received. Both applicants were parents but differed in the length of leave they had taken after their child was born. The findings from this field experiment suggest that mothers in Germany experience normative discrimination when deviating from traditional childcare roles, that is, they are penalized for deviating from the image of the “ideal mother” (Hays, 1996) but do not benefit from adhering to the stereotype of the “ideal worker” (Acker, 1990). Mothers who took a short period of family leave (two months) were almost 40 percent less likely to be invited to a job interview than mothers who took a long period of family leave (12 months). For fathers, however, the likelihood of being invited to a job interview was unaffected by the length of family leave.
To further explore the mechanisms that underlie such normative discrimination against mothers with short family leave, I conducted a laboratory study in which student raters evaluated the same application materials that were sent out in the field study. The results from the laboratory experiment provide additional indication that violations of prevailing gender norms affect mothers’ but not fathers’ likelihood of getting a job interview. Mothers who “leaned in” and only took a short period of leave experienced penalties in hiring recommendations. Rather than being perceived as more competent and committed than mothers who took longer leave, they were perceived as having poorer parental qualities and as being less warm, less communal, less interpersonally appealing, and more hostile than mothers who took longer leave. There were no differences between fathers who took a short versus long leave regarding perceived competence and commitment. Fathers who took longer periods of leave were perceived to be more likeable and better parents than fathers who took shorter leave, but they were also perceived to be less agentic. These findings suggest that being both a good employee and a good parent is possible for men but not for women. Employers do not just evaluate job applicants as potential employees but also as parents. While this is beneficial for men, it actually harms women.
This study contributes to the existing literature in several ways. First, in light of ongoing debates about paid family leave for new parents, the findings are of great importance in understanding how different degrees of childcare involvement affect men's and women's labor market chances. This is the first study to experimentally examine how parents differ in their likelihood of getting a job by the length of leave taken, and to test different motives for discrimination related to nontraditional childcare roles with both field and lab experimental data. Second, the results of this study advance our theoretical understanding of the motives for discrimination. The finding that mothers who signal great commitment to paid work are less likely to be invited to a job interview and are not perceived as more committed to their careers or more competent than mothers who signal a strong family orientation suggests that employer discrimination may arise from both productivity-related and non-utility-related considerations. The finding that fathers are not penalized for having taken a long leave suggests that in the workplace context, high-status actors—a group to which fathers belong (Ridgeway & Correll, 2004)—generally have more leeway in their behavior than low-status actors—a group that includes women and mothers (ibid.). For employers, it seems, women who violate the “ideal mother” norm and do not prioritize their family but instead “lean in” to adhere to the stereotype of being the “ideal worker” are not a good “fit” (Nichols et al., 2023). However, they assess whether fathers are a good fit based on a broader range of evaluation criteria and give men more leeway in their behaviors. Finally, the findings of this study are of practical importance as they challenge the widespread belief that women can avoid motherhood penalties if they spend as little time as possible at home after childbirth and the assumption that fathers face career repercussions when they take time off to become more involved in childcare.
Background: Gendered Labor Market Effects of Parenthood
Scholarship on gender and work has repeatedly shown that parenthood has different effects on men's and women's hiring chances and wages. Field experiments have shown that mothers have greater difficulties in finding jobs than childless women (e.g., Baert et al., 2018; Bygren et al., 2017; Correll et al., 2007; González et al., 2019; Hipp, 2020). Findings from cross-country comparative survey research, moreover, suggest that mothers earn lower wages than both childless women and men (Budig et al., 2012, 2016; Cukrowska-Torzewska & Lovasz, 2020; Cukrowska-Torzewska & Matysiak, 2020). This motherhood wage penalty is attributable not only to discrimination but also women's and men's self-selection into different jobs, firms, and networks (Brick et al., 2023; Fuller, 2018; Fuller & Hirsh, 2019; Mickey, 2022; Wuestenenk & Begall, 2022). Being a parent, by contrast, does not necessarily harm men's careers. In the U.S., Germany, and Spain, for example, fathers are as likely to be invited to a job interview as non-fathers (Bygren et al., 2017; Correll et al., 2007; González et al., 2019; Hipp, 2020). Becoming a father, moreover, does not decrease men's wages; if anything, fathers tend to earn higher wages than childless men (e.g., Glauber, 2018; Mari, 2019).
One strategy mothers might adopt to reduce these penalties and disadvantages is to minimize their time out of the labor market and signal their strong commitment to paid work. Changes in women's traditional caregiver roles, however, may also affect fathers. If childcare responsibilities are not fully outsourced to relatives or professional childcare providers, mothers’ increased engagement in paid work will require greater childcare involvement from their partners (e.g., Andersen, 2018; Dobrotić et al., 2022; Frodermann et al., 2023; Margolis et al., 2019).
For fathers, the existing research on the professional consequences of nontraditional childcare roles has yielded ambiguous results. Laboratory studies for the U.S. and the Netherlands found that “active fathers” were appraised more negatively and perceived to be less engaged than fathers who did not take leave (Vinkenburg et al., 2012; Wayne & Cordeiro, 2003). By combining field and survey experimental data, Weisshaar (2018) also found that “opt-out” fathers in the U.S. had lower chances of re-entering the labor market than men who had lost their jobs. Likewise, observational studies from the Finland, Norway, and Sweden found (some) wage penalties for leave-taking fathers (Albrecht et al., 2015; Morosow & Cooke, 2022; Rege & Solli, 2013). Other studies, by contrast, either found no effect or showed that fathers even benefited from greater involvement in family life. Studying the effects of family leave on student raters’ perceptions of fathers in Germany, Fleischmann and Sieverding (2015) found no differences in the likelihood of being recommended for hiring between fathers who had taken a long versus short period of leave. Analyzing German panel data, Bünning (2016) found that fathers did not encounter wage penalties for taking family leave. In a series of survey and laboratory experiments, Denny (2013) and Krstic and Hideg (2019) even found positive effects of childcare involvement and leave take-up on evaluations of fathers in the U.S. among different rater populations. Similarly, in her analysis of European panel data, Koslowski (2011) found that fathers who reduced their working hours to have more time with their children earned higher hourly wages than fathers who did not.
For mothers, little empirical evidence on the effects of assuming nontraditional childcare roles on labor market outcomes currently exists. Findings from a laboratory study with U.S. undergraduate students suggest that mothers do not benefit from demonstrating high competence and commitment to paid work: Female raters perceived highly successful mothers as less likeable and warm than otherwise identical fathers, and penalized successful mothers in recommendations on salary, hiring, and promotions (Benard & Correll, 2010). A laboratory experiment on business students in Spain, by contrast, found that fictitious job candidates who expressed a strong commitment to paid work were more likely to receive hiring recommendations and opportunities for professional advancement, and that this effect was stronger for female than male applicants (Aranda & Glick, 2014).
Theoretical Influences: Discrimination, Status Characteristics, and Caregiving Norms
To explain the potentially adverse effects of nontraditional caregiving roles on men and women, I propose the following two mechanisms: First, employers may penalize both mothers and fathers for having taken a long leave, because this indicates low productivity and low workplace commitment (statistical and status-based discrimination). Second, employers may penalize men for taking a long leave and women for taking a short leave, because these behaviors violate prevailing gender norms and do not conform to the gendered image of a good parent (normative discrimination).
Theories of statistical discrimination (Arrow, 1973; Phelps, 1972) and status-based discrimination (Correll et al., 2007; Correll & Benard, 2006) both suggest that mothers and fathers with more extensive childcare responsibilities face discrimination because employers expect them not to be “ideal workers” and hence less productive than individuals with fewer childcare commitments. The two theories differ, however, in their assumptions about employers’ underlying motives. Theories of statistical discrimination argue that employers behave in a rational and risk-averse manner when offering jobs to new employees. Because pre-hire data on future productivity is usually unavailable or costly to obtain, employers base their decisions on the perceived mean performance of a given demographic group (or the variance in performance). This reduces their risk of hiring a low-performing worker. Because both mothers and fathers who signal greater involvement in childcare may, on average, be less productive and less committed to their work than parents with fewer childcare duties (or have greater variance in productivity and commitment), employers perceive them to be “risky hires” and are therefore less likely to offer them a job (e.g., Birkelund et al., 2021; Bygren et al., 2017; Galos & Coppock, 2023; Kübler et al., 2018; Wenz & Hoenig, 2020).
Theories of status-based discrimination challenge the assumption that discrimination occurs on rational grounds and question whether employers draw on accurate information in their decision-making. Status-based discrimination theories argue instead that employers rely on biased priors when evaluating workers and tend to hold low-status individuals to higher standards than high-status individuals. Characteristics of low-status individuals in the paid work context include being a caregiver and being female (Ridgeway & Correll, 2004). Although employers may perceive mothers and fathers to be more competent at care and nurturing than nonparents, these prototypical female skills are of lower value to them than skills culturally ascribed to men (see Benard & Correll, 2010 on laboratory evidence on such biased standards). As stronger commitments to childcare make the characteristic of being a caregiver more salient, employers should therefore also be less likely to hire applicants who signal more intense childcare involvement.
Theories of normative discrimination (Benard & Correll, 2010), which can be thought of as a subcategory or specification of taste-based discrimination (Becker, 1957), attribute preferences for or against certain demographic groups to prevalent cultural norms. Employers may discriminate against mothers and fathers for engaging in gender-nonconforming behaviors (e.g., Rudman & Fairchild, 2004; Rudman & Phelan, 2008). Because women traditionally assumed the primary caregiving role, employers may judge them based not only on their professional skills and work commitment but also on their qualities as mothers (Acker, 1990; Blair-Loy, 2004; Rudman & Mescher, 2013). Women may be applauded when they conform to the norm of being a caring and hence “ideal mother” and penalized when they do not. Likewise, men may be judged according to the gendered expectations of what being a man “should” entail. In this sense, being responsible for a family's material well-being and having a strong career orientation may be taken as a signal of masculinity (Cha & Thébaud, 2009; Vandello et al., 2008). Consequently, engaging in gender-nonconforming behaviors and signaling intensive caregiving responsibilities could potentially be detrimental to men's careers because it violates both the “ideal worker” and the norm of the “ideal father” who is a committed breadwinner.
Empirical Context and Predictions: Employer Discrimination and Nontraditional Family Roles in Germany
To empirically test the predictions derived from the different discrimination theories, I collected experimental data in Germany. Despite major advances in reducing gender inequalities in recent decades, Germany is still characterized by a strong prevalence of “intensive mothering” and “male breadwinner” norms (e.g., Blome, 2017; Gangl & Ziefle, 2015). The division of unpaid work in couples still follows traditional patterns (Deuflhard, 2023; Zoch & Heyne, 2023), and women's labor force participation and weekly working hours still lag considerably behind men's (Eurostat, 2023; OECD, 2023). Mothers, moreover, suffer comparatively high wage and hiring penalties (Cukrowska-Torzewska & Matysiak, 2020; Frodermann & Müller, 2019; Hipp, 2020).
The German Policy Context: Childcare Shortages and Gender Disparities in Parental Leave
These gender differences are not surprising given Germany's ambivalent mixture of incentives and disincentives to maternal employment—a policy context that is similar to those in many other rich democracies (von Gleichen & Seeleib-Kaiser, 2018). Despite Germany's expansion of publicly-provided childcare, only around one third of children aged 0–2 are enrolled in childcare, compared to more than 90 percent among children aged 3–6 (OECD, 2023). Likewise, although the proportion of fathers taking parental leave has increased considerably after a Swedish-style system was established in 2007, the distribution of the available leave has remained very gendered. Out of the up to 14 months paid parental leave per child available to mothers and fathers in two-parent families (remunerated at 65 percent of the leave taker's previous income, up to a cap of 1,800 euros per month), which can be divided between the two parents as desired, mothers typically take 12 months and fathers take either no leave at all (around two thirds) or the two months that would otherwise be lost (BMFSFJ, 2016; Mari & Cutuli, 2021).
In my study, I used the length of paid parental leave to signal parents’ childcare involvement. Employers were likely to interpret not taking any leave or only taking short leave as a signal that an employee is a dedicated worker and likely saw taking long leave as a sign that an employee is a dedicated father. Using the length of family leave as the experimental treatment ensures both the internal and external validity of my study's findings. Longer leave, particularly among fathers, has been found to be associated with greater parental childcare engagement even after the leave ends (Bünning, 2015). Indicating the length of leave in a CV in Germany, moreover, is an unobtrusive way to signal childcare involvement in written application materials.
Empirical Predictions
The theoretical considerations described above lead to two sets of opposing empirical predictions for the effects of short versus long leave-taking on mothers’ and fathers’ hiring chances in Germany. First, theories of statistical and status-based discrimination suggest employer discrimination against both mothers and fathers who take long family leave for violating the norm of the “ideal worker.” Employers should perceive parents with long leave as primary caregivers and hence as less productive and less committed than parents who take short leave. Long leave-takers, moreover, may take an equally long period of parental leave if they have another child, which may further reduce productivity expectations. Hypotheses 1 & 2 therefore predict that employers should prefer fathers and mothers who take short family leave to fathers and mothers who take long leave. Second, and by contrast, theories of normative discrimination suggest employer discrimination against fathers with long leave and against mothers with short leave. This is because employers’ decisions are not based on productivity expectations but on aversions to norm-violating behavior. Being an active caregiver is tightly linked to femininity; therefore, using parental leave to take on a more active role in caregiving violates masculinity norms. Employers may therefore perceive fathers who have previously taken long leave as less manly and less agentic than fathers with short leave. Likewise, they may perceive mothers with short leave as less warm, less communal, and less likeable than mothers who have taken long leave. Hypothesis 3 therefore predicts that employers prefer fathers who take short family leave to fathers who take long family leave, and Hypothesis 4 predicts that employers prefer mothers who take long family leave to mothers who take short family leave.
Experimental Design
To test these hypotheses, I conducted a field and a laboratory experiment. With the field experiment, I sought to examine the extent to which mothers and fathers who signal stronger versus weaker family involvement experience employer discrimination when applying for real jobs. With the laboratory experiment, I sought to further examine the underlying mechanisms of potential discrimination against parents who signal different degrees of childcare involvement. Both experiments employed a 2-by-2 within-subject factorial design (within-subject designs require smaller sample sizes than between-subject designs in “reasonable” labor markets, see Hipp, 2020).
Experimental Materials and Treatment
In the field experiment, comparable application materials from either two fictitious female or two fictitious male job seekers were sent in response to actual job advertisements, and in the laboratory study, student raters were asked to rate the same application materials on a variety of dimensions. Both of the fictitious applicants were parents of a 3-year-old child. Because of the wide range of information included in job applications in Germany, the duration of family leave could be easily and explicitly incorporated into the job applicants’ CVs.
The application materials used in both studies were designed for the position of an event manager. 1 This choice was guided by several considerations. First, in Germany, event management is considered to be a medium-status occupation in which men and women are fairly similarly represented (Wünsch, 2008). This minimizes the risk of occupational gender compositions driving potential differences in discrimination (Fuller & Kim, 2023; Galos & Coppock, 2023; Landivar, 2020; Mai, 2022). Second, event managers are hired in various sectors (e.g., hospitality, banking, insurance, leisure, public service, etc.) and by organizations of different sizes (Wünsch, 2008; see Bjørnshagen, 2022; Yu & Hara, 2021 on the relevance of company size for hiring discrimination). This increases the generalizability of the study's findings. Third, the tasks and activities that event managers perform are similar across different sectors and types of organizations. This makes it easier to use the same application materials for different types of job advertisements and avoid introducing potential bias due to changes in application materials. Moreover, some of the task and skill requirements in event management tend to be seen as “typically male,” whereas others are seen as “typically female.” This gender balance in task requirements ensures that preferential treatment of short versus long leave-takers is not inherent in the job. Fourth, the number of job openings for event managers advertised per week on the most widely used job search websites in Germany is limited (between six and 10 per week at the time I collected my data). Therefore, I was able to submit applications for all suitable jobs and could thus interpret differences in invitation rates in substantive terms. Finally, the fact that the tasks and job requirements of event managers are easily understood by laypersons made it possible for student raters in the lab to evaluate the fictitious job applicants
The application materials used in the field and the laboratory experiment consisted of a cover letter, a CV, and a reference letter from the current employer. To create realistic and suitable applications, the CVs were modeled on those of real event managers found on professional network platforms.
Both applicants stated in their CVs that they had been born in the mid-to-late 1980s, were married, and had one 3-year-old child for whom they had taken either two or 12 months of parental leave. At the time of data collection (and to this very day), two months was the median leave duration for fathers who took leave (BMFSFJ, 2016) 2 , while 12 months was typical for mothers (at the time this study was conducted, around 15 percent of mothers of young children who lived with a partner took no leave beyond the mandatory two months, which was the second most frequent duration of leave for mothers in Germany after the 12 months taken by 75% of mothers, see Bernhardt et al., 2016). Including such personal details in job applications is standard practice in Germany and does not raise suspicion.
I slightly varied the fictitious job applicants’ birth dates and timelines for transitioning between education, trainee programs, and internships to ensure that the 10-month difference in family leave did not affect their total work experience. As childcare enrollment rates for children aged three and older are extremely high (> 90%, OECD, 2023), employers and raters had good reason to assume that the applicants’ care responsibilities would not interfere with their work responsibilities. In addition, given that mothers in Germany typically have their second child in their early 30s and that the typical age difference between children in Germany is 2.5 years (Human Fertility Database, 2024), employers may have also expected that the applicants would have another child in the near future.
To reduce suspicion and risk of detection, the fictitious applicants had very common first and last names (first names were chosen from the top-20 list in the respective age cohort). As is typically the case in Germany, the CVs included photos of the applicants. To ensure that our applicants all looked equally attractive, competent, trustworthy, and so on, numerous photos were pretested in the lab with student raters to select comparable pictures (i.e., pictures with no statistically significant differences in mean ratings on any of the dimensions that were of interest to this study). The names of applicants’ current and previous employers (with whom they had completed an internship) were modeled after real (small) firms.
The application materials were reviewed by HR managers from an event management agency and a hotel and modified based on their suggestions. In addition, the application materials were tested informally (without the experimental condition) with colleagues and students in lectures, and systematically in the laboratory to ensure that the design of the applications was perceived to be of similar quality (although the experimental conditions across the pairs of CVs were counterbalanced in both the field and the lab experiment to ensure that only the effect of the treatment was measured). As the data collection ran for almost 1.5 years, the applicants’ age, dates of graduation, children's dates of birth, and so on, were updated in the CVs and the letters of recommendation, ensuring that the application materials were comparable over time.
Field Experiment
In the field experiment, a total of 718 applications were submitted in response to 359 actual job advertisements between November 2014 and February 2016. Each employer received two applications from either two fathers (n = 358) or two mothers (n = 360) who had either taken two or 12 months of parental leave. Job openings were randomly assigned to either the male or female pair of applicants; random assignment was also used for the treatment (two versus 12 months of leave), the application template (including the photo and the layout of the CV and cover letter), and whether the application from the parent with the two or the 12 months of leave was sent first. To minimize disruption to employers in their hiring processes and ensure that fictitious applications did not disadvantage real job seekers, all applications were promptly withdrawn after receiving any response from an employer (these precautions were also a key requirement for obtaining ethical approval for the study). In order to minimize suspicion, the second applications were sent out at a time lag of two to four days. A balance check showed that the experimental conditions were approximately evenly distributed between the applicants and the order in which the applications were sent out. The number of applications sent out to employers was based on a priori power calculations and results from previous research.
The outcome variable in the field experiment was the invitation to a job interview. This choice was motivated by both substantive and methodological considerations. Invitations are a more unequivocal sign of employer interest than callbacks and de facto require a smaller sample size to correctly reject a false null hypothesis (Hipp, 2020; Vuolo et al., 2018). Sensitivity analyses using callbacks as the outcome variable showed that the substantive findings of the current study remained unchanged; only the level of statistical significance decreased due to higher sample size requirements (Appendix Table A1).
Laboratory Experiment
In the laboratory experiment, student raters from different disciplines (mainly business, economics, and engineering) were asked to assess the application materials from one of the two sets of fictitious job applicants—either two male or two female applicants, who differed regarding their family leave but were otherwise comparable. Students were told to rate the applications based on their first impressions. No cover story or other form of deception was used, and student raters were not given any specific information about parental leave regulations to avoid revealing the main purpose of the study. The laboratory data were collected between November 2014 and February 2015. Student raters’ gender was equally distributed across conditions. Random assignment was used for both the experimental treatment and the application template (whether the application from the parent with the two or the 12 months of leave was shown first and which type of CV and cover letter).
My decision to work with student raters in a highly controlled setting was guided by the following considerations: First, previous studies have found high agreement between student and manager samples in rating job applicants (Cleveland & Berman, 1987; Olian & Schwab, 1988; Trottier & Gordon, 2011, 2018); student samples have therefore also been used in similar studies (Correll et al., 2007). Second, the highly controlled and incentivized lab setting ensures serious engagement with the application materials; this is impossible in online experiments, which typically use self-selected subjects who declare that they have HR responsibilities and which may in fact yield even less reliable findings (Kees et al., 2017). Third, as student raters tend to be less experienced in reviewing hiring materials, they may also be less likely than experienced HR managers to exhibit social desirability bias when asked to assess two similar sets of application materials.
The main outcome variable in the laboratory experiment was the hiring recommendation for one of the two applicants. To explore the causal mechanisms of discrimination against parents with short versus long leave, study participants were also asked to rate the applicants on a variety of dimensions. To capture discrimination for statistical and status-based reasons, they were asked to assess the applicants’ competence and commitment. To capture discrimination for normative reasons, they were asked to assess applicants’ parental qualities, their warmth, likability, and scores on both stereotypical male and female traits. These measures were closely modeled after previous research (e.g., Benard & Correll, 2010; Cuddy et al., 2004; Okimoto & Heilman, 2012; Rudman & Mescher, 2013) and are presented in detail in Appendix Table A2. Moreover, participants also completed a short survey consisting of demographic and attitudinal questions and had the option to leave comments in the open field entries to justify their hiring recommendations.
Analytic Strategy
To investigate whether prospective employers and student raters treat fathers and mothers who took short family leave differently to those who took long leave, I first conducted a McNemar's test, which is a nonparametric test that allows researchers to examine differences between paired proportions (Vuolo et al., 2018). Essentially, the test assesses statistically significant differences in those instances in which only one of the applicants was invited (i.e., the differences between the discordant cells π10 and π01); the information regarding the instances in which either both or none of the applicants were invited is used to determine the size of the difference (i.e., the information provided in the concordant cells π00 and π11). In the next step, I estimated multivariable regressions (linear probability models with employer fixed effects and robust standard errors) to reduce the variability introduced by “applicant characteristics”, such as the photograph or the template of the CV (and student rater characteristics when analyzing the laboratory data). As sensitivity tests, I also ran logistic regressions with clustered robust standard errors (see Appendix Table A3). Finally, I used the measures collected in the laboratory study on the fictitious job applicants’ competence, commitment, likeability, etc., to conduct t-tests and mediation analyses to further explore the nature of the relationship between length of leave and hiring discrimination and to test the validity of the various theoretical explanations.
Findings
Findings from the Field Experiment
To examine whether employers treat parents differently based on the duration of family leave, Table 1 presents descriptive evidence on differences in interview invitations by gender and duration of leave. Overall, the invitation rate for fictitious male applicants was around 20 percent. Independent of the length of leave, fathers in an applicant pair were treated equally by the large majority of employers. No applicant was invited in around 74 percent of cases, and both applicants were invited in around 15 percent. Unequal treatment by employers, which occurred in around 12 percent of cases, was evenly distributed between applicants who took long leave and applicants who took short leave. The length of leave taken, therefore, had no bearing on the fathers’ likelihood of being invited to a job interview—contrary to Hypothesis 1 and Hypothesis 4. There is hence no support for the predictions derived from theories of statistical, status-based, and normative discrimination for men.
Interview Invitations in the Field Experiment by Gender and Length of Leave.
N = 718
For the female applicants, the overall invitation rate was around 18 percent. As was the case for fathers, mothers in an applicant pair were also treated equally by the large majority of employers. No applicant was invited in around 74 percent of cases and both applicants were invited in around 10 percent of cases. However, in contrast to what was found for the male applicants, for the female applicants, short and long leave were treated unequally. Only around 4 percent of mothers who had taken short leave were invited to a job interview, whereas approximately 12 percent of mothers who had taken long leave were invited. The results of the McNemar's test show that these differences in the invitation rates for applicants with long versus short leave are statistically significant at p < 0.01 (McNemar's chi-square of 6.53). Based on this finding, Hypothesis 2, which predicted that mothers who took short parental leave would be more likely to be invited to a job interview than mothers who took long parental leave, is also rejected (statistical and status-based discrimination). Instead, this finding provides support for the competing Hypothesis 3, which predicted that mothers who took short parental leave would be subject to discrimination for being “poor” mothers (normative discrimination).
These results hold when controlling for the design of the application materials and the order in which the applications were sent out. Figure 1 compares the predicted marginal probabilities of receiving an invitation to an interview for mothers and fathers with different family leave durations based on linear probability models with employer fixed effects and robust standard errors (Appendix Table A4 displays the full regression results). Fathers’ probability of being invited to a job interview was not affected by the duration of family leave, whereas mothers’ probability of being invited to an interview increased by 50 percent (8 percentage points, p < 0.05) if they had taken 12 instead of two months of family leave in the past. As the study design ensured that the applicants were equally qualified and as the treatment was randomly assigned, these findings provide causal evidence that the duration of family leave is, on average, not relevant for fathers when applying for jobs. For mothers, by contrast, having previously taken no family leave beyond the legally mandatory time negatively affects their likelihood of getting an invitation to a job interview.
Probability of Being Invited to a Job Interview for Mothers and Fathers by Leave Duration (Field Experiment). N = 718. Note: The figure shows the marginal probabilities derived from a linear probability model with employer-fixed effects and robust standard errors; covariate adjustment for the order in which applications were sent out and application template; given that the application materials included a picture, the invitation rates between men and women cannot be compared (see Hipp, 2020 for an extensive discussion on this issue).
Given that around two thirds of fathers in Germany do not take any family leave at all (BMFSFJ, 2016) and a study for the U.S. found that applying for a job while taking care of a child was more detrimental to men's likelihood of receiving a callback than being unemployed (Weisshaar, 2018), I sought to rule out the possibility that merely taking leave—rather than the duration of leave—affects fathers’ job prospects. To do so, I combined my data with data from a companion study on hiring discrimination against parents versus non-parents, which used the same application materials. The combination of these two data sets allowed me to examine whether fathers who did not mention any leave in their CVs were more likely to be invited to a job interview than fathers who mentioned having taken either two or 12 months of leave. The results of these sensitivity analyses (Appendix Table A5) show no significant difference in invitation rates between fathers who did not mention any leave and fathers who mentioned that they had taken either two or 12 months of leave.
Because cultural, institutional, and economic differences between East and West Germany still persist more than 30 years after reunification—particularly in women's employment rates and childcare availability—I also explored whether employers in East and West Germany differed in their treatment of parents (Appendix Table A4). The results of these additional analyses essentially show that there is no difference in the results for fathers applying to jobs in East and West Germany. Fathers with both short and long leaves are equally likely to be invited to a job interview in both parts of the country. For mothers, however, I find different results for East and West Germany: In West Germany, mothers are more likely to be invited to a job interview when having taken 12 months of leave; in East Germany, there is no difference in invitation rates between mothers with short and long leaves (neither the coefficient for East Germany nor the interaction effect between length of leave and East Germany are statistically significant). It is important to note, though, that the cell sizes for East Germany are very small (only 30 job openings to which female and 30 to which male applicants had applied), which requires a careful interpretation of these within-country differences. I hence only find robust evidence of employer discrimination against mothers with short family leave for West Germany.
Findings from the Laboratory Experiment
While the field experiment showed that employers in Germany discriminate against mothers who had taken short family leave, it cannot fully explain why employers treat mothers differently depending on the length of their family leave, but not fathers. I therefore now turn to the findings from the laboratory study. Both the McNemar test and the multivariable linear probability and logistic regressions are in line with the findings from the field experiment (Appendix Table A6-A8). Student raters gave preferential treatment to mothers who had taken long leave. Mothers who had taken short leave were considerably less likely to be recommended for a job interview than mothers who had taken long leave (−20 percentage points, p < 0.001). Fathers who had taken long leave were also more likely to be recommended for a job interview than fathers who had taken short leave, but this difference was not statistically significant. 3
To identify the mechanisms underlying the unequal treatment of mothers with different lengths of leave, I next examined how student raters perceived fictitious job applicants with short versus long parental leave. Based on theories of statistical and status-based discrimination, both mothers and fathers who took long family leave should be perceived as less competent and committed to their work than the parents who took short leave. Theories of normative discrimination, by contrast, predict negative evaluations of the parents who exhibit gender-nonconforming behaviors. Thus, the fathers who took long family leave should be perceived as weaker and less agentic than the fathers who took short leave, and the mothers who took short family leave should be perceived as less warm, less appealing, and less communal, and more hostile and dominant than mothers who took long family leave. According to prevailing norms of the “good father” being the breadwinner and the “good mother” being the homemaker, mothers with long periods of leave and fathers with short periods of leave, moreover, should have higher ratings on the “good parent” index than men and women who deviated from the gendered parental leave norms (e.g., Petts et al., 2023 for the U.S.).
Table 2 presents the mean student ratings of male and female applicants with different leave lengths. The upper part of the table presents the ratings of the different perceived characteristics of the applicants connected to statistical/status-based discrimination, and the lower part presents the ratings for characteristics connected to normative discrimination. The differences in the mean ratings of the composite measures between parents who took short and long leave were assessed with two-sided test statistics.
Means of Applicants’ Evaluations in the Laboratory Experiment.
+ p < 0.1, *p < 0.05, ** p < 0.01, *** p < 0.001.
Note: The underlying items for each composite measure and their respective source can be found in Appendix Table A2. Assignment of mothers’ and fathers’ application materials was between-sessions. Sample sizes for the fathers’ (n = 94) and mothers’ (n = 128) application materials differ, because I had to exclude ratings from one lab session in which the treatment assignment was not properly recorded. Evaluations should only be compared within genders, as the application materials also contained a picture, and the perceived physical attractiveness of male applicants and female applicants may hence be a confounding factor with gender.
Given that both the field and the laboratory experiment showed that parents who had taken short leave were not preferred over parents who had taken long leave, it is not surprising that the two groups were evaluated similarly with regard to their competence and commitment. However, regarding the different measures that were designed to capture normative discrimination, we see considerable differences between evaluations of mothers and fathers with different lengths of leave. The fathers who had taken long leave were perceived to be warmer, to have more interpersonal appeal, and to be better parents (e.g., Petts et al., 2023 for similar findings), but were also seen as less agentic than the fathers who had taken short family leave (e.g., Denny, 2013 and Triana, 2011). The additional personal qualities and interpersonal skills of fathers who take long family leave, hence, seem to be offset by the perception that they are less masculine, which may explain why the length of leave did not affect fathers’ employment chances in either the field or the laboratory experiment. Mothers, by contrast, were always evaluated more positively if they had previously taken long instead of short leave. They were perceived to be warmer, less hostile, more communal, more interpersonally appealing, and better parents than mothers who had taken short leave (e.g., Petts et al., 2023).
To further explore the relationship between the evaluations of mothers who had taken short versus long leave and their likelihood of being recommended for hiring, I ran mediation analyses for those indicators that showed statistical significance in Table 2. The results of these analyses suggest that mothers' perceived warmth, hostility, interpersonal appeal, communality, and parental qualities indeed explain the preferential treatment of mothers who took long leave. Parental qualities explain around 44 percent, perceived warmth around 24 percent, interpersonal hostility and appeal both around 17 percent, and communality almost 30 percent of the total variance of the effect of length of family leave on mothers’ likelihood of being recommended for a job interview (Table 3). Although the results of these mediation analyses cannot be interpreted causally due to a lack of random assignment of the mediator variables, they nonetheless lend further support to my interpretation of the field experimental data and suggest that normative discrimination may explain why employers prefer to work with mothers who took long leave—women whose behaviors are in line with prevailing stereotypes, e.g., by being caring and “motherly” may hence be likely to be perceived a better “fit” (Nichols et al., 2023).
Results of a Mediation Analysis to Assess for the Role of Normative Discrimination Against Mothers with Short Leave in the Laboratory Experiment.
Note: N = 256 (128 pairs of observations). Separate analyses were performed for each mediator; 95% confidence intervals are in brackets. Estimates were derived from 1,500 simulations; robust standard errors are clustered by respondent.
Overall, the results from the laboratory experiment provide further evidence that normative discrimination explains the unequal treatment of mothers with different lengths of family leave by employers in Germany. Mothers who only take short leave seem to violate the ideal of a “good mother.” In the open-field entries in the survey questionnaire, mothers who had taken short leave either were described in a neutral way (“she only took two months of family leave, which means that work is really important to her”) or were judged very harshly (“no community service and only two months of family leave is very little—might she be too ambitious and too dominant?”). Mothers who had taken long family leave, by contrast, were applauded for their choices (“the fact that she took one year of leave shows that she has her priorities straight”, “her family leave of one year is indicative of her sense of responsibility, her sensitivity, empathy, and family orientation. This is certainly helpful when dealing with clients”). While one should not extrapolate from student assessments to actual employers, these quotes from the laboratory study nonetheless illustrate the normative judgments that career-oriented mothers in Germany may face.
Discussion
By showing that different degrees of childcare involvement affect men and women unequally when applying for jobs, this study contributes to the existing literature on persistent gender inequalities and ongoing debates about paid family leave programs for parents of newborn children. In contrast to the widespread belief that discrimination occurs primarily due to direct or indirect utility expectations, my study shows that gender norm violations can also lead to discrimination—and that women, in particular, are penalized for such norm-violating behaviors.
Mothers in Germany who return to the labor market relatively soon after the arrival of a child are less likely to be invited to a job interview than mothers who stay home for an entire year, which suggests that mothers experience discrimination for normative reasons. That is, when seeking to be a “good worker,” they experience discrimination for not being a “good mother.” Mothers hence seem to be required to enact traditional family roles even in the workplace. If women violate the ideal of being a nurturing and caring mother so that they can live up to the image of the “ideal worker,” they seem to not be perceived as a good “fit” (Nichols et al., 2023). This suggests that gender norms around women's role as primary caregivers in Germany may be stronger than the normative expectation of mothers being “ideal workers.”
Fathers, by contrast, have more leeway in their behaviors and are evaluated on a more flexible range of criteria. For men, I find no sanctions for deviating from culturally prescribed behaviors by taking longer family leave. This is in line with the predictions of status characteristics theory that high-status actors may experience no or lower penalties for engaging in low-status activities (Correll & Benard, 2006; Ridgeway & Correll, 2004). Hence, being seen as an overly dedicated father is less of a norm violation than being seen as an uncaring mother. These insights add an additional layer to the extensive literature on fatherhood premiums (Abendroth et al., 2011; Cooke, 2014; Correll et al., 2007), which have been attributed to men's increased professional commitment and their financial obligations as fathers. As being an active caregiver contradicts both attributions, the findings of my study suggest that men's higher social status in workplace contexts helps them to compensate for violations of the “ideal worker norm.”
Before concluding this paper, I would like to raise a few issues that warrant discussion and may inspire future research. First, this study was designed to test the effects of the length of parental leave for applicants with the same amount of employment experience (by adjusting birth dates and educational transitions). However, it is likely that the positive effects that mothers encounter when taking long leave could be offset by having less work experience and resulting losses in human capital (e.g., Evertsson & Duvander, 2010 who found for Sweden that mothers who took more than 15 months of leave were less likely to experience upward occupational moves than mothers with shorter leaves). Hence, to derive concrete advice for mothers regarding the length of leave that is least detrimental to their hiring chances, future research should examine how employers evaluate women with varying lengths of leave and varying levels of work experience. Kalb (2018), for example, suggests that leave durations between seven and twelve months are optimal for allowing mothers to recover from childbirth without experiencing significant declines in human capital.
Second, examining variations in leave effects depending on the timing, the duration, and the frequency on mothers’ and fathers’ employment chances would be another relevant avenue for future research. Weisshaar (2018) found hiring penalties for U.S. mothers and fathers who applied for jobs when they were out of the labor force due to childcare responsibilities. The “scar” of being actively involved in parenting might fade over time, but it is unclear how applying for a job immediately after a care-related work interruption may play out in the German context and elsewhere. Likewise, the length of leave may have differing impacts if parents have several children or if the length of leave is indicative of an egalitarian family model (e.g., indicated by taking seven months of leave in the German context). Another promising area for further research is the comparison of the effects of different types of leave—such as family vs. medical leave—on labor market outcomes (see Holbrow, 2025 on recent survey experimental evidence from Japan). Additionally, future research could explore how signaling intensive childcare responsibilities affects mothers’ and fathers’ hiring chances in post-pandemic context (Chavez et al., 2022).
Third, this study examined the effects of family leave in the German context for one particular occupation and one particular point in time. Although Germany shares similarities with other wealthy democracies across various dimensions, the findings of this study cannot generalize to other contexts. In many countries, job applications do not contain private information to the same extent as in Germany, which means that hiring discrimination (for normative reasons) against leave takers should be less likely to occur. Likewise, in countries with no universal paid leave system, such as the United States, employer reactions to mothers’ and fathers’ leave may also be different. Moreover, even though the selected occupation is well represented across multiple sectors and organization sizes, employers may nevertheless react differently to parents’ leave taking behaviors in other occupations. Lastly, employer discrimination against mothers who took short parental leave may have changed since I collected the data for this study—for instance, because employers today may no longer assume that being married necessarily implies having a different-sex partner (which was the case at the time I collected my data) or because gender role attitudes have become more progressive and skilled workers more scarce. It is therefore important to update my data, although a more recent field experiment on nonacademic jobs in Germany prior to the Covid-19 pandemic (Schmieder et al., 2024) and a survey experiment on work and family-related evaluations in marketing jobs in the aftermath of the pandemic (Hipp et al., 2025) suggest that the penalties I found in my study seem to have persisted.
Conclusion
This paper examined how nontraditional childcare roles impact mothers’ and fathers’ likelihood of getting a job. In field and laboratory experiments conducted in Germany, a country in which both fathers and mothers are legally entitled to take up to 12 + 2 months of paid parental leave per child, I manipulated the lengths of family leave taken by parents to signal differences in their childcare involvement.
The results of the study provide evidence of normative discrimination against mothers who exhibit a strong attachment to the labor market by returning to work relatively soon after childbirth. The field experiment showed that mothers who took nothing beyond the mandatory leave of two months were 40 percent less likely to be invited to a job interview than mothers who took 12 months of leave, which is still the most common leave duration for mothers in Germany. Fathers’ likelihood of getting an invitation to a job interview, by contrast, was unaffected by the length of leave. Men may only be penalized after repeatedly engaging in low-status and norm-violating behaviors, whereas for women, “small behavioral clues” may be sufficient to trigger sanctions (see Ridgeway & Correll, 2004).
The findings from the laboratory experiment confirm the interpretation of the field experimental data that mothers who took short leave—and adhered to the ideal of a “good worker” who prioritizes their job over their family while violating the ideal of the “good mother”—experienced employer discrimination for normative reasons. The results from the laboratory experiment, moreover, suggest that employers value different traits in fathers who took short leave and fathers who took long leave. When taking longer leaves, fathers violate the “ideal worker” norm and therefore experience poorer professional evaluations; this perceived flaw, however, is compensated by the more favorable evaluations they receive as individuals and parents.
The practical and—either discouraging or empowering—implication of this study is that adhering to the “ideal worker” norm by minimizing the time spent away from work after childbirth is not necessarily beneficial for mothers. My findings challenge the widespread belief that young women can avoid motherhood penalties by “leaning in” to their career. A quick return to paid work signals strong commitment, helps to minimize the deterioration of human capital, and increases work experience (Gumy et al., 2022). These factors should, it is believed, help women's careers and promote their professional advancement. This advice is, however, not necessarily well-founded, as my study shows. For men, by contrast, the duration of parental leave seems to have little effect on their labor market chances. In line with findings from previous studies for the U.S. and Germany (Bünning, 2016; Samtleben et al., 2019; Thébaud & Pedulla, 2022), men hence can afford to be more courageous regarding their involvement in family life.
Supplemental Material
sj-docx-1-wox-10.1177_07308884251360325 - Supplemental material for Damned if You Do, Damned if You Don’t? Experimental Evidence from Germany on Hiring Discrimination Against Mothers with Short Family Leave
Supplemental material, sj-docx-1-wox-10.1177_07308884251360325 for Damned if You Do, Damned if You Don’t? Experimental Evidence from Germany on Hiring Discrimination Against Mothers with Short Family Leave by Lena Hipp in Work and Occupations
Footnotes
Acknowledgements
This research was funded by a grant from the German Ministry of Education and Research (Grant # 01UG1206 The Social Psychology of Care Work). I am indebted to Christina Botros, Nora Schneck, Jan Eilrich, Sophia Hess, Jakob Simonson, Marcel Lumkowsky, Friederike Molitor, and Ferdinand Kriesche for their thorough work in collecting the data, to Shelley Correll, Stephen Bernard, and In Paik for sharing the materials from their 2007 study with me, to Ruth Ditlmann for her advice on experimental design, as well as to Irene Boeckmann, Kristin Kelley, Lynn Prince Cooke, David Pedulla, David Brady, and the members of the WZB writing workshop for providing me with feedback on earlier versions of this paper.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the German Ministry of Education and Research, (grant number Grant # 01UG1206 The Social Psychology of Care Work).
Supplemental material
Supplemental material for this article is available online.
Notes
Author Biography
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
