BUSINESS MANAGEMENT GREAT WORK, ON TIME, NO PLAGARISM,
Pelicans!!322
wk4EBSCO-FullText-01_25_2026-1-1.pdf
Received: 8 November 2021 Revised: 9 December 2022 Accepted: 19 January 2023
DOI: 10.1111/peps.12577
OR I G I N A L A RT I C L E
Examining the role of maternity benefit comparisons and pregnancy discrimination in women’s turnover decisions
Samantha C. Paustian-Underdahl1 LauraM. Little2
AshleyM.Mandeville3 Amanda S. Hinojosa4 AndrewKeyes1
1Florida State University, Tallahassee, Florida,
USA
2Terry College of Business, University of
Georgia, Athens, Georgia, USA
3Florida Gulf Coast University, FortMyers,
Florida, USA
4Howard University,Washington, DC, USA
Correspondence
Samantha C. Paustian-Underdahl, Florida
State University, 821 AcademicWay,
Tallahassee, FL 32306, USA.
Email: [email protected]
Abstract
Retaining pregnant women and mothers is a prevalent chal-
lenge for companies in the United States. In this paper, we
highlight the importance of favorable maternity benefits.
Specifically, we argue that maternity benefits can signal how
pregnant workers are treated within the organization, par-
ticularly as women compare their own benefits to referent
others’. Drawing from identity threat response theory, we
propose a conceptual framework that explains the influence
of maternity benefit comparisons on perceptions of discrim-
ination and, subsequently, turnover. Upon evaluating two
studies using multi-wave survey data and two vignette stud-
ies, our results indicate that when women perceive their
maternity benefits to be less favorable than referent oth-
ers’ benefits, they perceive more pregnancy discrimination.
In turn, perceptions of pregnancy discrimination influence
their subsequent turnover decisions. Consistent with iden-
tity threat response theory, our results also suggest that
perceived supervisor support is a significant moderator,
weakening the impact of maternity benefit comparisons on
perceptions of pregnancy discrimination.
KEYWORDS
perceived supervisor support, pregnancy, pregnancy discrimination, turnover
Personnel Psychology. 2024;77:819–846. © 2023Wiley Periodicals, Inc. 819wileyonlinelibrary.com/journal/peps
820 PAUSTIAN-UNDERDAHL
1 INTRODUCTION
Research suggests that family-supportive policies can help organizations retain women (Baum, 2003; Grover &
Crooker, 1995; Lee&Kim, 2010;Piszczek, 2020), but fewstudies have investigated the influenceofmaternity benefits,
in particular. Maternity benefits are a type of benefit specifically provided for pregnant women and newmothers. The
most commonmaternity benefit is maternity leave; however, in addition to maternity leave, some organizations offer
othermaternity benefits such asmaternity “concierge services” (Best Upon Request, 2020), cash stipends (Ball, 2015;
Miller, 2016), educational materials, nurse advice lines, special screenings for high-risk pregnancies, and other mater-
nity health programs (Lumen Technologies, n.d.; United Healthcare, n.d.). Previous research on maternity leave and
benefits has primarily focused on differences in maternity leave policies across various countries (e.g., Budd &Mum-
ford, 2006; Sterling & Allan, in press), the well-being of women and their infants after maternity leave (e.g., Cabeza
et al., 2011; Carneiro et al., 2015; Hewitt et al., 2017; Feldman et al., 2004; Jou et al., 2018; Sterling & Allan, in press;
Thomas, 2015), the economic ramifications of offering or requiringmaternity leave (e.g., Dahl et al., 2016; Fallon et al.,
2017; Gruber, 1994), and the role of expectations and negotiations with supervisors (e.g., Buzzanell & Liu, 2007; Liu
& Buzzanell, 2004; Miller et al., 1996). Surprisingly, little focus has been paid to how maternity benefits influence
women’s retention.
Understanding how maternity benefits influence retention in the United States is both important and timely as
companies continue to lament the labor shortage (Cheng, 2021; Spiggle, 2021), and women, in particular, are leav-
ing their organizations (i.e., “the shecession”; Gupta, 2020; Deloitte, 2021; Jones et al., 2020). On average, 3.7 million
babies are born in the United States annually (Hamilton et al., 2020), meaning millions of working women become
mothers yearly. Of these women, 24% do not return to their organizations within 12months of having their first baby
(Sandler & Szembrot, 2019), and nearly one in threeworkingmotherswith young children have considered downshift-
ing their jobs or dropping out of the workforce entirely (Huang et al., 2021). We argue that these statistics, paired
with the considerable variability in maternity benefits offered in the United States, highlight the need for additional
research to understand the role of maternity benefits on women’s turnover.
Perhaps due to this lack of extant knowledge, organizations varywildly regarding thematernity benefits they offer,
particularly in the United States—the only developed country lacking a federally mandated, paid maternity leave ben-
efit. Although the Family and Medical Leave Act (FMLA) stipulates access to unpaid leave for women in the United
States (SHRM, 2019), only 56% of private-sector employees are eligible (Brown et al., 2020). Thus, organizations gen-
erally determine the level of benefits they want to offer. As a result, some organizations have no maternity leave or
other benefits. Others offer unpaid, partially paid, or fully paid maternity leave of varying lengths with other fringe
benefits. Even large, well-known corporations have substantially varied offerings. Microsoft offers 20 weeks of paid
leave to all new birth mothers (Microsoft, n.d.); Netflix offers unlimited time off for parental leave (Netflix, n.d.); and
Starbucks’ baristas receive 6weeks of paid leave (Starbucks, 2017). Research suggests that, when evaluating benefits,
employees use comparisons with referent others (Williams et al., 2002 ). Thus, we argue that this significant variation
in the maternity benefits in the United States creates a context ripe for social comparisons regarding one’s maternity
benefits.
This paper draws on identity threat response theory (Petriglieri, 2011) to suggest that maternity benefits reflect a
pregnant woman’s social identity—pregnant worker. Specifically, we argue that as pregnant employees consider their
benefits and compare them to referent others’, they will develop assumptions related to how their pregnant identity
will be valued at their organization. In other words, maternity benefit comparisons (i.e., perceptions of how favorable
or unfavorable one’s ownmaternity benefits are compared to referent others’ benefits) should influence perceptions
of pregnancy discrimination (i.e., an individual’s perception that they are subject to unfair treatment because of their
social identity groupmembership–pregnancy). Perceptionsofpregnancydiscriminationwill, in turn, increase turnover.
Identity threat response theory also points to a key moderator—perceived supervisor support (PSS), or the degree
to which an employee perceives that their immediate supervisor values their contributions and cares about their
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 821
well-being. Petriglieri (2011) suggests that social support canhelpbuffer theextent towhichapotential identity threat
leads to the devaluation of one’s identity. We chose PSS because social support from one’s supervisor plays a critical
role in identity-related perceptions of workplace treatment and organizational commitment for working parents and
pregnant women (Little et al., 2017; Ng & Sorensen, 2008). It has also been found to function as amoderator between
organizational policies and attitudinal outcomes (Dysvik & Kuvaas, 2013).
Our research makes multiple contributions to theory and practice related to the importance of maternity benefit
comparisons in retaining pregnant women. First, guided by identity threat response theory, we provide a theoret-
ical explanation for how, why, and under what conditions (Sutton & Staw, 1995; Whetten, 1989) maternity benefit
comparisons relate to turnover decisions.We extend previous research examining the effects of general benefit com-
parisons on employee attitudes (Williams et al., 2002). Specifically, we uncover an explanatorymechanism—perceived
pregnancy discrimination—by which maternity benefit comparisons relate to turnover decisions. We propose that
maternitybenefit comparisons serveas a cue that is especiallymeaningfulwhenawomanbecomespregnant and seeks
to gauge if her new identity is threatened at work. Additionally, by examining how maternity benefit comparisons
and PSS interact to relate to perceived pregnancy discrimination and, ultimately, turnover, we answer a call to bet-
ter understand under what conditions (i.e., low PSS) this relationship is particularly strong. Further, by introducing the
idea of supervisor support as ameans toweaken the positive relationship betweenmaternity benefit comparisons and
perceived pregnancy discrimination, we answer a call for research examining relational ways to reduce perceptions of
discrimination (Green &Kalev, 2007).
Second, we extend the identity literature, which has examined how organizational identity threats are associated
with various health and career outcomes, including turnover, because they represent concerns regarding the organi-
zation’s treatment of one’s identity (Petriglieri, 2011; Rothausen et al., 2017). Our study builds on this literature by
positioning unfavorable maternity benefit comparisons as a potential threat—one that organizations may not have
considered. Further, research on identity management has investigated how pregnant women and other employees
use signaling—or behaviors used by stigmatized individuals to gauge if others will be supportive of their identity—to
discover if they are in a safe space among supportive coworkers (e.g., Jones & King, 2014). In addition to signaling, our
paper suggests that potentially stigmatized individuals would also interpret information already present in the work
environment as either threatening or supportive.
Below, we outline our theoretical framework, including background on identity threat response theory, arguing
that perceived pregnancy discrimination is a mechanism through which less favorable maternity benefit comparisons
relate to turnover decisions among pregnant women andmothers.
2 THEORETICAL FRAMEWORK
2.1 Identity threat response theory
Although research has long acknowledged the existence of identity threats in organizational life—particularly for
those with marginalized identities (Ashforth & Kreiner, 1999; Steele et al., 2002), Petriglieri (2011) offers a syn-
thesized and coherent framework to explain the process by which individuals recognize an experience as identity
threatening and decide how to respond to it. To understand how identities can be threatened, it is first important
to understand what an identity is. An identity is a self-referential description that provides contextually appropriate
answers to the question “Who am I?” (Ashforth et al., 2008: p. 327). Identities can be based on group membership,
roles, and unique characteristics and traits. Many social identity categorizations, such as race, are stable; however,
throughout an employee’s tenure at an organization, major life events, such asmarriage, divorce, pregnancy, or illness,
mayprovokenewsocial identity categorizations. Becausepeople are socially embedded, theynegotiate their identities
within social relationships and interactions. This identity development process assigns value to identities and defines
and shapes their meanings (Swann, 1987; White, 1992). These values are important because they relate directly to
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
822 PAUSTIAN-UNDERDAHL
perceptions of self-worth (Gecas, 1982). The more positive an identity, the more self-worth an individual draws from
it, and vice versa.
An identity threat, then, is any experience appraised as indicating potential harm to the value, meanings, or enact-
ment of an identity (Petriglieri, 2011). Because people tend to value their identities positively (Gecas, 1982), identity
threats occur when individuals encounter people, experiences, or situations that suggest their identity is not valued
(Dutton, 2010). A vast array of experiences can be appraised as potentially harmful (Petriglieri, 2011). In the work-
place, experiences that “signal threat by providing evidence that one’s identity may be a liability or source of stigma,
devaluation, or mistreatment” constitute identity threats (Emerson &Murphy, 2014; p. 509). Identity threats are par-
ticularly influential when they suggest that identity devaluation may be ongoing and extend into the future and, as
such, represent “present cues of future harm” (Petriglieri, 2011; p. 644).
When individuals experience a threat to their identity, they respond by engaging in coping behaviors. Petriglieri
(2011) suggests that these coping responses are cognitive and behavioral efforts aimed at decreasing the likelihood
or severity of potential identity harm. These responses include removing the identity threat—which, she suggests, can
include anorganizational exit. These types of responses aremore commonwhen individuals are in the process of nego-
tiating the meanings of new or nascent identities. Moreover, because identities are socially construed, an individual’s
social environment should be considered in his or her assessment of the identity threat. Specifically, she suggests that
social relationships can shape individual identities, and they can also lend social support to sustain them. Social sup-
port can act as “social buffering,”whichpartially insulates individuals from the identity threats theyencounter because
they provide a source of identity affirmation (Ashforth et al., 2007).
2.2 Maternity benefit comparisons as a potential identity threat
When a woman learns of her pregnancy, she engages in an identity transition from working woman to pregnant
working woman (Ladge et al., 2012). This time can be uncertain for working women as this new identity—pregnant
worker—is often stigmatized and devalued in organizations (Jones, 2017; Jones et al., 2020; King & Botsford, 2009;
Ladge et al., 2012; Little et al., 2015). Research suggests that pregnancy initiates a cross-domain identity transition
for working women characterized by fear of stigmatization at work because of their pregnancy and impending moth-
erhood (Jones, 2017, Jones et al., 2020; King & Botsford, 2009; Ladge et al., 2012; Little et al., 2015). Coworkers and
supervisorsoftenperceivepregnantworkers as fragile, needingassistance, and less competent andcommitted towork
(Hebl et al., 2007; Jones, 2017; Jones et al., 2020). These notions are based on gendered assumptions and stereotypes
that women tend to be more involved in family roles than men (Baltes & Heydens-Gahir, 2003; Jones, 2017; King,
2008). Thus, pregnancy represents a social identity that has the potential to be threatened at work.
Aswe suggest above, identity threat response theory proposes that as individuals take on new identities, they scan
their environments to understand how these identities will be valued (Petriglieri, 2011). Maternity benefits are likely
among the first and most readily available indicators of how the organization treats and values pregnant employees.
Thus, pregnant workers may look to their maternity benefits to understand how their new identity may be received in
their organization. Indeed, maternity benefits are unique in relation to general benefits in that they reflect support of
a particular social identity. Less favorablematernity benefit comparisons can signal a devaluation of one’s identity as a
pregnant worker.
A plethora of research suggests that, when evaluating benefits, employees use comparisons with referent others
(Williamset al., 2002 ). Likewise, compensation researchers have longused comparisons todirectlymeasure theextent
to which employees believe their pay or benefits are equitable (Rice et al., 1989; Sweeney et al., 1990). As pregnant
women consider their benefits and compare them to referent others’, they will develop assumptions related to how
their pregnant identity will be valued at their organization. These assessments “set and color” later experiences and
thus, will anchor their expectations about how they believe they will be treated at work (Lind et al., 2001, p. 192).
According to identity threat response theory, these assumptions about treatment from the organization are likely to
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 823
remain consistent once developed (Petriglieri, 2011) and highlight potential unfair treatment and subsequent iden-
tity devaluation, leading to perceptions of discrimination or an individual’s perception that they are subject to unfair
treatment because of their social identity groupmembership (Sanchez & Brock, 1996).
Here, we argue that the identity transition that occurs for working women during pregnancy is likely to heighten
pregnant women’s search and attention to cues related to their identity. Drawing from theory on identity threats
(Petriglieri, 2011), we suggest that as women consider their benefits and compare them to referent others’, they will
reach conclusions about whether their pregnant identity is valued at their organization. Because these benefits are
linkedexplicitly to their pregnant identity, unfavorablematernity benefit comparisonswill lead to feelings of perceived
pregnancy discrimination at work.
Hypothesis 1: Less favorable maternity benefit comparisons are positively related to perceived pregnancy
discrimination.
2.3 Identity threat response: Perceived pregnancy discrimination and turnover
Identity threat response theory proposes that as individuals experience threats to their identity, they will withdraw
from the source of devaluation (Petriglieri, 2011). Doing so eliminates the source of threat and, thus, the likelihood of
potential harm (Ashforth, 2001; Ebaugh, 1988). Identity exit occurs when an individual abandons the context in which
their identity is threatened—physically disengaging from roles or groups associated with the threat (Ashforth, 2001;
Ebaugh, 1988). Empirical research substantiates this notion. When employees sense social identity threats and feel
that desired outcomes are unattainable in their current workplace, they seek other employment options (Howard &
Cordes, 2010).Moreover,multiple studieshave reported significant relationshipsbetweenperceiveddiscriminationat
work and turnover intentions (e.g., Madera et al., 2012; Ragins & Cornwell, 2001), as well as turnover (e.g., McDonald
et al., 2008). Based on identity threat response theory and this previous empirical work, we suggest that as a preg-
nant woman perceives discriminatory treatment at work because of her pregnancy, she will bemore likely to consider
leaving her organization during late pregnancy, and actually leave her workplace late after maternity leave.
Hypothesis 2a: Perceived pregnancy discrimination is positively related to turnover intentions in late pregnancy.
Hypothesis 2b: Perceived pregnancy discrimination is positively related to turnover after maternity leave.
2.4 Perceived supervisor support
PSS reflects employees’ perceptions as to how much they believe their supervisor values them and cares about their
welfare. Importantly for the current investigation, Dirks and Ferrin (2002) suggested that employees can distin-
guish relations with their immediate supervisor from policies and procedures associated with the organization. Thus,
although employees view supervisors as agents of the organization, they can still feel value and support from them
even in light of unsupportive policies. Little empirical research has evaluated the moderating effect of PSS on percep-
tions of discrimination (Crouse, 2020); however, studies have shown that PSS canmoderate the relationship between
various organizational policies and attitudinal outcomes (e.g., Kuvaas & Dysvik, 2010). In particular, these studies
suggest that supervisor support can reduce (or exacerbate) the influence of unsupportive organizational policies by
sending signals that employees are valued (or not) if these policies do not support them.
Considering these issues, we propose that the degree to which employees feel their supervisor cares for and
supports them during pregnancy should weaken the effects of unfavorable benefits on perceptions of pregnancy dis-
crimination in the workplace.When that support is not available, the influence of unfavorable policies on perceptions
of pregnancy discrimination will be greater. Whereas distributive policies surrounding maternity benefits may be
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
824 PAUSTIAN-UNDERDAHL
constrained by bureaucracy and the like, supervisors still have the opportunity to demonstrate that they value (or
don’t) an employee.We argue that supervisors seen as lacking in supportmay exacerbate the threat pregnant employ-
ees feel when they perceive more unfavorable benefits. The more unfavorable maternity benefits are, the more
pregnant employees feel their organizations do not support them and their new identity. In these cases, unsupportive
supervisors are particularly threatening to one’s identity because—and as opposed to cases when benefits are more
favorable—unsupportive supervisors may appear to have organizational backing regarding their lack of support. This
combination of more unfavorable policies and a lack of supervisor support will lead to greater perceptions of preg-
nancy discrimination.When supervisors are supportive,whether or not the policies are favorablematters less because
support from one’s supervisor will provide more proximal evidence for how well one will be treated at work during
pregnancy and beyond.
As such, we extend our argument above—wherein we suggested that maternity benefit comparisons are an impor-
tant indicator to pregnant women as to how valued their identity is at work—to include the moderating effects of
support from their supervisor. Social support within one’s workplace is thought to override perceptions of discrimina-
tion due to its influence on positive interpersonal relationships (Green & Kalev, 2007). This is consistent with identity
threat response theory which positions social support as a buffer to the extent to which a potential identity threat
leads to an actual threat or devaluation of one’s identity (Petriglieri, 2011). Thus, we suggest that when PSS is lacking,
the negative influence of unfavorable maternity benefits will be stronger. When PSS is strong, this negative influence
will be attenuated.
Hypothesis 3: PSSwill moderate the relationship between less favorablematernity benefit comparisons and per-
ceived pregnancy discrimination such that the positive effect of less favorablematernity benefit comparisons
on perceived pregnancy discrimination will be stronger (weaker) when employees perceive less (more) PSS.
2.5 Overall model
Altogether, we propose a moderated mediation in which maternity benefits perceived as less favorable than refer-
ent others’ benefits influence turnover intentions and turnover via pregnancy discrimination, with PSS serving as a
first-stage moderator. Hence, we expect that less favorable maternity benefit comparisons will interact with PSS to
influence perceived pregnancy discrimination, subsequently relating to women’s turnover decisions.
Hypothesis 4a: The positive indirect effect of less favorable maternity benefit comparisons on turnover inten-
tions through perceived pregnancy discrimination is weaker (stronger) when employees perceive more (less)
PSS.
Hypothesis 4b: The positive indirect effect of less favorable maternity benefit comparisons on turnover through
perceived pregnancy discrimination is weaker (stronger) when employees perceivemore (less) PSS.
3 OVERVIEW OF STUDIES
We tested our theoretical model across four studies using complementary methodologies. In Studies 1 and 2 (data
collected from 2013–2016), we collected multi-wave field data, which allowed us to test our model in an orga-
nizational context and offers desirable external validity. Study 1 investigates the influence of maternity benefits
comparisons on turnover intentions via perceived discrimination and PSS. Study 2 builds upon Study 1 in multi-
ple ways. First, we measured the actual turnover of women from their organizations 12 months following their
maternity leaves—rather than turnover intentions. This allows us to temporally separate the measurements of per-
ceived pregnancy discrimination and turnover and provides for a more objective outcome of unfavorable benefits
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 825
and perceived pregnancy discrimination—actual turnover. We also included a measure of discrimination in Study 2
that more explicitly captures perceptions of impeded career progress based on stigma related to being a pregnant
worker. Please see Figure 1 for our conceptual model.
Given the correlational nature of Study 1 and 2, we could not conclude that maternity benefit comparisons cause
perceived discrimination, as there could be other unmeasured variables that could also explain this effect. Thus, we
conducted two scenario-based experiments (Vignette Studies 1 and 2, data collected in 2022) to better establish
causality and obtain better internal validity. In Vignette Study 1, we manipulated less favorable maternity benefits
and examined their influence on perceived pregnancy discrimination to further test Hypothesis 1. In Vignette Study
2, we manipulated perceived pregnancy discrimination to investigate its relationship with turnover intentions (i.e.,
Hypothesis 2a). Following best practices in assessing experimental mediation, we used the experimental causal chain
approach outlined in Spencer et al. (2005). This design allowed us to utilize the power of experiments to demonstrate
causality (i.e., manipulate both the independent variable and the mediator) while still making strong inferences about
the causal chain of events (i.e., mediation). Vignette Study 11 and Vignette Study 22 were preregistered. Additionally,
these studies were approved through the IRB (PROJECT00005158; entitled: Benefits Vignette).
4 STUDY 1
4.1 Participants and procedure
Participants were recruited via advertisements placed on various online parenting communities and websites (e.g.,
BabyCenter.com). The current samplewas part of a larger data collection effort focusedonpregnantworkers and their
partners indual-earner couplesworking at least 30hperweekandmarried and/or cohabiting in theUnitedStates. This
data collection was approved by the Institutional Review Board (IRB; protocol #IRB-15-0054; entitled: Dual-Career
Couples and Pregnancy Disclosure at Work). We collected demographic data in the first survey from 162 pregnant
women who were 10.8 weeks pregnant, on average. We measured perceived maternity benefit comparisons a week
later in the second survey (n = 155). Perceived supervisor support was measured at Time 3 (within 1 week of the
pregnancy disclosure at work) from 150 women. Finally, perceived pregnancy discrimination and turnover intentions
were assessed in the final survey from 105 women between 30 and 35 weeks pregnant.3 We removed two partic-
ipants because they stated they were self-employed and could not adequately respond to workplace-related items.
Participants were compensated with a $5 gift card after completing each survey.
At the time of the first survey, the women were, on average, 30.62 (SD= 3.22) years old, and most had no children
(46.2% no children, 36.8% one child, 11.3% two children, and 5.6% had three or more children). They represented
various races (7.5%African American, 4.7%Asian, 79.2%White, 6.6%Hispanic, and 1.9% other). Their average tenure
with their current organization was 3.88 years (SD = 3.11), and they worked, on average, 39.61 h a week (SD = 5.09).
They also held a wide range of jobs and were predominantly in non-management positions (70% non-management
positions, 16% supervisory, 9% manager, and 5% senior managers). The majority of women were from the following
five industries: healthcare (n=23), education (n=30), business services (n=6), government (n=8), and other (n=38).
4.2 Measures
Unless otherwise indicated, itemsweremeasured using a 5-point Likert scale, from1 (strongly disagree) to 5 (strongly
agree).
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
826 PAUSTIAN-UNDERDAHL
4.2.1 Maternity benefits in comparison to referent others
We adapted the general benefit comparisons scale from Williams et al. (2002) to measure comparisons of mater-
nity benefits rather than general benefits, using a 5-point scale from 1 (much less) to 5 (much more). The four items
included: “Compared with others I know with similar abilities and training, the level of maternity benefits I currently
have access to is”; “Compared with others with my level of seniority, the level of maternity benefits I currently have
access to is”; “Compared to my friends and family, the level of maternity benefits I currently have access to is”; “Com-
pared with the benefits I need to meet my financial needs, the level of maternity benefits I currently have access
to is.” We reverse-scored these items such that larger values reflect less favorable maternity benefit comparisons
(α= .96).
4.2.2 Perceived supervisor support
We used Eisenberger et al. (2002) validated, three-item version of the Eisenberger et al. (1986) perceived supervisor
support scale. The items are: “My supervisor is willing to extend him or herself in order to help me perform my job to
the best of my ability”; “My supervisor takes pride in my accomplishments at work”; “My supervisor tries to make my
job as interesting as possible” (α= .82).
4.2.3 Perceived pregnancy discrimination
In Study 1, we used seven items from James et al.’s (1994) Workplace Prejudice/Discrimination Inventory adapted
to focus on pregnancy discrimination. The items are: “I am unfairly singled out at work because of my pregnancy”; “I
feel socially isolated at work because of my pregnancy,” “I am seen as less capable of doing my job when I disclose my
pregnancy at work,” “I am treated poorly at work because of my pregnancy,” “It was safe to disclose my pregnancy at
work” (reverse-scored), “People at my work reacted negatively to my pregnancy,” “People at my work considered me
less committed tomy job upon learning of my pregnancy” (α= .92).
4.2.4 Turnover intentions
Wemeasured turnover intentionswhen thewomenwerebetween30–35weeks pregnant usingO’Reilly et al.’s (1991)
three-itemmeasure: “Towhatextentwouldyouprefer anothermore ideal job than theoneyounowwork in?”; “Towhat
extent have you thought seriously about changing organizations since beginning towork here?”; “If you have your own
way, will you be working for this organization three years from now (reverse scored)”? (α = .87). We used a 5-point
response scale with 1 (definitely not) to 5 (definitely yes).
4.2.5 Control variables
Following best practices related to control variables (Bernerth &Aguinis, 2016), we included several logically and the-
oretically derived control variables. First, we controlled for number of children. Our focus is on pregnancy, pregnant
worker identity, and perceptions of pregnancy discrimination. We argue that each time a woman becomes pregnant,
they will engage in an identity transition process (e.g., Little et al., 2015); however, we acknowledge that the num-
ber of children a woman has could play a role in perceptions of discrimination and turnover cognitions, and thus, we
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 827
TABLE 1 Descriptive statistics and correlations among variables for study 1
Variables M SD 1 2 3 4 5 6
1. Number of children .80 1.03
2. Age 30.62 3.22 .24**
3. Race .79 .40 .16 .14
4. Less favorable benefits 3.33 .83 –.11 –.05 .08
5. PSS 3.90 .75 .02 .02 –.01 –.17
6. Perceived pregnancy
discrimination
1.60 .68 .18 –.06 .05 .23* –.39**
7. Turnover intentions 3.05 1.26 –.16 –.22* –.11 .35* –.33** .27**
Note: n = 103. *p < .05; **p < .01; Less favorable benefits = less favorable maternity benefit comparisons. PSS = per-
ceived supervisor support. Race:White= 1; Non-White= 0.
controlled for it. Further, maternal age is related to a greater risk of pregnancy complications; therefore, older women
may require additional medical appointments during and following pregnancy (Luke & Brown, 2007). Research has
shown that such accommodations could relate to discrimination and turnover (McDonald et al., 2008). As such, we
also control for maternal age. We also controlled for race (White = 1; Non-White = 0) because discrimination can
be heightened when a person has more than one stigmatized identity (e.g., Berdahl &Moore, 2006). All hypothesized
effects remained consistent and significant whenwe conducted the analyses without controls.
4.3 Preliminary analysis
The descriptive statistics and correlations for Study 1 are displayed in Table 1. We utilized the full information max-
imization likelihood (FIML) estimation in Mplus 8.1 because it estimates model parameters based on all available
information in the variance−covariance matrix and results in unbiased parameter estimates under various levels of
missing data (Enders, 2001, 2010; Enders & Bandalos, 2001; Graham, 2009; Larsen, 2011). To examine the factor
structure of our study variables, we conducted a confirmatory factor analysis (CFA), which included the four contin-
uous scales (maternity benefit comparisons, PSS, perceived pregnancy discrimination, and turnover intentions). The
expected four-factor model, with correlated factors, provided moderate fit to the data (chi-squared (χ2) = 293.03,
df= 113, comparative fit index (CFI)= .85, standardized root mean square residual (SRMR)= .085, root mean square
error of approximation (RMSEA)= .12).We followed recommendations to allow theerror termsof highly related items
to correlate (Brown, 2015; Cole et al., 2007). Two perceived pregnancy discrimination items (the only two that began
with “People at work. . . ”) and two benefits comparison items (“Compared with others, with my level of seniority, the
level of parental leave benefits I currently have access to is:” and “Compared with others, I know with similar abilities
and training, the level of parental leave benefits I currently have access to is:”) loaded more strongly on each other
than the other scale items. In survey research, correlated errors “may arise from items that are very similarly worded,
reverse-worded, differentially prone to social desirability, or the like” (Brown, 2015, pg. 157). In these cases, it is rec-
ommended to correlate error terms. Thus, we correlated these error terms, resulting in an improved fit. This model
provided better fit to the data (χ2 = 200.97, df= 109; CFI= .92; SRMR= .07; RMSEA= .09). Moreover, this model fit
better than a one-factor model where all items loaded onto one factor (χ2 = 762.62, df= 119, CFI= .47, SRMR= .18,
RMSEA = .23) and a three-factor model, where Time 2 variables loaded onto a factor, Time 3 variables loaded onto
a second factor, and Time 4 variables loaded onto a third factor (χ2 = 454.43, df = 116, CFI = .72, SRMR = .14,
RMSEA= .17) (Figure 1).
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
828 PAUSTIAN-UNDERDAHL
Less Favorable
Maternity
Benefit
Comparisons
Perceived Pregnancy
Discrimination
Turnover Decisions: Turnover Intentions
(Study 1; Vignette 2)
Turnover (Study 2)
Perceived Supervisor
Support
F IGURE 1 Conceptual model.
TABLE 2 Study 1: Unstandardized results
Perceived Pregnancy
Discrimination Turnover Intentions
Model 1 Model 2 Model 3 Model 1 Model 2 Model 3
Intercept 2.12** (.61) 2.86**(.71) 2.09**(.56) 5.58**(1.12) 3.19**(1.19) 4.63**(1.13)
Controls
Number of children .14*(.06) .15* (.06) .13*(.06) –.11 (.12) –.14 (.12) –.14 (.12)
Age –.02 (.02) –.02 (.02) –.02 (.02) –.08* (.04) –.06*** (.04) –.06*** (.04)
Race .07 (.16) .01 (.15) –.02 (.15) –.17 (.30) –.32 (.28) –.32 (.28)
Predictors
Less favorable benefits .15* (.07) .18*(.07) .43**(.14) .43**(.14)
PSS –.32**(.08) –.29**(.08)
Benefits X PSS –.20* (.09)
Mediator
Perceived pregnancy
discrimination
.40*(.17) .40*(.17)
R2 .05 .23** .27** .07 .22** .22**
Conditional Indirect Effects Est. 95% CI
Low PSS: Benefit Comparisons→ Perceived Discrimination→ Turnover Intentions .13* [.019, .329]
High PSS: Benefit Comparisons→ Perceived Discrimination→ Turnover Intentions .01 [–.065, .117]
Abbreviation: CI, confidence interval.
Note: M1 = model with control variables only; n = 106, M2 = main-effect model; n = 103, and M3 = full model; n = 103.
Numbers outside of parentheses are unstandardized estimated coefficients and numbers inside of parentheses are estimated
standard errors. Less favorable benefits = less favorable maternity benefit comparisons. PSS = perceived supervisor sup-
port. Race:White= 1; Non-White= 0. ***p< .10. *p< .05. **p< .01 (two-tailed).
4.4 Hypothesis testing
Table 2 presents the unstandardized coefficient estimates for three separate models. Model 1 (M1) includes the con-
trol variables, Model 2 (M2) includes the control variables, PSS, and perceptions of unfavorable benefits, and Model
3 (M3) includes the variables in M2 in addition to the interaction term. We calculated model fit indices for each of
thesemodels (M1: fully-identified by nature of the model; M2: χ2 = 6.22, df= 1, CFI= .89, SRMR= .05, RMSEA= .22;
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 829
More Favorable Benefits (-1 SD) Less Favorable Benefits (+1 SD)
1
1.5
2
2.5
3
3.5
4
4.5
5
Low PSS (-1 SD)
High PSS (+1 SD)
Pe rc
ei ve
d Pr
eg na
nc y
D isc
ri m
in at
io n
F IGURE 2 Study 1 interaction. More/Less favorable benefits=More/less favorable maternity benefit comparisons.
M3: χ2 = 6.25, df = 2, CFI = .92, SRMR = .05, RMSEA = .14). These indices suggested that our hypothesized model
fit the data moderately well. Supporting Hypothesis 1, less favorable maternity benefit comparisons were positively
related to perceived pregnancy discrimination (B = .15, p = .034; see Model 2 in Table 2). Additionally, we found
that perceived pregnancy discrimination was positively related to turnover intentions, thus supporting Hypothesis
2a (B = .40, p = .018). To test our moderated-mediated model, we used path analysis inMplus 8.1 (Stride et al., 2015)
and followed best practices provided by Edwards and Lambert (2007). The variables included in interactions were
centered prior to the analyses. We found a significant interaction between maternity benefit comparisons and PSS in
influencing perceived pregnancy discrimination, supporting Hypothesis 3 (B= –.20, p= .022; seeModel 3 in Table 2).
The effect of less favorable benefits on perceived discrimination was significantly positive when PSS was low (–1 SD;
B = .33, p = .002), and was not significant when PSS was high (+1 SD; B = .03, p = .765; see Figure 2). To investi-
gate themoderated-mediated relationships proposed inHypothesis 4a,we examined themagnitude of the conditional
indirect effect of less favorable maternity benefit comparisons on turnover intentions through perceived pregnancy
discrimination using an empirical bootstrapping approach with bias-corrected confidence intervals (CIs). Consistent
with our hypothesis, at lower levels of PSS, the indirect effect is stronger and significant (–1 SD B = .13, 95% CI:
.019, .329) and at higher levels of PSS, the indirect effect becomes weaker and non-significant (+1 SD B = .01, 95%
CI: –.065, .117; see Table 2). For a formal test of the hypothesized moderated mediation effect, we followed the pro-
cedures outlined by Hayes (2015) to compute the index of moderated mediation (a3*b). To obtain its 95% CI using
the Monte Carlo resampling method with 20,000 repetitions, we used an Rweb utility (Selig & Preacher, 2008). The
index of moderated mediation was significant (index = –.08, 95% CI [–.195, –.002]). Thus, overall, our results support
Hypothesis 4a.
4.5 Study 1 discussion
Our first study found support for our proposed model. As women perceived that their maternity benefits
were unfavorable compared to referent others, they experienced more pregnancy discrimination and stronger
turnover intentions. Additionally, PSS helped buffer the effects of unfavorable benefits on perceived pregnancy
discrimination.
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
830 PAUSTIAN-UNDERDAHL
5 STUDY 2
5.1 Participants and procedures
Participants were recruited using the same method as Study 1. This data collection was approved through the IRB
(protocol #2013102001; entitled: PregnantWomen andMothers in theWorkplace).Wemeasured demographic vari-
ables,maternity benefit comparisons, and PSS in the first survey.Overall, 879women completed the first surveywhen
they were, on average, 26 weeks pregnant (SD = 6.45). Time 2 surveys were emailed to participants at 41 weeks so
they could reflect on the pregnancy discrimination they experienced toward the end of their pregnancy. Of the initial
879 participants, 459women completed Time2 surveys. Time3 surveyswere emailed to participants approximately 1
year following childbirth (M= 12.74months; SD= 1.68) to capturewomen’s turnover from their organizations follow-
ingmaternity leave. The Time 3 survey was completed by 295 participants.4 This survey askedwomen to report when
and if they returned to the organization they worked at while pregnant. Participants were entered into a drawing for
Amazon.com gift cards for their participation in the study.
At the time of the first survey, the women were, on average, 31.00 (SD= 4.61) years old, and most had no children
(53% had no children, 32% had one child, 8% had two children, and 6% had three or more children). They represented
various races (2% African American, 2% Asian, 86% White, 4% Hispanic, 2% Native American, and 3% other). Their
average tenure with their current organization was 4.54 years (SD = 3.91), and they worked, on average, 39.32 h a
week (SD = 8.81). They also held a wide range of jobs and were predominantly in non-management positions (69%
non-management position, 21% supervisory, and 9% manager/senior manager). Women in Study 2 gave birth when
they were between 28 and 42 weeks pregnant (M = 39.07, SD = 1.93), and 13 women in this study gave birth before
they were 37weeks pregnant, which is considered premature.
5.2 Measures
Unless otherwise indicated, itemsweremeasured using a 5-point Likert scale, from1 (strongly disagree) to 5 (strongly
agree).
5.2.1 Maternity benefits in comparison to referent others
The adapted general benefit comparisons scale from Williams et al. (2002) from Study 1 was again used to mea-
sure comparisons of maternity benefits. We reverse-scored these items such that larger values reflect less favorable
maternity benefit comparisons (α= .89).
5.2.2 Perceived supervisor support
As in Study 1, we used the perceived supervisor support scale from Eisenberger et al. (2002; α= .90).
5.2.3 Perceived pregnancy discrimination
The pregnancy discrimination scale in Study 1measures perceptions that pregnantwomen feel they are being treated
selectively and differentially because of their pregnancy. In Study 2, we used a measure that more explicitly captured
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 831
perceptions of impeded career progress based on stigma related to social identity (i.e., pregnant worker). Following
Little et al. (2015), we adapted seven items from Sanchez and Brock’s (1996) measure of discrimination to reflect
perceived discrimination against pregnant workers, using a 5-point scale from strongly disagree to strongly agree.
Items that could not easily be adapted to reflect pregnancy were removed (e.g., At work, people look down upon me
if I practice customs of my culture). The other items had minor adaptations to reflect pregnancy rather than ethnic
background. The items are: “At work, I sometimes felt that my pregnancy was viewed as a limitation”; “At work, many
people have stereotypes about pregnant women and treated me as if they were true”; “At work, I sometimes felt that
people actively tried to stop me from advancing because of my pregnancy”; “At work, I did not get enough recognition
because of my pregnancy”; “At work, I felt that others exclude me from their activities because of my pregnancy”; “At
work, people looked downuponme if I needed special consideration because ofmypregnancy”; “Atwork, peoplemade
comments that indicatedmy pregnancy would negatively impact my career or my current job status” (α= .92).
5.2.4 Turnover after maternity leave
Following Little et al. (2015),we askedwomen if theyhad returned towork aftermaternity leave. Theywere also asked
if they returned to work for the same company. This construct was collapsed into a dichotomous variable in which “0”
represented returning to their job (n= 215), and “1” represented that they left their organization (n= 80).
5.2.5 Control variables
In Study 2, we again controlled for number of children, maternal age, and race. Additionally, prior work indicates that
pregnancy discrimination is associated with premature birth rates (Hackney et al., 2021). Having a premature baby
could result in mothers deciding to leave their job (altogether or for a less demanding one). Thus, we controlled for
the number of weeks that women were pregnant when they gave birth in Study 2. All hypothesized effects remained
consistent and significant whenwe conducted the analyses without controls.
5.3 Preliminary analysis
The descriptive statistics and correlations for Study 2 are displayed in Table 3. Again, we used FIML in Mplus 8.1
to account for missing data. We conducted a CFA to examine the factor structure of our study variables. The CFA
included the three continuous scales (maternity benefit comparisons, PSS, and perceived pregnancy discrimination).
We first, ran aCFAwith no correlated error terms, which provided good fit to the data (χ2= 222.85, df= 74, CFI= .95,
SRMR = .05, RMSEA = .08). Moreover, this model fit better than a one-factor model where all of the items loaded
onto one factor (χ2 = 1468.93, df = 77, CFI = .51, SRMR = .18, RMSEA = .24) and a two-factor model, where Time
1 variables loaded onto a factor and Time 2 variables loaded onto a second factor (χ2 = 796.52, df = 76, CFI = .75,
SRMR= .14, RMSEA= .17). Study 2 utilized a larger sample and a different measure of perceived pregnancy discrimi-
nation than study1. Various studies have shown the impact of sample size onmodel fit (e.g., Du et al., 1989), and thus, it
was not surprising that the fit wasmuch better in Study 2. Still, given our arguments regarding similarly-worded items,
we thought it appropriate to investigate the model fit with the two similarly-worded maternity benefits comparison
errors correlated. Doing so did not improve the fit (χ2= 209.29, df= 73, CFI= .95, SRMR= .04, RMSEA= .08).
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
832 PAUSTIAN-UNDERDAHL
TABLE 3 Descriptive statistics and correlations among variables in study 2
Variables M SD n 1 2 3 4 5 6 7
1. Number of children .71 .96 295
2. Age 31.00 4.61 295 .32**
3. #wks pregnant at birth 39.07 1.93 295 –.12* –.03
4. Race .85 .35 295 –.04 .02 .09
5. Less favorable benefits 3.29 .87 295 –.00 –.08 .00 .05
6. PSS 3.65 1.03 295 –.06 –.10*** .01 –.03 –.21**
7. Perceived pregnancy
discrimination
2.27 .95 178 .22** .06 –.01 –.00 .21** –.20**
8. Turnover .27 .45 295 .05 –.05 .03 .04 .16** –.21** .20*
Note: n = 178–295. Less favorable benefits = less favorable maternity benefit comparisons. PSS = perceived supervisor sup-
port. Race:White= 1; Non-White= 0. ***p< .10. *p< .05. **p< .01.
5.4 Hypothesis testing
Table 4 presents the unstandardized coefficient estimates for the three estimated models (following the same struc-
ture as Study 1). Logistic regression was used because turnover is a binary outcome. Given that our model uses
logit modeling for the dichotomous outcome variable, this precludes the calculation of standardized coefficients and
related fit statistics. Thus, we assessedmodel fit with the Bayesian Information Criterion (BIC). Lower BIC values indi-
cate better model fit. First, we tested our hypothesizedmodel (M3; BIC= 876.46) against an alternativemodel where
PSS predicts unfavorable benefits (BIC = 1636.56), as well as an alternative model where PSS predicts turnover and
unfavorable benefits (BIC= 1637.52). Our hypothesizedmodel has the lowest BIC and provides the best fit.
Consistent with Study 1 and in support of Hypothesis 1, less favorable maternity benefit comparisons were pos-
itively related to perceived pregnancy discrimination (B = .20, p = .036; see M2 in Table 4). Additionally, we found
that perceived pregnancy discrimination was positively related to turnover, thus supporting Hypothesis 2b (B = .56,
p = .005; see M2 in Table 4). Because turnover is a binary variable, these results suggest that for a one-increment
change in discrimination, the log odds of turnover increase by .56. The logistic regression odds ratio estimate indicates
that for a one-unit increase in discrimination, the odds of exiting one’s organization increase by a factor of 1.75. To test
our moderated-mediation model, we again used path analysis inMplus 8.1 (Stride et al., 2015). The variables included
in interactions were centered prior to the analyses (see M3 in Table 4). We tested whether the interaction between
maternity benefit comparisons and PSS influenced perceived pregnancy discrimination.
In support of Hypothesis 3, the interaction was significant, (B= –.15, p= .037; seeM3 in Table 4). The effect of less
favorable benefits on perceived discrimination was significantly positive when PSS was low (–1 SD; B= .31, p= .001),
and was not significant when PSS was high (+1 SD; B= .01, p= .942; see Figure 3). To further explore the moderated-
mediated relationships proposed in Hypothesis 4a, we examined the magnitude of the conditional indirect effect of
less favorablematernity benefit comparisonson turnover throughperceivedpregnancydiscriminationusing anempir-
ical bootstrapping approach with bias-corrected CIs. In line with our hypothesis, at lower levels of PSS, the indirect
effect was stronger and significant (–1 SD B = .19, 95% CI: .044, .421) and at higher levels of PSS, the indirect effect
becameweaker and non-significant (+1 SDB= .01, 95%CI: –.213, .208; Table 4). As in Study 1, we computed the index
of moderated mediation (a3*b) and used an Rweb utility with 20,000 repetitions to obtain the confidence intervals
(Selig & Preacher, 2008). The index of moderated mediation was significant (index= –.09, 95% CI [–.214, –.003]). Our
results support Hypothesis 4b.
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 833
T A B L E 4
St u d y 2 :U
n st an d ar d iz ed
re su lt s
P er ce iv ed
P re gn an cy
D is cr im
in at io n
Tu rn ov er
M o d el 1
M o d el 2
M o d el 3
M o d el 1
M o d el 2
M o d el 3
In te rc ep
t 2 .0 4 (1 .3 8 )
1 .3 3 (1 .5 2 )
1 .4 6 (1 .5 5 )
2 .6 7 (3 .4 6 )
3 .5 1 (3 .5 7 )
2 .6 8 (3 .1 6 )
C on tr ol s
N u m b er
o fc h ild
re n
.2 5 * ( .1 0 )
.2 4 ** (.0
9 )
.2 2 ** (.0
8 )
.1 7 (.1
5 )
.0 2 (.1
7 )
.0 1 (.1
6 )
A ge
.0 0 (.0
2 )
– .0 0 (.0
2 )
– .0 0 (.0
2 )
– .0 3 (.0
3 )
– .0 3 (.0
3 )
– .0 3 (.0
3 )
R ac e
– .0 2 (.2
3 )
– .0 8 (.2
1 )
– .0 9 (.1
9 )
.2 5 (.3
9 )
.2 8 (.4
5 )
.2 7 (.4
2 )
# w ks
p re gn
an t at
b ir th
.0 0 (.0
4 )
.0 2 (.0
4 )
.0 2 (.0
4 )
.0 6 (.0
8 )
.0 2 (.0
8 )
.0 2 (.0
8 )
Pr ed ic to rs
Le ss fa vo ra b le b en
ef it s
.2 0 * (.1
0 )
.1 6 ** * (.0
9 )
.2 5 (.1
8 )
.2 4 (.1
7 )
P SS
– .1 8 * ( .0 7 )
– .1 7 * ( .0 7 )
B en
ef it s X P SS
– .1 5 * (.0
7 )
M ed ia to r
P er ce iv ed
p re gn
an cy
d is cr im
in at io n
.5 6 ** (.2
0 )
.5 9 ** (.2
0 )
R 2
.0 6
.1 5 **
.1 8 **
.0 1
.1 1 *
.1 2 *
C on di ti on al In di re ct Ef fe ct s
Es t.
95 % C I
Lo w P SS :B
en ef it C o m p ar is o n s →
P er ce iv ed
D is cr im
in at io n →
Tu rn ov er
In te n ti o n s
.1 9 *
[.0 4 4 ,. 4 2 1 ]
H ig h P SS :B
en ef it C o m p ar is o n s →
P er ce iv ed
D is cr im
in at io n →
Tu rn ov er
In te n ti o n s
.0 1
[– .2 1 3 ,. 2 0 8 ]
A b b re vi at io n :C I=
co n fi d en
ce in te rv al .
N ot e: M 1 = m o d el w it h co n tr o lv ar ia b le s o n ly ;n
= 3 1 1 ,M
2 = m ai n -e ff ec t m o d el ;n
= 2 9 5 ,a n d M 3 = fu ll m o d el ;n
= 2 9 5 .N
u m b er s o u ts id e o f p ar en
th es es
ar e es ti m at ed
co ef fi ci en
ts an
d
n u m b er s in si d e o fp
ar en
th es es
ar e es ti m at ed
st an
d ar d er ro rs .P SS
= p er ce iv ed
su p er vi so r su p p o rt .R ac e: W h it e = 1 ;N
o n -W
h it e = 0 .
** * p < .1 0 .
* p < .0 5 .
** p < .0 1 (t w o -t ai le d ).
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
834 PAUSTIAN-UNDERDAHL
More Favorable Benefits (-
1 SD)
Less Favorable Benefits (+1
SD)
1
1.2
1.4
1.6
1.8
2
2.2
2.4
2.6
2.8
3
Low PSS (-1 SD)
High PSS (+1 SD)
Pe rc
ei ve
d Pr
eg na
nc y
D isc
ri m
in at
io n
F IGURE 3 Study 2 interaction. More/Less favorable benefits=More/less favorable maternity benefit comparisons.
5.5 Study 2 discussion
Our second study replicated the support for our proposed model, which is important for confirming our understand-
ing of the phenomenon of interest (Köhler & Cortina, 2021). As women perceived that their maternity benefits were
unfavorable compared to referent others, they experienced more pregnancy discrimination and stronger turnover.
Additionally, we again found that PSS helped buffer the effects of unfavorable benefits on perceived pregnancy
discrimination.
6 VIGNETTE STUDIES
Asmentionedabove,weused theexperimental-causal-chaindesignproposedbySpencer et al. (2005),which consisted
of two separate vignettes. In the first vignette, we randomly assigned participants to conditions representing the inde-
pendent variable (i.e., unfavorablematernity benefits) andmeasured themediating variable (i.e., perceived pregnancy
discrimination). In the second experiment, we randomly assigned participants to conditions representing themediator
(i.e., we manipulated perceived pregnancy discrimination) and measured the outcome variable (i.e., turnover inten-
tions). The experimental-causal-chain design providesmore rigorous evidence ofmediation than correlational studies
because it tests the causal relationship between the mediating and dependent variables (Lee & Feeley, 2018; Spencer
et al., 2005).
6.1 Vignette 1
For Vignette Study 1, we sought to support causality between less favorable maternity benefits comparisons and
perceived pregnancy discrimination. We conducted an a-priori power analysis using G*Power (Faul et al., 2007) with
power= .95 and alpha= .05 (two-tailed) to determine the required sample size. Our analysis revealed a sample of 246
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 835
is needed todetect aneffect sizeof .225,which is theaverage correlationbetweenmaternitybenefits comparisons and
perceivedpregnancydiscrimination fromStudy1andStudy2.Hoping to get a final sample of roughly 250participants,
we recruited 270 female employees from the United States on Prolific Academic–269 of whom completed the survey
and passed the attention check items. In exchange for participation, participants were compensated $2.00. Interested
participants were directed to an online survey, which provided the vignette information and measures. We incorpo-
rated a carelessness check item (‘answer this question with a 2″) into the survey. No participants failed to answer this check correctly. We also asked participants to rate the quality of the data provided (from 1 poor to 5 excellent) and to
rate the degree towhich theywere able to pretend that theywere truly part of the scenario (from 1 not at all to 5 com-
pletely). Those who did not rate above a three (19 respondents) were removed. Of the remaining final sample of 250
participants, the average agewas 36.8 (SD= 10.7); 66.4% self-identified asWhite, 7.2%Asian, 8.0%African American,
5.6%Hispanic, 12.4% indicatedmore than one race/ethnicity, and 52% other of whom had been pregnant before. Par-
ticipants were asked to read a workplace scenario, treating it as realistically as possible. Participants were asked to
imagine that they were pregnant and wondered about the company’s maternity benefits. We defined maternity ben-
efits, and at random, participants were told they had either much better (n = 123) or much worse (n = 127) benefits
than others (for full text, see Appendix). To foster a sense of realism in the study, we asked participants to describe in
their ownwords how they would feel in this situation. Participants completed the samemeasures for maternity bene-
fits comparisons (manipulation check; coefficient α= .99) and perceived pregnancy discrimination (coefficient α= .97)
described in the second field study above.
We ran an ANOVA using SPSS and found support for the effectiveness of our manipulations, with subjects in the
less favorable benefits condition perceiving significantly less favorable benefits (M = 4.72, SD = .56) than those in
the favorable benefits condition ((M = 1.31, SD = .43), F(1, 249) = 2916.02, p < .001). We also used an ANOVA to
test Hypothesis 1. Consistent with this hypothesis, those given the less favorable benefits conditionwere significantly
more likely toperceivepregnancydiscrimination (M=4.00, SD= .67) compared to those in themore favorable benefits
condition (M= 1.68, SD= .66); F(1, 249)= 758.20, p< .001).
6.2 Vignette 2
ForVignette Study2,we sought to support causality betweenperceived pregnancy discrimination and turnover inten-
tions. We conducted an a-priori power analysis using G*Power (Faul et al., 2007) with power = .95 and alpha = .05
(two-tailed) to determine the required sample size. Our analysis revealed a sample of 215 is needed to detect an
effect size of .24, which is the average correlation between perceived pregnancy discrimination and turnover inten-
tions/turnover from Study 1 and Study 2. Hoping to get a final sample of around 250 participants, we recruited 290
female employees from the United States on Prolific Academic—282 of whom completed the survey. In exchange for
participation, participants were compensated $1.59. Interested participants were directed to an online survey, which
provided the vignette information andmeasures. Again, we incorporated a carelessness check item (‘answer this ques-
tion with a 2″) into the survey; but no participants failed to answer this check correctly. We also asked participants to
rate the quality of the data provided (from 1 poor to 5 excellent) and to rate the degree to which they were able to pre-
tend that they were truly part of the scenario (from 1 not at all to 5 completely). Those who did not rate above a three
(35 respondents) were removed, leaving a final sample of 247. Participants had an average age of 36.7 (SD = 10.9);
70% self‘identified asWhite, 7%Asian, 8%African American, 2%Hispanic, and 53% had been pregnant before.
Participants were asked to read a workplace scenario and to treat it as realistically as possible. Participants were
randomly assigned to one of two conditions perceived pregnancy discrimination conditions. Following the experimen-
tal causal-chain approach (Spencer et al., 2005), we conceptually transformed the perceived pregnancy discrimination
measure (from Study 2) into a manipulation (for full text, see Appendix). Thus, the high perceived pregnancy discrim-
ination condition asked them to imagine they were pregnant and that their pregnancy was negatively impacting their
job status (n= 118). In contrast, the low discrimination condition stated there was no negative impact (n= 129).
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
836 PAUSTIAN-UNDERDAHL
To foster a sense of realism in the study, participants were asked to describe in their own words how they would
feel in this situation. Participants completed the samemeasure for perceived pregnancy discrimination (manipulation
check; coefficient α = .99) described in the second field study, and the same measure of turnover intentions (coef-
ficient α = .96) described in the first field study. Using SPSS, we ran an ANOVA to examine and find support for the
effectiveness of our manipulations with subjects in the high perceived pregnancy discrimination condition perceiving
significantlymore discrimination (M= 4.69, SD= .50) than those in the low perceived pregnancy discrimination condi-
tion (M=1.61, SD= .71), F(1, 245)=1572.41, p< .001). Furthermore, consistentwithHypothesis 2a, anANOVA found
that those given the high pregnancy discrimination condition were significantly more likely to have turnover inten-
tions (M= 6.57, SD= .71) compared to those in the low pregnancy discrimination condition ((M= 2.21, SD= 1.19); F(1,
246)= 1241.63, p< .001).
7 GENERAL DISCUSSION
In this paper, we sought to understand how maternity benefit comparisons relate to women’s turnover intentions
and turnover, which is of growing concern for organizations (Pinsker, 2019; Schulte, 2017). Consistent with identity
threat response theory (Petriglieri, 2011), we find that less favorable benefits lead to perceptions of pregnancy dis-
crimination and, ultimately, turnover intentions and actual turnover across two field studies and two vignette studies
of women in the United States. As further suggested by Petriglieri (2011), we find that social support in the form of
PSS helps buffer the degree towhich thematernity benefit comparisons relate to perceived pregnancy discrimination,
influencing turnover.
7.1 Implications for theory and research
Our studies contribute to theory and research in several ways. First, we contribute to the maternity benefits litera-
ture, which has generally neglected the influence of maternity benefits on women’s retention. We extend prior work
on maternity benefits, primarily examining national policies, economic outcomes, and the health of women and their
babies by illustrating the importance ofmaternity benefit comparisons in the retention of workingmothers. Given the
lack of federally mandatedmaternity leave and the considerable variability of maternity benefits offered to women in
the United States, women’s perceptions of how their benefits compare to others are quite salient as they form work-
place impressions of how their identity is valued, how they are treated, and how they will respond to that treatment
(i.e., turnover decisions).
Relying on identity threat response theory, we find that perceived pregnancy discrimination is a core mechanism
explaining why maternity benefit comparisons relate to turnover decisions of pregnant women and mothers. Mater-
nity benefits (unlike most general benefits) signal how a woman’s organization values her new social identity as a
pregnant employee—representing a potential “cue of future harm” (Petriglieri, 2011). Importantly, we also find that
perceived social support from one’s supervisor can act as “social buffering,” which partially insulates individuals from
the identity threats theyencounterbecause theyprovidea sourceof identity affirmation (Ashforthet al., 2007).On the
other hand, a lack of PSS amplifies the effects of unfavorable benefits, resulting in higher levels of perceived pregnancy
discrimination, and ultimately, turnover.
Our work also contributes to research on social identities. First, regarding theory on identity threats, posi-
tioning unfavorable maternity benefit comparisons as a potential identity threat—that organizations may not have
considered—opens new avenues toward understanding how perceptions of discrimination might develop. Second,
identity management research has investigated how pregnant women and other employees use signaling—or behav-
iors used by stigmatized individuals to gauge if otherswill support their identity—to discover if they are in a safe space
among supportive coworkers (e.g., Jones & King, 2014). In addition to signaling, we find that potentially stigmatized
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 837
individuals also interpret information already present in the work environment as either threatening or supportive.
Our work suggests that employees seeking to understand how their stigmatized identity will be valued at work may
look at benefit comparisons related to that identity as a cue of support or devaluation of one’s identity.
7.2 Implications for practice
Our findings have several implications for practice related tomaternity benefits and retention, particularly within the
United States and notably at a timewhenwomenhave left theworkforce in record numbers (Calvan&Rugaber, 2021).
The notion that maternity benefits comparisons influence retention is not radical, and still, only 21% of US workers
have access to paid family leave through private employers (U.S. BLS, 2020). Additionally, only 56% of private-sector
employees are eligible for FMLA, which provides up to 12 weeks of unpaid leave (Brown et al., 2020). A few states
offer paid-family leave—namely, California, Massachusetts, New Jersey, New York, Rhode Island, Washington, Ore-
gon, Colorado, and Connecticut plus the District of Columbia—typically funded through employee-paid payroll taxes
and administered through disability insurance programs. Given the wide variation in maternity benefits offered to
pregnant women in the United States, many women struggle to patch together their maternity leave using state reg-
ulations, employer policies such as vacation time and sick leave, and short-term disability coverage. This context is
ripe for social comparisons regarding one’s maternity benefits. Our research suggests that these comparisons are
particularly important to pregnant women’s turnover decisions.
Interestingly, opponents of improving maternity benefits for women in the United States have claimed that doing
so will increase discrimination because fewer organizations will hire women of child-bearing age due to the expenses
associated withmaternity benefits (Pinsker, 2019; Schulte, 2017). However, our findings show that whenwomen per-
ceive their maternity benefits to be less favorable than referent others’, such perceptions may drive experiences of
pregnancy discrimination and subsequent turnover. As such, organizations should carefully consider the costs of not
providing favorable maternity benefits.
Consistent with our findings, when Google expanded its paid parental leave policy from 12 to 18 weeks in 2007,
post-maternity leave retention rates increasedby50% (Michelson, 2021). Additionally, a survey conductedbyDeloitte
found that 77% of respondents agreed that the amount of parental leave offered by an employer could sway their
decision when choosing one employer over another (Deloitte, 2016). Other family-friendly benefits such as flexible
care accounts, flexible schedules, childcare assistance, and fertility-related benefits are also important towin the “war
for talent” (Ball, 2016). Given such findings, we recommend that organizations provide paid parental leave policies and
other family-friendly benefits that will be seen as competitive by job applicants and incumbents.
In addition to providing favorable maternity benefits, our findings also suggest that having supportive supervi-
sors is important for retaining pregnant women and working mothers. Importantly, we found that PSS interacts with
maternity benefit comparisons influencing perceived pregnancy discrimination, and ultimately, turnover. Given that
supervisors often have limited control over the maternity benefits their organizations offer, they should understand
that their supportive behaviors toward pregnant employees can help counteract potential negative consequences.
This support mechanism is especially critical given that women in the United States—and in countries that pro-
vide extensive pregnancy and maternity-related benefits (European Commission, 2012)—are increasingly reporting
pregnancy-related discrimination. Considering that robust organizations need to attract and retain employees, and
given the considerable costs and disruption associated with turnover, organizations should train supervisors to be
more supportive. Such training is important for supporting pregnantwomen andmothers atwork.However, given that
PSS is experienced at the interpersonal level between the employee and their supervisor, we recognize that it is not
a viable substitute for more widespread and institutionalized supportive maternity benefits offered by organizations
and/or mandated by local and federal governments.
Finally, with the rise of corporate social justice initiatives (Zheng, 2020), many corporations havemade promises to
increase diversity, equity, and inclusion (DEI) within their ranks (Stevens, 2020). To fulfill these pledges, organizations
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
838 PAUSTIAN-UNDERDAHL
must remove workplace barriers for the millions of expectant mothers in the workforce, which is essential for helping
women break through the glass ceiling and succeed in leadership roles (e.g., Cabrera, 2009). One way to begin to do
so is through maternity benefits. Awareness of maternity benefit comparisons, and the signal they send to expectant
mothers, is essential for organizations that wish to improveDEI efforts by increasing the number of women andmoth-
erswithin their organizational ranks (Blau&Kahn, 2013; Falk&Grizard, 2005).Human resourcemanagers should seek
to clearly inform all employees of the maternity benefits provided to them, and highlight any positive characteristics
of the benefits, to increase the likelihood of a positive reception.
7.3 Limitations and future research
Aswith all studies, ours is subject to limitations thatmay guide future research.We focus our analysis on samples from
the United States because of the wide variation in maternity benefits offered throughout the country, in part due to
the lack of a federally-mandated paidmaternity leave.Nonetheless, we do expect our results to apply to other national
contexts, given our focus on social comparisons. Although countries within the EuropeanUnion, Asia/Pacific, theMid-
dle East, the Americas, and across Africa have different policies around maternity leave length (ILO, 2014), individual
companies within these regions can choose to provide additional maternity benefits beyond what is federally man-
dated. As such, there may be considerable social comparisons made among women within these regions, and PSS is
likely to be an important buffer against unfavorable benefits across these contexts.
As such, it is important to investigatematernity benefit comparisons in countries outside of theUS Future research
should consider how a country’s regulatory environment influences the interpretation and comparison of benefits,
perceived pregnancy discrimination, and ultimately, turnover intentions and behaviors. In addition to the influence
of national regulations, individual factors such as race, sexual orientation, or family structure could also influence our
model. A limitation of Studys 1 and2was a lack of racial diversity among the participants (79%and86%White, respec-
tively). While the vignette studies achieved greater racial diversity (66% and 70% White, respectively), we believe
that future research should explore pregnant women’s experiences with an intersectional lens to better understand
howwomen with multiple marginalized identities perceive their benefits in comparison to others and how those per-
ceptions relate to discrimination and turnover intentions (Ruggs et al., 2013; Salter et al., 2021; Sawyer et al., 2013).
Indeed, pregnant women of color likely face a double stigma in the workplace which may compound the negative
consequences of identity devaluation on perceived discrimination.
Despite the consistent relationship between pregnancy discrimination and turnover decisions found in all of our
studies, it is difficult to rule out all possible alternative mechanisms. Indeed, the inclusion of two vignette studies
strengthens our claims of causality; however, we encourage future research to examine the influence of unfavor-
able benefits and perceptions of other related variables such as perceived inclusion and job satisfaction to show their
effects on perceived pregnancy discrimination and turnover intentions. Finally, the focus of our study was maternity
benefit perceptions rather than objective maternity benefits—such as access to FMLA. Future research should inves-
tigate objective maternity benefits to understand how they relate to the comparison-based heuristics that pregnant
workers perceive. Relatedly, fairness perceptions ofmaternity benefits may be due, in part, to deviation from industry
or jobnorms. Paidparental leave ismoreprevalent for employees inmanagerial roles, full-timeworkers, andworkers in
large organizations (Donovan, 2019). Further, industry is seemingly an important variable to consider, as employees in
technical services and financial services tend to have access tomore paid parental leave compared to employeeswork-
ing in industries such as construction and hospitality (Donovan, 2019). Thus, fairness judgments may depend partly
upon one’s industry or job type and those of comparison others. Future research should determine the importance of
industry and job typewhen considering differences in perceptions ofmaternity benefits andpregnancydiscrimination.
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 839
8 CONCLUSION
Our study explains how,why, and under what conditions (Sutton & Staw, 1995;Whetten, 1989) maternity benefit com-
parisons relate to turnover. Opponents of increasingmaternity benefits for women suggest that doing sowill increase
discrimination againstwomenatwork.However, data from two studies usingmulti-wave survey data and twovignette
studies indicate thatwhenwomenperceive theirmaternity benefits to be less favorable than referent others’ benefits,
they perceive more pregnancy discrimination. In turn, perceptions of pregnancy discrimination influence their sub-
sequent turnover decisions. Consistent with identity threat response theory, our results also suggest that perceived
supervisor support is a significant moderator, weakening the impact of maternity benefit comparisons on perceptions
of pregnancy discrimination. Offering competitive maternity benefits and having supportive supervisors may be key
for retaining women andmothers in the workplace.
ORCID
SamanthaC. Paustian-Underdahl https://orcid.org/0000-0003-4041-1472
LauraM. Little https://orcid.org/0000-0002-0554-9030
AshleyM.Mandeville https://orcid.org/0000-0003-0894-4172
AmandaS.Hinojosa https://orcid.org/0000-0003-4572-9172
ENDNOTES 1Vignette Study 1 was preregistered at https://osf.io/c2rxh/?view_only=1f2eac8fafbe4227aff0ef9e5dc58ec1Additionally,
this study was approved through the IRB (PROJECT00005158; entitled: Benefits Vignette). 2Vignette Study 2was preregistered at: https://osf.io/w285d/?view_only = 345806c5c9aa4ba69a2643f0b8f767c4. This pre-
registration includes two vignette studies. The first vignette study was originally included in this paper but was removed
during the review process and replaced with the current “Vignette Study 1.” Reviewers expressed concern regarding the
causal relationship between less favorable benefits and perceived discrimination. Including and manipulating the modera-
tor of PSS in this vignette study did not help support causality between benefits and discrimination and thus, in the next
revision, we included a study (Vignette Study 1) that includes only the IV and the mediator. Results suggested that favor-
able benefits with a supportive supervisor had clear positive outcomes (as we found in all studies), however, manipulated
supervisor support did not attenuate the influence of unfavorable benefits as strongly as perceived supervisor support
from the field studies. It is likely difficult to understand the impact of supportive supervision when one also has unfa-
vorable benefits in a short, vignette study. We believe this issue is related to the differences in our results and speaks
to the inappropriateness of including the moderator in our vignette. The removed vignette study is summarized here:
https://osf.io/gmn9h/?view_only=1099a09763e34e3e8414e9b1583e4a70 Additionally, this study was approved through
the IRB (PROJECT00005158; entitled: Benefits Vignette). 3We conducted a sensitivity power analysis using G*Power (Faul et al., 2007) with an alpha= .05 (two-tailed) and six predic-
tors for a linear regression model. Our analysis revealed that our research design and sample size had 97% power to detect
an effect size of R2 = .13 or higher.
4We conducted a sensitivity power analysis using G*Power (Faul et al., 2007) with an alpha= .05 (two-tailed) and seven pre-
dictors for a linear regressionmodel.Our analysis revealed that our researchdesign and sample sizehad99%power todetect
an effect size of R2 = .13 or higher.
REFERENCES
Ashforth, B. E. (2001). Role transitions in organizational life: An identity based perspective. Lawrence ErlbaumAssociates.
E Ashforth, B. E., Kreiner, G. A., Clark, M., & Fugate, M. (2007). Normalizing dirty work: Managerial tactics for countering
occupational taint. Academy of Management Journal, 50(1), 149–174. https://doi.org/10.5465/amj.2007.24162092
Ashforth, B. E., Harrison, S. H., & Corley, K. G. (2008). Identification in organizations: An examination of four fundamental
questions. Journal of Management, 34(3), 325–374. https://doi.org/10.1177/0149206308316059 Ashforth, B. E., & Kreiner, G. E. (1999). “How can you do it?”: Dirty work and the challenge of constructing a positive identity.
Academy of Management Review, 24(3), 413–434. https://doi.org/10.5465/amr.1999.2202129
Ball, P. (2015, November 23). 7 companies with innovative parental leave policies. Care @Work. https://benefits.care.com/7-
companies-with-innovative-parental-leave-policies
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
840 PAUSTIAN-UNDERDAHL
Ball, P. (2016, August 23). Are companies really betting on family-friendly benefits to win the talent wars? Care @ Work.
https://benefits.care.com/are-companies-really-betting-on-family-friendly-benefits-to-win-the-talent-wars
Baltes, B. B., & Heydens-Gahir, H. A. (2003). Reduction of work-family conflict through the use of selection, optimization, and
compensation behaviors. Journal of Applied Psychology, 88(6), 1005–1018. https://doi.org/10.1037/0021-9010.88.6.1005 Baum, C. L. (2003). The effects of maternity leave legislation on mothers’ labor supply after childbirth. Southern Economic
Journal, 69(4), 772–799. https://doi.org/10.2307/1061651 Berdahl, J. L., &Moore, C. (2006).Workplace harassment: Double jeopardy for minority women. Journal of Applied Psychology,
91(2), 426. http://doi.org/10.1037/0021-9010.91.2.426 Bernerth, J. B., & Aguinis, H. (2016). A critical review and best-practice recommendations for control variable usage. Personnel
Psychology, 69(1), 229–283. https://doi.org/10.1111/peps.12103 Best Upon Request. (2020, August 31). Themost powerful way to attract and retain female talent. https://www.bestuponrequest.
com/the-most-powerful-way-to-attract-and-retain-female-talent/
Blau, F. D., & Kahn, L. M. (2013). Female labor supply: Why is the United States falling behind? American Economic Review, 103(3), 251–56. https://doi.org/10.1111/peps.12103
Brown, T. A. (2015). Confirmatory factor analysis for applied research (2nd ed.). Guilford Publications. Brown, S., Herr, J., Roy, R., & Klerman, J. A. (2020). Employee and worksite perspectives of the Family and Medical Leave Act:
Results from the 2018 surveys. ABTAssociates andU.S. Department of Labor. https://www.dol.gov/sites/dolgov/files/OASP/
evaluation/pdf/WHD_FMLA2018SurveyResults_FinalReport_Aug2020.pdf
Budd, J.W., &Mumford, K. A. (2006). Family-friendly work practices in Britain: Availability and perceived accessibility.Human ResourceManagement, 45(1), 23–42. https://doi.org/10.1002/hrm.20091
Buzzanell, P., & Liu, M. (2007). It’s ‘give and take’ Maternity leave as a conflict management process. Human Relations, 60(3), 463–495. https://doi.org/10.1177/0018726707076688
Cabrera, E. F. (2009). Fixing the leaky pipeline: Five ways to retain female talent. People & Strategy, 32(1), 40–46. https://doi. org/10.1257/jep.33.1.43
Cabeza, M. F., Johnson, J. B., & Tyner, L. J. (2011). Glass ceiling and maternity leave as important contributors to the gender
wage gap. Southern Journal of Business and Ethics, 3, 73–82. Calvan, B. C., &Rugaber, C. (2021).Manywomen have left theworkforce.Whenwill they return?AssociatedPress. https://apnews.
com/article/coronavirus-pandemic-business-lifestyle-health-careers-075d3b0ab89baffc5e2b9a80e11dcf34
Carneiro, P., Løken, K. V., & Salvanes, K. G. (2015). A flying start? Maternity leave benefits and long-run outcomes of children.
Journal of Political Economy, 123(2), 365–412. https://doi.org/10.1086/679627 Cheng, M. (2021, June 29). CEOs can’t stop talking about the U.S. labor shortage. Quartz. https://qz.com/2025987/from-
fedex-to-chewy-ceos-are-talking-about-the-labor-shortage/
Cole, D. A., Ciesla, J. A., & Steiger, J. H. (2007). The insidious effects of failing to include design-driven correlated residuals in
latent-variable covariance structure analysis. Psychological Methods, 12(4), 381. https://doi.org/10.1037/1082-989X.12.4. 381
Crouse, S. (2020). Supervisor and coworker support: Their moderating roles on the relationship between diversity climate perceptions and retention-related outcomes [Unpublished dissertation]. San Jose State University.
Dahl, G. B., Løken, K. V., Mogstad, M., & Salvanes, K. V. (2016). What is the case for paid maternity leave? Review of Economics and Statistics, 98(4), 655–670. https://doi.org/10.1162/REST_a_00602
Deloitte. (2016). Parental leave survey. https://www2.deloitte.com/content/dam/Deloitte/us/Documents/about-deloitte/us-
about-deloitte-paternal-leave-survey.pdf
Deloitte. (2021, July 1). Why women are leaving the workforce after the pandemic — and how to win them back. Forbes. https://www.forbes.com/sites/deloitte/2021/07/01/why-women-are-leaving-the-workforce-after-the-pandemic-and-
how-to-win-them-back/?sh=33546ab2796e
Dirks, K. T., & Ferrin, D. L. (2002). Trust in leadership:Meta-analytic findings and implications for research and practice. Journal of Applied Psychology, 87(4), 611–628. https://doi.org/10.1037//0021-9010.87.4.611
Dutton, J. E., Roberts, L. M., & Bednar, J. (2010). Pathways for positive identity construction at work: Four types of positive
identity and the building of social resources.Academy ofManagement Review,35(2), 265–293. https://doi.org/10.5465/amr.
35.2.zok265
Dysvik, A., & Kuvaas, B. (2013). Perceived job autonomy and turnover intention: The moderating role of perceived supervisor
support. European Journal ofWork andOrganizational Psychology,22(5), 563–573. https://doi.org/10.1080/1359432X.2012. 667215
Ebaugh, H. R. F. (1988). Becoming an ex. University of Chicago Press. Edwards, J. R., & Lambert, L. S. (2007).Methods for integratingmoderation andmediation: A general analytic framework using
path analysis. Psychological Methods, 12(1), 1–22. https://doi.org/10.1037/1082-989X.12.1.1 Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D. (1986). Perceived organizational support. Journal of Applied
Psychology, 71(3), 500–507. https://doi.org/10.1037/0021-9010.71.3.500
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 841
Eisenberger, R., Stinglhamber, F., Vandenberghe, C., Sucharski, I. L., & Rhoades, L. (2002). Perceived supervisor support: Con-
tributions to perceived organizational support and employee retention. Journal of Applied Psychology, 87(3), 565–573. https://doi.org/10.1037/0021-9010.87.3.565
Emerson, K. T., & Murphy, M. C. (2014). Identity threat at work: How social identity threat and situational cues contribute to
racial and ethnic disparities in the workplace. Cultural Diversity and Ethnic Minority Psychology, 20(4), 508–520. http://doi. org/10.1037/a0035403
Enders, C. (2001). The impact of nonnormality on full information maximum-likelihood estimation for structural equation
models withmissing data. Psychological Methods, 6(4), 352–370. https://doi.org/10.1037/1082-989X.6.4.352 Enders, C. K. (2010). Applied missing data analysis. Guilford. Enders, C. K., & Bandalos, D. L. (2001). The relative performance of full informationmaximum likelihood estimation formissing
data in structural equation models. Structural Equation Modeling: A Multidisciplinary Journal, 8(3), 430–457. https://doi.org/ 10.1207/S15328007SEM0803_5
European Commission. (2012). http://ec.europa.eu/justice/gender-equality/files/your_rights/discrimination__pregnancy_
maternity_parenthood_final_en.pdf
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: Aflexible statistical power analysis program for the social,
behavioral, and biomedical sciences. Behavior ResearchMethods, 39, 175–191. http://doi.org/10.3758/BF03193146 Falk, E., &Grizard, E. (2005). The “glass ceiling” persists:Women leaders in communication companies. Journal ofMediaBusiness
Studies, 2(1), 23–49. https://doi.org/10.1080/16522354.2005.11073426 Fallon, K. M., Mazar, A., & Swiss, L. (2017). The development benefits of maternity leave. World Development, 96, 102–118.
https://doi.org/10.1016/j.worlddev.2017.03.001
Feldman, R., Sussman, A. L., & Zigler, E. (2004). Parental leave and work adaptation at the transition to parenthood: Individ-
ual, marital, and social correlates. Journal of Applied Developmental Psychology, 25(4), 459–479. https://doi.org/10.1016/j. appdev.2004.06.004
Gecas, V. (1982). The self-concept. Annual Review of Sociology, 8(1), 1–33. https://doi.org/10.1146/annurev.so.08.080182. 000245
Graham, J.W. (2009).Missing data analysis:Making itwork in the realworld.Annual Reviewof Psychology,60, 549–576. https:// doi.org/10.1146/annurev.psych.58.110405.085530
Green, T. K., & Kalev, A. (2007). Discrimination-reducing measures at the relational level. Hastings Law Journal, 59(6), 1435– 1461.
Grover, S. L., & Crooker, K. J. (1995). Who appreciates family-responsive human resource policies: The impact of family-
friendly policies on the organizational attachment of parents and non-parents. Personnel Psychology, 48(2), 271–288. https://doi.org/10.1111/j.1744-6570.1995.tb01757.x
Gruber, J. (1994). The incidence of mandatedmaternity benefits. The American Economic Review, 84(3), 622–641. Gupta, A. H. (2020). Why some women call this recession a ‘shecession’: In her words. The New York Times. https://www.
nytimes.com/2020/05/09/us/unemployment-coronavirus-women.html
Hackney, K. J., Daniels, S. R., Paustian-Underdahl, S. C., Perrewé, P. L., Mandeville, A., & Eaton, A. A. (2021). Examining the
effects of perceived pregnancy discrimination on mother and baby health. Journal of Applied Psychology, 106(5), 774–783. https://doi.org/10.1037/apl0000788
Hamilton, B. E., Martin, J. A., & Osterman, M. J. K. (2020). Births: Provisional data for 2019. Centers for Disease Control and
Prevention. https://www.cdc.gov/nchs/data/vsrr/vsrr-8-508.pdf
Hayes, A. F. (2015). An index and test of linear moderatedmediation.Multivariate Behavioral Research, 50(1), 1–22. https://doi. org/10.1080/00273171.2014.962683
Hebl, M. R., King, E. B., Glick, P., Singletary, S. L., & Kazama, S. (2007). Hostile and benevolent reactions toward preg-
nant women: complementary interpersonal punishments and rewards that maintain traditional roles. Journal of Applied Psychology, 92(6), 1499–1511. https://doi.org/10.1037/0021-9010.92.6.1499
Huang, J., Krivkovich, A., Rambachan, I., & Yee, L. (2021). For mothers in the workplace, a year (and counting) like no other.
McKinsey. https://www.mckinsey.com/featured-insights/diversity-and-inclusion/for-mothers-in-the-workplace-a-year-
and-counting-like-no-other
Howard, L. W., & Cordes, C. L. (2010). Flight from unfairness: Effects of perceived injustice on emotional exhaustion and
employeewithdrawal. Journal of Business and Psychology, 25(3), 409–428. https://doi.org/10.1007/s10869-010-9158-5 Hewitt, B., Strazdins, L., & Martin, B. (2017). The benefits of paid maternity leave for mothers’ post-partum health and well-
being: Evidence fromanAustralian evaluation. Social Science&Medicine,182, 97–105. https://doi.org/10.1016/j.socscimed.
2017.04.022
ILO. (2014). Maternity and paternity at work: Law and practice across the world. https://www.ilo.org/global/publications/
books/WCMS_242615/lang–en/index.htm
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
842 PAUSTIAN-UNDERDAHL
James, K., Lovato, C., & Cropanzano, R. (1994). Correlational and known-group comparison validation of a workplace prej-
udice/discrimination inventory. Journal of Applied Social Psychology, 24(17), 1573–1592. https://doi.org/10.1111/j.1559- 1816.1994.tb01563.x
Jones, K. P. (2017). To tell or not to tell? Examining the role of discrimination in the pregnancy disclosure process at work.
Journal of Occupational Health Psychology, 22(2), 239. https://psycnet.apa.org/doi/10.1037/ocp0000030 Jones, K. P., Clair, J. A., King, E. B., Humberd, B. K., &Arena,D. F. (2020).Howhelp during pregnancy canundermine self-efficacy
and increase postpartum intentions to quit. Personnel Psychology, 73(3), 431–458. https://doi.org/10.1111/peps.12365 Jones, K. P., & King, E. B. (2014).Managing concealable stigmas atwork: A review andmultilevelmodel. Journal ofManagement,
40(5), 1466–1494. https://doi.org/10.1177/0149206313515518 Jou, J., Kozhimannil, K. B., Abraham, J. M., Blewett, L. A., & McGovern, P. M. (2018). Paid maternity leave in the United States:
Associations with maternal and infant health.Maternal and Child Health Journal, 22(2), 216–225. https://doi.org/10.1007/ s10995-017-2393-x
King, E. B. (2008). The effect of bias on the advancement of working mothers: Disentangling legitimate concerns from inaccu-
rate stereotypes as predictors of advancement in academe.Human Relations, 61(12), 1677–1711. https://doi.org/10.1177/ 0018726708098082
King, E. B., & Botsford,W. E. (2009).Managing pregnancy disclosures: Understanding and overcoming the challenges of expec-
tant motherhood at work. Human Resource Management Review, 19(4), 314–323. https://doi.org/10.1016/j.hrmr.2009.03.
003
Köhler, T., & Cortina, J. M. (2021). Play it again, Sam! An analysis of constructive replication in the organizational sciences.
Journal of Management, 47(2), 488–518. https://doi.org/10.1177/0149206319843985 Kuvaas, B., & Dysvik, A. (2010). Exploring alternative relationships between perceived investment in employee development,
perceived supervisor support and employee outcomes. Human Resource Management Journal, 20(2), 138–156. https://doi. org/10.1111/j.1748-8583.2009.00120.x
Du, L., Terence, J., & Tanaka, J. S. (1989). Influence of sample size, estimationmethod, andmodel specification on goodness-of-
fit assessments in structural equationmodels. Journal of Applied Psychology,74(4), 625–635. https://doi.org/10.1037/0021- 9010.74.4.625
Ladge, J. J., Clair, J. A., & Greenberg, D. (2012). Cross-domain identity transition during liminal periods: Constructing multiple
selves as professional and mother during pregnancy. Academy of Management Journal, 55(6), 1449–1471. https://doi.org/ 10.5465/amj.2010.0538
Larsen, R. (2011). Missing data imputation versus full information maximum likelihood with second level dependencies.
Structural EquationModeling, 18, 649–662. https://doi.org/10.1080/10705511.2011.607721 Lee, S., & Feeley, T. H. (2018). The identifiable victim effect: Using an experimental-causal-chain design to test for mediation.
Current Psychology, 37(4), 875–885. https://doi.org/10.1007/s12144-017-9570-3 Lee, K. Y., & Kim, S. (2010). The effects of commitment-based human resource management on organizational citizenship
behaviors. Themediating role of the psychological contract.World Journal of Management, 2(1), 130–147. Lind, E. A. (2001). Fairness heuristic theory: Justice judgments as pivotal cognitions in organizational relations. In J. Greenberg
& R. Cropanzano (Eds.), Advances in organizational justice (pp. 56–88). Stanford University Press. Lind, E. A., Kray, L., & Thompson, L. (2001). Primacy effects in justice judgments: Testing predictions from fairness heuristic
theory.Organizational Behavior and Human Decision Processes, 85(2), 189–210. https://doi.org/10.1006/obhd.2000.2937 Little, L., Hinojosa, A., & Lynch, J. (2017). Make them feel: How the disclosure of pregnancy to a supervisor leads to changes in
perceived supervisor support.Organization Science, 28(4), 618–635. https://doi.org/10.1287/orsc.2017.1136 Little, L. M., Major, V. S., Hinojosa, A. S., & Nelson, D. L. (2015). Professional image maintenance: How women navigate
pregnancy in the workplace. Academy of Management Journal, 58(1), 8–37. https://doi.org/10.5465/amj.2013.0599
Liu, M., & Buzzanell, P. M. (2004). Negotiating maternity leave expectations: Perceived tensions between ethics of justice and
care. The Journal of Business Communication, 41(4), 323–349. https://doi.org/10.1177/0021943604268174 Luke, B., & Brown, M. B. (2007). Elevated risks of pregnancy complications and adverse outcomes with increasing maternal
age.Human Reproduction, 22(5), 1264–1272. https://doi.org/10.1093/humrep/del522
Lumen Technologies. (n.d.). LiveWell at Lumen. https://jobs.lumen.com/global/en/benefits
Madera, J. M., King, E. B., & Hebl, M. R. (2012). Bringing social identity to work: The influence of manifestation and suppres-
sion on perceived discrimination, job satisfaction, and turnover intentions. Cultural Diversity and Ethnic Minority Psychology, 18(2), 165–170. http://doi.org/10.1037/a0027724
McDonald, P., Dear, K., & Backstrom, S. (2008). Expecting the worst: Circumstances surrounding pregnancy discrimination
at work and progress to formal redress. Industrial Relations Journal, 39(3), 229–247. https://doi.org/10.1111/j.1468-2338. 2007.00486.x
Michelson, J. (2021). How small companies can offer great paid-leave programs. Harvard Business Review. https://hbr.org/
2021/01/how-small-companies-can-offer-great-paid-leave-programs
Microsoft. (n.d.). Make themost of life. U.S. Benefits atMicrosoft. https://careers.microsoft.com/us/en/usbenefits
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 843
Miller, S. (2016, October 24). Facebook, Coca-Cola, Braun Medical offer paid parental leave tips. Society for Human
Resource Management. https://www.shrm.org/resourcesandtools/hr-topics/benefits/pages/facebook-coca-cola-braun-
medical-paid-parental-leave.aspx
Miller, V.D., Jablin, F.M., Casey,M.K., Lamphear-VanHorn,M., &Ethington,C. (1996). Thematernity leave as a role negotiation
process. Journal of Managerial Issues, 8(3), 286–309. Netflix. (n.d.). Work-life philosophy. Netflix Jobs. https://jobs.netflix.com/work-life-philosophy
Ng, T. W., & Sorensen, K. L. (2008). Toward a further understanding of the relationships between perceptions of sup-
port and work attitudes: A meta-analysis. Group & Organization Management, 33(3), 243–268. https://doi.org/10.1177/ 1059601107313307
O’Reilly, III, C. A., Chatman, J., & Caldwell, D. F. (1991). People and organizational culture: A profile comparison approach to
assessing person-organization fit. Academy of Management Journal, 34(3), 487–516. https://doi.org/10.5465/256404 Petriglieri, J. L. (2011). Under threat: Responses to and the consequences of threats to individuals’ identities. Academy of
Management Review, 36(4), 641–662. https://doi.org/10.5465/amr.2009.0087
Pinsker, J. (2019, July 25). The conservative argument over paid family leave. The Atlantic. https://www.theatlantic.com/
family/archive/2019/07/family-leave-conservatives/594838/
Piszczek, M. M. (2020). Reciprocal relationships between workplace childcare initiatives and collective turnover rates of men
andwomen. Journal of Management, 46(3), 470–494. https://doi.org/10.1177/0149206318799480 Ragins, B. R., & Cornwell, J. M. (2001). Pink triangles: Antecedents and consequences of perceived workplace discrimination
against gay and lesbian employees. Journal of Applied Psychology, 86(6), 1244–1261. https://doi.org/10.1037/0021-9010. 86.6.1244
Rice, R.W., McFarlin, D. B., & Bennett, D. E. (1989). Standards of comparison and job satisfaction. Journal of Applied Psychology, 74(4), 591–598. https://doi.org/10.1037/0021-9010.74.4.591
Rothausen, T. J., Henderson, K. E., Arnold, J. K., & Malshe, A. (2017). Should I stay or should I go? Identity and well-
being in sensemaking about retention and turnover. Journal of Management, 43(7), 2357–2385. https://doi.org/10.1177/ 0149206315569312
Ruggs, E. N., Hebl, M. R., Law, C., Cox, C. B., Roehling, M. V., Wiener, R. L., & Barron, L. (2013). Gone fishing: I–O psychologists’
missed opportunities to understandmarginalized employees’ experienceswith discrimination. Industrial andOrganizational Psychology: Perspectives on Science and Practice, 6(1), 39–60. https://doi.org/10.1111/iops.12007
Salter,N. P., Sawyer, K., &Gebhardt, S. T. (2021).Howdoes intersectionality impactwork attitudes? The effect of layered group
memberships in a field sample. Journal of Business and Psychology,36(6), 1035–1052. https://doi.org/10.1007/s10869-020- 09727-y
Sanchez, J. I., & Brock, P. (1996). Outcomes of perceived discrimination amongHispanic employees: Is diversitymanagement a
luxury or a necessity? Academy of Management Journal, 39(3), 704–719. https://doi.org/10.2307/256660 Sandler, D., & Szembrot, N. (2019). Maternal labor dynamics: Participation, earnings, and employer changes. U.S. Census
Bureau. https://www2.census.gov/ces/wp/2019/CES-WP-19-33.pdf
Sawyer, K., Salter, N., & Thoroughgood, C. (2013). Studying individual identities is good, but examining intersectionality is
better. Industrial and Organizational Psychology, 6(1), 80–84. https://doi.org/10.1111/iops.12012 Schulte, B. (2017,May 18). The case againstmaternity leave. Better Life Lab. https://slate.com/human-interest/2017/05/the-
case-against-maternity-leave.html
Selig, J. P., & Preacher, K. J. (2008). Monte Carlo method for assessing mediation: An interactive tool for creating confidence
intervals for indirect effects [Computer software]. Available from http://quantpsy.org/
SHRM. (2019). 2019 employee benefits. Society for Human Resource Management Research & Surveys. https://www.shrm.
org/hr-today/trends-and-forecasting/research-and-surveys/pages/benefits19.aspx
Spencer, S. J., Zanna, M. P., & Fong, G. T. (2005). Establishing a causal chain: Why experiments are often more effective than
mediational analyses in examining psychological processes. Journal of Personality and Social Psychology, 89(6), 845–851. https://doi.org/10.1037/0022-3514.89.6.845
Spiggle, T. (2021, July 8). What does a worker want?What the labor shortage really tells us. Forbes. https://www.forbes.com/
sites/tomspiggle/2021/07/08/what-does-a-worker-want-what-the-labor-shortage-really-tells-us/?sh=25bf0a2a539d
Starbucks (2017,March 21). Starbucks parental leave benefits. Starbuck Stories&News. https://stories.starbucks.com/press/
2017/starbucks-to-expand-parental-leave-benefits/
Steele, C. M., Spencer, S. J., & Aronson, J. (2002). Contending with group image: The psychology of stereotype and social
identity threat. In Advances in experimental social psychology (Vol. 34, pp. 379–440). Academic Press.
Sterling, H. M., & Allan, B. A. (in press). Predictors and outcomes of U.S. quality maternity leave: A review and conceptual
framework. Journal of Career Development, 49(6) 1435-1453) https://doi.org/10.1177/08948453211037398
Stevens, P. (2020, June 11). Companies are making bold promises about greater diversity, but there’s a long way to
go. CNBC. https://www.cnbc.com/2020/06/11/companies-are-making-bold-promises-about-greater-diversity-theres-
a-long-way-to-go.html
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
844 PAUSTIAN-UNDERDAHL
Stride, C. B., Gardner, S., Catley, N., & Thomas, F. (2015). Mplus code for mediation, moderation, and moderated mediation
models. http://www.offbeat.group.shef.ac.uk/FIO/mplusmedmod.htm
Sutton, R., & Staw, B. (1995). What theory is not. Administrative Science Quarterly, 40(3), 371–384. https://doi.org/10.2307/ 2393788
Swann, W. B. (1987). Identity negotiation: Where two roads meet. Journal of Personality and Social Psychology, 53(6), 1038– 1051. https://doi.org/10.1037/0022-3514.53.6.1038
Sweeney, P. D.,McFarlin, D. B., & Inderrieden, E. J. (1990). Using relative deprivation theory to explain satisfactionwith income
andpay level: Amultistudy examination.Academy ofManagement Journal,33(2), 423–436. https://doi.org/10.5465/256332 Thomas,M. (2015). The impact of mandatedmaternity benefits on the gender differential in promotions [Unpublished dissertation].
University of Chicago.
United Healthcare. (n.d.). Taking care of your health during pregnancy. https://www.uhc.com/health-and-wellness/health-
topics/pregnancy
U.S. BLS. (2020). National compensation survey: Employee Benefits in the United States. https://www.bls.gov/ncs/ebs/
benefits/2020/employee-benefits-in-the-united-states-march-2020.pdf
Whetten, D. A. (1989). What constitutes a theoretical contribution? Academy of Management Review, 14(4), 490–495. https:// doi.org/10.5465/amr.1989.4308371
White, H. C. (1992). Cases are for identity, for explanation, or for control. In C. C. Ragin & H. S. Becker (Eds.),What is a case? Exploring the foundations of social inquiry (pp. 83–104). Cambridge University Press.
Williams, M. L., Malos, S. B., & Palmer, D. K. (2002). Benefit system and benefit level satisfaction: An expanded model of
antecedents and consequences. Journal of Management, 28(2), 195–215. https://doi.org/10.1177/014920630202800204 Zheng, L. (2020).We’re entering the age of corporate social justice. Harvard Business Review. https://hbr.org/2020/06/were-
entering-the-age-of-corporate-social-justice
How to cite this article: Paustian-Underdahl, S. C., Little, L. M., Mandeville, A., Hinojosa, A., & Keyes, A.
(2024). Examining the Role ofMaternity Benefit Comparisons and Pregnancy Discrimination inWomen’s
Turnover Decisions. Personnel Psychology, 77, 819–846. https://doi.org/10.1111/peps.12577
APPENDIX
Description ofmanipulations for Vignette Studies 1 and 2
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
PAUSTIAN-UNDERDAHL 845
Independent
Variable Level Text ofManipulation
Maternity Benefits
Comparisons
Vignette 1
Less Favorable Imagine that you are pregnant. You are concerned about how
your organizationmight view pregnant women and are
very curious about how supportive your organizationmay
be. Youwonder about thematernity benefits offered by
your organization.
Maternity benefits are a type of benefit specifically provided
for pregnant women and newmothers. Themost common
maternity benefit is maternity leave; however, in addition
tomaternity leave, some organizations offer other
maternity benefits such asmaternity “concierge services,”
cash stipends, educational materials, nurse advice lines,
special screenings for high-risk pregnancies, and other
maternity health programs. You investigate these options
at your organization. You assess the quality of your
maternity benefits by comparing them to thematernity
benefits of others (e.g., those whowork a similar
position/level of seniority, friends and family members, and
those with similar abilities and training as you). You also
consider whether thematernity benefits that your
organization offers tomeet your financial needs.
After you investigate yourmaternity benefits, you discover
that the level and quality of thematernity benefits you
have access to aremuch less than others (e.g., those who
work a similar position/level of seniority, friends and family
members, and those with similar abilities and training as
you) and they do notmeet your financial needs.
More Favorable Imagine that you are pregnant. You are concerned about how
your organizationmight view pregnant women and are
very curious about how supportive your organizationmay
be. Youwonder about thematernity benefits offered by
your organization.
Maternity benefits are a type of benefit specifically
provided for pregnant women and newmothers. Themost
commonmaternity benefit is maternity leave; however, in
addition tomaternity leave, some organizations offer other
maternity benefits such asmaternity “concierge services,”
cash stipends, educational materials, nurse advice lines,
special screenings for high-risk pregnancies, and other
maternity health programs. You investigate these options
at your organization. You assess the quality of your
maternity benefits by comparing them to thematernity
benefits of others (e.g., those whowork a similar
position/level of seniority, friends and family members, and
those with similar abilities and training as you). You also
consider whether thematernity benefits that your
organization offers tomeet your financial needs.
After you investigate yourmaternity benefits, you discover
that the level and quality of thematernity benefits you
have access to aremuchmore than others (e.g., those who
work a similar position/level of seniority, friends and family
members, and those with similar abilities and training as
you) and theymeet your financial needs.
(Continues)
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
846 PAUSTIAN-UNDERDAHL
Independent
Variable Level Text ofManipulation
Perceived Pregnancy
Discrimination
Vignette 2 High Imagine that you are pregnant. Since becoming pregnant, you
have seen signs that your organization does not value
pregnant workers. People in the organization often talk
about how your pregnancy will negatively impact you and
your job status. Theymake it clear that they look down on
pregnant women because they need special considerations
due to their pregnancy. Furthermore, pregnant women are
almost always excluded from activities and receive less
recognition and opportunities for advancement.
Low Imagine that you are pregnant. Since becoming pregnant, you
have seen no signs that your organization does not value
pregnant workers. People in the organization do not ever
talk about how your pregnancy will negatively impact you
and your job status. Theymake it clear that they do not
look down on pregnant women because they need special
considerations due to their pregnancy. Furthermore,
pregnant women are never excluded from activities and
receive equal recognition and opportunities for
advancement.
17446570, 2024, 2, D ow
nloaded from https://onlinelibrary.w
iley.com /doi/10.1111/peps.12577 by E
B SC
O SU
B SC
R IPT
IO N
SE R
V IC
E S - Journal D
igital L ic, W
iley O nline L
ibrary on [09/06/2024]. See the T erm
s and C onditions (https://onlinelibrary.w
iley.com /term
s-and-conditions) on W iley O
nline L ibrary for rules of use; O
A articles are governed by the applicable C
reative C om
m ons L
icense
Copyright of Personnel Psychology is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
