Social Forces 98(2) - Gender Inequality in Product Markets Research

This research paper, published in Social Forces, investigates gender inequality in product markets through three online experiments. The study, conducted by Elise Tak, Shelley J. Correll, and Sarah A. Soule at Stanford University, explores how gender status beliefs influence product evaluations. The authors develop a theory of status belief transfer, demonstrating that products made by women are disadvantaged in male-typed markets (craft beer) but not in female-typed markets (cupcakes). The research reveals an asymmetric negative bias, where products made by women receive lower evaluations in male-dominated markets. The study also examines the moderating effects of external status conferral and evaluator's product knowledge. The findings highlight the pervasive influence of gender-typing in product markets and suggest strategies to mitigate gender biases, drawing on status characteristics theory (SCT) to explain how these biases are perpetuated.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Gender Inequality in Product Markets: When an
How Status Beliefs Transfer to Products
Elise Tak, Stanford University
Shelley J. Correll and Sarah A. Soule, Stanford University
This paper develops and evaluates a theory of status belief transfer, the
by which gender status beliefs differentially affect the evaluations of pr
made by men and women.We conductthree online experiments to evaluate
this theory. In Study 1, we gathered 50 product categories from a large onli
and had participants rate each product’s association with femininity and ma
We find evidence ofthe pervasiveness ofgender-typing in productmarkets.In
Studies 2 and 3,we simulate male-typed and female-typed productmarkets (craft
beerand cupcakes,respectively).In the male-typed productmarket,a craftbeer
described as produced by a woman is evaluated more negatively than the s
product described as produced by a man. Consistent with our predictions, w
find that if the beer is conferred externalstatus via an award,the evaluation of the
beer made by a woman improves by a greater magnitude than the same be
by a man.In the female-typed productmarketof cupcakes,the producer’s gender
does not affect ratings. Together, the two studies provide evidence of an as
negative bias:products made by women are disadvantaged in male-typed ma
butproducts made by men are notdisadvantaged in female-typed markets.These
studies also provide compelling evidence of status belief transfer from prod
their products. We draw out the implications of these findings and suggest w
gender biases in product markets can be reduced.
Introduction
Research has consistently shown that women are disadvantaged by the sta
belief that they are less competent than men, which leads people to have lo
expectationsfor women’sperformance relativeto men’s (Ridgeway 2011).
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
This research wassupported bya grant from the Stanford GraduateSchool of Business,
“Collaboration and Innovation in the American Craft Beer Industry.” Stanford University appro
this IRB protocol (31730).An earlier version of this paper was presented atthe American
SociologicalAssociation Meetings in August2015 in Chicago,Stanford–Berkeley Organizational
Behavior Conference,EGOS 2016, and the Academy ofManagementMeeting 2017 in Atlanta,
Chicago,IL. The authors would like to thank the participants ofthe Stanford Organizational
Behavior Seminar for comments on an earlier version of this paper. Direct correspondence to
Tak, 655 Knight Way, Stanford, CA, 94305; email:[e-mail].
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
© The Author(s) 2019. Published by Oxford University Press on behalf of the
University of North Carolina at Chapel Hill. All rights reserved. For permissions,
please e-mail: [e-mail].
Social Forces 98(2) 548–577, November 2019
doi: 10.1093/sf/soy125
Advance Access publication on 22 January 2019
Gender Inequality in Product Markets
Social Forces 98(2)548
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

fined as culturally shared expectations about the competency or
worthiness of certain socialcategories (Correlland Ridgeway 2006),powerfully
shape evaluations and interactions in ways that disadvantage members of deva
groups.These beliefs have substantialrepercussions in the labor market—in the
case of gender, women,relative to men,tend to be judged by harsher standards
(Foschi1996,2000;Steinpreis,Anders and Ritzke 1999),are compensated less
(Castilla and Benard 2010),are less likely to get jobs (Goldin and Rouse 2000;
Turco 2010; Correll et al. 2007), and are less satisfied with their jobs (Chan and
Anteby 2015).These patterns are especially pronounced for women in male-
dominated jobs (Turco 2010). In contrast, there is no consistent evidence of suc
penalties for men in female-dominated domains (Taylor 2010; Williams 1992).
While there is extensive research on the disadvantaging effects of status beli
in the domain of labor markets, there is surprisingly little work on the effects of
these status beliefs on the evaluations of products.1 This is surprising because
many products are closely associated with those who produce them,as is the
case in entrepreneurship (e.g.,Thornton 1999),smalland medium businesses,
scholarly work (Simcoe and Waguespack2011), and cultural markets
(DiMaggio 1987). If, as we argue in more detail below, status beliefs about indi-
vidual producers transfer to their products, status beliefs affect the assessment
of the quality of products, in addition to affecting the assessments of the people
producing them. This suggests that the effects of gender status beliefs on gend
inequality in the economy are greater than previously shown.
In this paper, we draw on status characteristics theory (SCT) to develop and
evaluate a theory of status belief transfer, showing how gender status beliefs d
ferentially affect the evaluations of products made by men and women. We pre
dict that genderstatusbeliefsabout men and women willtransferto the
products they produce,such that there willbe a lower evaluation of products
made by women in certain product markets. To the extent that our predictions
hold, this implies that in many product markets,a female producer would be
rated lower than an otherwise equal male producer (e.g., being seen as less cap
ble of delivering a product) and that her product itself would also be seen as
lower quality.
We conductthree experimentalstudiesto examine these ideas.Study 1
shows the pervasiveness of gender-typing in product markets by testing for cul
tural associations with masculinity or femininity among fifty widely used pro-
ducts. Studies 2 and 3 test for the transfer of status beliefs from the producer t
their products by investigating how products made by men and women fare in
male-typed and female-typed markets, respectively. We find that women-made
products receive lower evaluation than the same product made by a man in a
male-typed market (Study 2), but that this penalty does not exist for men-made
products in a female-typed market(Study 3).Furthermore,we develop and
evaluate two mechanisms that moderate the effect of gender status beliefs on
product evaluations:externalstatus conferraland evaluator’s product knowl-
edge. We conclude by drawing out the implications of our findings for reducing
gender biases in product markets.
Status beliefs, de
Gender Inequality in Product Markets549 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

A status characteristic is a nominaldistinction,such as an ascribed attribute
(e.g., race or gender) or a role (e.g., professor or mother), with widely share
liefs aboutthe competence or worthiness ofdifferentstates ofthe distinction
(e.g., men versus women) (Berger, Cohen, and Zelditch 1972). These beliefs
important because they affect relative expectations individuals have for the
formance of individuals in evaluative contexts (Ridgeway and Correll 2004).
a consequence, in the case of gender, people tend to have higher performa
pectations for men compared to women, leading men to be judged as more
petent than women (Foschi 1996, 2000). SCT has been widely used to expla
the differences in evaluative outcomes of people in labor markets (Foschi 19
2000;Steinpreis,Anders,and Ritzke 1999)and in organizations (Bunderson
2003; Turco 2010).
We argue that status beliefs also transfer to the products that men and w
make, and we theorize about the mechanism of status belief transfer. There
both theoretical and empirical reasons to expect that status belief transfer i
vasive and is also an important mechanism by which inequality is perpetuat
outside of the domain of labor markets. From a theoretical perspective, the
that status beliefs may spread to “non-status elements” was an assumption
the foundationalwork on SCT by Berger et al.(1972,p. 245).However,this
assumption has notbeen systematically tested,which has limited our under-
standing of the extent to which status beliefs transfer from individuals to pr
ducts. In addition to providing a test of the mechanism, we also develop a n
argument for why status beliefs transfer to products. We do so by investigat
how status beliefs shape evaluations of products in male- and female-typed
kets. Empirically, in many product markets, there are close associations bet
the productand its producer.Examples include scholarly work (Simcoe and
Waguespack 2011),cultural products(DiMaggio 1987) and products of
entrepreneurialendeavors (Thornton 1999).Understanding the mechanism of
status belief transfer will illuminate how status-based gender inequality is p
uated in these product markets.
Status Beliefs and the Evaluations of Products in Male- and Fem
Typed Markets
Conceptually, there are two types of status characteristics that have been s
to affect the evaluation of individuals:specificand diffuse (Correll and
Ridgeway 2006). Specific status characteristics, such as an occupational rol
possession ofspecific skillset,carry culturalexpectations for a well-defined
range of tasks, thus forming beliefs and aiding evaluations in a limited rang
contexts (Kunda and Spencer 2003). For instance, without specific informat
about an individual, we expect a software engineer to be competent at prog
ming a computer and a kindergarten teacher to be adept at teaching childre
the extent that those characteristics are relevant to the task at hand, specifi
tus characteristics influence performance expectations.
A Theory of Status Belief Transfer
Social Forces 98(2)550
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

eral expectations about the competence of members of a category. In the case
gender,widely shared beliefs continue to contain expectations thatmen are
more capable than women in a wide array of settings, independent of the speci
task at hand (Correll and Ridgeway 2006). In tandem, there are also specific ex
pectations thatmen are more capable atstereotypicalmasculine tasks (e.g.,
sports),while women are more capable atstereotypicalfeminine tasks (e.g.,
caretaking). Research has shown that these combinations of specific and diffus
status beliefs shape the evaluations ofmen and women in self-fulfilling ways
that,on balance,disadvantage women in various labor market outcomes (e.g.,
Castilla and Benard 2010;Chan and Anteby 2015;Goldin and Rouse 2000;
Steinpreis, Anders, and Ritzke 1999; Turco 2010).
We argue that these specific and diffuse status beliefs combine to produce re
ative devaluation of women’s products, leading to the mechanism of status bel
transfer. Concretely, we argue that the diffuse status beliefs associated with pro
ducer’sgendercombine with the specific statusbeliefsassociated with the
gender-typing of the product market, which jointly determine the overall evalua
tion of the product.When the producer’s gender aligns with the gender-typing
of the product market, such as men in male-typed product markets or women i
female-typed product markets, specific status beliefs will favor these producers
leading to higher performance expectations. However, when a product is made
by someone whose gender does not align with the gender-typing of the produc
market, specific status beliefs will be disadvantaging. In other words, specific st
tus beliefs vary with gender-typing of the market, whereas diffuse status beliefs
are consistent across contexts.
We differentiate between male-typed markets (i.e., those culturally associate
with men, such as sports markets) and female-typed markets (i.e., those cultur
ally associated with women, such as markets for children’s products) and make
predictions about each of these markets. First, in male-typed product markets,
man’s product will benefit both from the diffuse status belief that men are gene
ally more capable and from the specific status belief that men are better at ma
typed tasks,such as those required to make or consume male-typed products.
Both the specific and diffuse status characteristics are associated with more po
tive performance expectations for a man’s product than for a woman’s product
leading to more favorable assessments of men’s products. If the identical produ
were instead made by a woman,her product would be disadvantaged both by
the diffuse status belief that women are overall less capable and by the specific
status belief that women are not as good at the tasks required to produce male
typed products. This results in a lower evaluation for the same product made by
a woman than by a man in such markets.2
Hypothesis 1. In a male-typed market, a product made by a woman will
be evaluated more negatively than the same product made by a man.
In female-typed product markets, the diffuse expectation that men are generall
competent is at odds with the specific expectation that men are less skilled at t
tasks required to make or consume female-typed products. Based on the conte
Diffuse status characteristics, such as gender and race, carry broad and gen-
Gender Inequality in Product Markets551 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

woman would be evaluated more positively than an identical product produ
by a man. However, even in female-typed markets, the diffuse status belief
women are less competent or worthy may persist, thus disadvantaging wom
products. This tension between diffuse and specific status beliefs complicate
prediction.Status characteristics theory has shown thatspecific status beliefs
have a bigger impact on performance expectations and evaluations than do
fuse status beliefs because specific status beliefs are more relevant to the t
hand (Correlland Ridgeway 2006).Thus, we expectthatthe advantage wo-
men’s products gain from specific status beliefs is larger than the advantag
men’s products gain from diffuse status beliefs.This means that women’s pro-
ducts should have an advantage over men’s products, but the magnitude of
advantageshould be smallerthan that experienced by men in male-typed
markets.
Hypothesis 2. In a female-typed market, a product made by a man will
be evaluated more negatively than the same product made by a woman
although the magnitude of the gap will be smaller than that found in a
male-typed market.
External Status Conferral
Sociologists have shown that status beliefs powerfully influence judgments
conditions ofuncertainty (Gould 2002;Podolny 1993;Salganik,Dodds,and
Watts 2006).Similarly,socialpsychologists have shown that stereotypes func-
tion as cognitive shortcuts when there is ambiguity about how to evaluate p
formances (cf. Correll 2017). Applying these insights to product evaluations
expect that if there is uncertainty about the quality of a product, individuals
on status beliefs to reduce the uncertainty, thereby forming evaluations tha
consistent with status beliefs associated with the product (Hypotheses 1 an
However, to the extent that there is a reduction in uncertainty about the pro
uct, we should expect the effects of status beliefs to diminish.
One way to reduce uncertainty around the quality of a product is external
tus conferral. Research shows that status conferral, such as winning an awa
receiving a certification, is associated with more positive evaluations becau
signalsexternaljudgmentsof quality (Podolny 1993;but seeKovács and
Sharkey 2014).However,externalstatus conferralcould have different effects
for products made by women and men. For example, recent empirical work
shown that there may be greater benefits from status conferral for low-statu
tors (e.g.,women) in the context of entrepreneurship (Thébaud 2015;Tinkler
et al. 2015). Building on this work, we explore how these predictions play ou
product markets and, importantly, how these predictions might differ depen
on the gender-typing of the product market.
We develop an argument for why there may be different effects of extern
status conferralby theorizing about how multiple pieces of status information
are synthesized when forming performance evaluations and expectations.We
of specific status beliefs, we might expect that female-typed products made
Social Forces 98(2)552
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

attenuation principle,
additionalinformation about status has more impact on evaluations when the
information isnovel than when the information isredundant(Correll and
Ridgeway 2006).Additionalpositive status information (such as winning an
award) adds incremental value on performance expectations. Second, the incon
sistency principle predicts that a single piece of positive status information will
have a greater impact on the evaluative outcome if it is conventionally associat
with negative status information (Correll and Ridgeway 2006). Based on these
principles, we predict that a product made by a high-status actor will benefit les
from additional status than a product made by a low-status actor.
Applying these principles to a male-typed product market, products made by
men already benefit from both diffuse (being produced by a man) and specific
(men’s higher competency in a male-typed task)status beliefs.For a product
made by a man,winning an award adds more positive information to a field
already imbued with positive status. Thus, the attenuation principle leads us to
predict that the evaluation of a man’s product would be raised only slightly afte
winning an award. However, as argued in Hypothesis 1, a woman’s product in
a male-typed market typically experiences negative expectations because of bo
lower diffuse and specific status beliefs. Thus, for women’s products, the positiv
signal from winning an award should have a large positive increase on the eval
ation of her product,as predicted by the inconsistency principle.We predict a
gender-differentiated impact of status conferral on product evaluations.
Hypothesis 3. In a male-typed market, the evaluations of products made
by women willexperience greater gains from status conferralthan the
evaluations of products made by men will.
In female-typed product markets, we expect that status conferral will have a les
gender-differentiated effect on product evaluations. In these settings, specific a
diffuse status beliefs lead to inconsistent information about status, such that th
gain thatwomen experience in female-typed markets is relatively smallcom-
pared to the gain men experience in male-typed markets (Hypothesis 2).The
partially offsetting effects of diffuse and specific status beliefs mean that the le
of uncertainty about product quality should also be more equal for women and
men in female-typed markets compared to male-typed markets. Thus, we predi
that status conferralwill have a similar effect for men and women in female-
typed markets.
Hypothesis 4. In a female-typed market, status conferral will lead to pos-
itive improvements in the evaluations ofboth products made by men
and products made by women.
Product Knowledge
Anotherway thatuncertainty aboutproductquality could be reduced,and
thereforedecreasean evaluator’srelianceon statusbeliefs,is by assessing
whether the evaluator is knowledgeable about the product. Lack of knowledge
draw on two principles in SCT.First, according to the
Gender Inequality in Product Markets553 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

’ uncertainty abouthow to
evaluate its quality. This prediction is based on research in consumer marke
which finds that there are differences in how experts and novices gather an
information about products, which results in meaningful differences in how
conduct product evaluations (Brucks 1985). The more knowledge the consu
possess, the more likely they are to make evaluations based on a larger num
of product attributes, and the less likely they are to rely on brand names, w
convey statusinformation (Bettman and Park 1980;Park and Lessig 1981;
Sujan 1985).
This literature also points to differences in how experts respond to atypica
features of a product. Knowledgeable consumers generate more product-re
thoughts and spend more time when evaluating atypicalproducts,a difference
not observed among the less-knowledgeable consumers (Sujan 1985). Relat
Peracchio and Tybout (1996) show that evaluators who possess a high level
knowledge are less affected by cultural schemas, such as status beliefs, com
with those who have a low level of knowledge. Since those with more knowl
are better able to attend to product features, they are less likely to need to
on status information as a signal of quality. While these papers theorize bro
about atypical attributes of products, we specifically focus on one type of at
cality, which is when a product is made by someone whose gender is incons
with the gender-typing of the particular product. We predict that those who
knowledge about a product will be more likely to discriminate based on the
der of the person producing the product.Therefore,as the evaluator’s levelof
knowledge about the product increases, the less likely the evaluator is to re
status beliefs about gender when evaluating the product.
Hypothesis 5. The lower the level of an evaluator’s product knowledge,
the greater the gender gap in product evaluation for a product made by
a producer whose gender is inconsistent with the gender-typing of the
market.
Overview of Studies
We conducted three experimental studies that test our arguments about ge
inequality in product markets. In Study 1, we demonstrate the pervasivenes
gender-typing in product markets by investigating people’s gendered perce
of a wide range of consumer products. Theoretically, Study 1 forms the basi
our argument that these specific status beliefs could operate in ways that d
vantage producers whose gender is inconsistent with these beliefs. Studies
3 ask participants to evaluate a product based on a product label, where sp
information was changed to manipulate the gender of the producer and whe
the product was conferred external status. These studies test the claim that
ple arrive at different evaluations because of the gender of the producer an
gender-typing of the market.We offer these studies as evidence of status belief
transferto product markets.In all studies,we recruited participantsfrom
Amazon’s Mechanical Turk (MTurk).3
abouta product,conversely,increases evaluators
Social Forces 98(2)554
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

There is a lay perception that some products are more strongly associated with
femininity or with masculinity,such as baby products or golf clubs.However,
no sociologicalwork has systematically investigated the extentto which pro-
ducts are gender-typed. Study 1 demonstrates the pervasiveness of gender-typ
in productmarkets by testing for gender associations for 50 commonly used
consumer and household products. In order to compile an exhaustive list of pro
ducts, we collected the complete list of product categories on Jet.com (now par
of Walmart), a generalist online retailer that offers branded household and con-
sumer products. The complete list resulted in 362 products. After sorting alpha-
betically,we used a random number generator in the statisticalsoftware R to
choose fifty products for inclusion in our study.4
Participants and Task
We recruited 150 paid participants who were residing in the United States and
were at least twenty-one years old (see appendix A for a summary of partici-
pants’ demographic characteristics across three studies). We filtered participan
by gender before the survey started to achieve equal representation of men an
women, resulting in seventy-six women and seventy-four men. For each partici-
pant, the survey (created in Qualtrics) showed a random set of twenty products
out of fifty possible products,then asked the participant to rate the extent to
which the product was gender-typed. This resulted in each product having ~60
ratings ([150 participants× 20 product ratings/participant]/50 products).
Participants answered one question about each product’s overallfemininity or
masculinity: “Most people think (product) is _____” with a 7-point Likert scale
that ranged from “Very feminine” (7) to “Very masculine (1).” We used these
ratings to construct a variable called gender-typing. The means and 95 percent
confidence intervals are shown in Figure1.
Results
We find that many products have stronger associations with one gender than th
other. As shown in figure 1, examples of strongly male-typed products (defined
as ≤3 on the 7-point Likert scale) include hand tools, lawn mower, shovel, bas-
ketballequipment,camping equipment,Sci-Fi books,and car roofranks (22
percent of allproducts surveyed).Examples of strongly female-typed products
(defined as ≥5 on the 7-pointLikert scale)include baking mix,ring, stroller,
body moisturizer, and high heels (18 percent of all products surveyed). Thus, w
see evidence of strong associations with one gender or another across in 40 pe
cent of product categories.While the percentage of products that are gender-
typed is quite high, we expect these effects would be even larger if we included
products that are more explicitly marketed to men or women (e.g.,razors tar-
geted at women vs. men; vitamin pills for men vs. women). There are also pro-
ducts that did not have strong associationswith either gender,such as
Study 1. Gender-Typing in Product Markets
Gender Inequality in Product Markets555 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

humidifier, table lamp or comforter. But even here, we see evidence that it
not take much for a product category to become gender-typed.For example,
consider two very similar products: “comforter” and “duvet cover.” The form
was rated as gender neutraland the latter,as feminine even though both pro-
ducts are bed covers. Further,in the pretest of consumer products in Studies 2
and 3 described below,we find that 60 percent of the products were rated as
either stereotypically masculine or feminine. In sum, we find that many prod
have gendered associations, which offers empiricalevidence that gender-typing
is pervasive in product markets.
Studies 2 and 3 Overview and Pretest
Studies 2 and 3 evaluate the mechanism of status belief transfer by evaluat
the effects of producers’ gender on the assessment of their products. We do
simulating a male-typed product market of craft beer (Study 2) and a femal
typed productmarketof cupcakes (Study 3).We recruited participants who
were residing in the United States, were at least twenty-one years old, and h
had the experience of consuming craft beer/cupcakes. However, we did not
participantsbased on gender,as doing so could prime participantsto con-
sciously think about gender.
We acknowledge that in order for a status characteristic to have an effect
performance expectationsand evaluations,the statuscharacteristic mustbe
salientand/ortask relevant.Thus, we carefully selected productsthat were
gender-typed, meaning that people culturally associate them with either ma
linity or femininity, as we show in Study 1. Moreover, because we are intere
in the effects of status beliefs,which are widely held culturalexpectations,we
took care to identify products that are broadly available to a general audien
Figure 1.Gender-typing of products listed on Jet.com, with means and 95 percent con
intervals
Female-Typed Products
Male-Typed Products
dence
Social Forces 98(2)556
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

test (N = 50) that tested for the gender-typing of ten widely available and affor
able food products in a way that is similar to Study 1.5 The test identified four
products that were considered gender-neutral(dinner rolls,soup,spinach,and
coffee), three that were more male-typed (craft beer, chicken wings, and bacon
and three thatwere more female-typed (dietcola, cupcakes,and macarons).
Next, we narrowed down our choices by considering monthly consumption,
which ranged from 1 (never)to 5 (more than twenty timesper month).
Cupcakes and craft beer were consumed with similar frequency, with a mean o
1.88 for craft beer and 1.54 for cupcakes, a difference that is not statistically sig
nificant (t = −1.70, 95 percent CI = [−0.74, 0.06]). Further, we needed product
for which we could reasonably make cases for describing the identity of the pro
ducer (i.e.,brewer,baker),including their gender,withoutraising suspicion
about these studies’ focus on gender. Based on these criteria, we chose craft b
and cupcake for our two studies.6
Study 2. Male-Typed Product Market of Craft Beer
Procedure
Study 2 utilized a between-subject design where 272 participants were random
assigned to one of the four conditions, which crossed brewer’s gender (man or
woman) with externalstatus conferral(with or without status).To encourage
participants to pay attention to the information provided on the label,their
screen was locked for forty-five seconds before proceeding to the evaluation
task.
Manipulation
Brewer’s gender
We manipulated brewer’s gender by changing the name of the brewer on the
label. In the condition where the brewer was a woman, the labeldescribed the
producer as Sarah and included a brief description of her career in brewing. All
pronouns referred to Sarah as a she.In the condition where the brewer was a
man,the labelwas identicalexcept that the producer was David and allpro-
nouns referred to David as a he.
External status conferral
In the status conferral condition, the label described the beer as the winner of a
gold medal at the Great American Beer Festival. This is a prestigious award pre
sented annually by the Brewer’s Association,a nationaltrade association for
craft and home brewers. While we used a real award, the beer label described a
fictitious beer. In the no-status condition, there was no mention of an award. To
keep the text length similar to the status condition, we included some text abou
the kind of glassware that is appropriate for the beer. The suggested glassware
was described as a Pilsner glass, a common style of glassware.7
In order to select products that met these requirements, we conducted a pre-
Gender Inequality in Product Markets557 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

The dependent variable,overallevaluation,was constructed by creating a com-
posite of four items.The first item in the composite was taken from a question
where participants were asked how likely they would be to buy the beer, wh
had answer choices ranging from “very unlikely” (1) to “very likely” (7).The
mean was 4.67 (SD = 1.42). The second item was how much the participant
willing to pay for a pint of this beer at a brewpub. Participants could enter a
amount but if they recorded a value higher than $10, which is unusually hig
a pint of beer, the survey alerted the respondent to reconsider. The average
ingness to pay was $4.53 (SD = 1.84), with a minimum of $0.00 and a maxi
of $11.00.Third, we measured the participants’expected taste.Participants re-
sponded to,“overall,what do you think the beer willtaste like?” with answer
choices ranging between “terrible” (1) and “excellent” (7) (mean = 4.90,SD =
1.02).Lastly,we measured expected quality with answer choices ranging from
“lowest quality” (1) to “highest quality” (7),which had the mean of 5.20 and
standard deviation of 1.21. To construct the composite variable, overall eva
tion, we first standardized each of these four variables to a standard norma
bution, then averaged these z-scores. Cronbach’s alpha for the measure is 0
Evaluator’s Product Knowledge
We draw on the literature on consumer product knowledge to construct our
uct knowledge measures. This literature classifies product knowledge as ob
and subjective knowledge (Carlson et al.2009).Objective knowledge describes
the actualinformation that consumers possess,while subjective knowledge de-
scribes consumers’perceptions of their own knowledge.We mirrored a widely
cited measurement methodology to measure both dimensions of product kn
edge (Mitchelland Dacin 1996). The first question was, “in your opinion,how
would you evaluate your level of expertise in craft beer?” and the answer ch
ranged between “no expertise” (1) to “connoisseur” (5), and the mean was
(SD = 1.07). This measure captures one’s own assessment of expertize—su
knowledge (Carlson et al. 2009). We also asked about their familiarity with c
beer in two ways. First we asked whether the participant had ever been call
“beer snob,” a term in the craft beer community to describe those who are
edgeable about craft beer. Twenty-three percent of the participants reporte
they had been called a beer snob.The second question was,“on average,how
many days in a week do you drink beer?” and answer choices ranged from 0
days.The mean was 2.13 days a week (SD = 1.67).This measure was useful
because consumption frequency increases familiarity with the product and f
iarity is correlated with consumer knowledge (Alba and Hutchinson 1987; Pa
and Lessig 1981).
Descriptive Statistics
Table 1 presents summary statistics for the four conditions and the results o
paired t tests that compare the overall evaluations of man-made and woma
Dependent Variable
Social Forces 98(2)558
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

beers within the status conferraland the no-status conferralconditions.For the
beers that did not win an award, the overall evaluation of men’s craft beer (mea
= 0.03)is approximately 0.30 points higher than thatof women’s craftbeer
(mean = −0.26), a significant difference (p value <0.05). In contrast, there are
discernable differences between man-made and woman-made beers if the beer
won an award (mean of 0.06 and 0.18, respectively; ns).8
Multivariate Analyses
Next, we use ordinary least squares models to estimate the effects of producer
gender and status conferral on the beers’ overall evaluation (table 2). Models 1
and 2 test the main effects of the gender of the producer and status conferral,
respectively.In Model 1, the coefficient estimate for the woman-made dummy
variable is negative but not significant (βˆ = −0.39). This is not surprising,as
Model 1 does not take into account the effect of the status manipulation, which
we hypothesize willraise the evaluations of women produced beers more than
men produced beers. Model 2 evaluates the main effect of status. The coefficie
for the status dummy variable (βˆ = 0.23, p value <0.01) is significant and posi-
tive, suggesting that status conferral leads to higher evaluations of beers overa
No-status Status
Man-
made
Woman-
made
Man-
made
Woman-
made Min Max
Dependent variables
Overall evaluation
(normalized)
0.03• −0.26 0.06 0.18 −3.02 1.48
(0.63) (0.87) (0.72) (0.63)
Likeliness to buy 4.78+ 4.28 4.64 4.98 1 7
(1.36) (1.71) (1.38) (1.10)
Willingness to pay ($) 4.41• 4.30 4.59 4.83 0 11
(1.77) (1.89) (1.61) (2.08)
Taste 4.99+ 4.54 4.93 5.15 1 7
(1.14) (1.38) (1.19) (1.04)
Quality 5.19 4.90 5.34 5.38 2 7
(1.12) (1.11) (1.03) (0.93)
Moderator—product knowledge
Beer snob 0.23 0.29 0.16 0.23 0 1
Weekly consumption 2.36 2.15 1.80 2.21 0 7
(1.72) (1.65) (1.54) (1.78)
Self-reported expertize 2.78 2.68 2.37• 2.79 1 5
(1.15) (1.02) (1.04) (1.03)
SDs in parentheses; tests of the difference in means between men- and women-made within
and no-status conditions; + p < 0.10; • p < 0.05.
Table 1.Descriptive Statistics, Study 2 (N = 272)
Gender Inequality in Product Markets559 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Model 3 offers a direct test of Hypothesis 1.As predicted by Hypothesis 1,
the coefficient estimate for woman-made is negative and significant (βˆ = −0.29,
p value <0.05),suggesting thata woman’s beer performs significantly worse
than an identical beer made by a man.
Model 3 also allows us to investigate whether externalstatus conferralpro-
vides larger improvements in the evaluations of a woman-made beer than it
for a man-made beer. As discussed above, the attenuation principle leads u
predict that external status will have less of an effect on the rating of beer p
duced by men,since being produced by a man already confers positive status
Our results support this prediction. In Model 3, the coefficient estimate for s
tus is positive but insignificant and the coefficient estimate for the interactio
term (status× woman-made)is positive and significant(βˆ = 0.41; p value
<0.05).That is,winning an award raises the evaluation of women’s beers by
0.41 points on our composite measure. In contrast, for men’s beers, status c
ferral does not significantly improve beer ratings. The sizes of these coeffici
are noteworthy: the improvement in the overall evaluation from status conf
(βˆ = 0.41) completely overcomes the penalty arising from being produced by
woman brewer (βˆ = −0.29) to the point that their products are actually per-
ceived to be even better than those made by men whose beer has won an a
Consistent with Hypothesis 3,externalstatus conferralhas a greater effect on
women’s products, at least in male-typed domains.
Appendix B presents a replication of Model3 for each of the four variables
that comprisethe overall evaluationmeasureand find similar patterns.
However,the pattern is more pronounced in some measures than others.In
terms of willingness to pay,there is no a significant difference between men’s
and women’s beers. This makes sense since typically there is not a great de
variation in the price of beers, and prices are usually not correlated with qua
In contrast,thereis more gender-based and status-based differencein the
Beer snob
(1) (2) (3) (4) Yes (5) No
Status 0.23•• 0.03 0.05 0.04
(0.09) (0.12) (0.24) (0.14)
Woman-made −0.09 −0.29• −0.26 −0.32•
(0.09) (0.12) (0.21) (0.15)
Woman-made × status 0.41• 0.30 0.45•
(0.17) (0.32) (0.21)
Intercept 0.04 −0.12+ 0.03 0.15 −0.01
(0.06) (0.06) (0.09) (0.15) (0.10)
Degrees of freedom 270 270 268 58 206
Sample size 272 272 272 62 210
+ p < 0.10; • p < 0.05; •• p < 0.01; standard errors in parentheses.
Table 2.OLS Models of Overall Evaluation of Craft Beer (N = 272)
Social Forces 98(2)560
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

multifaceted and thatstatusbelieftransferis more pronounced along some
dimensions.
Panel A of figure 2 is a graphical representation of Model 3 from table 2. The
solid line represents the evaluations of man-made beers with and without statu
The relatively flat line shows that, consistent with the attenuation principle, eva
luations of beers produced by men change very little after having won an award
In contrast,the upward-sloping dashed line in PanelA shows that the evalua-
tions of beers made by women improve considerably from the status conferral
associated with winning an award.
Are Evaluators with Low Levels of Product Knowledge More
Discriminatory?
Finally, we evaluate whether those with lower levels of product knowledge are
more discriminatory.We collected three measuresof productknowledge—
whether the participant has been called a beer snob,self-reported expertize in
craft beer, and weekly consumption. In the interest of space, we only discuss th
results of the beer snob measure here but analyses of the weekly consumption
and expertize measures are included in appendix C, and these analyses yield si
ilar patterns.9
Whether the participant reports having been called a beer snob is a dichoto-
mous variable. Among those who have been called a beer snob (N = 62, Model
4, table 2), the coefficient estimates for status, woman-made, and woman-mad
status are all insignificant. That is, for these knowledgeable evaluators, the ove
evaluation is not affected by whether the beer is made by a man or a woman, n
by whether the beer has won an award.
However, those who have relatively less product knowledge and have not bee
called a beersnob evaluate productsdifferently depending on the genderof
brewer and external status (see Model 5 of table 2). There is no main effect of s
tus, a negative main effectof being made by a woman (βˆ = −0.32, p value
<0.05), and a positive interaction term (woman-made × statusβˆ = 0.45, p value
<0.05).This meansthat those who are notknowledgeable aboutbeersrate
woman-made beers significantly lower than they rate man-made beers. Howeve
if a woman-made beer has won an award, this raises their ratings to the point o
completely overcoming penalty associated with being a woman-made beer.
These results are consistent with the prediction that less-knowledgeable eva-
luators are more likely to discriminate against a producer whose gender is incon
sistent with the gender-typing of the market, as is the case for products made b
women in the craft beer market. However, we should be cautious in interpreting
these results. We conducted a power analyses and determined that we lacked t
sample size in the beer snob condition to make a strong claim about the non-
significantresults for those possessing productknowledge.Nonetheless,it is
encouraging that the analyses in appendixC, which presents our other two mea-
sures of product knowledge,consumption frequency,and self-reported expert-
ize, show similar patterns. Thus, we conclude that we find tentative support for
likeliness to buy and expected taste. These measures suggest that evaluations
Gender Inequality in Product Markets561 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Hypothesis 5: those evaluators with lower product knowledge tend to discrim
nate against products made by women in male-typed markets.
Study 2 Summary
Building on prior evidence that status beliefs disadvantage women in labor
kets, these results lend support to the idea that the devalued status beliefs
ated with women transfer to the output of their labor, that is the products th
they make.In an experiment simulating a male-typed market of craft beer,we
percent confidence intervals
A
B
Study 2 (Craft Beer, N = 272)
Study 3 (Cupcakes, N = 259)
Figure 2.Effect of status conferral and gender of producer on overall evaluation, mea
Social Forces 98(2)562
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

compared to the same products made by men. Moreover, external status confe
ral, in the form of an award,has a larger impact on raising the evaluations of
products produced by women than men. Finally, we find no significant effect of
the evaluator’s gender on the evaluation of craft beer (and for cupcakes in Stud
3). This is consistent with status characteristic theory, which demonstrates that
these are beliefs that are widely held at the societal level and therefore are pre
dicted to have similar effects for men and women (Correll and Ridgeway 2006).
Theoretically, status conferral reduces the uncertainty associated with evaluati
products produced by women in male-typed markets, thus bringing evaluations
of their products at parity with same products made by men.
Study 3. Female-Typed Product Market of Cupcakes
Procedure
Similar to Study 2, Study 3 utilized a between-subject design where 259 paid pa
pants were randomly assigned to one of the four conditions, which crossed prod
cer’s gender (man or woman) with external status conferral (with or without sta
Study 3 simulated a female-typed product market of cupcakes.We again manipu-
lated gender and status conferral by making specific changes to the informatio
vided on the product’s label. In this way, the design mirrored that of Study 2, ex
the participants viewed a cupcake label instead of a beer label. Again, they did
actually taste any cupcakes.We used the same manipulations and procedures as
Study 2. To encourage participants to pay attention to the information provided
the label, their screen was locked for forty-five seconds before proceeding to th
uation task. Participants of Study 2 were not allowed to participate in Study 3.
Study 3 evaluates two predictions. The first is that the magnitude of the gend
gap in product evaluations willbe smaller in a female-typed market than in a
male-typed market (Hypothesis 2). Second, we predict that the effect of status
ferral will not be gender differentiated in a female-typed market (Hypothesis 4)
Manipulation
Baker’s gender
In the condition where the baker was a woman, the baker was described as Sar
followed by a brief description of her career in baking. All pronouns referred to
as a she. In the condition where the baker was a man, the label was identical ex
that the baker was described as David and all pronouns referred to David as a h
External status conferral
In the externalstatus conferralcondition,the labeldescribed the cupcake as the
winner of the “Classic Cupcakes” episode of Cupcake Wars,a popular cooking
contest on the Food Network. In the no-status condition, there is no mention of
award. In order to keep the text length similar to the status condition, we includ
innocuous text about the bakery’s seasonal flavors in the no-status condition.
find that these status beliefs lead to lower ratings of products made by women
Gender Inequality in Product Markets563 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

We use identical questions and scales as Study 2 to construct the dependen
able, overall evaluation. The four items that comprise this measure are expe
taste (mean = 5.85, SD = 1.04), expected quality (mean = 5.81, SD = 1.03)
ness to buy (mean = 5.10, SD = 1.60) and willingness to pay (mean = 2.57,
1.24, minimum = $0.00, maximum = $8.00). As before, we scaled each of t
measures to a normal distribution, then took the average of the four standa
scores. This composite variable has a Cronbach’s alpha of 0.74. We only rep
analyses of the composite measure, overall evaluation, but the results are s
tively unchanged when the four measures are separately analyzed (append
Evaluator’s Product Knowledge
We asked six questions to measure evaluators’ objective and subjective kno
edge about cupcakes.To get at objective knowledge,we asked the following:
has anyone evercalled you a cupcake fanatic? (Yes/No;shown ascupcake
fanatic).Which of the following is a famous cupcake store known forits
Cupcake ATM? (choiceswere Magnolia Bakery,Sprinkles [correctanswer],
Georgetown Cupcake;shown as Cupcake ATM and coded as 1 if correct, 0 if
not). To get at subjective knowledge, we asked, in your opinion, how would
evaluateyour level of expertizein cupcakes(1 = No expertize, …, 5 =
Connoisseur; shown as self-reported expertize). Since frequency of consump
is correlated with knowledge, we also asked the following: In the past year,
many times did you bake a cupcake or a cake? (1 = 0 times, 2 = 1–2 times,
3–4 times, 4 = 5–10 times, 5 = more than 10 times; shown as baking freque
in table3). In the past year,how many times did you watch Cupcake Wars?
(same scale;shown as Cupcake Wars frequency).In the past year,how many
times did you buy a cupcake? (same scale; shown as buying frequency).
Descriptive Statistics
Table 3 presents summary statistics for our composite variable (overall eval
tion), the four variables that make up the composite variable (likeliness to b
willingness to pay,taste,quality) and the six measures of product knowledge.
We ran paired t tests that compare the evaluations of woman-made and ma
made cupcakes within the status conferraland within the no-status conferral
conditions. No significant differences emerge, except for one product knowl
measure, cupcake fanatic/expert. This table begins to show that when our s
manipulation is held constant(i.e., when we compare within award-winning
cupcakes or within non-award-winning cupcakes),the gender of the producer
does not meaningfully affect how the product is evaluated.This diverges from
the results of Study 2 where we find that for the craft beer without status co
ral, men’s craft beer receives more favorable evaluations. We now turn to m
variate models to formally test Hypotheses 2 and 4.
Dependent Variable
Social Forces 98(2)564
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Multivariate Analyses
We ran ordinary least squares regression models to understand how external st
tus conferral and producer’s gender affect overall evaluation. Model 1 of table 4
estimates the main effect for man-made,which is the producer category that is
inconsistent with the gender-typing of this product market. We find that the ma
made dummy variable is negative and insignificant, which is unsurprising becau
the model does not yet take into account the status manipulation. Model 2 test
for the main effect of status conferral and finds that the coefficient estimate for
tus is positive and significant (βˆ = 0.25), indicating that award-winning cupcakes
are evaluated more positively than cupcakes that did not win an award.10
Model 3 estimates the effects of status conferral,producer gender,and the
interaction term on overallevaluation.The main effect of status continues to
be positive and significant(βˆ = 0.27, p value <0.05),but the main effectof
No-status Status
Woman-
made
Man-
made
Woman-
made
Man-
made Min Max
Dependent variable
Overall evaluation (normalized)−0.11 −0.13 0.16 0.09 −1.99 2.28
(0.69) (0.76) (0.73) (0.79)
Likeliness to buy 5.06 4.93 5.33 5.11 1 7
(1.75) (1.73) (1.28) (1.58)
Willingness to pay ($) 2.48 2.34 2.90 2.67 0 8
(1.13) (1.15) (1.56) (1.38)
Taste 5.63 5.63 5.84 5.92 3 7
(0.98) (1.14) (0.92) (1.09)
Quality 5.59 5.70 6.00 5.97 2 7
(0.97) (1.09) (0.98) (1.05)
Moderator—product knowledge
Baking frequency 2.48 2.47 2.75 2.50 1 5
(1.20) (1.07) (1.46) (1.23)
Cupcake wars frequency 1.78 1.90 1.60 1.82 1 5
(1.06) (1.16) (1.06) (1.09)
Buying frequency 2.91 2.84 2.81 2.53 1 5
(1.26) (1.14) (1.20) (1.17)
Self-reported expertize 2.53 2.45 2.47 2.47 1 5
(0.94) (0.93) (1.07) (0.96)
Cupcake fanatic (Yes/No) 0.16• 0.03 0.10 0.11 0 1
Cupcake ATM (% correct) 0.66 0.60 0.60 0.60 0 1
SDs in parentheses; tested for difference in means between man and woman within sta
no status conditions; • p < 0.05.
Table 3.Descriptive Statistics, Study 3 (N = 259)
Gender Inequality in Product Markets565 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

man-made and the interaction term man-made × status are both insignifica
This means that cupcakes baked by a man, whose gender is inconsistent wi
gender-typing of the market,are not evaluated more negatively than cupcakes
baked by a woman. Moreover, the effect of status conferral on the evaluatio
of cupcakes does not differ between cupcakes produced by men and cupcak
produced by women. External status conferral elevates evaluations of cupca
but it does so equally for all products.
These results are presented graphically in figure 2, Panel B. The solid and
ted lines correspond to the evaluations ofman-made and woman-made cup-
cakes, respectively. In the no-status condition, the average evaluations of m
and women’scupcakesare nearly identicaland their confidenceintervals
completely overlap. In the status condition, the same pattern emerges. Whi
hypothesized (Hypothesis 2) that the gender gap in product evaluations wo
be smaller in female-typed markets, our results show that there is no signifi
gender gap.Previously we found a significant gap favoring men in male-typed
markets.We further hypothesized that the benefit from status conferralwould
not be gender differentiated in female-typed markets (Hypothesis 4). We fin
gender gap in status conferral benefit in the female-typed markets. Finally,
there were no differences between man-made and woman-made products i
setting, we omit the analysis of the moderating role of evaluator knowledge
Limitations and Future Research
While ours is the first study to develop and evaluate how gender status beli
transfer to products made by women and men, there are several limitations
should be addressed in future research. A potential limit to our design is tha
participants did not actually taste the products they evaluated (i.e., blind ta
(1) (2) (3)
Status 0.25•• 0.27•
(0.09) (0.24)
Man-made −0.04 −0.02
(0.09) (0.13)
Man-made × status −0.05
(0.18)
Intercept 0.02 −0.12+ −0.11
(0.07) (0.06) (0.09)
Degrees of freedom 257 257 255
Sample size 259 259 259
+ p < 0.10; • p < 0.05; •• p < 0.01; standard errors in parentheses.
Table 4.OLS Models of Overall Evaluation of Cupcakes (N = 259)
Social Forces 98(2)566
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

test; (Goldin and Rouse 2000) conduct a similar analysis that has been done in
the context of orchestras), making them less fit to measure dimensions like qua
ity or taste. However, our current design actually better mirrors the assessmen
process involved in buying many products.In the productscreening process,
prospective consumers mustmake an investmentdecision up-front,typically
without actually tasting or sampling the product. Therefore, asking the partici-
pants to evaluate the product based on the information on a label or packaging
is actually similar to the process by which consumers select products or service
on a daily basis. Moreover, if the objective quality of these products can also be
manipulated, it would be interesting to understand how perceived quality of the
Second, while our theory describes how status beliefs affect the evaluations o
products produced by women and men, we test our arguments in the context o
food products.Future research should establish whether the results hold for
other types of products to test the generalizability of status belief transfer as a
mechanism.Relatedly,we explored status belief transfer between the producer
and the product. Yet, it is also possible that status transfer may arise from othe
relevant actors in the market, such as investors, intermediaries, founders or oth
salient status-conferring individuals who are associated with a product.
Finally, our studies show that evaluators are less likely to purchase women’s
beers but how much they are willing to pay is not significantly different by the
gender of the brewer. This may be due to the fact that these are markets where
pricesdo not vary considerably by the (perceived)quality of the product.
However,we may find variations in people’s willingness to pay by producer’s
gender in product markets where there are larger variations in price by the stat
of the product (e.g., handbags, cosmetic products). How and when status belief
transfer occurs from these actors are interesting directions for future work.
Summary and Implications
This paper develops and evaluates a theory ofstatus belieftransfer,showing
how gender status beliefs also affect the evaluations of products made by men
and women.In particular,we develop an integrated theory for why there is a
spread of status beliefs from producers to their products by specifying how dif-
fuse status characteristics associated with individuals combine with the specific
statuscharacteristicsassociated with the gender-typing ofproductmarkets.
While the idea that status beliefs may spread to “non-status elements” was an
early assumption of status characteristic theory,the assumption has not been
explicitly theorized or empirically validated. We develop such a theory here. Th
results of three studies confirm many of our predictions.
Our argument predicts that gender-typing of product markets will affect the
direction and magnitude of the status transfer process. For this to be correct, w
needed to establish that products are gender-typed (i.e.,some product markets
are culturally associated with men and/or masculinity and some with women
and/or femininity). Study 1 collect data that provided evidence that product ma
kets are,in fact, gender-typed,albeitto varying degrees.We gather product
product interacts with these status signals.
Gender Inequality in Product Markets567 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

each product with masculinity or femininity.We find that many products are
indeed gendered. Studies 2 and 3 provide key test for our arguments aroun
tus belief transfer. Collectively, they provide evidence of an asymmetric neg
bias toward women’s products in male-typed markets. In a male-typed prod
market of craft beer (Study 2),a beer described as produced by a woman re-
ceives a lower evaluation than the same beer described as produced by a m
However,if these beers are conferred status by winning an award,the gap is
eliminated. This is because lower status individuals have more to gain from
tus conferral.Thus,the evaluations of woman-made beers improve more than
the evaluations of man-made beers do. Moreover, we find some support for
claim thatevaluators who lack knowledge abouta productare more heavily
influenced by gender status beliefs. The implication is that the effects of ge
status beliefs on gender inequality are greater than previously thought,leading
to lower evaluations of women in labor markets and to lower evaluations of
men’s products in product markets.
In contrast, in the female-typed product market of cupcakes (Study 3), th
cakes made by men and women producers receive similar evaluations. Furt
tus conferral in the form of winning an award raises the evaluations of cupc
produced by both a man and a woman and the magnitude of the improvem
undifferentiated between the two groups.To explain these differences across the
two markets, we rely on SCT’s attenuation and inconsistency principles, and
that in a female-typed product market,specific and diffuse status beliefs lead to
inconsistent status information for both the male and female producers.Women
benefit from the specific status belief that women are better at making fem
products and men benefit from the diffuse belief that men are generally mo
petent than women.When neither group has a clear status advantage,the status
value of the award is either completely inconsistent or redundant with prior
information,suppressing both inconsistency and attenuation effects.Therefore,
there are undifferentiated benefits from winning an award to men’s and wom
cupcakes. Theoretically, we illustrate how the inconsistency and attenuation
ples of status characteristics theory operate in product markets.
While we test for status belief transfer in product markets,our theory may
also generalize to the assessments of gender-typed objects.Some recent papers
provide evidence thatis consistentwith this notion. For instance,Thébaud
(2015) finds that in entrepreneurship, where founders are predominantly m
women’s pitches are evaluated more negatively than the identical pitch ma
a man.Sarsons (2017) finds that in top economics departments,where a vast
majority of the faculty are male, women’s coauthored papers get less recog
in tenure decisions than men’s coauthored papers do.
While we generally find that women are disadvantaged in male-typed ma
our results also show that women in these domains might benefit uniquely f
awards, certification, and other forms of external status conferral. If so, this
a route to reduce gender gaps in perceptions of product quality among the
public who makes purchasing decisions.This suggests a practicalimplication of
our findings:award-granting organizations and committees (such as professio
categories from a large online retailer and test for the culturalassociations of
Social Forces 98(2)568
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

biases from their award selection process,so that women receive a fair share of
awards. By conferring status to women via an award, organizations can block th
tendency of the larger public to rate women’s products lower than men’s in ma
typed markets. For this to happen, however, award-granting organizations mus
motivated to remove biases from their award selection process,since the award
selection process itself is susceptible to the very same status belief transfer me
nism we test in this paper. There is some evidence that providing evidence that
award process is unfair to women can provide this motivation. For example, wh
the Association for Women in Science (AWIS) found that women scientists were
disproportionately underrepresented among winners of major awards in thirtee
scientific societies, they successfully leveraged this finding to get the scientific
ties to change their reward processes to be more gender fair (Lincoln et al. 2012).
Encouragingly, we also find some evidence to suggest that those with greate
product knowledge were not biased against women’s products in our study—
those who lacked product knowledge exhibit significant biases against women.
Past sociologicalresearch has focused on how status beliefs affect evaluators’
quality judgmentsand decision-making underconditionsof uncertainty.To
date,this uncertainty has been described as resulting from ambiguities in the
product being evaluated (Podolny 1993), or in the process of evaluation,or in
the criteria used in decision-making(Correll 2017; Correll et al. 2014;
Ridgeway and Correll2004).Our findings show thatuncertainty stemming
from lack of domain-specific knowledge on the part of decision-makers can also
induce decision-makers to rely on status beliefs when making decisions.
To the extent that this argument gains further support in future work, there
are two practicalimplications for this finding.First,for women’s products (or
products made by other low-status groups) to receive a fair evaluation,expert
evaluators are needed. Within organizations, personnel decisions and review of
deliverables should be done by people with experience and knowledge on the
topic.Second,since decision-makers who lack knowledge about products are
especially prone to biases, organizations should invest in raising knowledge and
familiarity of the product.Not only would this mitigate bias against women’s
products, it would also allow for more accurate assessments of product quality
that are not tainted by non-meritocratic factors, like the gender of the producer
These results show that status belief transfer can disadvantage women’s pro
ducts, but it also suggests that these beliefs are malleable. Credible signs of sta
tus,such as awards,lead to positive improvements for both men and women.
Moreover, for a subset of people who care and know about the product, these
status signals are less important to how they assess and evaluate the product.
While showing how status beliefs differently influence our assessmentof pro-
ducts made by women and men, this paper also proposes mechanisms that can
lead to more fair and accurate assessments.To that end,we hope to see more
research on mechanisms thatcan remedy biases and stereotypes to putlow-
status actors on an equitable playing field with higher status actors.
societies or university award committees) should actively work to remove gend
Gender Inequality in Product Markets569 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Notes
1. Research in socialpsychology associated with the Goldberg paradigm (Goldberg
1968) has provided preliminary evidence in support of this prediction, showing t
rather than a man (Olian, Schwab and Haberfeld 1988; Top 1991). However, the
experiments do not develop a theoretical argument about why and when wome
likely to experience an evaluative penalty in product markets
2. One reviewer asked whether this prediction is similar to the well-established fin
in labor markets that female-typed jobs (such as nursing) are themselves devalu
such that both women and men who work in these jobs earn lower wages than i
they had worked in an otherwise similar male-typed job. If so, this would lead us
predict that products that are female-typed are devalued relative to otherwise s
products that are male-typed for both male and female producers. This is an int
ing idea but is not the focus of the current study. Instead, we expect that the ge
of producer affects product evaluation in gender-typed markets,such that the pro-
ducts of gender atypicalproducers (e.g.,women in male-typed markets) are evalu-
ated more negatively than gender typical producers.
3. For the studies, we recruited participants through MTurk, a crowdsourcing platfo
that allows work requesters to hire workers to complete short, paid jobs. Resear
has found thatthese workersare more demographically diverse than American
undergraduate volunteers,who are often employed for experimentalresearch,and
that the data obtained are as reliable as those obtained by traditional experime
methods (Buhrmester,Kwang, and Gosling 2011).MTurk was deliberate choice
over other survey-administering platforms for severalreasons.First,we wanted to
investigate status beliefs,which is a culturally shared beliefamong members ofa
society. Thus, a diverse audience was useful. Second, the tasks and questions d
require any particular skillor expertize in a given domain.Importantly,they rated
products thatwere familiar and highly accessible,which gives us confidence that
these participants’ answers will mirror those of actual consumers.
4. Water,honey,baking mix,salad dressing,paper towel,facialtissue,cleaning pro-
ducts (detergent, soap, etc.), food storage products (bags, foil, etc.), over-the-co
allergy medicine, body moisturizer, tweezers, potties, potty-training pants for ba
baby stroller, baby proofing gates and rails, cat treat, futon, living room chair, h
tool, wrench,table lamp,baking sheet,shower curtain,shower caddy,comforter,
duvet cover, shovel, lawn mower, computer monitor, printer ink and toner, scan
tabletcomputer,air conditioner,humidifier,vacuum,women’s shoes –heels,wo-
men’s sunglasses,men’s activewear,earring,ring, camping equipment,basketball
equipment,Star Wars merchandise,doll, play vehicle (toy),tricycle and scooter
(toy), comic book, science fiction book, car roof rack, auto parts.
5. The products are dinner rolls, soup, spinach, coffee, craft beer, chicken wings, b
diet cola, cupcakes, and macarons.
6. While craft beer and gourmet cupcakes first appeared in trendy urban areas like
Francisco and New York, both industries have experienced considerable growth
are now available across the United States, with one author noting that cupcake
are taking the place of ice cream stores (Olson 2009) and another describing the grow-
ing popularity ofcraftbeer in ruralareas across the United States (Siegler 2017).
While these products may be more heavily consumed by younger,urban,college-
educated people, there are several advantages of these products for our purpos
the demographics of MTurk (from which we draw our sample) are similar to thos
the very same poem or essay is rated lower if it is believed to be written by a w
Social Forces 98(2)570
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

cal argument. Second, even if not all Americans purchase craft beer and gourmet cu
cakes at equalrates, most people are nonetheless aware of cupcakes and beer, more
generally, and are also likely to have been exposed to the craft version of these pro
ducts via TV shows like Cupcake Wars and House Hunters (which commonly shows
diverse couples from across the country discussing housing options while sampling
craft beer) and by the increasing availability of such products in grocery stores. And
finally, there is no reason to believe our argument would only apply to upper class p
ducts.Indeed, we can imagine even larger gender biases for products like baby pro-
ducts and power tools.While our goalwas to test that our theory holds under the
conditions we lay out, future research should test the argument with a wider array o
consumer products to establish the boundary conditions of the theory.
7. For both Studies 2 and 3, we ran attention checks to ensure that the participants
carefully read the beer and cupcake labels.In Study 2, we asked two multiple
choice questions—one aboutthe city thatthe brewery was located in and one
about the graphic that was on the beer label.Ninety percent answered the first
question correctly and 99 percentanswered the second question correctly.In
Study 3,we again asked two multiple choice questions—one about the city that
the bakery was located in and one about the type of frosting that was on the cup-
cake,and 94 percentof participants passed both attention checks.We did not
specifically ask about brewer/baker gender as we did not want to raise suspicion.
Instead,we assumed that if they recalled other details of the labelthey also at-
tended to gender,given the prominence of the brewer/baker name on the label.
For both studies,restricting the analyses to only include those who passed both
tests did not substantively change the results.
8. Table1 summarizes how participants’ product knowledge varied across the four cond
tions as measured by whether they have been called a beer snob, weekly consumpt
and self-reported expertize. Interestingly, participants that are randomly assigned t
award-winning woman-made beers condition report that they consume more beer, h
higher level of expertize about craft beer, and more of them have been called a bee
than those assigned to award-winning man-made beers.As one reviewer suggested,it
could be that evaluators see themselves as experts when evaluating a lower status
(a woman-made beer) than a higher status project (a man-made beer). However, th
ference is significant only for one of the three measures, self-reported expertize.
9. We considered creating a single measure ofproductknowledge by collapsing the
three variables but decided to keep the three measures as separate because the th
measures’ Cronbach’s alpha was slightly low having a value of 0.62. However, the
slightly low Cronbach’s alpha should not be a source of alarm. Indeed, the consume
marketing literature has reported substantialvariations across studies in regard to
the relationship among subjective and objective knowledge (Carlson et al. 2009). On
the higher end of the correlation was 0.54 (Brucks 1985),whereas others reported
non-significant relationship between objective and subjective knowledge (r = 0.08).
This suggests that the three measures are allmeasures of product knowledge,but
they are picking up differentaspects ofproductknowledge,such thatit is more
appropriate to keep them as separate measures.
10. The magnitude of the coefficient for the status conferral variable (0.25) is very simi-
lar to that found in Study 2 for beer (0.23), indicating that status conferral had simi-
lar effects in raising the evaluations of beer and cupcakes.
mostly like to consume these products, so they are well suited for testing our theore
Gender Inequality in Product Markets571 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

All
Woman-made
× no status
Woman-made
× status
Man-made
× no status
Man-made
× status
Gender (% women)
Study 1 51% N/A N/A N/A N/A
Study 2 42% 35% 45% 45% 44%
Study 3 42% 42% 45% 39% 42%
Education
High school or less
Study 1 21% N/A N/A N/A N/A
Study 2 16% 19% 17% 20% 7%
Study 3 18% 19% 16% 22% 14%
Some undergraduate
Study 1 30% N/A N/A N/A N/A
Study 2 29% 23% 26% 33% 32%
Study 3 28% 25% 26% 31% 30%
Undergraduate degree
Study 1 38% N/A N/A N/A N/A
Study 2 42% 42% 42% 33% 51%
Study 3 42% 42% 48% 37% 42%
Master’s or higher
Study 1 11% N/A N/A N/A N/A
Study 2 13% 16% 15% 13% 9%
Study 3 12% 14% 10% 9% 14%
Income
≤$50K
Study 1 62% N/A N/A N/A N/A
Study 2 71% 76% 65% 71% 71%
Study 3 71% 77% 65% 76% 65%
$50,001–$100K
Study 1 34% N/A N/A N/A N/A
Study 2 26% 24% 27% 29% 25%
Study 3 27% 20% 31% 22% 33%
(Continued)
Appendix A.Gender, Education and Income Distribution Across Studies and Condition
Social Forces 98(2)572
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

All
Woman-made
× no status
Woman-made
× status
Man-made
× no status
Man-made
× status
$100,001–$150K
Study 1 3% N/A N/A N/A N/A
Study 2 2% 0% 5% 0% 3%
Study 3 2% 2% 3% 1% 2%
>$150K
Study 1 2% N/A N/A N/A N/A
Study 2 1% 0% 3% 0% 1%
Study 3 1% 2% 2% 0% 0%
Appendix B.OLS Models of Four Evaluation Measures (Craft Beer)
Likeliness
to buy
Willingness
to pay ($) Taste Quality
Status −0.14 0.18 −0.06 0.15
(0.24) (0.31) (0.20) (0.18)
Woman-made −0.51• −0.10 −0.45• −0.29
(0.24) (0.32) (0.20) (0.18)
Status × woman-
made
0.85• 0.35 0.67• 0.33
(0.34) (0.45) (0.29) (0.25)
Intercept 4.78• 4.41• 4.99 5.19•
(0.17) (0.22) (0.14) (0.13)
Degrees of freedom 268 268 268 268
Sample size 272 272 272 272
+ p < 0.10; • p < 0.05; standard errors in parentheses.
Appendix A.continued
Gender Inequality in Product Markets573 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Appendix C.OLS Models of Overall Evaluation of Craft Beer: Moderating Role of Evalua
Weekly consumption Self-reported expertize
High Low High Low
Status 0.24 0.01 0.14 0.05
(0.19) (0.93) (0.56) (0.76)
Woman-made −0.20 −0.30+ −0.38+ −0.25+
(0.25) (0.16) (0.08) (0.10)
Status × woman-made 0.19 0.43+ 0.31 0.39+
(0.46) (0.06) (0.34) (0.21)
Intercept 0.14 −0.05 0.16 −0.03
(0.20) (0.66) (0.23) (0.80)
Degrees of freedom 84 180 61 203
Sample size 88 184 65 207
+ p < 0.10; • p < 0.05; standard errors in parentheses.
Evaluators in the “high” weekly consumption condition are those who reported ththey
drink beer 3 or more times a week. They make up 32 percent of the response. Evalu
“high” self-reported expertize condition are those who said 4 or more on the questio
had a scale of 1 through 7. They made up 24 percent of the response.
Appendix D.OLS Models of Four Evaluation Measures (Cupcakes)
Likeliness to buy Willingness to pay ($) Taste Quality
Status 0.26 0.42+ 0.23 0.41•
(0.28) (0.23) (0.18) (0.18)
Man-made −0.14 −0.14 0.00 0.11
(0.24) (0.23) (0.18) (0.18)
Status × man-made −0.14 −0.09 0.07 −0.14
(0.28) (0.33) (0.25) (0.25)
Intercept 5.06• 2.48• 5.63 5.59•
(0.20) (0.16) (0.13) (0.13)
Degrees of freedom 255 255 255 255
Sample size 259 259 259 259
+ p < 0.10; • p < 0.05; standard errors in parentheses.
Product Knowledge Measured as Weekly Beer Consumption and Self-Reported Expe
Social Forces 98(2)574
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

Elise Tak is a researcherat Airbnb. Her PhD dissertation explored gender
inequality and homophily in labor markets. She received her PhD from Stanford
University in 2018.
Shelley J. Correll is the Michelle Mercerand Bruce Golden Professorof
Women’s Leadership at Stanford University where she also serves as the Barba
D. Finberg Directorof the Clayman Institute forGenderResearch and the
founder and director of the VMware Women’s Leadership Innovation Lab. Her
current research evaluates how organizationalpolices and practices can effec-
tively reduce gender inequalities in modern workplaces.
Sarah A. Soule is the Morgridge Professor ofOrganizationalBehavior at
Stanford University, Graduate School of Business. Her research examines state-
and organizational-level policy change and diffusion, and the role social move-
ments have on these processes.Recentpublished work has appeared in the
American Journal of Sociology, Administrative Science Quarterly, the American
Sociological Review, Organizational Studies, the Strategic Management Journal,
and the Annual Review of Sociology.
Supplementary Material
Supplementary material is available at Social Forces online.
References
Alba,Joseph W.,and J.Wesley Hutchinson.1987.“Dimensions ofConsumerExpertise.” Journalof
Consumer Research 13(4):411–54.
Berger,Joseph,Bernard P.Cohen,and Morris Zelditch Jr.1972.“Status Characteristics and Social,
Interaction.” American Sociological Review 37(3):241–255.
Bettman,James R.,and C.Whan Park.1980.“Effects of Prior Knowledge and Experience and Phase
the Choice Process on ConsumerDecision Processes:A ProtocolAnalysis.” Journalof Consumer
Research 7(3):234–48.
Brucks, Merrie. 1985. “The Effects of Product Class Knowledge on Information Search Be
of Consumer Research 12(1):1–16.
Buhrmester,Michael,Tracy Kwang,and SamuelD.Gosling.2011.“Amazon’s MechanicalTurk:A New
Source of Inexpensive, Yet High-Quality Data?” Perspectives on Psychological Science
Bunderson, J. Stuart. 2003. “Recognizing and Utilizing Expertise in Work Groups: A Statu
Perspective.” Administrative Science Quarterly 48(4):557–591.
Correll,Shelley J.,Stephen Benard,and In Paik.2007.“Getting a Job:Is there a Motherhood Penalty?”
American Journal of Sociology 112(5):1297–133.
Carlson,Jay P.,Leslie H.Vincent,David M.Hardesty,and William O.Bearden.2009.“Objective and
Subjective Knowledge Relationships:A Quantitative Analysisof ConsumerResearch Findings.”
Journal of Consumer Research 35(5):864–76.
Castilla,EmilioJ., andStephenBenard.2010.“TheParadoxof Meritocracyin Organizations.”
Administrative Science Quarterly 55(4):543–76.
About the Authors
Gender Inequality in Product Markets575 Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

“Task Segregation as a Mechanism for Within-job Ine
Women and Men ofthe Transportation Security Administration.” Administrative Scien
https://doi.org/10.1177/0001839215611447.
Correll, Shelley J. 2017. “SWS 2016 Feminist Lecture: Reducing Gender Biases In Mo
Small Wins Approach to Organizational Change.” Gender & Society 31(6):725–750
Correll,Shelley J.,Erin L.Kelly,Lindsey Timble O’Connor,and Joan C.Williams.2014.“Redesigning,
Redefining Work.” Work and Occupations 41(1):3–17.
Correll,Shelley J.,and Cecilia L.Ridgeway.2006.“Expectation States Theory.” In Handbook ofSocial
Psychology. New York, NY: Springer USA.
DiMaggio, Paul. 1987. “Classification in Art.” American Sociological Review, 440–455
Foschi,Martha.1996.“Double Standards in the Evaluation ofMen and Women.” SocialPsychology
Quarterly 59(3):237–54.
———. 2000.“Double Standards for Competence:Theory and Research.” AnnualReview of Sociology
26:21–42.
Goldberg, Philip. 1968. “Are Women Prejudiced Against Women?” Society 5(5):28–3
Goldin,Claudia,and Cecilia Rouse.2000.“Orchestrating Impartiality:The Impact of ‘Blind’Auditions on
Female Musicians.” American Economic Review 90(4):715–41.
Gould, Roger V. 2002. “The Origins of Status Hierarchies: A Formal Theory and Emp
Journal of Sociology 107(5):1143–1178.
Kovács,Balázs,and Amanda J.Sharkey.2014.“The Paradox ofPublicity How Awards Can Negativel
Affect the Evaluation of Quality.” Administrative Science Quarterly 59(1):1–33.
Kunda,Ziva,and Steven J.Spencer.2003.“When Do Stereotypes Come to Mind and When Do
ColorJudgment? A Goal-Based TheoreticalFramework forStereotype Activation and Application
Psychological Bulletin 129(4):522–44.
Lincoln,Anne E.,Stepahnie Pincus,JanetBandows Koster,and Phoebe S.Leboy.2012.“The Matilda
Effect in Science:Awards and Prizes in the US,1990s and 2000s.” SocialStudies of Science 42(2):
307–320.
Mitchell,Andrew A.,and Peter A.Dacin.1996.“The Assessment of Alternative Measures of Con
Expertise.” Journal of Consumer Research 23(3):219–39.
Olian,Judy D.,Donald P.Schwab,and Yitchak Haberfeld.1988.“The Impactof ApplicantGender
Compared to Qualifications on Hiring Recommendations:A Meta-Analysis of ExperimentalStudies.”
Organizational Behavior and Human Decision Processes 41(2):180–95.
Olson, Elizabeth. 2009. “The Latest Entrepreneurship Fantasy is Selling Cupcakes.”
Park,C.Whan, and V.Parker Lessig.1981.“Familiarity and Its Impact on Consumer Decision Bia
Heuristics.” Journal of Consumer research 8(2):223–31.
Peracchio,Laura A.,and Alice M.Tybout.1996.“The Moderating Role of Prior Knowledge in Sch
Based Product Evaluation.” Journal of Consumer Research 23(3):177–92.
Podolny,JoelM. 1993.“A Status-Based Modelof Market Competition.” American Journalof Sociology
98(4):829–72.
Ridgeway,Cecilia L.2011.Framed by Gender:How GenderInequality Persists in the Modern World
Oxford, UK: Oxford University Press.
Ridgeway,Cecilia L.,and ShelleyJ. Correll.2004.“Unpacking the GenderSystem a Theoretical
Perspective on Gender Beliefs and Social Relations.” Gender & Society 18(4):510–
Salganik, Matthew J., Peter S. Dodds, and Duncan J Watts. 2006. “Experimental Stud
Unpredictability in an Artificial Cultural Market.” Science 311(5762):854–856.
Sarsons,Heather.2017.“Recognition forGroup Work:GenderDifferences in Academia.” American
Economic Review 107(5):141–45.
Chan,Curtis K.,and MichaelAnteby.2015.
Social Forces 98(2)576
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020

“Tapping RuralAmerica:Craft Brewers Pour New Life into SmallTowns.” National
Public Radio.
Simcoe,Timothy S.,and Dave M.Waguespack.2011.“Status,Quality,and Attention:What’s in a
(Missing) Name?” Management Science 57(2):274–290.
Steinpreis,Rhea E.,Katie A.Anders,and Dawn Ritzke.1999.“The Impact of Gender on the Review of
the Curricula Vitae of Job Applicants and Tenure Candidates:A NationalEmpirical Study.” Sex Roles
41(7):509–528.
Sujan,Mita.1985.“ConsumerKnowledge:Effectson EvaluationStrategiesMediatingConsumer
Judgments.” Journal of Consumer Research 12(1):31–46.
Taylor,Catherine J.2010.“OccupationalSex Composition and the Gendered Availability ofWorkplace
Support.” Gender & Society 24(2):89–212.
Thébaud,Sarah.2015.“Status Beliefs and the Spiritof Capitalism:Accounting forGenderBiases in
Entrepreneurship and Innovation.” Social Forces 94(1):61–86.
Thornton,Patricia H.1999.“The Sociology ofEntrepreneurship.” AnnualReview ofSociology 25(1):
19–46.
Tinkler,Justine E., Kjersten Bunker Whittington,Manwai C.Ku, and Andrea Rees Davies.2015. “Gender
and Venture CapitalDecision-Making:The Effects ofTechnicalBackground and SocialCapitalon
Entrepreneurial Evaluations.” Social Science Research 51:1–16.
Top,Titia J.1991.“Sex Bias in the Evaluation ofPerformance in the Scientific,Artistic,and Literary
Professions: A Review.” Sex Roles 24(1–2):73–106.
Turco,Catherine J.2010.“CulturalFoundations ofTokenism Evidence From the Leveraged Buyout
Industry.” American Sociological Review 75(6):894–913.
Williams,ChristineL. 1992.“TheGlassEscalator:HiddenAdvantagesfor Men in the‘Female’
Professions.”Social Problems 39(3):253–267.
Siegler,Kirk.2017.
Gender Inequality in Product Markets5 7 7
Downloaded from https://academic.oup.com/sf/article-abstract/98/2/548/5298604 by guest on 11 April 2020