Women and Leadership Project

§ I C H I C A G O J O U R N A L S G e t t i n g a J o b : Is T h e r e a M o th e r h o o d P e n a lty ? A u th o r(s ): S h e lle y T. C o r r e ll, S te p h e n B e n a rd , a n d In P aik Source: A m e r ic a n J o u r n a l o f S o cio lo g y, Vol. 112, No. 5 (M arch 2007), pp. 1297-1339 P u b lis h e d b y : The U n iv ersity o f Chicago Press S tab le URL: http://www.jstor.org/stable/10.1086/511799 A ccessed: 0 1 /0 8 /2 0 1 4 16:08

Your use of the JSTOR archive indicates your acceptance of the Tenns & Conditions of Use, available at http://www.jstor.org/page/info/about/policies/tenns.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use infonnation technology and tools to increase productivity and facilitate new fonns of scholarship. For more infonnation about JSTOR, please contact [email protected].

I I The University o f Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to American Journal o f Sociology. http://www.jstor.org

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTOR Terms and Conditions

G e t t i n g a J o b : I s T h e r e a M o t h e r h o o d P e n a l t y ? 1

Shelley J. C orrell, S tep h en B e n a rd , a n d In P a ik Cornell U n iversity

S u rv ey research finds th a t m o th ers suffer a s u b sta n tia l w age penalty, a lth o u g h the cau sal m e c h a n ism p ro d u c in g it re m a in s elusive. T h e a u th o rs em ployed a la b o ra to ry e x p erim en t to e v alu ate the h y p o th ­ esis th a t sta tu s-b a se d d iscrim in atio n p lays an im p o rta n t role a n d an a u d it s tu d y of a c tu a l em ployers to assess its real-w o rld im plications. In b o th studies, p a rtic ip a n ts e v a lu a te d ap p lic a tio n m a te ria ls for a p a ir of sam e-g en d er eq u ally q ualified jo b c a n d id a te s w ho differed on p a re n ta l statu s. T h e la b o ra to ry ex p e rim e n t fo u n d th a t m o th ers w ere p e n a liz e d on a h o st of m easu res, in clu d in g p erceiv ed com pe­ tence a n d reco m m en d ed sta rtin g salary. M en w ere n o t p e n alized for, a n d som etim es benefited from , bein g a p a re n t. T h e a u d it stu d y show ed th a t a c tu a l em ployers d iscrim in ate a g a in st m o th ers, b u t n o t a g a in st fathers.

M o th e rs experience d isa d v a n ta g e s in the w o rk p lace in a d d itio n to those com m only asso ciated w ith gender. F o r exam ple, tw o re c e n t stu d ies find t h a t em ployed m o th e rs in the U n ite d S tates suffer a per-child w age p e n a lty of a p p ro x im a te ly 5% , on av erag e, after co n tro llin g for the u su al h u m a n ca p ita l a n d o ccu p atio n al facto rs th a t affect w ag es (B udig a n d E n g la n d 2001; A n d erso n , B inder, a n d K rau se 2003). In a s u m m a ry of econom ic research , C ritte n d e n (2001) concludes th a t, for those u n d e r the age of 35, the p a y gap b etw een m o th e rs a n d n o n m o th e rs is larg er th a n the p a y gap b etw een m en a n d w om en. As G lass (2004) notes, em ployed m o th ers are

1 W e th a n k P i-C h u n H s u , D e v a h P ager, C e cilia R idgew ay, C a te T aylor, L is a Troyer, K im W eeden, a n d R o b b W ilier fo r h e lp fu l c o m m e n ts a n d sug g e stio n s, a n d M a y s h a A rtis, M o n ic a C eledón, H e a th e r F e rg u so n , A d rie n n e G a lle t, K im G illece, K a th ry n H eley, S h a ri M oseley, S h a n a P la tz , C o n n o r P uleo, K ristin Seeger, a n d M ic h a e l Stein fo r c a p a b le re se a rc h a ssista n ce . S u p p o rt for th is re se a rc h w a s p ro v id e d b y a g ra n t to th e first a u th o r fro m th e A lfred P. S loan F o u n d a tio n . D ire c t c o rre sp o n d e n c e to Shelley J. C o rre ll, D e p a rtm e n t of Sociology, 323 U ris H a ll, C o rn e ll U n iv e rsity , Ith a c a , N e w Y o rk 14853. E -m ail: sjc62@ cornell.edu

© 2007 b y T h e U n iv e rs ity of C hicago. All rig h ts re serv e d . 0 0 02-9602/2007/11205-0001$10.00

A J S V olum e 112 N u m b e r 5 (M arch 2007): 1297-1338 1297

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

the gro u p of w o m en th a t n o w ac c o u n t for m o st of the “g en d er g a p ” in w ages.

T h e d isa d v a n ta g e s are n o t lim ited to pay. C uddy, F iske, a n d C lick (2004) show th a t d escrib in g a c o n s u lta n t as a m o th e r leads e v a lu a to rs to ra te h e r as less co m p e te n t th a n w h en she is d escrib ed as n o t h a v in g children. Sim ilarly, o th er stu d ies show th a t visib ly p re g n a n t w o m en m a n ­ agers are ju d g e d as less c o m m itted to th e ir jo b s, less d ep en d ab le, a n d less a u th o rita tiv e , b u t w arm er, m ore em otional, a n d m ore irra tio n a l th a n o th ­ erw ise eq u al w om en m a n a g e rs w h o are n o t v isib ly p re g n a n t (H alp ert, W ilson, a n d H ic k m a n 1993; C orse 1990). W hile the p a tte rn is clear, the u n d e rly in g m e c h a n ism re m a in s opaque. W h y w o u ld bein g a p a r e n t lead to d isa d v a n ta g e s in the w o rk p lace for w om en? A n d w h y m ig h t sim ilar d isa d v a n ta g e s n o t occur for m en?

T h is article p re se n ts a la b o ra to ry e x p erim en t a n d an a u d it stu d y of ac tu a l em ployers. T h e la b o ra to ry ex p erim en t ev a lu a te s the hy p o th esis th a t the “m o th e rh o o d p e n a lty ” on w ag es a n d ev a lu a tio n s of w o rk p lace p e r­ fo rm an ce a n d s u ita b ility occurs, a t le a st p artially, because c u ltu ra l u n ­ d e rsta n d in g s of the m o th e rh o o d role exist in tension w ith the c u ltu ra l u n d e rsta n d in g s of the “id eal w o rk e r” role. We propose th a t this perceiv ed tension b etw een in co m p atib le c u ltu ra l u n d e rsta n d in g s or schem as leads ev alu ato rs, p e rh a p s unconsciously, to expect m o th e rs to be less c o m p eten t a n d less c o m m itted to th e ir jo b s (B lair-L oy 2003; R id g ew ay a n d C orrell 2004). To the e x te n t th a t m o th e rs are believ ed to be less c o m m itted to the w o rk p lace, we arg u e th a t em ployers w ill su b tly d iscrim in ate a g a in st m o th e rs w h e n m a k in g ev a lu a tio n s th a t affect hirin g , p ro m o tio n , a n d sal­ a ry decisions. We do n o t expect th a t fa th e rs w ill experience these types of w o rk p lace d isa d v a n ta g e s since u n d e rsta n d in g s of w h a t it m ean s to be a good fa th e r are n o t seen in o u r cu ltu re as in co m p atib le w ith u n d e r­ sta n d in g s of w h a t it m ean s to be a good w o rk e r (T ow nsend 2002). By h a v in g p a rtic ip a n ts ra te jo b ap p lic a n ts, we expect th a t a p p lic a n ts p re ­ sen ted as w o m en w ith ch ild ren w ill be v iew ed as less co m p e te n t a n d less co m m itted to w o rk , w ill n eed to p re se n t evidence th a t th ey are m ore qualified for the jo b , w ill be ra te d as less p ro m o ta b le , a n d w ill be offered low er sta rtin g salaries co m p a re d w ith otherw ise sim ilar a p p lic a n ts p re ­ sen ted as w o m en w ith o u t children. We also expect th a t the m o th e rh o o d p e n a lty w ill be m e d ia te d b y ev a lu a tio n s of com petence a n d co m m itm en t. W hile the la b o ra to ry ex p e rim e n t allow s us to isolate a n d exam ine the m e c h a n ism of d iscrim in atio n , the a u d it s tu d y p ro v id e s ex tern al v a lid ity b y e v a lu a tin g w h e th e r a c tu a l em ployers d iscrim in ate a g a in st m others.

In the follow ing p a ra g ra p h s we rev iew the em p irical lite ra tu re on the m o th e rh o o d w age p e n a lty a n d existing e x p lan atio n s for it. We th en d e­ velop o u r th eo retical a rg u m e n t b y d ra w in g on s ta tu s ch ara c teristic s th eo ry a n d the lite ra tu re on c u ltu ra l concep tio n s of m o th erh o o d . Finally, we

1298

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

describe the la b o ra to ry ex p e rim e n t a n d the a u d it study, a n d sum m arize evidence for the m o th e rh o o d p e n a lty fro m these tw o studies.

WAGE PENALTY FOR M OTHERHOOD

A v a rie ty of facto rs h av e been p ro p o sed as ex p lan atio n s for the m o th ­ erh o o d w age gap, in clu d in g re d u c e d in v e stm e n t in h u m a n cap ita l by m o th ers, low er w o rk effo rt b y m o th e rs c o m p a re d w ith n o n m o th ers, u n ­ ob serv ed h etero g en eity b etw een m o th e rs a n d n o n m o th ers, a n d discrim i­ n a tio n a g a in st m o th e rs b y em ployers. In general, e x p la n a tio n s for the m o th e rh o o d w age p e n a lty can be classified as those th a t seek to id en tify im p o rta n t differences in the tra its, skills, a n d b e h a v io rs b e tw een m o th ers a n d n o n m o th e rs (i.e., w o rk e r ex p lan atio n s) a n d those th a t rely on the d ifferen tial preferen ce for or tre a tm e n t of m o th e rs a n d n o n m o th e rs (i.e., d iscrim in atio n explanations). E m p iric a l e v a lu a tio n s of these e x p lan atio n s h a v e larg ely focused on the former.

F o r exam ple, B u d ig a n d E n g la n d (2001) exam ine differences in w o rk p a tte rn s b etw een m o th e rs a n d n o n m o th e rs a n d find th a t in te rru p tio n s fro m w o rk , w o rk in g p a rt-tim e , a n d d ecreased seniority/experience collec­ tively explain no m ore th a n a b o u t o n e -th ird of the m o th e rh o o d penalty. T h e y also show th a t “m o th e r-frie n d ly ” jo b ch ara c teristic s (i.e., differences in the type of jo b s m o th e rs a n d n o n m o th e rs choose) explain v e ry little of the penalty. Sim ilarly, A n d erso n e t al. (2003) find th a t h u m a n capital, occu p atio n al, a n d h o u seh o ld resource v a ria b le s (e.g., n u m b e r of a d u lts in the household) collectively ac c o u n t for 24% of the to ta l w age p e n a lty for one child a n d 44% for w om en w ith tw o or m ore children. L ikew ise, W aldfogel a n d M ey er (2000) find th a t o ccu p atio n al controls do n o t elim ­ in ate the penalty. As B u d ig a n d E n g la n d (2001) conclude, the rem ain in g w age gap likely arises eith er because em ployed m o th e rs are som ehow less p ro d u c tiv e a t w o rk th a n n o n m o th e rs or because em ployers discrim in ate a g a in st m o th e rs (or som e c o m b in atio n of the tw o processes).

B e c k e r’s (1985) “w o rk effo rt” h y p o th esis is p e rh a p s the b e st-k n o w n p ro d u c tiv ity ex p lan atio n . A cco rd in g to B ecker, m o th e rs m a y in fa c t be less p ro d u c tiv e a t w o rk because th e y h a v e d issip ated th eir reserve of energy carin g for th e ir children. In an in d ire c t a tte m p t to ev a lu a te this claim , A nd erso n et al. (2003) co m p are the m o th e rh o o d w age p e n a lty for m o th e rs in d iffe re n t ed u c a tio n a l g ro u p s— hig h school d ro p o u ts, high school g ra d u a te s, those w ith some college, a n d college g rad u ates. T h e y hypothesize th a t if jo b s th a t req u ire m ore e d u catio n re q u ire m ore effort, th en the m o th e rh o o d w age p e n a lty shou ld be g re a te r for m o th e rs w ith h ig h e r levels of ed u catio n . C o n tra ry to th is p red ictio n , A n d erso n a n d colleagues fo u n d th a t m o th e rs w ho w ere h ig h school g ra d u a te s actu a lly

1299

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

experienced the la rg e st w age penalty. T h e y in te rp re t this non m o n o to n ic re latio n sh ip b etw een level of e d u catio n a n d the m ag n itu d e of the w age p e n a lty as evidence c o n tra d ic tin g p ro d u c tiv ity ex p lan atio n s of the m o th ­ erh o o d w age penalty. H ow ever, the a u th o rs la c k d ire c t m easu res of p ro ­ ductivity, lim itin g th e ir a b ility to rule o u t p ro d u c tiv ity explanations.

P ro d u c tiv ity a n d D iscrim in atio n

A logical w a y to distin g u ish b e tw een d iscrim in atio n a n d p ro d u c tiv ity ex p la n a tio n s w o u ld be to co m p are the w o rk p lace outcom es (e.g., salaries, hirin g , prom o tio n s) of em ployed m o th ers a n d n o n m o th e rs w ho h a v e equal levels of w o rk p lace produ ctiv ity . I f differences in p a y or p ro m o tio n ra te s w ere fo u n d b etw een g roups of m o th ers a n d n o n m o th e rs w hose p ro d u c ­ tiv ity levels w ere equal, th is finding w o u ld suggest th a t d iscrim in atio n facto rs w ere a t w ork. H ow ever, the d a ta sets an aly zed in the studies d escrib ed above la c k d ire c t m easu res of w o rk e r p ro d u ctiv ity . O ne likely reason for the la c k of w o rk p lace p ro d u c tiv ity m easu res is th a t it is in ­ h e re n tly p ro b le m a tic to fully specify w h a t m ak es som eone a good or p ro d u c tiv e em ployee. T h is difficulty lead s to an o th er: u n ex p lain ed gaps in w ages b e tw een tw o g ro u p s (e.g., em ployed m o th ers a n d n o n m o th ers) can alw ay s be a ttrib u te d to u n m e a su re d p ro d u c tiv ity differences b etw een the tw o groups. F o r exam ple, if the w ages of a tto rn e y s w ere co m p ared a n d p ro d u c tiv ity w as m e a su re d in term s of b illable ho u rs, a n d it w as fo u n d th a t co n tro llin g for this m easu re of p ro d u ctiv ity , fem ale a tto rn e y s w ith ch ild ren e a rn e d less th a n fem ale a tto rn e y s w ith o u t ch ildren, w e could n o t k n o w w h e th e r the w age gap fo u n d w as the re su lt of d iscrim in atio n a g a in st em ployed m o th ers or w as in ste a d the re su lt of som e o th er u n ­ m e a su re d fo rm of p roductivity.

To a d d re ss these p roblem s, in b o th the la b o ra to ry a n d a u d it studies, we e x p e rim en tally h o ld c o n sta n t the w o rk p lace p e rfo rm a n c es a n d o th er re le v a n t c h a ra c teristic s of a p a ir of fictitious jo b a p p lic a n ts a n d v a ry only th e ir p a re n ta l statu s. In the la b o ra to ry exp erim en t, we m easu re how ev ­ a lu a to rs ra te the a p p lic a n ts in term s of perceiv ed com petence, w o rk p lace co m m itm en t, hireability, prom o tab ility , a n d reco m m en d ed salary. In the a u d it study, we m easu re p ositive responses to a p p lic a n ts b a se d on the n u m b e r of callb ack s fro m a c tu a l em ployers. B y e x p erim en tally hold in g c o n sta n t w o rk p la c e -re le v a n t c h a ra c teristic s of the ap p lic a n ts, a n y differ­ ences b e tw een the ra tin g s of m o th e rs a n d n o n m o th e rs c a n n o t be a ttrib u te d to p ro d u c tiv ity or skill differences. W hile th is design c a n n o t rule o u t the possib ility th a t p ro d u c tiv ity differences a c c o u n t for p a r t of the w age p e n ­ alty th a t h a s been show n to exist, the la b o ra to ry s tu d y w ill isolate a p o te n tia l sta tu s-b a se d d iscrim in atio n m e c h a n ism b y e v a lu a tin g w h e th e r bein g a p a r e n t d isa d v a n ta g e s m o th e rs in the w o rk p lace even w h e n no

1300

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

p ro d u c tiv ity d ifferences exist b etw een th e m a n d w o m en w ith o u t children. In the n e x t section, we d ra w on sta tu s ch a ra c teristic s th e o ry to develop an e x p lan atio n for h o w m o th e rh o o d s ta tu s could lead to ev alu a tiv e biases a g a in st em ployed m others.

PERFORM ANCE EXPECTATIONS AND EVALUATIONS OF WORKPLACE COM PETENCE

S ta tu s C h a ra c te ristics T h e o ry

T h e th eo retical claim to be a d v a n c e d a n d e v a lu a te d is th a t m o th e rh o o d is a “s ta tu s c h a ra c te ristic ” th a t, w h en salient, resu lts in b iased ev a lu a tio n s of com petence a n d co m m itm en t, the use of a stric te r s ta n d a rd for ev a l­ u a tin g the w o rk p lace p e rfo rm a n c es of m o th e rs th a n of n o n m o th ers, a n d a b ias a g a in st m o th e rs in hirin g , p ro m o tio n , a n d sa la ry decisions. As defined b y sta tu s ch a ra c teristic s theory, a sta tu s ch ara c teristic is a cate­ gorical distin ctio n am o n g people such as a p erso n al a ttrib u te (e.g., race, gender) or a role (e.g., m o th erh o o d , m an ag er), th a t h a s a tta c h e d to it w idely h eld beliefs in the cu ltu re th a t associate g re a te r sta tu s w o rth in e ss a n d com petence w ith one categ o ry of the d istin ctio n th a n w ith o th ers (B erger et al. 1977). A sta tu s ch ara c teristic becom es sa lie n t w h e n it d ifferen tiates those in the settin g or because the ch ara c teristic is b elieved to be directly re le v a n t to the ta s k a t h a n d . T h e th e o ry arg u es th a t acto rs th en im p licitly use the sa lie n t ch ara c teristic to guide th e ir b e h a v io rs a n d ev alu atio n s.

T h e th e o retical c o n stru c t lin k in g s ta tu s ch aracteristics, such as gender or race, to differences in b eh a v io rs a n d ev a lu a tio n s is “p erfo rm an ce ex­ p e c ta tio n s.” A ccording to the theory, acto rs im p licitly expect m ore com ­ p e te n t ta sk p e rfo rm an ces from those w ith the m ore v a lu e d state of a ch ara c teristic (m en, m an ag ers) co m p a re d w ith those w ith the less v a lu e d state (w om en, n o n m an ag ers). T h ese d iffe re n tia ted p erfo rm an ce ex p ecta­ tions op erate in a self-fulfilling w ay — since th e y are expected to offer m ore co m p e te n t p erfo rm an ces, h ig h -sta tu s acto rs are given m ore o p p o rtu n itie s to p a rtic ip a te , h a v e m ore influence over o th ers in a group, an d , im p o r­ ta n tly for the c u rre n t p ro je c t, h a v e th e ir p e rfo rm a n c es e v a lu a te d m ore po sitiv ely (see C orrell a n d R id g ew ay 2003). T h ese effects are p re d ic te d except w h e n the ta s k or settin g is one for w h ich lo w er-statu s in d iv id u a ls are believ ed to be “n a tu r a lly ” b etter, such as a ta s k re q u irin g n u rtu rin g ab ility in the case of gender. E x p e rim e n ts confirm th a t a w ide v a rie ty of sta tu s ch aracteristics, in clu d in g race, gender, level of e d u catio n , a n d p h y s­ ical attra c tiv e n ess, system atically organize the a p p e a ra n c e of com petence, influence, a n d deference in this m a n n e r (L o v ag lia e t al. 1998; R id g ew ay 2001; T royer a n d Y ounts 1997; W eb ster a n d F osch i 1988).

T h e o ry a n d em p irical research suggest th a t in a d d itio n to th e ir im p a c t

1301

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

on p erfo rm an ce e v alu atio n s, sta tu s c h a ra c teristic s also affect the s ta n d a rd in d iv id u a ls use to d eterm in e w h e th e r a given p erfo rm an ce is in d icativ e of a b ility (Foschi 1989). T h e c en tral id e a is th a t a b ility s ta n d a rd s are stric te r for those w ith low er p erfo rm an ce expectations, th a t is, those w ith d ev a lu e d sta tu s ch aracteristics. T h e logic b e h in d th is p re d ic tio n is th a t good p e rfo rm a n c es are in c o n siste n t w ith ex p ectatio n s for lo w er-statu s a c ­ tors; th erefo re w h e n lo w er-statu s acto rs p e rfo rm w ell a t a task , th e ir p e r­ fo rm an ces are critically scrutinized. W h en h ig h e r-sta tu s acto rs p erfo rm eq u ally as w ell, th eir p e rfo rm a n c es are c o n sisten t w ith ex p ectatio n s a n d are th erefo re less scrutinized. Since p e rfo rm a n c es of lo w er-statu s actors are m ore h e av ily scru tin ized , th eir p e rfo rm an ces are ju d g e d b y a stricter s ta n d a rd co m p a re d w ith h ig h e r-sta tu s actors. T h erefo re, the p erfo rm an ces of lo w -sta tu s acto rs— even w h e n “o b je c tiv e ly ” eq u al to th a t of th e ir high- sta tu s c o u n te rp a rts — are less likely to be ju d g e d as d e m o n stra tin g ta sk ab ility or com petence. A “d ouble s ta n d a r d ” benefiting h ig h -sta tu s in d i­ v id u a ls is p re d ic te d except w h e n the ta sk or settin g is c u ltu ra lly associated w ith the lo w -sta tu s g ro u p (e.g., a ta s k re q u irin g n u rtu rin g a b ility m ig h t a d v a n ta g e m o th e rs over childless employees). E m p iric a l evidence s u p ­ p o rts these p re d ic tio n s for b o th g en d er a n d race, a n d the p re d ic tio n s hold b o th w h e n in d iv id u a ls e v a lu a te o th ers a n d w h e n th ey ev a lu a te them selves (F oschi 1996; B ie rn a t a n d K obry n o w icz 1997; C orrell 2001, 2004).

I f m o th e rh o o d is a d ev a lu e d s ta tu s in w o rk p lace settings, w e p re d ic t th a t m o th e rs w ill be ju d g e d b y a h a rs h e r s ta n d a rd th a n n o n m o th ers. T h e y w ill h a v e to p re s e n t evidence of g re a te r a b ility before bein g seen as com ­ p eten t. W hile this a rg u m e n t sh ares some sim ilarities w ith econom ic th e ­ ories of statistic a l d iscrim in atio n (P helps 1972; A rro w 1973; B ielb y a n d B a ro n 1986), sta tu s-b a se d d iscrim in atio n differs in th a t it claim s th a t the sta n d a rd u sed to ev a lu a te w o rk e rs is sy stem atically b iased in fav o r of h ig h -sta tu s g ro u p s.2

3 I n its o rig in a l fo rm u la tio n , th e th e o ry o f sta tis tic a l d isc rim in a tio n a ssu m e s th a t one g ro u p o f p e o p le (e.g., A fric a n -A m e ric a n s, w o m e n , o r m o th e rs) a re less p ro d u c tiv e th a n a n o th e r g ro u p (e.g., w h ites, m en , n o n m o th e rs), a n d t h a t o b ta in in g in fo rm a tio n a b o u t th e p r o d u c tiv ity o f in d iv id u a ls is p ro h ib itiv e ly ex p en siv e (P h e lp s 1972). R a tio n a l e m ­ p lo y ers th e re fo re p re fe r to h ire w o rk e rs fro m th e m ore p ro d u c tiv e g ro u p . W h ile th e o rie s o f s ta tis tic a l d isc rim in a tio n a ssu m e t h a t e m p lo y e rs a p p ly a n u n b ia s e d s t a n d a r d to a c c u ra te e stim a te s o f w o rk e r p ro d u c tiv ity , th e o rie s o f s ta tu s -b a s e d d isc rim in a tio n a rg u e th a t th e s ta n d a r d o f e v a lu a tio n is sy s te m a tic a lly b ia se d in fa v o r of th e h ig h -s ta tu s g ro u p . F u rth e rm o re , s ta tis tic a l d isc rim in a tio n th e o rie s a ssu m e r a tio n a l a c to rs relying on b ia se d in fo rm a tio n , w hile s ta tu s th e o rie s a ssu m e th a t c u ltu ra l beliefs d is to rt cog­ nitio n , e v en w h e n in fo rm a tio n is p e rfe c t. T h e s e d iffere n ce s a p p e a r su b tle , b u t lea d to d iffe rin g p re d ic tio n s . C o n sid e r th e h y p o th e tic a l case o f tw o e q u a lly p ro d u c tiv e e m ­ ployees, a fa th e r a n d a m o th e r, w h o h a v e e ac h left w o rk e a rly tw ice w ith in th e last m o n th . S ta tistic a l d isc rim in a tio n th e o rie s p re d ic t no d iffere n ce s b y g e n d e r w h e n a su p e rv is o r e v a lu a te s th e tw o em ployees. S ta tu s th eo ries, in c o n tra s t, p re d ic t t h a t the m o th e r w ill be h e ld to a h a rs h e r s ta n d a rd . F o r ex am p le, a su p e rv is o r m a y a ssu m e

1302

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

P re lim in a ry E m p iric a l S u p p o rt

R ecen t stu d ies p ro v id e some evidence c o n sisten t w ith sta tu s-b a se d dis­ c rim in atio n a n d illu stra te th a t the d isa d v a n ta g e s associated w ith m o th ­ erh o o d are n o t lim ited to pay. In one ex p erim en t, p a rtic ip a n ts w ere asked to im agine th a t th ey w ere clients choosing a c o n s u lta n t from a consulting firm (C u d d y et al. 2004). T h e research ers fo u n d th a t sim ply ad d in g the p h ra se “h a s a tw o-year-old c h ild ” to the d escrip tio n of the c o n su lta n t lead e v a lu a to rs to ra te the c o n s u lta n t as less c o m p e te n t th a n an otherw ise eq u al c o n s u lta n t n o t p re se n te d as h a v in g a child. L ikew ise, F u eg en e t al. (2004) fo u n d th a t w h e n e v a lu a to rs ra te d fictitious a p p lic a n ts for an a t­ to rn ey position, fem ale a p p lic a n ts w ith ch ild ren w ere h eld to a slightly, alth o u g h insignificantly, h ig h er s ta n d a rd th a n fem ale a p p lic a n ts w ith o u t children. F a th e rs w ere a c tu a lly held to a significantly low er s ta n d a rd th a n m ale n o n p a re n ts.

N o t all the resu lts in these tw o stu d ies w ere c o n sisten t w ith the a u th o r s ’ em p irical pred ictio n s. F o r exam ple, C u d d y et al. (2004) fo u n d no differ­ ence in the com petence ra tin g s b e tw een em ployed fa th e rs a n d m others, a n d F u eg en e t al. (2004) fo u n d no reliable effects of g en d er a n d p a re n ta l sta tu s on e v a lu a to rs ’ im p ressio n s of the a p p lic a n t’s co m m itm e n t in one of th e ir tw o sam ples of u n iv e rsity stu d en ts. M ore generally, m a n y of th eir resu lts w ere in c o n siste n t across th e ir d e p e n d e n t v a ria b le m easures. W hile these stu d ies suggest th a t a m o th e rh o o d p e n a lty m a y exist, the in co n sisten t p a tte rn of resu lts fails to conclusively d e m o n stra te sy stem atic d iscrim i­ n a tio n on the basis of m o th erh o o d . S ta tu s c h a ra c teristic s th e o ry offers reasons for these inconsistencies an d , m ore im p o rta n t, allow s us to gen­ erate precise p red ictio n s a b o u t w h en a n d to w h a t e x te n t m o th e rh o o d will

th a t th e m o th e r lea v es e arly to a tte n d to c h ild re n , w h ile th e fa th e r le a v e s e a rly to m eet clients. I n o u r la b o ra to ry e x p e rim e n t w e a re a b le to m e a su re th e a b ility s ta n d a rd s e v a lu a to rs use to ra te a p p lic a n ts , a llo w in g fo r a n e v a lu a tio n o f th e s ta tu s -b a s e d d is­ c rim in a tio n a rg u m e n t. F o r a m o re d e ta ile d d isc u ssio n o f th e d iffere n ce s b e tw e e n s ta tu s - b a se d a n d sta tis tic a l d isc rim in a tio n , see C o rre ll a n d B e n a rd (2006). A v a ria tio n of sta tis tic a l d isc rim in a tio n th e o ry d isp e n se s w ith th e a s s u m p tio n th a t g ro u p s of w o rk e rs v a ry in th e ir a v e ra g e m a rg in a l p r o d u c tiv ity (A igner a n d C a in 1977). In s te a d , th is fo rm u la tio n a ssu m e s t h a t th e v a ria n c e o f e m p lo y e rs ’ e stim a te s of w o rk e r p ro d u c tiv ity is g re a te r fo r w o m e n a n d m in o ritie s, a n d t h a t e m p lo y e rs a re risk a v erse . As a re su lt, ra tio n a l e m p lo y e rs a re a g a in p re su m e d d is p ro p o rtio n a te ly to h ire w h ite m en. T h e noisier sig n a l of p r o d u c tiv ity fo r w o m e n a n d m e m b e rs of m in o rity g ro u p s is a ssu m e d to o b ta in b e ca u se (1) m is c o m m u n ic a tio n is m o re likely to o c c u r b e tw e e n m e m b e rs of d iffe re n t g ro u p s th a n m e m b e rs o f th e sa m e g ro u p , so th a t (u su a lly w h ite a n d m ale) e m p lo y e rs re ce iv e c le a re r in fo rm a tio n fro m w h ite m ale a p p lic a n ts t h a n fro m o th e r a p p lic a n ts , a n d (2) w h ite m a le s a re m o re likely to use p e rs o n a l c o n ta c ts to a c q u ire jo b s , a n d these c o n ta c ts a re a ssu m e d to p a ss a c c u ra te in fo rm a tio n a b o u t th e a p p lic a n t to th e e m p lo y e r (O e ttin g e r 1996). As o u r la b o ra to ry e x p e rim e n t h o ld s th e q u a lity a n d source o f in fo rm a tio n a b o u t a p p lic a n ts c o n s ta n t a cro ss c o n d itio n s, w e a re also a b le to e lim in a te th is v e rsio n o f s ta tis tic a l d isc rim in a tio n a s a p o ssible e x p la n a tio n .

1303

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

lead to e v a lu a tiv e biases. B y testin g these p red ictio n s, this s tu d y c o n trib ­ utes an e x p lan atio n for the in c o n siste n t evidence yield ed in p a s t studies a n d offers a m ore com plete ac c o u n t of w h e th e r a n d h o w m o th e rh o o d serves as a source of d isa d v a n ta g e .

O ne stre n g th of s ta tu s ch a ra c teristic s th e o ry is th a t it d elin eates a set of p ro p o sitio n s th a t specify the circu m stan ces u n d e r w h ich sta tu s ch ar­ acteristics h a v e th e ir effects a n d the relativ e stre n g th of th e ir im p a c t u n d e r differing conditions. F o r exam ple, acco rd in g to the “salien ce” pro p o sitio n , m o th e rh o o d w ill only lead to ev a lu a tiv e biases w h e n it d ifferen tiates those in the settin g (some are m o th e rs a n d some are not) or if it is believ ed to be re le v a n t to the ta s k a t h a n d . Ironically, b u t c o n sisten t w ith the salience p ro p o sitio n , som e w o rk -fam ily policies th a t are in te n d e d to am elio rate the effects of m o th e rh o o d on w o rk p lace outcom es m a y a c tu a lly lim it the career m o b ility a n d w ag es of w o m en w ho tak e a d v a n ta g e of th e m b y m a k in g m o th e rh o o d sta tu s h ig h ly salient. G lass (2004) fo u n d th a t m o th ers em ployed in professio n al a n d m a n a g e ria l jo b s w ho p a rtic ip a te d in p ro ­ gram s such as teleco m m u tin g ex perienced low er w age g ro w th co m p ared w ith otherw ise sim ilar m o th e rs w h o av o id ed such p ro g ra m s (b u t see also W eeden 2005). In the F u eg en et al. (2004) s tu d y d escrib ed above, ev a l­ u a to rs e v a lu a te d only one ap p lic a n t. Since m o th e rh o o d w as n o t a differ­ e n tia tin g ch ara c teristic in th is study, it is likely th a t m o th e rh o o d w as n o t salien t w h e n a p p lic a n ts w ere e v alu ated .'5

In the C u d d y e t al. (2004) stu d y e v a lu a to rs d id rate m ore th a n one “c o n s u lta n t,” a n d the co n su lta n ts differed in p a re n ta l sta tu s, th e re b y m a k ­ ing p a re n ta l s ta tu s salient. In all conditions, an “ex p erim en tal c o n s u lta n t” w as p re se n te d as bein g 32 y ears old w ith an M B A a n d a long com m ute. D e p e n d in g on condition, th is c o n su lta n t w as fu rth e r d escrib ed as either m ale or fem ale a n d as a p a r e n t or not. P a rtic ip a n ts also ra te d tw o “filler” co n su lta n ts w h o h a d no ch ild ren a n d a BS degree. O ne w as d escrib ed as a m id d le-ag ed w o m an w ith a long com m ute a n d the o th er as a you n g m a n w ith a sh o rt com m ute. B e tw e e n -su b je c t c o m p ariso n s w ere m ad e to co m p are the ra tin g s of the ex p erim en tal co n su lta n ts across condition. In c lu d in g the filler profiles serv ed to m ak e p a re n ta l sta tu s salient, b u t it also m ad e several o th er sta tu s ch ara c teristic s sa lie n t sim u ltan eo u sly (gen­ der, age, p a re n ta l statu s, level of education). A ccording to s ta tu s ch ar­ acteristics theory, the a g g reg ated exp ectatio n s in d iv id u a ls fo rm for actors

3 S alience— th e re q u ire m e n t t h a t a c h a ra c te ris tic , su c h a s p a re n t s ta tu s , d iffe re n tia te s a p p lic a n ts — is a scope c o n d itio n o f o u r theory. T h is n a rro w s th e ra n g e o f se ttin g s to w h ic h o u r th e o ry a p p lie s, b u t n ot in a w a y th a t su b s ta n tia lly lim its th e a p p lic a b ility o f th e th e o ry w h e n c o n sid e rin g h irin g decisions. T h is is b e ca u se th e o v e rw h e lm in g m a jo rity of h irin g p ro c esses in clu d e c o m p a ris o n s of m u ltip le ré su m é s r a th e r th a n a n e x a m in a tio n o f a p p lic a tio n s s e q u e n tia lly (i.e., e v a lu a tin g a p p lic a tio n s o ne a t a tim e w ith o u t e v e r d ire c tly c o m p a rin g them ).

1304

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

in a settin g are im p a c te d b y all of the salien t s ta tu s in fo rm atio n , b u t a d d itio n a l pieces of c o n sisten t sta tu s in fo rm atio n h av e a declining m a r­ ginal im p a c t on overall ex p ectatio n s (B erger et al. 1977). T h erefo re, if we first notice th a t a perso n h a s an M B A , a n d th en notice th a t she also does n o t h a v e ch ild ren (an a d d itio n a l piece of p ositive s ta tu s in fo rm atio n ) the effect of bein g childless on o u r overall ex p ectatio n s for h e r w o u ld be less th a n it w o u ld h a v e been if p a re n ta l sta tu s w ere the only salien t sta tu s in fo rm a tio n in the setting. T h u s , an e x p erim en t th a t sim u ltan eo u sly v aries m u ltip le s ta tu s ch a ra c teristic s is n o t the m o st efficient for d etecting w h e th e r m o th e rh o o d sta tu s p ro d u ces the p re d ic te d effects.

C o n siste n t w ith th e o ry p re se n te d here, we m ak e p a re n ta l s ta tu s salien t b y h a v in g e v a lu a to rs rate a p a ir of ap p lic a n ts, one p a r e n t a n d one n o n ­ p a re n t. All o th er s ta tu s in fo rm a tio n is k e p t co n stan t. B efore describing the e x p erim en tal design in fu rth e r detail, we first tu rn to the lite ra tu re on c u ltu ra l conceptions of m o th e rh o o d to p ro v id e p re lim in a ry evidence th a t m o th e rh o o d is a s ta tu s ch ara c teristic in U.S. society.

M o th e rh o o d as a S ta tu s C h a ra c te ristic

To u n d e rs ta n d how m o th e rh o o d m ig h t fun ctio n as a d ev a lu e d sta tu s ch ar­ acteristic in w o rk p lace settings, it is h elp fu l to b ro a d e n the co n v en tio n al usage of “p erfo rm an ce e x p ectatio n s.” T h e th e o ry arg u es th a t since high- s ta tu s acto rs are expected to offer m ore co m petent p e rfo rm an ces, th e y are often given b e h a v io ra l a n d e v a lu a tiv e a d v a n ta g e s co m p a re d w ith low- s ta tu s actors. H ow ever, the th e o ry im plies th a t an y facto r th a t increases the relativ e exp ectatio n a b o u t the c a p acity of a p erso n to p e rfo rm in a settin g sh o u ld a d v a n ta g e h e r/h im in th a t settin g (Berger, C ohen, a n d Z eld itch 1966, 1972). E x p e c ta tio n s a b o u t p e rfo rm a n c e c a p a c ity h a v e a t le a st tw o dim ensions: com petence (or ability) a n d effo rt (H eid er 1958). W hile re search ers typically focus on the com petence dim ension, c u ltu ra l beliefs a b o u t the relativ e effort th a t social g roups ex ert in ta s k situ atio n s can also be the b asis for form in g d iffe re n tia ted p e rfo rm a n c e expectations. In d eed , som e of the earliest d escrip tio n s a n d exam ples of sta tu s c h a ra c ­ teristics relied on the id e a th a t a n tic ip a te d effo rt im p a c ts p erfo rm an ce exp ectatio n s (B erger e t al. 1966, 1972). F o r exam ple, w h e n explain in g w h y social class is a s ta tu s ch ara c teristic , B erger et al. (1966, pp. 33-34), d e­ scribe beliefs th a t the “w h ite collar class” is “m ore in d u s trio u s ” a n d “m ore en erg etic” th a n the low er class.

In co n sidering evidence for w h y m o th e rh o o d m ig h t o p erate as a sta tu s ch a racteristic, it is logically d ifficult to see w h y ta k in g on the m o th e r role sh o u ld affect a p e rs o n ’s u n d e rly in g a b ility or com petence, alth o u g h there is som e evidence th a t c u ltu ra l beliefs do associate m o th e rh o o d w ith a lessening of a b ility (see C ritte n d e n 2001). T h e re is, how ever, considerable

1305

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

evidence th a t c o n te m p o ra ry c u ltu ra l beliefs inclu d e a ssu m p tio n s th a t em ­ p loyed m o th e rs are less c o m m itted to w o rk th a n n o n m o th e rs an d , con­ sequently, p u t less effort into it (for a d e ta ile d rev iew see R id g ew ay a n d C orrell 2004). W hile co m m itm e n t a n d effo rt are n o t synonym ous, w hen e v a lu a tin g p o te n tia l em ployees, em ployers likely use perceiv ed co m m it­ m e n t as a p ro x y for a n tic ip a te d fu tu re effort.

M o th e rh o o d affects p ercep tio n s of com petence a n d co m m itm e n t b e ­ cause c o n tra d ic to ry schem as govern conceptions of “fam ily d e v o tio n ” a n d “w o rk d e v o tio n ” (B lair-L oy 2003, p. 5). C o n te m p o ra ry c u ltu ra l beliefs a b o u t the m o th e r role include a n o rm a tiv e exp ectatio n th a t m o th e rs will a n d sh o u ld engage in “in te n s iv e ” m o th e rin g th a t p rio ritizes m eetin g the needs of d e p e n d e n t ch ild ren above all o th er activ ities (H ay s 1996; Ko- b ry n o w icz a n d B ie rn a t 1997; B lair-L oy 2003). T h e c u ltu ra l n o rm th a t m o th e rs sh o u ld alw ay s be on call for th eir ch ild ren coexists in tension w ith a n o th e r w id ely h eld n o rm a tiv e b elief in o u r society th a t the “ideal w o rk e r” be u n e n c u m b e re d b y co m p etin g d e m a n d s a n d be “alw ay s th e re ” for his or h e r em ployer (A cker 1990; H a y s 1996; W illiam s 2001; B lair- L oy 2003). A cco rd in g to th is “id eal w o rk e r” belief, the b e st w o rk e r is the “c o m m itte d ” w o rk e r w h o d e m o n stra te s inten siv e effo rt on the jo b th ro u g h a ctions th a t a p p e a r to sacrifice all o th er concerns for w o rk (E p stein e t al. 1999; W illiam s 2001). E x a m p le s inclu d e a w illingness to d ro p e v ery th in g a t a m o m e n t’s notice for a new w o rk d e m a n d , to devote en orm ous h o u rs to “face tim e ” a t w o rk , a n d to w o rk late n ig h ts or w eekends. W hile it h as often been o b serv ed th a t “face tim e ” a n d long h o u rs are n o t necessarily asso ciated w ith a c tu a l w o rk e r p erfo rm an ce or p ro d u c tiv ity (e.g., E p ste in et al. 1999), in the c o n te m p o ra ry o rg an izatio n of w o rk , th ey fu n ctio n as a c u ltu ra l sign of the effo rt c o m p o n e n t of p erfo rm an ce capacity. N o r­ m a tiv e concep tio n s of the “id eal w o rk e r” a n d the “good m o th e r” create a c u ltu ra l tension b etw een the e n a c tm e n t of the m o th e rh o o d role a n d the e n a c tm e n t of the c o m m itted w o rk e r role. T h e c u ltu ra l logic of “in te n siv e ” m o th e rin g in U.S. society to d a y assum es th a t the “good m o th e r” w ill d ire c t h e r tim e a n d em o tio n al energy to w a rd h e r ch ild ren w ith o u t lim it (H ays 1996; B lair-L oy 2003). B y th is c u ltu ra l definition, then, a good m o th e r m u s t give less effo rt a n d p rio rity to w o rk d e m a n d s a n d therefore be a less c o m m itted w orker.

I t is im p o rta n t to keep in m in d th a t the tension b etw een these tw o roles occurs a t the level of n o rm a tiv e c u ltu ra l a ssu m p tio n s, a n d n o t necessarily a t the level of m o th e rs ’ ow n co m m itm e n t to w o rk roles. In fact, if w o rk co m m itm e n t is m e a su re d b y the im p o rta n c e people a tta c h to th e ir w o rk id en tities— eith er ab so lu tely or relativ e to o th er iden tities, such as fam ily id en tities— no difference is fo u n d in co m m itm e n t b etw een m o th ers a n d n o n m o th e rs (B ielby a n d B ielb y 1984). I t is the p erceived c u ltu ra l tension

1306

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

b etw een these tw o roles th a t lead s us to suggest th a t m o th e rh o o d is a d ev a lu e d s ta tu s in w o rk p lace settings.

EMPIRICAL PREDICTIONS

M o th e rh o o d P e n a lty

O u r m a in em p irical p re d ic tio n s are th a t jo b a p p lic a n ts w h o are p resen ted as m o th e rs w ill be ra te d as less co m p eten t, less c o m m itted to p a id w ork, less su itable for hire a n d p ro m o tio n , a n d d eserv in g of low er s ta rtin g sal­ aries co m p a re d w ith otherw ise e q u a l w om en w ho are n o t m o th ers. We expect th a t the com petence a n d co m m itm e n t ra tin g s w ill m ed iate the e v a lu a tio n v ariab les. In o th er w ords, we p re d ic t e v a lu a to rs w ill offer m o th e rs low er salaries a n d o th er re w a rd s because th ey assum e th a t th ey are less c o m p e te n t a n d c o m m itted th a n o th er k in d s of w o rk ers. We also expect th a t m o th ers w ill be ju d g e d b y a h a rs h e r s ta n d a rd .

A d d itio n a l F a c to rs A ffecting W o rk er E v a lu a tio n s

The effect o f fa th e r h o o d .— U nlike the m o th e rh o o d role, bein g a good fa th e r is n o t seen as c u ltu ra lly in c o m p a tib le w ith bein g an ideal w orker. In fact, as T o w n sen d (2002) describes, bein g a good fa th e r a n d a good em ployee are p a r t of the “p ack ag e d e a l” defining w h a t it m e a n s to be a m an. T h e re ­ fore, since the “good f a th e r” a n d “ideal w o rk e r” are n o t p erceiv ed to be in tension, bein g a p a r e n t is n o t p re d ic te d to lead to low er w o rk p lace ev a lu a tio n s for fathers.

In fact, research on the “m a rria g e p re m iu m ” for m e n ’s w ages, one of the m o st ro b u s t em p irical findings in la b o r econom ics, suggests th a t fa ­ th ers m ig h t experience a d v a n ta g e s in la b o r m a rk e t outcom es (H ersch a n d S tra tto n 2000; L oh 1996; K o re n m a n a n d N e u m a r k 1991; H ill 1979). L a b o r econom ists fre q u e n tly re p o rt th a t m a rrie d m en earn h ig h er w ages th a n u n m a rrie d m en, a n d speculate th a t th is m a y occur for one of the follow ing reasons: (1) m ore p ro d u c tiv e m en m a rry a t g re a te r ra te s (a ttrib u tin g the m a rria g e p re m iu m to selection bias), (2) m en becom e m ore p ro d u c tiv e follow ing m a rria g e (due to la b o r m a rk e t specialization b y m en a n d do ­ m estic sp ecialization b y w om en), or (3) em ployers fav o r m a rrie d m en (due to g en d er bias). E m p iric a l in v estig atio n s c a st d o u b t on the selection bias (Loh 1996; K o re n m a n a n d N e u m a r k 1991) a n d p ro d u c tiv ity (H ersch a n d S tra tto n 2000; L o h 1996) explan atio n s.

O u r th e o ry suggests th a t the m a rria g e p re m iu m m a y a c tu a lly be, in p a rt, a fa th e rh o o d p rem iu m . C u ltu ra l concep tio n s of fath e rh o o d in the U n ite d S tates often inclu d e the rig h t to a “fam ily w a g e ” b o n u s to ensure th a t m a rrie d m en serve as b re a d w in n e rs for th e ir fam ilies (O rloff 1996;

1307

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

H ill 1979). T h e re is a lre a d y some evidence for this: w h e n ch ild ren u n d e r six are in c lu d e d into m e n ’s w age eq u atio n s, th e y ex ert a positive a n d significant effect on w ages a n d reduce the m a g n itu d e of the m a rria g e p re m iu m (e.g., H ersch a n d S tra tto n 2000). W hile o u r th eo ry focuses on the h y p o th esized w age p e n a lty for m o th ers, it is c o n sisten t w ith a w age p re m iu m for fathers.

T he effects o f race a n d gen d er.— Since race a n d g en d er are also sta tu s c h aracteristics, w e w o u ld expect th a t if race a n d /o r g en d er w ere salien t w h e n a p p lic a n ts w ere ev a lu a te d , a p p lic a n ts w ho are fem ale a n d /o r A f­ rican -A m erican w o u ld be ra te d as less h ira b le a n d p ro m o ta b le a n d offered low er salaries th a n w h ite m ale a p p lican ts. H ow ever, since p a rtic ip a n ts w ill e v a lu ate a p a ir of a p p lic a n ts w ho are the sam e g en d er a n d race, race a n d gender sh o u ld n o t be salient. I t is of course possible th a t p a rtic ip a n ts w ill im plicitly com pare the a p p lic a n ts to o th ers w h o m th ey im agine are bein g e v a lu a te d an d , in so doing, th a t th e y w ill d ra w on sta tu s beliefs a b o u t g en d er a n d race, lead in g to b iased ratings. To the e x ten t th a t this occurs, it is possible th a t a m a in effect of race a n d g en d er on ev a lu a tio n s w ill be found, alth o u g h these effects sh o u ld be w e a k e r a n d less reliable th a n the effects of m o th e rh o o d status.

We m a n ip u la te g en d er so th a t we can e v a lu ate the claim ab o v e th a t m en are n o t p e n a liz e d in the w o rk p lace for bein g fathers. We m a n ip u la te race in the la b o ra to ry e x p erim en t so th a t w e can e v a lu ate w h e th e r our a rg u m e n t h olds for b o th w h ite a n d A frican -A m erican a p p lic a n ts.4 O ur th eo ry p re d ic ts th a t m o th ers, b o th A frican -A m erican a n d w h ite, w ill ex­ p erience ev a lu a tiv e biases in w o rk p lace settings. H ow ever, A n d erso n et al. (2003) re p o rt th a t p a s t stu d ies u sing su rv ey d a ta are inconclusive a b o u t h o w the m ag n itu d e of the m o th e rh o o d p e n a lty differs for A frican -A m er­ ican a n d w h ite w om en. G iven th is incon sisten cy a n d the differences in w o rk p lace h istories a n d experiences of A frican -A m erican a n d w hite w om en, it is im p o rta n t to e v a lu a te w h e th e r sta tu s-b a se d d iscrim in atio n w o rk s sim ilarly for b o th groups, r a th e r th a n assum e th a t race does n o t im p a c t this process. O u r la b o ra to ry e x p erim en t design w ill allow us to assess w h e th e r the m o th e rh o o d p e n a lty accrues to b o th A frican -A m erican a n d w h ite w o m en a n d to co m p are the m a g n itu d e of a n y p e n a lty found. T h e design w ill also allow us to co m p are the ra tin g s of childless w om en to childless m en a n d the ra tin g s of m o th ers a n d fathers. W hile n o t the ce n tra l focus, these co m p ariso n s are im p o rta n t since responses to lab o r m a rk e t d a ta on the g en d er w age gap com pare w ages for m en a n d w om en, often co n tro llin g for p a re n ta l sta tu s (e.g., C ritte n d e n 2001; V enable 2002).

4 W e do n o t v a ry ra ce in th e a u d it study, a s w e d e sc rib e below .

1308

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

TH E LABORATORY E X PE R IM E N T

P a id u n d e rg ra d u a te v o lu n te e rs ra te d a p a ir of e q u ally qualified, sam e- g en d er (either m ale or fem ale), sam e-race (either A frican -A m erican or w hite) fictitious jo b a p p lic a n ts, p re se n te d as real, w ho differed on p a re n ta l statu s. P a irin g ap p lic a tio n m a te ria ls b y race a n d g en d er gen erates four ex p erim en tal cond itio n s w h ere p a rtic ip a n ts ra te one p a r e n t a n d one n o n ­ p a r e n t a p p lic a n t w h o are eith e r A frican -A m erican m en, A frican -A m erican w om en, w h ite m en, or w h ite w om en. M ale a n d fem ale p a rtic ip a n ts w ere ra n d o m ly assigned to one of these fo u r conditions, a n d p a re n ta l sta tu s w as c o u n te rb a la n c e d across the tw o m em b ers of the a p p lic a n t p a ir (i.e., one m e m b e r of the p a ir w as p re se n te d as a p a r e n t to h a lf of the p a rtic ­ ip a n ts in each cond itio n , a n d the o th er m e m b e r w as the p a r e n t for the o th er h a lf of the p a rtic ip a n ts in th a t sam e condition). T h u s, the design consisted of three b etw e e n -su b je cts factors (gender of p a rtic ip a n t, gender of a p p lic a n t pair, a n d race of a p p lic a n t pair) a n d one w ith in -su b je c ts facto r (p a re n ta l statu s).5 We m a k e no p re d ic tio n s a b o u t the effect of p a r­ tic ip a n t gender on a p p lic a n t ratin g s, a lth o u g h the design allow s us to assess if m ale a n d fem ale e v a lu a to rs re a c t d ifferen tly to p a re n ta l statu s.

T h e s tu d y in c lu d e d 192 p a rtic ip a n ts (84 m en a n d 108 w om en), b etw een 19 a n d 28 p e r condition. F o u r p a rtic ip a n ts (2.1%) w ere suspicious a b o u t some asp e c t of the study, a n d co n seq u en tly th e ir d a ta w ere excluded p rio r to analysis, c reatin g an effective sam ple size of 188. R ejection rules w ere c o n serv ativ e a n d e stab lish ed b efo reh an d . All analyses w ere also con­ d u c te d w ith all a v ailab le d a ta , a n d no su b sta n tiv e differences w ere found.

T h e Use of U n d e rg ra d u a te s

T h e la b o ra to ry e x p erim en t featu res a h ig h ly contro lled settin g w ith a diverse set of m easu res, allo w in g us to g enerate d a ta th a t are w ell su ited for e v a lu a tin g o u r th eo retical m echanism . In p artic u la r, the la b o ra to ry

5 “B e tw e e n -s u b je c ts ” fa c to rs a re th o se th a t do n o t v a ry w ith in a n e x p e rim e n ta l c o n ­ d itio n , su c h as p a rtic ip a n t gender, w h e re a s “w ith in -s u b je c ts ” fa c to rs do vary. F o r ex­ a m p le , in th is e x p e rim e n t p a r e n ta l s ta tu s is a w ith in -s u b je c t fa c to r since it v a rie s a cro ss th e tw o m e m b e rs of th e a p p lic a n t pair. Since th e p rim a ry p u rp o se of th is p ro je c t is to assess th e effect o f p a r e n ta l s ta tu s on th e ra tin g s a n d e v a lu a tio n s o f a p p lic a n ts , it is im p o rta n t th a t p a r e n ta l s ta tu s be m e a su re d a s a w ith in -s u b je c ts c o m p a riso n , w h ic h is m o re efficient th a n b e tw e e n -p a ir c o m p a ris o n s (C ohen 1988). T h is efficiency ra tio n a le m ig h t suggest a n a lte rn a te design, w h e re a p p lic a n t ra ce a n d g e n d e r w e re also m e a su re d w ith in su b je c t. I n th is a lte rn a te d e sig n m ale a n d fem ale p a rtic ip a n ts w o u ld e v a lu a te eig h t sets o f a p p lic a tio n m a te ria ls — a p a re n t a n d n o n p a re n t a p p lic a n t fro m e ac h ra c e / sex c o m b in a tio n . P re te s tin g e sta b lish e d t h a t th is a lte rn a te d e sig n a ro u se d suspicion since p a rtic ip a n ts w e re re q u ire d to e x am in e eig h t sets o f v e ry sim ila r m a te ria ls . E v e n th o u g h b e tw e e n -su b je c t c o m p a ris o n s a re less efficient e s tim a to rs th e y a re n o n e th e less u n b ia se d .

1309

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

settin g ensures th a t w e can m a in ta in sufficient co ntrol o v er facto rs th a t w o u ld in terfere w ith tests of o u r h y p o th eses (e.g., no o th er people in the ro o m to p rim e o th er sta tu s ch aracteristics, no telephone or o th er d istra c ­ tions present). H ow ever, it is n o t feasible to convince ap p ro x im a te ly 200 ac tu a l em ployers to v isit the la b o ra to ry a n d sp en d one h o u r p a rtic ip a tin g in an experim ent. T h erefo re, b y n ecessity we m u s t rely on a sam ple of u n d e rg ra d u a te s in o rd er to o b ta in a com plete te st of ou r th eo retical claim s reg a rd in g the d iscrim in atio n m ech an ism . U n d e rs ta n d in g this m ech an ism is im p o rta n t if the goal is to find w ay s to red u ce the d isa d v a n ta g e s m o th ers face.

T h e th eo ry p re se n te d here im plies th a t to the e x ten t th a t em ployers sh are the belief th a t m o th e rs are less c o m m itted to or c o m p e te n t in w o rk ­ place settings, th ey too w ill su b tly d iscrim in ate a g a in st m o th ers. Q u ali­ ta tiv e research suggests th a t em ployers do, in fact, sh are this b elief a n d d iscrim in ate a g a in st m o th e rs in a ran g e of settings (B lair-L oy 2003; C rit­ te n d e n 2001; K en n elly 1999). F o r exam ple, B lair-L oy q u o tes a chief fi­ n a n c ia l officer w ho ack n o w led g es d e lib e ra te ly rejectin g w om en a p p lic a n ts on the b asis of p a r e n t statu s. H e r source conceded, “I find m yself choosing m en here ev ery d a y o v er a w o m a n w ith a child. I f I h a d kids, I m ig h t n o t h a v e m ad e the sam e co m m itm e n t to m y j o b ” (B lair-L oy 2003, pp. 119-20). Q u a n tita tiv e research also suggests th a t m a n a g e rs a n d s tu d e n t e v a lu a to rs offer sim ilar a p p ra isa ls of a p p lic a n ts a n d th a t e v a lu a to rs re ­ sp o n d sim ilarly to real a n d h y p o th e tic a l a p p lic a n ts (C lev elan d a n d B er­ m a n 1987; C le v e la n d 1991; O lian a n d S ch w ab 1988).

To assess m ore d irectly the e x te n t to w h ich em ployers d iscrim in ate a g a in st m o th ers, w e c o n d u cted an a u d it s tu d y of a c tu a l em ployers. We describe th is s tu d y a n d its resu lts a fte r p re se n tin g the resu lts from the la b o ra to ry experim ent. W hile the a u d it stu d y c a n n o t assess the m ech an ism of d iscrim in atio n , it w as designed to allow for c o m p a ra b ility w ith the la b o ra to ry exp erim en t, a n d p ro v id es evidence of the re a l-w o rld im p li­ cations of the a rg u m e n t e v a lu a te d here.

Pro ced u re

P a rtic ip a n ts cam e to the lab in d iv idually, re a d a d escrip tio n of a co m p an y t h a t w as p u rp o rte d ly h irin g for a m idlevel m a rk e tin g p osition, a n d ex­ a m in ed ap p lic a tio n m a te ria ls for tw o a p p lic a n ts for the positio n w ho differed on p a re n ta l sta tu s b u t w ere oth erw ise sim ilar. T h e y ex am in ed the a p p lic a n t files one a t a tim e, a n d w e c o u n te rb a la n c e d w h ich file, the p a r e n t or n o n p a re n t, th e y v iew ed first. A fter rev iew in g an a p p lic a n t’s file, p a rtic ip a n ts im m e d ia te ly co m p leted an “in itial im p re ssio n s” su rv ey for th a t a p p lic a n t, w h ich c o n ta in e d item s th a t allow us to assess w h e th e r the p a rtic ip a n ts in th is s tu d y view m o th e rh o o d as a sta tu s c h aracteristic

1310

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

(see “d e p e n d e n t m e a su re s” below). O n the sam e in stru m e n t, p a rtic ip a n ts w ere ask ed to p ro v id e a list of p ro s a n d cons for each a p p lic a n t, a ta sk in te n d e d to entice th e m to look m ore closely a t the a p p lic a n ts ’ m a te ria ls before p ro ceed in g to the n e x t stage of e v alu atio n . P a rtic ip a n ts w ere n e x t in stru c te d to look a t the ap p lic a tio n m a te ria ls m ore closely a n d com plete an “a p p lic a n t e v a lu a tio n sh e e t” for each can d id ate. T h is in s tru m e n t con­ ta in e d o u r a b ility s ta n d a rd a n d e v a lu a tio n m easu res, d escrib ed below. B efore le av in g the lab, p a rtic ip a n ts a n sw e re d a series of free response item s to assess w h e th e r the ex p erim en tal m a n ip u la tio n s w ere successful a n d to d eterm in e if th ey w ere suspicious of som e asp e c t of the study. T h e y w ere also briefly in te rv ie w e d as a fu rth e r check on ex p erim en tal m a n ip ­ ula tio n s a n d suspicion, a n d th en th ey w ere deb riefed a n d p aid.

C o v er S tory

P a rtic ip a n ts w ere told th a t a C a lifo rn ia -b a se d sta rt-u p co m m u n icatio n s c o m p a n y w as co n d u ctin g an em p lo y m en t search for a p erso n to h e a d up its new E a s t C o a st m a rk e tin g d e p a rtm e n t. T h e y h e a rd th a t the com ­ m u n ic a tio n s co m p an y w as in te re ste d in receiving fe ed b ack fro m younger ad u lts since y o u n g people are h e a v y consum ers of c o m m u n icatio n s te c h ­ nology. To fu rth e r increase th e ir ta sk o rie n ta tio n , p a rtic ip a n ts w ere told th a t th e ir in p u t w o u ld be in c o rp o ra te d w ith the o th er in fo rm a tio n the c o m p a n y collects on a p p lic a n ts a n d w o u ld im p a c t a c tu a l h irin g decisions.6 P a rtic ip a n ts th e n re a d a b o u t the re q u ire m e n ts of the m a rk e tin g position a n d the p ro p o sed sa la ry ran g e ($135,000-$180,000).

A p p licatio n M a te ria ls

P a rtic ip a n ts in sp ected an a p p lic a n t file for each of the tw o a p p lican ts. O th e r th a n v a ry in g first n a m e s to m a n ip u la te race a n d gender (see below),

6 T h e d e cision to use d e c e p tio n in e x p e rim e n ta l se ttin g s re q u ire s a c a re fu l w e ig h in g of th e p o te n tia l co sts a n d benefits. A s o th e rs h a v e n o ted , “E x p e rim e n te rs o ften find th e m ­ selves in th e tro u b le so m e p o sitio n o f co n ce alin g th e t r u th fro m th e ir su b je c ts in o rd e r to re v e a l a t r u th a b o u t h u m a n b e h a v io r ” (A ronson et al. 1990, p. 89). I f w e d id not use d e c e p tio n in th is study, p a rtic ip a n ts m a y h a v e p ro v id e d socially d e sira b le a n sw e rs (to a v o id a p p e a rin g to d isc rim in a te a g a in st m o th e rs) o r m a y n ot h a v e ta k e n th e ta s k se rio u sly (a n d th u s a n sw e re d carelessly, r a th e r th a n in a w a y t h a t re v e a le d th e ir tru e preferences). B o th o f these w o u ld h a v e s u p p re ss e d o u r a b ility to d e te c t re a l d isc rim ­ in a tio n a g a in st m o th e rs. We b elieve th a t th e costs o f a re la tiv e ly b rie f (the d e ce p tio n a n d th e re a so n in g b e h in d it w a s fully e x p la in e d to p a rtic ip a n ts a t th e e n d o f th e study, m e a n in g th a t p a rtic ip a n ts w e re d e ce iv e d fo r less th a n a n h o u r in m o st cases) a n d m ild (the d e c e p tio n d id n o t d istre s s th e p a rtic ip a n ts or v io la te th e ir p riv a c y ) u se o f d e ce p tio n w e re o u tw e ig h e d b y th e in c re a s e d risk o f c o n c lu d in g th a t m o th e rs do n o t face d is­ c rim in a tio n w h e n th e y a c tu a lly do.

1311

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

the files w ere id en tical across condition. T h e files c o n ta in e d three items: a sh o rt m em o, a “fa c t sh e e t,” a n d a résum é. T h e m em os w ere sim ilar to those used b y C u d d y et al. (2004) a n d co n ta in e d a few b rie f n o tes p u r­ p o rte d ly fro m a h u m a n resources sta ff m e m b e r a t the h irin g co m p an y w ho c o n d u cted a sh o rt screening in te rv ie w w ith the ap p lic a n t. T h e m em os a n d the résum és w ere used to m a n ip u la te p a re n ta l sta tu s as d escribed below. T h e “fa c t s h e e t” su m m a riz e d re le v a n t in fo rm a tio n a b o u t the p o ­ ten tia l em ployees (such as college g rad e p o in t average) th a t w as n o t p re ­ sen ted on the résum é. T h e résum és listed the a p p lic a n t’s career goals, ed u c a tio n a l history, p a s t w o rk experience, a n d o th er re le v a n t activities. T h e résum és in d ic a te d th a t the a p p lic a n ts h a d b a c h e lo r’s degrees from one of tw o large m id w e ste rn un iv ersities a n d h a d a p p ro x im a te ly seven y ears of w o rk experience. B o th a p p lic a n ts w ere p re se n te d as h ig h ly p ro ­ du ctiv e b y in c lu d in g “re su lts” on the résum é, such as “in creased division sales b y 10% b etw een 2000 a n d 2002.” T h e fa c t sh eet a n d the résum és w ere used to estab lish th a t the c a n d id a te s w ere e q u ally p ro d u c tiv e in th eir p a s t jo b s a n d th a t th ey h a d e q u iv a le n t skills a n d b a c k g ro u n d s. O ne ch a l­ lenge of this s tu d y w as to create tw o sets of m a te ria ls th a t w ere of e q u iv ­ a le n t q u a lity w ith o u t bein g suspiciously sim ilar.

P rio r to the a c tu a l exp erim en t, we p re te ste d the tw o versio n s of the m a te ria ls to assess w h e th e r th e y w ere of e q u iv a le n t quality. A t th is stage, race, gender, a n d p a re n ta l s ta tu s in fo rm a tio n w as n o t a v a ilab le to ev a l­ u a to rs so th a t we could d eterm in e w h e th e r the résum és w ere perceiv ed to be e q u a l in the absence of the sta tu s m a n ip u la tio n s to be em ployed in the a c tu a l experim ent. A d iffe re n t sam ple (A = 60) d ra w n fro m the sam e p o p u la tio n as in the a c tu a l e x p erim en t ra te d these tw o “te m p la te ” résum és, one a t a tim e, u sing sev en -p o in t scales ra n g in g fro m “n o t a t all” to “ex­ tre m e ly ” cap ab le, efficient, skilled, intelligent, in d e p e n d e n t, self-confident, w a rm , a n d sincere. N o significant differences w ere fo u n d b e tw een p a r­ tic ip a n ts ’ ra tin g s of the tw o résum és on a n y of these eight traits. P a rtic ­ ip a n ts also in d ic a te d w h ich of the tw o a p p lic a n ts a p p e a re d m ore qualified for a m a rk e tin g position w ith a n ew sta rt-u p com pany, a n d no significant a ssociation w a s fo u n d b etw een résum é te m p la te a n d bein g m ore qualified ( X 2 = 1.79, d f= 3 ) , in d ic a tin g th a t one a p p lic a n t d id n o t a p p e a r significantly m ore q ualified th a n the other. N o n eth eless, to ensure th a t differences in résum és w ere n o t sy stem atically im p a c tin g the results, p a re n ta l sta tu s w as co u n te rb a la n c e d in the a c tu a l e x p erim en t across the tw o versio n s of the résum és for each condition.

Race a n d G en d er M a n ip u la tio n s

Fo llo w in g B e rtra n d a n d M u lla in a th a n (2003), the race a n d gender of a p p lic a n ts w ere m a n ip u la te d b y a lte rin g first na m e s on the a p p lic a n t files.

1312

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

B e rtra n d a n d M u lla in a th a n (2003) u sed b irth certificate d a ta to generate a list of n am es com m only given to w h ite a n d A frican -A m erican children in the m id-1970s a n d th e n p re te ste d these n am es to estab lish th a t th ey ev o k ed the race a n d g en d er a ttrib u tio n s p re d ic te d (see B e rtra n d a n d M u l­ la in a th a n 2003, app. A). We u sed the follow ing first n am es ta k e n from this list: A llison a n d S a ra h (w hite fem ales), E b o n y a n d L a to y a (A frican- A m erican fem ales), T yrone a n d J a m a l (A frican-A m erican m ales), a n d B ra d a n d M a tth e w (w hite males). E a c h m e m b e r of the a p p lic a n t pair, reg ard less of race or gender, h a d one of tw o la st n am es th a t w ere con sisten t across condition.

P a re n ta l S ta tu s M a n ip u la tio n

P a re n ta l sta tu s w as m a n ip u la te d on the résum é a n d on the h u m a n re ­ sources m em o. T h e résum é for the p a r e n t m e m b e r of the a p p lic a n t p a ir listed “P a re n t-T e a ch e r A ssociation c o o rd in a to r” u n d e r the h e a d in g “o th er re le v a n t activ itie s.” T h e n o n p a re n t w as in ste a d d escrib ed as the fu n d ra ise r for h is/h er n e ig h b o rh o o d association. F u eg en et al. (2004) successfully used a sim ilar m a n ip u la tio n to in d icate p a re n ta l statu s. F o llo w in g C u d d y et al. (2004), the m em o for the p a r e n t m e m b e r of the p a ir in c lu d e d the follow ing phrase: “M o th e r/fa th e r to T om a n d Em ily. M a rrie d to J o h n / K a re n .” T h e n o n p a re n t w as d escrib ed as sim ply “m a rrie d to J o h n /K a re n .”7

D e p e n d e n t M easu res

A ccording to the theory, if m o th e rh o o d o p erates as a s ta tu s c h a racteristic, th en m o th e rs w ill be p erceiv ed as less co m p e te n t or c o m m itted th a n n o n ­ m o th ers. As a resu lt, m o th e rs w ill be ju d g e d b y a h a rs h e r s ta n d a rd th a n o th er p o te n tia l em ployees a n d w ill be v iew ed as less h ira b le , less pro- m o tab le, a n d d eserv in g of low er sta rtin g salaries. To e v a lu ate this a r­ gu m en t, th ere are eight d e p e n d e n t m easures: tw o th a t m easu re the tra its of com petence a n d co m m itm en t, tw o th a t m easu re the a b ility s ta n d a rd p a rtic ip a n ts u sed to ju d g e the ap p lic a n ts, a n d fo u r th a t serve as our key ev a lu a tio n m easures. T h e eight m easu res are m o d e ra te ly co rrelated , w ith c o rrelatio n s ra n g in g .25—.45.

C om petence a n d c o m m itm e n t m easu res.— D u rin g the first p h ase of eval-

7 W e d id n o t give th e sp o u ses a n d c h ild re n of th e A fric a n -A m e ric a n a p p lic a n ts “b la c k ” n am es. T h e goal w a s to m a k e ra ce o f a p p lic a n t sa lie n t, b u t h o ld all o th e r fa c to rs c o n sta n t. W h ile o ne o f th e a p p lic a n ts w a s p re s e n te d a s a p a re n t, th e a p p lic a tio n m a ­ te ria ls for th e o th e r m e m b e r o f th e p a ir m a d e no m e n tio n o f c h ild re n , a lth o u g h she o r he w a s d e sc rib e d as m a rrie d . W e use th e p h ra s e s “n o n p a r e n t” o r “c h ild less” for c o n v en ien c e e v en th o u g h it w o u ld be m o re c o rre c t, e v en if a w k w a rd , to re fe r to these a p p lic a n ts a s “w o m e n (or m en) w h o d id n ot give ev id e n ce of b e in g a p a r e n t.”

1313

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

u atio n , p a rtic ip a n ts ra te d a p p lic a n ts on a series of item s, in clu d in g item s in te n d e d to m easu re the e x te n t to w h ich th e y v iew ed the c a n d id a te s as com petent. F o llo w in g C u d d y et al. (2004) we cre a te d a com posite com ­ p etence m easu re b y c a lcu latin g a w e ig h te d av erag e of p a r tic ip a n ts ’ ratin g s of the a p p lic a n ts on se v e n -p o in t scales ra n g in g fro m “n o t a t a ll” to “ex­ tre m e ly ” cap ab le, efficient, skilled, intelligent, in d e p e n d e n t, self-confident, aggressive, a n d org an ized (m ean = 5.48, SD = .69, a lp h a = .85).8

To m easu re perceiv ed c o m m itm en t, we in c lu d e d a single-item q u estion on the “a p p lic a n t e v a lu a tio n s h e e t” w h ich a sk e d p a rtic ip a n ts h ow com ­ m itte d th ey th o u g h t the a p p lic a n t w o u ld be relativ e to o th er em ployees in sim ilar positio n s a t the com pany. T h e y w ere given 10 choices ran g in g fro m “m ore c o m m itted th a n 0% of o th e r em ployees” to “m ore c o m m itted th a n 99% of o th er em ployees.” O n av erag e p a rtic ip a n ts v iew ed the a p ­ p lic a n ts as m ore c o m m itted th a n 74.7% of o th er sim ilar em ployees (SD = 18.0%). T h e m a in em p irical p re d ic tio n is th a t if m o th e rh o o d is a sta tu s c h a racteristic, the m ean com petence a n d co m m itm e n t ratin g s sh o u ld be low er for m o th e rs th a n for n o n m o th ers.

A b ility sta n d a rd m easu res.— P a rtic ip a n ts a n sw e re d tw o item s on the second p h ase of e v a lu a tio n designed to p ro v id e a b ility s ta n d a rd m easures. T h e first q u estio n w as “in w h a t p ercentile w o u ld the a p p lic a n t n eed to score on h is/h er m a n a g e m e n t profile exam in o rd e r for y o u to consider h im /h e r for e m p lo y m e n t? ” P a rtic ip a n ts w ere given choices ra n g in g from the fifth to the n in e ty -n in th percentile. T h e “m a n a g e m e n t profile ex a m ” w as d escrib ed to p a rtic ip a n ts as p ro v id in g evidence a b o u t p o te n tia l for ad v a n c e m en t. T h e m e a n for the sco re-req u ired ite m w a s 68.7 (S D = 3 0 .3 ), in d ic a tin g th a t a p p lic a n ts w o u ld need, on av erag e, to score in a p p ro x i­ m a te ly the six ty -n in th p ercentile or h ig h er in o rd er to be hired.

P a rtic ip a n ts w ere fu rth e r ask ed , “H o w m a n y days could this a p p lic a n t a rriv e late or leave early p e r m o n th before y o u w o u ld no longer recom ­ m e n d h im /h e r for h ire ? ” T h e m ean for the d ays late ite m w as 3.43 days (SD = 2.12). A ccording to the sta tu s-b a se d d iscrim in atio n m ech an ism , if m o th e rh o o d o p erates as a sta tu s c h a racteristic, m o th e rs w ill be ju d g e d b y a h a rs h e r s ta n d a rd th a n n o n m o th ers. T h e y w ill be re q u ire d to score in a h ig h er p ercen tile th a n n o n m o th e rs before bein g con sid ered h irab le a n d w ill be allo w ed fe w e r days of bein g late or leav in g early.

E v a lu a tio n m ea su res.— F o u r e v a lu a tio n m easu res w ere in c lu d e d on the a p p lic a n t e v a lu a tio n sheet. P a rtic ip a n ts w ere ask ed w h a t salary in dollars th ey w o u ld reco m m en d for each a p p lic a n t if the a p p lic a n t w ere hired.

8 W h ile w e follow ed C u d d y et al. in c re a tin g o u r c o m p e ten c e co m p o site v a ria b le , th e re su lts p re s e n te d in th is a rtic le do n o t h in g e on th e e x ac t c o m p o sitio n o f th e v a ria b le . T h a t is, in d iv id u a l ite m s p ro d u c e q u a lita tiv e ly sim ila r re su lts, a s d o c o m p o site s c re a te d fro m sm a lle r su b s e ts of th e item s.

1314

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

R ecall th a t the p a rtic ip a n ts w ere told th a t the p ro p o sed sa la ry ran g e w as $135,000-$180,000. T h e m e a n sa la ry reco m m en d ed w as $145,000 (SD = $22,400). P a rtic ip a n ts w ere also ask ed to estim ate the likelihood th a t an a p p lic a n t w o u ld be su b se q u e n tly p ro m o ted if h ired . R esponses w ere on a fo u r-p o in t scale ra n g in g fro m “m o st ce rta in ly w ill N O T be p ro m o te d ” to “m o st ce rta in ly w ill be p ro m o te d ,” w ith a m e a n response of 3.14 (SD = .67), suggesting th a t a p p lic a n ts, if h ired , w ere generally v iew ed as m o d e ra te ly p ro m o tab le. P a rtic ip a n ts w ere fu rth e r a sk e d if th ey th o u g h t the a p p lic a n t, if h ired , sh o u ld be reco m m en d ed for a m anagem ent- tra in in g course designed for those w ith stro n g a d v a n c e m e n t poten tial. O verall, 83.5% of a p p lic a n ts w ere reco m m en d ed for th is course. Finally, p a rtic ip a n ts w ere a sk e d for each a p p lic a n t if th ey w o u ld reco m m en d h er/ h im for hire. O verall, 66.5% of a p p lic a n ts w ere reco m m en d ed for hire. Since th e y w ere e v a lu a tin g tw o a p p lic a n ts (one a t a tim e), it w as possible a t this stage for p a rtic ip a n ts to reco m m en d h irin g b o th ap p lic a n ts, a l­ th o u g h m o st d id n o t— 112 p a rtic ip a n ts reco m m en d ed only one of tw o a p p lic a n ts for hire, w h e re a s 7 reco m m en d ed none a n d 69 reco m m en d ed b o th . T h e m a in em p irical p re d ic tio n s are th a t m o th e rs w ill be offered low er sta rtin g salaries, be ra te d as less p ro m o ta b le , be less likely to be reco m m en d ed for m a n a g e m e n t, a n d be less likely to be reco m m en d ed for hire.

LABORATORY E X PE R IM E N T RESULTS

We first ad d re ss the ce n tra l q u estio n of w h e th e r m o th ers face u n iq u e d isa d v a n ta g e s in w o rk p lace e v alu atio n s. We th e n c o n d u c t an c illa ry a n a l­ yses to see if p a rtic ip a n t g en d er or a p p lic a n t race q u alify the results. Finally, w e conclude w ith an analysis d esigned to assess w h e th e r the com petence a n d co m m itm e n t ra tin g s of a p p lic a n ts m ed iate the effect of m o th e rh o o d on w o rk p lace ev alu atio n s.

Is T h e re a M o th e rh o o d P enalty?

T able 1 p ro v id e s m ean s or p ro p o rtio n s of the p a r tic ip a n ts ’ ra tin g s of the ap p lic a n ts, along w ith co rresp o n d in g p a ire d /-tests to co m p are m ean s a n d Z -te sts to co m p are p ro p o rtio n s. A t th is stage we pooled the d a ta for m ale a n d fem ale su b jects a n d for A frican -A m erican a n d w h ite a p p lic a n ts to h ig h lig h t the m ain co m p ariso n m o tiv a tin g this stu d y — the co m p ariso n of m o th e rs to n o n m o th ers. In the m u ltiv a ria te m odels below, w e decom pose the resu lts b y su b je c t g en d er a n d race of a p p lican t.

R a tin g s o f m others a n d n o n m o th e rs.— T h e first tw o colum ns of table 1 co m p are the ra tin g s of fem ale a p p lic a n ts w ho are m o th e rs w ith those

1315

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

T A B L E 1 M e a n s o r P r o p o r t i o n s o f S t a t u s , S t a n d a r d s , a n d E v a l u a t i o n Va r i a b l e s b y

G e n d e r a n d P a r e n t a l S t a t u s o f A p p l i c a n t

F e m a l e A p p l i c a n t s M a l e A p p l i c a n t s

M others N onm others F athers N onfathers

C om petence .......................... S.19** 5.75 5.51 5.44 (.73) (.58) (.68) (.66)

C om m itm ent ........................ 67.0** 79.2 78.5** 74.2 (19.1) (15.2) (16.3) (18.6)

Davs allow ed late ............... 3.16** 3.73 3.69** 3.16 (1.98) (2.01) (2.55) (1.85)

% score required on exam .................................. 72.4** 67.9 67.3 67.1

(27.5) (27.7) (32.7) (33.0) Salary recom m ended ($) ... 137,000** 148,000 150,000** 144,000

(21,000) (25,000) (23,000) (20,700) Proportion recom m end for

m a n a g e m e n t..................... ,691 + + .862 .936 + .851 L ikelihood of prom o­

tion ..................................... 2.74** 3.42 3.30* 3.11 (.65) (.54) (.62) (.70)

Proportion recom m end for hire ..................................... ,468++ .840 .734+ .617

N o t e .—SD s in parentheses. 94 p a rtic ip a n ts ra te d fem ale ap p lican ts, a n d 94 ra te d m ale applicants. F o r th is table, the d a ta for m ale a n d fem ale subjects are pooled, as are th e d a ta by race o f applicant. All values rep o rted to th ree significant digits. See te x t for v ariab le descriptions.

+ Z < .10, te st for difference in p ro portions b etw een p a re n ts a n d nonparents. ++ Z < .05. * P < .10, te st for difference in m eans b etw een p aren ts a n d n o n p arents. ** P < . OS.

w ho are n o n m o th ers. As p re d ic te d , m o th e rs w ere ju d g e d as significantly less c o m p e te n t a n d c o m m itted th a n w om en w ith o u t children. T h e com ­ p etence ra tin g s are ap p ro x im a te ly 10% low er for m o th e rs th a n for n o n ­ m o th ers, a n d the co m m itm e n t ra tin g s are a b o u t 15% lower. M o th e rs w ere also h eld to h a rs h e r p erfo rm an ce a n d p u n c tu a lity sta n d a rd s. M o th e rs w ere allow ed significantly few er tim es of b ein g late to w o rk , a n d th e y need ed a significantly h ig h e r score on the m a n a g e m e n t exam th a n n o n m o th e rs before bein g consid ered h irab le.

Sim ilarly, the e v a lu a tio n m easu res show significant a n d s u b sta n tia l p e n ­ alties for m o th erh o o d . T h e reco m m en d ed sta rtin g sa la ry for m o th e rs w as $11,000 (7.4%) less th a n th a t offered to n o n m o th ers, a significant differ­ ence. M o th e rs w ere also ra te d as significantly less p ro m o ta b le a n d w ere less likely to be reco m m en d ed for m an a g e m e n t. Finally, w hile p a rtic ip a n ts reco m m en d 84% of fem ale n o n m o th e rs for hire, th e y reco m m en d a sig­ n ifican tly low er 47% of m o th ers. R ecall th a t w h e n the résum és for the

1316

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

tw o a p p lic a n ts w ere p re te ste d w ith o u t a n y p a re n ta l sta tu s m an ip u la tio n s, no significan t differences w ere fo u n d in how th ey w ere ra te d , suggesting th a t the m o th e rh o o d m a n ip u la tio n p ro d u c e d the low er ra tin g s fo u n d here.

R a tin g s o f fa th e r s a n d n o n fa th e rs .— T h e la st tw o colum ns of tab le 1 c o m p are the ra tin g s of m ale a p p lic a n ts w h o are fa th e rs w ith those w ho are n o n fa th e rs. O u r th e o ry p re d ic te d th a t fa th e rs w o u ld n o t experience a fath e rh o o d penalty, a n d o u r resu lts are co n sisten t w ith this pred ictio n . In fact, fa th e rs w ere a c tu a lly a d v a n ta g e d on som e of these m easures. F o r exam ple, a p p lic a n ts w ho w ere fa th e rs w ere ra te d significantly m ore com ­ m itte d to th e ir jo b th a n n o n fa th e rs. F a th e rs w ere allo w ed to be late to w o rk significantly m ore tim es th a n n o n fath ers. Finally, th e y w ere offered significantly h ig h e r salaries th a n n o n fath ers.

M u ltiv a ria te A nalysis

We n o w tu rn to m u ltiv a ria te m odels to e v a lu ate the m o th e rh o o d p e n a lty h y p o th esis b y e stim atin g the effects of g en d er of a p p lic a n t, p a re n ta l statu s, a n d the in te ra c tio n of g en d er of a p p lic a n t w ith p a re n ta l sta tu s on each of the eig h t d e p e n d e n t v ariab les. We refer to the in te ra c tio n te rm (gender of a p p lic a n t x p a re n ta l statu s) as the “m o th e rh o o d p e n a lty in te ra c tio n .” A p p lic a n t race a n d p a r tic ip a n t g en d er are in c lu d e d in all m odels, a n d sta n d a rd erro rs are clu stered b y p a rtic ip a n t ID to tak e into a c c o u n t the n o n in d e p e n d e n c e of o b se rv a tio n s th a t resu lts from ask in g p a rtic ip a n ts to ra te a p p lic a n ts in p a irs.9 L in e a r regression m odels are u sed for the con­ tin u o u s d e p e n d e n t v ariab les. L ogistic regression m odels are estim a te d for the b in a ry e v a lu a tio n v a ria b le s (recom m end for m a n a g e m e n t a n d rec­ om m en d for hire). O rd e re d logistic regression, w ith the p ro p o rtio n a l odds specification, is u sed for the o rd e re d categorical e v a lu a tio n v a ria b le , lik e­ lihood of pro m o tio n . P a re n ta l sta tu s, g en d er of a p p lic a n t, g en d er of p a r­ ticip a n t, a n d race of a p p lic a n t are d u m m y v ariab les, w ith p a re n ts, fem ales, a n d A frican -A m erican s coded as 1.

T h e e stim a te d regression coefficients are p re se n te d in tab les 2 , 3 , a n d 4. F o r all eig h t d e p e n d e n t v a ria b le s, the m o th e rh o o d p e n a lty in te ra c tio n is significant a n d is in the p re d ic te d d irection. B ased on this result, we conclude th a t there is stro n g s u p p o rt for the m a in p re d ic tio n th a t p a re n ta l sta tu s n eg ativ ely im p a c ts ra tin g s for fem ale, b u t n o t m ale, a p p lican ts. We n o w describe m ore precisely the effect of m o th e rh o o d sta tu s on each of the d e p e n d e n t v ariab les.

C o m m itm e n t a n d co m p eten ce.— T able 2 c o n tain s estim a te d regression

9 U s in g M p lu s 3.1, ro b u s t s ta n d a r d e rro rs a re c o m p u te d u sin g th e s a n d w ic h e stim a to r, w h ic h ta k e s in to a c c o u n t n o n in d e p e n d e n c e o f o b se rv a tio n s (M u th é n a n d M u th é n 1 998- 2004).

1317

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

T A B L E 2 E s t i m a t e d R e g r e s s i o n C o e f f i c i e n t s f o r t h e E f f e c t s o f G e n d e r ,

P a r e n t a l S t a t u s , a n d R a c e o n A p p l i c a n t ’s P e r c e i v e d C o m p e t e n c e

a n d C o m m i t m e n t

In d e p en d e n t Variables Competence C om m itm ent

P a re n t .......................................... .089 5.15 *** (.088) (1.73)

Fem ale app lic an t ....................... .376 *** 5.68 ** (.104) (2.51)

A frican-A m erican ..................... -.0 3 8 - 2 . 0 1 (.090) (2.27)

Fem ale p a rtic ip a n t ................... .060 -2 .6 1 (.094) (2.26)

M otherhood interaction" .......... -.7 5 5 0 *** — 17 3 *** (.132) (2.32)

I n t e r c e p t ....................................... 5.42 *** 75.8 *** (.1 0 0 ) (2.55)

N o t e .—Ro b u st SE s in parentheses; clustered by p a rtic ip a n t ID . iV = 1 8 8 p articip an ts. A ll values rep o rted to three significant digits. See te x t for v ariab le descriptions.

a P a re n t x fem ale applicant. * P c . 1 0 .

** P < .05. *** P < .001.

coefficients a n d ro b u s t s ta n d a rd erro rs for the effects of the in d e p e n d e n t v a ria b le s on com petence a n d co m m itm en t. In o rd er to d eterm in e the m a g n itu d e of the effect of m o th e rh o o d sta tu s on the d e p e n d e n t v a ria b le s it is n ecessary to consider the a d d itiv e effects of p a re n ta l sta tu s, gender of a p p lic a n t, a n d the m o th e rh o o d p e n a lty in teractio n .

C o n firm in g o u r p red ictio n , m o th e rs are v iew ed as less c o m p e te n t th a n n o n m o th ers. As sh o w n in the le ft-h a n d colum n of tab le 2, the m o th e rh o o d p e n a lty in te ra c tio n is significan t a n d n eg ativ e, in d ic a tin g th a t bein g a p a r e n t low ers the com petence ra tin g s for w om en, b u t n o t m en. T h e fem ale a p p lic a n t d u m m y v a ria b le is significant a n d positive, im p ly in g th a t w o m en w ith o u t ch ild ren are ra te d as m ore c o m p e te n t th a n m en w ith o u t children. W hile th is finding w as n o t p red icted , one can im agine several reasons w h y w o m en w ith o u t ch ild ren m ig h t be ra n k e d h ig h e r th a n m en w ith o u t ch ild ren in this setting. C u ltu ra l co n stru ctio n s of g en d er often inclu d e beliefs th a t w om en w a n t (or even “n e e d ”) ch ild ren to feel fulfilled. As a result, p a rtic ip a n ts m a y assum e th a t w o m en w ho h a v e a p p a re n tly forgone ch ild b e a rin g to en te r the la b o r m a rk e t are e x tra o rd in a rily com ­ m itte d to w ork. In co n tra st, because m en are n o t expected to “n e e d ” children, th is in fo rm a tio n does n o t c a rry the sam e im p a c t for m en.

A second possib ility is th a t p a rtic ip a n ts m a y ra te n o n m o th e rs m ore h ig h ly as a w a y to co m p en sate for th e ir d iscrim in atio n a g a in st m others.

1318

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

R esearch h a s show n th a t people te n d to follow an im p licit stra te g y of m a in ta in in g “m o ral c re d e n tia ls” in w h ich in d ulgence in d iscrim in atio n to w a rd one ta rg e t is coupled w ith a n tid isc rim in a to ry action to w a rd a n ­ o th er ta rg e t (M onin a n d M iller 2001). T h u s , because p a rtic ip a n ts ra te d a w o m an w ith o u t ch ild ren a n d a w o m a n w ith ch ild ren sim ultaneously, th ey m a y h a v e a tte m p te d to ju stify d isc rim in a tin g a g a in s t the m o th e r b y u n ­ consciously em bellishing th e ir ra tin g s of the non m o th er. T h e h ig h e r ra tin g for w om en w ith o u t ch ild ren m a y th u s occur because p a rtic ip a n ts are forced to choose b e tw een a m o th e r a n d a non m o th er. We are n o t able to a d ju d ic a te b etw een these or o th er possible acco u n ts w ith the av ailab le d a ta .10 H ow ever, these acco u n ts are em p irically testab le a n d m e rit con­ sid eratio n in fu tu re research. We also note th a t this re su lt does n o t con­ tra d ic t o u r p rim a ry claim th a t m o th e rs face e v a lu a tiv e biases in the w orkplace.

P a rtic ip a n ts also p erceiv ed m o th ers as less c o m m itted th a n o th er a p ­ plicants: the m o th e rh o o d p e n a lty in te ra c tio n is significant a n d n egative in the m odel p re d ic tin g co m m itm e n t ra tin g s (rig h t-h a n d colum n of table 2). In this m odel, the fem ale a p p lic a n t coefficient is ag ain significant a n d positive, p ro v id in g fu rth e r evidence th a t w om en w ith o u t ch ild ren expe­ rience w h a t m ig h t be called a “childless b o n u s .” T h e re is also a positive a n d sig nificant m a in effect for p a re n ta l sta tu s, im p ly in g th a t fa th e rs are a c tu a lly ra te d as m ore c o m m itted th a n n o n fa th e rs b y a b o u t 5 percen tag e points. M o th e rs, b y co n tra st, suffer a re d u c tio n of a b o u t 6.4 percen tag e p o in ts in th eir co m m itm e n t ra tin g s co m p a re d w ith childless m en (sum of the m ain effects of p a re n ta l sta tu s a n d a p p lic a n t g en d er a n d the in te ra c tiv e term ) or a b o u t 12.1 p ercen tag e p o in ts co m p a re d w ith childless w om en (sum of the m a in effect of p a re n ta l sta tu s a n d the in te ra c tiv e term ).

S ta tu s ch ara c teristic s trig g er beliefs a b o u t p e rfo rm a n c e capacity, a n d these beliefs derive fro m exp ectatio n s a b o u t a n tic ip a te d effo rt a n d ab ility (com petence). O u r resu lts show th a t m o th e rs are n o t only v iew ed as less c o m m itted to p a id w o rk , th ey are also seen as h a v in g less w o rk p lace ability. T h e decreased com petence a n d co m m itm e n t ra tin g s for m o th ers suggest th a t m o th e rh o o d o p erates as a s ta tu s ch aracteristic.

10 A n a n o n y m o u s re v ie w e r su g g e ste d a n a d d itio n a l c o m p o n e n t of s ta tu s -b a s e d d is­ c rim in a tio n a g a in st m o th e rs. I t m a y be th e case t h a t em p lo y e rs p e rce iv e c u rre n t fe rtility a s a sig n a l o f fu tu re fertility, a n d th u s a p p a r e n t c o m m itm e n t to th e la b o r m a rk e t. T h e larg e g a p b e tw e e n w o m e n w ith c h ild re n a n d w o m e n w ith o u t c h ild re n m a y o c c u r in p a r t b e ca u se em p lo y e rs e x p ec t w o m e n w ith c h ild re n to h a v e m o re c h ild re n a n d d e ­ crease a tta c h m e n t to th e la b o r m a rk e t a s a re su lt. F u rth e rm o re , a s th is re v ie w e r p o in te d o u t, if c u rre n t fe rtility fu n c tio n s a s a sig n a l of fu tu re fertility, w e w o u ld n ot expect ra tin g s o f c o m p e ten c e a n d c o m m itm e n t to fully m e d ia te th e re su lts. T h is is b e ca u se e v a lu a to rs m a y e xpect p ro d u c tiv ity to d ecline m ore in th e fu tu re fo r m o th e rs (upon h a v in g m o re c h ild ren ) t h a n fo r n o n m o th e rs. E q u a l ra tin g s o f c u rre n t p ro d u c tiv ity th u s do n o t g u a ra n te e e q u a l e stim a te s o f fu tu re p ro d u c tiv ity .

1319

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

A b ility s ta n d a rd s.— A ccording to the sta tu s-b a se d d iscrim in atio n a r­ gu m en t, m o th ers w ill be h eld to stric te r s ta n d a rd s th a n o th e r k in d s of ap p lican ts. T able 3 c o n tain s the estim ates for the effects of the in d e p e n d e n t v a ria b le s on the a b ility s ta n d a rd item s. C o n siste n t w ith p red ictio n s, m o th ­ ers w ere h eld to a stric te r p erfo rm an ce sta n d a rd . T h e m o th e rh o o d in te r­ action is significant a n d p ositive in the m odel p re d ic tin g the re q u ire d te st score, w hile the m ain effects of gender of a p p lic a n t a n d p a re n ta l sta tu s are in sig n ifican t (rig h t-h a n d colum n of tab le 3). T h is show s th a t p a rtic ­ ip a n ts w o u ld req u ire m o th e rs (b u t n o t fath ers) to score h ig h er on a test of m a n a g e m e n t a b ility th a n o th er a p p lic a n ts before co n sidering th e m for a job.

P a rtic ip a n ts w ere also a sk e d h o w m a n y d ays an a p p lic a n t could be late or leave early before th e y w o u ld no longer consider h irin g them . We expected th a t p a rtic ip a n ts w o u ld allow m o th e rs less flexibility th a n o th er types of em ployees. As can be seen in the le ft-h a n d colum n of tab le 3, this p re d ic tio n is s u p p o rte d b y the significan t a n d n eg ativ e m o th e rh o o d p e n a lty in teractio n . T h e re is also a significant, p ositive m a in effect for bein g a p a r e n t a n d for bein g fem ale. T h u s, childless w om en a n d fath ers are allow ed to be late m ore fre q u e n tly w ith o u t its im p a c tin g th e ir p e r­ ceived s u ita b ility for hire. H ow ever, m o th e rs are e v id en tly h eld to a h ig h er s ta n d a rd of p u n ctu ality , bein g allow ed few er d ays of b ein g late.

Workplace e v a lu a tio n s.— In tab le 4, the m o th e rh o o d p e n a lty in te ra c tio n is significant a n d neg ativ e across all four m odels, in d ic a tin g th a t m o th ers, relativ e to o th e r ap p lic a n ts, are believ ed to deserve low er salaries a n d to be less su itab le for hirin g , p ro m o tin g , a n d tra in in g for m an a g e m e n t. In the m odel p re d ic tin g likelihood of p ro m o tio n , the m ain effect of p a re n ta l s ta tu s is m a rg in a lly significant a n d p ositive, w hile the m o th e rh o o d p e n a lty in te ra c tio n is significant a n d neg ativ e, in d ic a tin g th a t the n eg ativ e effect of p a re n ta l sta tu s on p e rcep tio n s of p ro m o ta b ility accrues only to w om en.

M o th e rs are also less likely th a n o th er types of a p p lic a n ts to be rec­ o m m en d ed for m a n a g e m e n t (second colum n of tab le 4). I f we c o n v e rt the regression coefficient for the p a re n ta l sta tu s v a ria b le to an odds ratio, fa th e rs are 1.83 tim es m ore likely to be reco m m en d ed for m a n a g e m e n t th a n childless m en, a difference th a t is m a rg in a lly significant. F o r fem ale a p p lic a n ts, childless w o m en are 8.2 tim es m ore likely th a n m o th e rs to be reco m m en d ed for m a n a g e m e n t.11

B eing a m o th e r also low ers the odds of an a p p lic a n t’s b ein g recom ­ m e n d e d for hire. T h e m a in effect of p a re n ta l sta tu s is n o t significant, in d ic a tin g th a t on th is m easu re fa th e rs are n o t a d v a n ta g e d over m en

11 T h e in v erse log of .605—2.716 = . 122. To sta te as th e o d d s fo r ch ild less w o m e n c o m ­ p a re d to m o th e rs, w e in v e rte d th is ra tio (1/.122 = 8.2). A sim ila r c a lc u la tio n is m a d e fo r th e o d d s of b e in g h ired.

1320

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

T A B L E 3 E s t i m a t e d R e g r e s s i o n C o e f f i c i e n t s f o r t h e E f f e c t s o f G e n d e r , P a r e n t a l

S t a t u s , a n d R a c e o n A b i l i t y S t a n d a r d Va r i a b l e s

Independent Variables Days Allowed Late T est Score R equired (%)

Parent ............................. .SIS *** 1.03 (.137) (.968)

Female applicant ............. .572 ** 1 .2 S (.294) (4.S2)

African-American ............ - .3 6 1 - 4 .0 6 (.294) (4.38)

Female participant .......... .234 — 9 . 4 4 ** (.289) (4.30)

Motherhood interaction3 ... 1 o * * * 3.56 *** (.213) (1 .2 1 )

Intercept........................... 3.22 *** 7 3 7 *** (.322) (4.27)

N o t e .—Ro b u st S E s in parentheses; clustered by p a rtic ip a n t ID. A = 1 8 8 particip an ts. A ll values re­ p o rte d to th ree significant digits. See te x t for v ariab le descriptions.

a P a re n t x fem ale applicant. * P c . 1 0 .

** P < . OS. *** P < .001.

w ith o u t children. T h e m ain effect for bein g a fem ale a p p lic a n t is again h ig h ly significant. S ta te d as an odds ratio , childless w o m en are 3.35 tim es m ore likely to be reco m m en d ed for hire th a n childless m en. H ow ever, childless w o m en are especially a d v a n ta g e d co m p a re d to m o th ers, being over six tim es m ore likely to be reco m m en d ed for hire.

M o th e rs are also offered low er sta rtin g salaries th a n o th er types of a p p lic a n ts, as in d ic a te d b y the significant, neg ativ e coefficient for the m o th e rh o o d in te ra c tio n term . U sing the v a lu e s in tab le 4 to calculate p re d ic te d valu es, w e find th a t childless m en w ere reco m m en d ed an a v ­ erage sa la ry of a p p ro x im a te ly $148,000.12 F a th e rs w ere offered a signifi­ c a n tly h ig h er sa la ry of ap p ro x im a te ly $152,000. In the p a st, em ployers legally p a id fa th e rs a “fam ily w a g e ” th a t w as h ig h e r to acco m m o d ate th eir su p p o sed b re a d w in n e r role. C u ltu ra l beliefs em p h asizin g the im p o rtan ce of m ale -h e a d ed h o u seh o ld s p ro v id e d the n o rm a tiv e u n d e rp in n in g s of this “fa th e rh o o d b o n u s ” (O rloff 1996). T h e resu lts suggest th a t, w hile the fam ­ ily w age is fo rm ally e x tin ct in the U n ite d S tates, it m a y in fo rm ally surv iv e in the fo rm of sa la ry p re m iu m s osten sib ly m o tiv a te d b y p ro d u ctiv ity . F o r

13 T h e se p re d ic te d v a lu e s a re for m ale su b je c ts (g e n d er o f s u b je c t = 0) a n d w h ite a p ­ p lic a n ts (A frica n -A m e ric an = 0). T h e sa m e p a tte rn of p re d ic te d v a lu e s is fo u n d w h e n c a lc u la tio n s a re m a d e w ith fem ale su b je c t a n d /o r A fric a n -A m e ric a n a p p lic a n t d a ta . T h a t is, re g a rd le s s o f g e n d e r o f s u b je c t o r ra ce of a p p lic a n t, m o th e rs a re offered sig n ific a n tly lo w e r s ta rtin g salaries.

1321

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

This content

dow nloaded

from 131.94.16.10

on Fri,

1 Aug

2014 16:08:30

PM

A ll use

subject to JSTO

R Term

s and

C onditions

T A B L E 4 E s t i m a t e d R e g r e s s i o n C o e f f i c i e n t s f o r t h e E f f e c t s o f G e n d e r , P a r e n t a l S t a t u s , a n d R a c e o n E v a l u a t i o n Va r i a b l e s

In d e p en d e n t Variables P rom otion Likelihood

(O rdered Logistic Estim ates) M anagem ent Training?

(Binary Logistic E stim ates) Hire?

(Binary Logistic E stim ates)

Recom m ended Salary in T housands

of Dollars (Linear E stim ates)

P a re n t .................................. 1.03 * .605 * .570 4 4 7 *** (.545) (.321) (.366) (1.84)

Fem ale app lic an t ............... .256 1.009 *** 1 ? 1 *** 2.56 (.425) (.319) (.365) (3.18)

A frican-A m erican ............. .309 - . 2 1 1 -.1 6 3 - 6 .8 0 ** (.299) (.218) (.197) (2.94)

Fem ale p a rticip a n t ........... .496 * .526 ** .606 *** .691 (.298) (.226) (.199) (2.82)

M otherhood interaction" .. — ? 14 *** — 2 72 *** -2 .3 8 *** _ 1 5 9 l b * * (.651) (.426) (.548) (2.42)

I n t e r c e p t ............................... b 4.56 *** . 2 1 0 148*** (.601) (.266) (2.55)

N o t e .—Ro b u st SE s in parentheses; clustered by p a rtic ip a n t ID . iV = 1 8 8 p articip an ts. A ll values rep o rte d to th ree significant digits. See te x t for variable descriptions.

a P a re n t x fem ale applicant. b Since ord ered logistic regression produces m ultiple intercepts, we do n o t p re sen t th e m here. * P c . 1 0 .

** P < . OS. *** P < .001.

M o th e rh o o d P e n a lty

w om en, p a re n th o o d h a s the opposite effect. W om en w ith o u t ch ild ren w ere offered ap p ro x im a te ly $151,000, w h e re a s m o th e rs w ere reco m m en d ed a significantly low er sa la ry of a b o u t $139,000, or a b o u t 7.9% less th a n otherw ise eq u al childless w om en. C o m p a re d to fath ers, m o th e rs w ere offered a p p ro x im a te ly 8.6% low er salaries.

Also of note in the sa la ry m odel, A frican -A m erican a p p lic a n ts w ere offered ap p ro x im a te ly $6,800 low er salaries, on av erag e, c o m p a re d w ith those offered to w hites. T h is difference is strik in g , especially since A frican- A m erican s w ere n o t ju d g e d to be less c o m p e te n t or c o m m itted to w o rk — th ey w ere sim ply offered low er salaries. T h e résum és in the A frican- A m erican cond itio n s w ere exactly the sam e, except for first nam es, as those u sed in the w h ite conditions, so differences in qualificatio n s do n o t explain this finding. I t a p p e a rs th a t a d iffe re n t m e c h a n ism of d iscrim i­ n a tio n is o p e ra tin g for A frican -A m erican s, or a t le a st for a p p lic a n ts w ith distin ctiv e A frican -A m erican first nam es.

In sum , across all eig h t d e p e n d e n t v a ria b le s, the m o th e rh o o d p e n a lty in te ra c tio n is significant, a n d its sign is in the p re d ic te d d irection. G iven the stre n g th of the effect across a diverse set of m easu res a n d the exper­ im e n ta l co ntrol of a p p lic a n t quality, we conclude th a t giving evidence of bein g a m o th e r lead s to d iscrim in atio n a g a in st m o th ers. B eing a fa th e r d id n o t le a d to sim ilar d isa d v a n ta g e s for m en an d , a t tim es, actu a lly led to a d v a n ta g e s. We n o w tu rn to a b rie f discussion of a few ad d itio n a l research q uestions, sta rtin g w ith w h e th e r g en d er of p a rtic ip a n t im p acts these results.

D oes P a rtic ip a n t G en d er Im p a c t the Size of the M o th e rh o o d P en alty ?

C o n siste n t w ith sta tu s c h a ra c teristic s theory, we expected b o th m ale a n d fem ale p a rtic ip a n ts to d iscrim in ate a g a in st m o th ers. U sing o u r d a ta to te st this claim show s th a t even th o u g h fem ale p a rtic ip a n ts ra te d ap p lic a n ts h ig h e r overall on som e m easures, b o th fem ale a n d m ale p a rtic ip a n ts ev a l­ u a te d m o th e rs significantly low er th a n n o n m o th e rs on all eight d e p e n d e n t v ariab les. O nly one significant difference w as fo u n d in the m a g n itu d e of the m o th e rh o o d p e n a lty b etw een m ale a n d fem ale p a rtic ip a n ts. In results n o t show n, we a d d e d the tw o -w ay in te ra c tio n of p a r tic ip a n t g en d er a n d a p p lic a n t g en d er a n d the th re e -w a y in te ra c tio n of p a rtic ip a n t gender, a p p lic a n t gender, a n d p a re n ta l s ta tu s to each of the m odels in tab les 2, 3, a n d 4. F o r m odels p re d ic tin g h o w m a n y d ays an a p p lic a n t w o u ld be allow ed to be late, the tw o -w ay in te ra c tio n of p a rtic ip a n t gender a n d a p p lic a n t g en d er w as m arg in a lly significan t a n d n eg ativ e, a n d the three- w a y in te ra c tio n of p a rtic ip a n t gender, a p p lic a n t gender, a n d p a r e n t sta tu s w as significant a n d positive. T h is m ean s th a t fem ale p a rtic ip a n ts h eld all fem ale a p p lic a n ts to a slightly h a rs h e r s ta n d a rd th a n m ale p a rtic ip a n ts

1323

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

did, allo w in g fem ale a p p lic a n ts few er days of bein g late th a n m ale a p ­ p lican ts, b u t th ey p en alized m o th e rs slightly less relativ e to childless w o m en th a n m ale p a rtic ip a n ts did. F o r all o th er d e p e n d e n t v ariab les, the m a g n itu d e of the m o th e rh o o d p e n a lty d id n o t differ significantly for m ale a n d fem ale p a rtic ip a n ts. We d id n o t expect th a t the sta tu s-b a se d d iscrim ­ in a tio n m e c h a n ism w o u ld w o rk d ifferen tly for m ale a n d fem ale p a rtic i­ p a n ts, a n d the resu lts are larg ely c o n sisten t w ith th a t pred ictio n .

D o A frican -A m erican s a n d W h ites B o th E x p erien ce a M o th e rh o o d P en alty ?

We p re d ic te d th a t b o th w h ite a n d A frican -A m erican m o th e rs w o u ld ex­ p erience a m o th e rh o o d p e n a lty co m p a re d w ith th e ir sam e-race, childless c o u n te rp a rts. To e v a lu a te th is p red ictio n a n d to co m p are the m a g n itu d e of the m o th e rh o o d p e n a lty for the tw o groups, w e a d d e d the th ree-w ay in te ra c tio n of a p p lic a n t race, p a re n ta l sta tu s, a n d a p p lic a n t g en d er to each of the m odels d escrib ed above (results n o t show n). T h e m o th e rh o o d p e n a lty in te ra c tio n re m a in s significant in each of the m odels w h e n the th re e -w a y in te ra c tio n is ad d ed , in d ic a tin g th a t reg ard less of race, m o th ers experience n eg ativ e biases in w o rk p lace e v alu atio n s. F u rth e r, the three- w a y in te ra c tio n w as significan t in only one of the m odels. A frican -A m er­ ican m o th e rs w ere ra te d as less likely to be p ro m o te d th a n w h ite m others, b u t none of the o th e r th re e -w a y in te ra c tio n s w ere significant. T h u s, d a ta fro m the ex p e rim e n t su ggest th a t A frican -A m erican w om en a n d w hite w o m en b o th experience a m o th e rh o o d penalty, a n d the m a g n itu d e of th a t p e n a lty is larg ely the sam e for b o th groups.

D o C om petence a n d C o m m itm e n t R a tin g s M e d ia te W orkplace E v a lu a tio n s?

T h u s far, we h a v e show n (1) th a t m o th e rh o o d is a s ta tu s ch ara c teristic (a tr a it w ith differen tially v a lu e d states th a t im p a c ts p e rfo rm a n c e expec­ tations) a n d (2) th a t m o th e rh o o d d isa d v a n ta g e s jo b a p p lic a n ts across d i­ verse m easures. To com plete o u r a rg u m e n t, we n eed to give evidence th a t m o th e rh o o d d isa d v a n ta g e s jo b a p p lic a n ts because it is a s ta tu s c h a ra c ­ teristic. To ev a lu a te this a rg u m e n t, we a d d e d the com petence a n d com ­ m itm e n t m easu res as in d e p e n d e n t v a ria b le s to the m odels p re d ic tin g w o rk p lace e v a lu a tio n s (see tab le 5). A ccording to the theory, em ployers h a v e low er exp ectatio n s for the w o rk p lace com petence a n d c o m m itm en t of m o th ers, a n d it is th is low er exp ectatio n th a t lead s th e m to d iscrim in ate a g a in st m o th e rs in hirin g , p ro m o tio n , a n d sa la ry decisions. I f the th eo ry is correct, th e n e v a lu a tio n s of com petence a n d co m m itm e n t shou ld m e­ diate the m o th e rh o o d penalty. N o t surprisingly, h ig h er com petence ratin g s

1324

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

This content

dow nloaded

from 131.94.16.10

on Fri,

1 Aug

2014 16:08:30

PM

A ll use

subject to JSTO

R Term

s and

C onditions

T A B L E 5 E s t i m a t e d R e g r e s s i o n C o e f f i c i e n t s f o r t h e M e d i a t i o n o f C o m p e t e n c e a n d C o m m i t m e n t o n t h e I m p a c t o f P a r e n t a l S t a t u s

o n W o r k p l a c e E v a l u a t i o n s

In d e p e n d e n t Variables

Prom otion L ikelihood

(O rdered Logistic Estim ates)

M anagem ent Training

(Binary Logistic E stim ates)

Hire? (B inary Logistic

E stim ates)

R ecom m ended Salary in T housands

of Dollars (L inear Estim ates)

C om petence ............................................... .628 ** 1.263 *** 1 ? 1 lb** o o * * *

(.295) (.281) (.258) (1.99) C o m m itm e n t ............................................. 237 *** .206 ** .308 *** 1.08

(.095) (.099) (.081) (.762) P a r e n t .......................................................... .901 * .508 .433 3.23 *

(.558) (.340) (.426) (1.78) Fem ale a p p lic an t ..................................... -.1 4 0 .661 ** .755 * -.8 1 7

(.426) (.332) (.410) (3.31) A frican-A m erican ..................................... .374 - .1 5 4 -.0 9 2 - 6 .3 0 ***

(.319) (.237) (.244) (2 .8 6 ) Fem ale p a rtic ip a n t .................................. .557 * .606 *** .755 *** .512

(.316) (.236) (.254) (2.81) M otherhood interaction 3 ........................ -1 .3 4 ** — 1.89 *** - 1 .3 9 ** -8 .5 2 ***

(.646) (.437) (.606) (2 .6 6 ) In te rc e p t ..................................................... b 3.64 *** —2 09 *** 140***

(.947) (.702) (6.37) R eduction o f m otherhood penalty (%) 37.4 30.5 41.6 46.4

N o t e . —R o b u st SE s in p arentheses, clustered by p a rtic ip a n t ID. N = 188 p articip an ts. All values rep o rte d to three significant digits. See te x t for v ariable descriptions. a P a re n t x fem ale applicant. b Since ord ered logistic regression produces m ultiple intercepts, w e do n o t p re sen t th e m here. * P c . 1 0 .

** P < . OS. * * * P c . 0 0 1 .

A m erican J o u rn a l of Sociology

lead to significantly h ig h e r reco m m en d ed sta rtin g salaries, h ig h er p e rc e p ­ tions of a p p lic a n t prom o tab ility , a n d in creased odds of reco m m en d in g the a p p lic a n t for m a n a g e m e n t a n d for hire. Sim ilarly, h ig h er c o m m itm en t ra tin g s w ere associated w ith significantly h ig h e r p e rcep tio n s of a p p lic a n t p ro m o ta b ility a n d in creased odds of reco m m en d in g the a p p lic a n t for m a n ­ ag e m e n t a n d for hire, alth o u g h the im p a c t of co m m itm e n t ra tin g on salary, w hile positive, w as n o t significant.

M o re im p o rta n t, w h e n the com petence a n d co m m itm e n t ra tin g s w ere a d d e d to the m odels, the n eg ativ e effect of m o th e rh o o d sta tu s on w o rk ­ place ev a lu a tio n s w as significantly re d u c e d .13 As can be seen on the la st row of tab le 5, the m ag n itu d e of the m o th e rh o o d p e n a lty w as red u ced b y 46% in the sa la ry m odel, b y 31% in the re c o m m en d -fo r-m an ag em en t m odel, b y 37% in the p ro m o tio n likelih o o d m odel, a n d b y 42% in the recom m end-for-hire m odel. C o n siste n t w ith th e o retical p red ictio n s, com ­ p etence a n d co m m itm e n t do m ed iate, a t le a st p artially, the neg ativ e effect of m o th e rh o o d sta tu s on w o rk p lace e v alu atio n s. In p a rt, m o th e rs are ra te d as less h irab le, less su itab le for p ro m o tio n a n d m a n a g e m e n t tra in in g , a n d d eserv in g of low er salaries because th e y are b elieved to be less c o m p eten t a n d less c o m m itted to p a id w ork.

T h e neg ativ e effects of m o th e rh o o d sta tu s w ere n o t com pletely elim i­ n a te d w h e n the com petence a n d co m m itm e n t m easu res w ere in c lu d e d in the m odels, how ever. P e rh a p s this is n o t su rp risin g given the m a g n itu d e of the m o th e rh o o d p e n a lty in the orig in al m odels. W hile we can only speculate on w h y th is resid u al effect rem ain s, it is possible th a t one or m ore a d d itio n a l d isc rim in a to ry m ech an ism s are a t w ork. F o r exam ple, some e v a lu a to rs m a y believe th a t m o th e rs shou ld n o t be in the w o rk p lace, b u t sh o u ld in ste a d be a t hom e w ith th e ir children. I f so, th ey m a y view m o th e rs as c o m p e te n t a n d c o m m itted to p a id w o rk , b u t still discrim in ate a g a in st them . T h a t is, in a d d itio n to sta tu s-b a se d d iscrim in atio n , some e v a lu a to rs m a y also engage in n o rm a tiv e d iscrim in atio n , in w h ich th ey recognize the com petence of m o th ers b u t believe th a t it is th e ir d u ty to rem ain a t hom e w ith th e ir children. F u tu re research th a t e x p erim en tally m a n ip u la te s com petence a n d co m m itm e n t levels is need ed to e v a lu a te this a lte rn a tiv e m echanism .

13 To c a lc u la te w h e th e r th e re d u c tio n s in th e m a g n itu d e s of th e coefficients a re sig n if­ ic a n t, w e follow th e p ro c e d u re s fo r c o m p a rin g re g re ssio n coefficients b e tw e e n n ested m o d els a s d e sc rib e d by C logg, P e tk o v a , a n d H a r ito u (1995). We in clu d e b o th m e d ia to rs in a single m o d el b e ca u se b o th a re in d ic a to rs th a t m o th e rh o o d is a s ta tu s c h a ra c te ris tic , o u r th e o re tic a l m e d ia tin g v a ria b le . E a c h m e d ia to r also re d u c e s th e m o th e rh o o d p e n a lty w h e n in c lu d e d in th e m o d el s e p a ra te ly (resu lts a v a ila b le u p o n request).

1326

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

D o A ctu al E m p lo y ers D iscrim in ate a g a in st M o th ers?

I t is w o rth re tu rn in g to the q u estio n of h o w the use of u n d e rg ra d u a te stu d e n ts as e v a lu a to rs im p a c ts o u r findings. One concern w ith stu d e n ts is th a t, lack in g w o rk p lace a n d h irin g experience, th e y m ig h t be m ore likely to rely on stereotypes w h e n m a k in g h irin g decisions. I f u n d e rg ra d u a te s are m ore likely to rely on stereo ty p es th a n em ployers, o u r s tu d y will ov erestim ate the m a g n itu d e of the m o th e rh o o d penalty. C onversely, b e ­ cause of th e ir la c k of w o rk p lace experience, stu d e n ts are less likely th a n people w ith m ore extensive experience to h a v e h a d a “b a d ex p erien ce” w ith an em ployed m o th e r a n d to generalize fro m this experience w h en ev a lu a tin g o th er em ployees w h o are m o th ers. F u rth e r, to the e x te n t th a t y o u n g er a d u lts h o ld m ore eg a lita ria n g en d er beliefs, th ey m ig h t be less likely th a n older a d u lts to d iscrim in ate a g a in st m o th ers. O u r stu d y m ig h t th u s u n d e re stim a te the m a g n itu d e of the m o th e rh o o d penalty. W hile the la b o ra to ry d a ta are id eally su ited to e v a lu ate the m e c h a n ism of d iscrim ­ in a tio n , th e y c a n n o t estab lish the e x te n t to w h ich a c tu a l em ployers p e ­ nalize m o th e rs in the h irin g process. To ad d re ss this questio n , we tu rn to the a u d it study.

TH E AUDIT STUDY

O v erv iew

T h e a u d it m eth o d o lo g y com bines e x p erim en tal design w ith real-life set­ tings. As in la b o ra to ry ex p erim en ts ex am in in g d iscrim in atio n , a u d it s tu d ­ ies isolate a ch ara c teristic of in te re st (e.g., race or gender) a n d test for d isc rim in a to ry behavior. D istin c t from m o st la b o ra to ry studies, a u d it s tu d y p a rtic ip a n ts are the people w ho m ak e im p o rta n t decisions a b o u t ac tu a l ap p lic a n ts, such as em ployers co n d u ctin g n e w em ployee searches. W hile la b o ra to ry ex p erim en ts occur in m ore h ig h ly co n tro lled settings, th e re b y p e rm ittin g closer in v e stig a tio n of the social a n d cognitive p ro ­ cesses in v o lv e d in an a c t of d iscrim in atio n , a u d it stu d ies p ro v id e g reater g en eralizab ility of the results. L ocal a n d n a tio n a l o rg an izatio n s h a v e con­ d u c te d a u d its to m easu re d iscrim in atio n b a se d on g en d er a n d race in the h o u sin g a n d jo b m a rk e ts since the e sta b lish m e n t of the F a ir H o u sin g A ct (Title V III of the C ivil R ig h ts A ct of 1968; Y in g er 1986) a n d the C ivil R ig h ts A ct of 1964 (Fix, G alster, a n d S tru y k 1993; P ag er 2003; B e rtra n d a n d M u lla in a th a n 2003). T h e c u rre n t s tu d y is the first to use the a u d it m eth o d o lo g y to m easu re em p lo y m en t d iscrim in atio n b a se d on p a re n ta l status.

To c o n d u c t th is study, résum és a n d cover le tters fro m a p a ir of fictitious, eq u ally qualified, sam e-g en d er a p p lic a n ts (both fem ale or b o th m ale) w ere

1327

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

se n t to em ployers ad v e rtisin g for entry- a n d m id lev el m a rk e tin g a n d b u si­ ness jo b openings a t a large, N o rth e a s te rn city n e w sp a p e r over an 18- m o n th p erio d of tim e. J o b openings w ere ra n d o m ly assigned to eith e r the m ale p a ir or fem ale p a ir condition. As w ith the lab exp erim en t, the sam e- sex p a ir co n ta in e d one p a r e n t a n d one n o n p a re n t, a n d p a re n ta l sta tu s w as c o u n te rb a la n c e d across m em b ers of the pair. We m o n ito re d w h e th e r g en d er a n d p a re n ta l sta tu s im p a c t the odds th a t an em ployer w ill call b a c k an ap p lic a n t. Tw o a d d itio n a l fe a tu re s of the a u d it stu d y design increase o u r a b ility to co m p are it to the la b o ra to ry experim ent: (1) a p ­ p lic a n t résum és in the a u d it s tu d y w ere b a se d on the te m p lates used in the la b o ra to ry exp erim en t, a n d (2) the a c tu a l jo b s to w h ich ap p licatio n s w ere sen t w ere sim ilar to the m a rk e tin g position d escrib ed in the la b o ­ ra to ry experim ent.

T h e design for this s tu d y w as p a tte rn e d a fte r a re cen t a u d it stu d y th a t ex am in ed the effect of race on em p lo y m en t decisions (B e rtra n d a n d M ul- la in a th a n 2003). T h e a u th o rs fo u n d th a t, d e p e n d in g on the race of the ap p lic a n t, 5 % -8 % of th e ir h y p o th e tic a l a p p lic a n ts received a c a llb ack w h e n ap p ly in g for executive, m a n a g e m e n t, a n d sales positions. B ased on this c a llb ack ra te a n d to ensure th a t we h a d sufficient statistical po w er to e v a lu ate the effect of p a re n ta l sta tu s, w e s u b m itte d 1,2 76 résum és a n d cover le tters to 638 em ployers over the 18-m onth p erio d follow ing p ro ­ ced u res d escrib ed below.

Pro ced u re

E a c h w eek all entry- a n d m idlevel m a rk e tin g a n d b u siness jo b s th a t m a tc h e d our a p p lic a n ts ’ q u alificatio n s w ere selected from the S u n d a y edition of a n e w sp a p e r in a large N o rth e a s te rn city. O n av erag e, 13 job o penings w ere selected p e r w eek, w ith a m in im u m of one jo b a n d a m a x im u m of 39 jo b s selected in one w eek over 18 m o n th s. We ran d o m ly assigned selected jo b s to one of tw o cond itio n s (m ale or fem ale a p p lic a n t pair), a n d th en g en e ra te d a p a ir of sam e-g en d er résum és a n d cover letters. R ésum és w ere b a se d on the sam e te m p lates u sed in the la b o ra to ry ex­ p e rim e n t, p re se n tin g tw o a p p lic a n ts w ith u n in te rru p te d w o rk histories a n d e q u ally stro n g ed u c a tio n a l cred en tials a n d professio n al experience. O ne m e m b e r of the a p p lic a n t p a ir w as p re se n te d as a p a re n t, as we describe below. P a re n ta l s ta tu s w as c o u n te rb a la n c e d across the tw o v e r­ sions of the ap p lic a tio n m aterials. R ésum és a n d cover letters w ere s u b ­ m itte d in the a p p ro p ria te fo rm at, as re q u ested b y the em ployer in the job a d v e rtise m e n t (i.e., e-m ail, fax, or p aper). W ith in each pair, ap p licatio n s w ere se n t to em ployers one d a y a p a rt, c o u n te rb a la n cin g w h e th e r the p a r­ en t or n o n p a re n t a p p lied first. T h e p u rp o se of this delay w as to avo id

1328

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

raisin g u n d o suspicion b y h a v in g o u r tw o sets of ap p lic a tio n m ate ria ls a rriv e sim ultaneously.

G e n d e r w as m a n ip u la te d b y assigning fem ale- a n d m ale-so u n d in g first n a m e s to the résum és a n d cover letters. Since race d id n o t im p a c t the m a g n itu d e of the m o th e rh o o d p e n a lty in the la b o ra to ry e x p erim en t a n d because of the large n u m b e r of ap p lic a tio n s a lre a d y re q u ire d to d etect the effects of g en d er a n d p a re n ta l sta tu s, race w as n o t m a n ip u la te d in the a u d it study. In s te a d a p p lic a n ts w ere given the sam e gender- a p p ro p ria te “w h ite n a m e s ” th a t w ere used in the la b o ra to ry experim ent.

P a re n ta l sta tu s w as m a n ip u la te d on the résum é a n d on the cover letter. C o n siste n t w ith the la b o ra to ry ex p erim en t, the résum é for the p a r e n t m e m b e r of the a p p lic a n t in d ic a te d th a t she or he w as an officer in an e le m e n ta ry school p a re n t-te a c h e r association, w hile the n o n p a re n t m e m ­ b e r of the p a ir w as p re se n te d as an officer in a college a lu m n i association. In the la b o ra to ry exp erim en t, we also m a n ip u la te d p a re n ta l sta tu s on a m em o p u rp o rte d ly from the h u m a n resources d e p a rtm e n t of the h irin g com pany. Since we could n o t use this m a n ip u la tio n in the a u d it study, we in ste a d m a n ip u la te d p a re n ta l sta tu s on the cover letter. T h e cover le tte r for the p a r e n t m e m b e r of the a p p lic a n t p a ir m en tio n e d th a t she or he w as relo catin g w ith his or h e r “fam ily ” to the city w h ere the jo b w as located. T h e cover le tte r for the n o n p a re n t m e m b e r of the p a ir also m e n ­ tio n ed th a t she or he w as relo catin g to the h irin g city, b u t d id n o t m en tio n a family.

We occasionally h a d to m ak e slig h t a d ju s tm e n ts to the ap p licatio n m a te ria ls in o rd e r to satisfy cond itio n s in the jo b ad v e rtise m e n ts. I f salary re q u ire m e n ts or h isto ries w ere also req u ested , for exam ple, a sentence w as ad d e d to b o th cover letters s ta tin g th a t the a p p lic a n t w as flexible a n d p re p a re d to discuss sa la ry if in te rv ie w e d for the position. I f a jo b req u ire d fluency in a lan g u ag e o th er th a n E n g lish , this w as a d d e d to b o th résum és. O f the 638 jo b ad v e rtise m e n ts, 83 (13.4%) re q u e ste d sa la ry re q u ire m e n ts or histories, a n d 34 (S.3%) re q u ire d fluency in a lan g u ag e o th er th a n E nglish.

A fter a p p lic a tio n s w ere su b m itte d , we m o n ito re d w h e th e r or n o t job c a n d id a te s received a c a llb a c k fro m p o te n tia l em ployers b y p h o n e or e- m ail. E a c h h y p o th e tic a l a p p lic a n t h a d h is/h er ow n voice m ail n u m b e r a n d e-m ail ad d ress, allo w in g us to tr a c k positive responses from em ploy­ ers. A ny in v ita tio n to an in te rv ie w on the p h o n e or a t a c o m p a n y office w as con sid ered a “c a llb a c k .” W h en an a p p lic a n t received a callb ack , we resp o n d ed n o n o b tru siv e ly (by e-m ail or le av in g a m essage d u rin g n o n w o rk hours) in d ic a tin g th a t the a p p lic a n t w as no longer in te re ste d in the position.

1329

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

R esults

D o real em ployers d iscrim in ate a g a in st m o th ers? We begin to a n sw e r this qu estio n b y ex am in in g the p ro p o rtio n s of a p p lic a n ts receiving a c a llb ack fro m em ployers b y g en d er a n d p a re n ta l sta tu s (show n in tab le 6). We c o m p are th is v a ria b le to the “w o u ld y o u hire this a p p lic a n t” v a ria b le from the la b o ra to ry s tu d y (see tab le 1).

T h e resu lts suggest th a t real em ployers do d iscrim in ate a g a in st m others. In tab le 6, we see th a t childless w om en received 2.1 tim es as m a n y call­ b a c k s as e q u a lly qualified m o th ers (6.6% co m p a re d w ith 3.1% ; P < .OS). T h is finding is sim ilar to the la b o ra to ry e x p erim en t (see tab le 1) in w hich childless w om en w ere reco m m en d ed for hire 1.8 tim es m ore freq u en tly th a n m o th ers. In the la b o ra to ry study, fa th e rs w ere reco m m en d ed for hire a t a slightly h ig h er rate, alth o u g h the difference w as only m arg in a lly significant. L ikew ise, in the a u d it study, fa th e rs w ere called b a c k a t a h ig h e r rate, alth o u g h the difference w as n o t significant.

We n o w consider a m u ltiv a ria te m odel for the effects of p a re n ta l statu s, a p p lic a n t gender, a n d the in te ra c tio n of p a re n ta l sta tu s a n d a p p lic a n t g en d er on the odds th a t an a p p lic a n t receives a c a llb ack fro m an em ployer. T able 7 c o n tain s the estim a te d logistic regression coefficients a n d ro b u st sta n d a rd erro rs fro m the m odel. S ta n d a rd erro rs are clu stered b y jo b ID to tak e in to acc o u n t the n o n in d e p e n d e n c e of o b se rv a tio n s th a t re su lt from sending a p a ir of ap p lic a tio n s for each job. As w ith the m odels from the la b o ra to ry ex p erim en t, w e focus on the m o th e rh o o d p e n a lty in te ra c tio n term , ask in g w h e th e r bein g a p a r e n t d ecreases the odds th a t a w o m an , b u t n o t a m an , receives a callback.

As can be seen in tab le 7, the m o th e rh o o d p e n a lty in te ra c tio n is sig­ n ific a n t a n d neg ativ e, w hile the m a in effect for p a re n ta l sta tu s is insig­ nificant, a n d the m a in effect for the fem ale a p p lic a n t v a ria b le is significant a n d positive. T h e significant n eg ativ e m o th e rh o o d p e n a lty in te ra c tio n te rm in d icates th a t bein g a p a r e n t low ers the odds th a t a w o m an , b u t n o t a m an , w ill receive a c a llb ack fro m em ployers. W hile w e find no evidence of a fath e rh o o d b o n u s in the a u d it study, as show n b y the insig n ifican t effect of p a r e n t statu s, the significan t a n d p ositive m a in effect for the fem ale a p p lic a n t v a ria b le m e a n s th a t childless w o m en are significantly m ore likely to receive a c a llb a c k fro m em ployers co m p a re d w ith equ ally qualified childless m en. T h ese resu lts are c o n sisten t w ith those fo u n d in the “W ould you hire this a p p lic a n t? ” m odel in the la b o ra to ry experim ent. R e tu rn in g to the m a in resu lt, the a u d it d a ta show th a t m o th ers are dis­ a d v a n ta g e d w h e n a c tu a l em ployers m ak e h irin g decisions. F u rth e rm o re , since the a p p lic a n ts bein g e v a lu a te d in this stu d y w ere e q u ally qualified b y e x p erim en tal design, w e conclude th a t em ployer d iscrim in atio n is re ­ sponsible for the d isa d v a n ta g e s we found.

1330

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

T A B L E 6 P r o p o r t i o n s o f A p p l i c a n t s R e c e i v i n g C a l l b a c k s b y G e n d e r a n d P a r e n t a l

S t a t u s

C allbacks/T otal Jobs Proportion Called Back

M others ................................ 10/320 .0313 Childless w om en ................ 21/320 ,06S6++ F ath ers .................................. 16/318 .0503 Childless m en ....................... 9/318 .0283

N o t e .—M o th e r s a n d c h ild le ss w o m e n a p p l ie d to th e s a m e 320 jo b s ; f a th e r s a n d c h ild le ss m e n a p p lie d to th e s a m e 318 jo b s . S ee t e x t fo r v a r i a b le d e s c rip tio n s .

++ Z < .05, t e s t fo r d iffe re n c e in p r o p o r tio n s b e tw e e n p a r e n t s a n d n o n p a r e n ts .

S tre n g th s a n d L im ita tio n s

T h e s tre n g th of the a u d it s tu d y is th a t it applies an ex p erim en tal design in a re a l-w o rld setting, allo w in g us to e v a lu a te w h e th e r a c tu a l em ployers d iscrim in ate a g a in st m o th ers in the h irin g process. W hile the la b o ra to ry s tu d y p e rm its careful ex am in atio n of the u n d erly in g m e c h a n ism of dis­ crim in atio n , the a u d it s tu d y allow s us g re a te r generalizability. T h e a u d it s tu d y does h a v e a few lim itatio n s, how ever. M o s t notably, the d e p e n d e n t v a ria b le — w h e th e r an a p p lic a n t is called b a c k — is a cru d e m easu re for testin g the sta tu s-b a se d d iscrim in atio n a rg u m e n t. U nlike the la b o ra to ry study, the a u d it s tu d y does n o t give us in sig h t into the m e c h a n ism u n ­ d erly in g d iscrim in atio n , because it w as n o t possible to collect em p lo y ers’ ra n k in g s of co m m itm en t, com petence, p erfo rm an ce s ta n d a rd s, a n d o th er re le v a n t variab les. T h e re is also fa r less co ntrol o v er the e v a lu a tio n setting. T h a t is, em ployers likely receive m a n y m ore ap p lic a tio n s th a n those from the p a r e n t a n d childless p a ir th a t w ere p a r t of the study, th e re b y in tro ­ du cin g a d d itio n a l sta tu s in fo rm a tio n in to the setting. T hese lim its m ean th a t w hile the a u d it s tu d y estab lish es th a t a c tu a l em ployers d iscrim in ate a g a in st m o th ers, it c a n n o t estab lish why. U n d e rs ta n d in g the m ech an ism u n d e rly in g d iscrim in atio n is im p o rta n t if the goal is to find w ay s to reduce or elim in ate the d isa d v a n ta g e s m o th e rs face.

B y co n sidering the resu lts of these tw o co m p an io n stu d ies sim u lta ­ neously, how ever, we find s u p p o rt for the sta tu s-b a se d d iscrim in atio n m e c h a n ism using the la b o ra to ry d a ta , a n d we see the real-w o rld im p li­ cations of the a rg u m e n t w ith d a ta gen e ra te d fro m the a u d it study. F u rth e r, these resu lts are c o n sisten t w ith q u a lita tiv e w o rk sh o w ing th a t em ployers d iscrim in ate a g a in st m o th e rs (B lair-L oy 2003; C ritte n d e n 2001; K en n ed y 1999) a n d w ith su rv e y research th a t co n sistently finds a w age p e n a lty for m o th e rh o o d (B udig a n d E n g la n d 2001; A nd erso n et al. 2003; W aldfogel a n d M ey er 2000). T h u s, across a w ide ran g e of m eth o d o lo g ical a p ­ p ro a c h e s— each of w h ich h a s its u n iq u e stre n g th s a n d w eak n esses— we

1331

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

T A B L E 7 E s t i m a t e d B i n a r y L o g i s t i c R e g r e s s i o n C o e f f i c i e n t s f o r t h e

E f f e c t s o f P a r e n t a l S t a t u s a n d G e n d e r o n t h e O d d s o f

R e c e i v i n g a C a l l b a c k

In d e p e n d e n t V ariable C allback? R o b u st SE

P a re n t ................................................. .598 .433 Fem ale app lic an t ............................. .887** .407 M otherhood interaction 3 ................ -1 .3 8 * * .590 I n t e r c e p t ............................................. 1 C/0 £ * * * .338

N o t e .—Clu stered b y jo b . M others a n d childless w om en a p p lied to the sam e 320 jobs; fath ers a n d childless m en ap p lied to th e sam e 318 jobs, for a to tal o f 1,276 applications to 638 jobs. All values rep o rte d to three significant digits. See te x tfo r v ariab le descriptions.

a P a re n t x fem ale applicant. * P c . 1 0 .

** P < .05. *** P < .001.

find evidence th a t m o th e rs experience d isa d v a n ta g e s in w o rk p lace settings a n d th a t d iscrim in atio n play s a role in p ro d u c in g these d isad v an tag es.

SUMMARY AND CONCLUSIONS

In this p ro ject, w e m ak e tw o m a in co n trib u tio n s. F irst, we isolate a n d ex p erim en tally e v alu ate a sta tu s-b a se d d iscrim in atio n m e c h a n ism th a t is p ro p o sed to explain som e of the d isa d v a n ta g e s m o th e rs experience in the p a id la b o r m a rk e t. W hile su rv e y research h a s e stab lish ed th a t em ployed m o th e rs experience a per-ch ild w age penalty, n e t of the u su al h u m a n ca p ita l a n d o ccu p atio n al facto rs th a t affect w ages, this research h a s been u n ab le to assess w h e th e r d iscrim in atio n is in p a r t responsible for this w age penalty.

T h e ex p e rim e n t p re se n te d here stro n g ly s u p p o rts the sta tu s-b a se d dis­ crim in atio n m ech an ism . T h is is the first s tu d y to show consistent, signif­ ic a n t evidence for the m o th e rh o o d p e n a lty over a b ro a d range of m easures. B y e x p erim en tally h o ld in g c o n s ta n t the qualificatio n s a n d b a c k g ro u n d experiences of a p a ir of fictitious jo b a p p lic a n ts a n d v a ry in g only th eir p a re n ta l sta tu s, we fo u n d th a t e v a lu a to rs ra te d m o th e rs as less c o m p eten t a n d c o m m itted to p a id w o rk th a n n o n m o th e rs, a n d consequently, dis­ crim in a te d a g a in st m o th e rs w h e n m a k in g h irin g a n d sa la ry decisions. C o n siste n t w ith o u r p red ictio n s, fa th e rs experienced no such discrim i­ n atio n . In fact, fa th e rs w ere a d v a n ta g e d o v er childless m en in several w ays, bein g seen as m ore c o m m itted to p a id w o rk a n d bein g offered h ig h er sta rtin g salaries. T h e fa c t th a t e v a lu a to rs offered h ig h er salaries to fath ers suggests th a t c u ltu ra l beliefs a b o u t g en d ered la b o r m a rk e ts a n d a fam ily w age still shape the allocation of o rg a n iz a tio n a l rew ard s. O ne u n ex p ected

1332

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

finding w as th a t childless w om en w ere a d v a n ta g e d over childless m en on several m easu res, in clu d in g bein g seen as m ore c o m p e te n t a n d bein g m ore likely to be reco m m en d ed for hire, alth o u g h th e y w ere n o t offered sig­ n ifican tly h ig h e r salaries. I t is possible th a t e v a lu a to rs p erceive childless w o m en as especially c o m m itted to p a id w ork.

T h e second c o n trib u tio n w e m ak e in th is p ro je c t is to show th a t real em ployers d iscrim in ate a g a in st m o th ers. O u rs is the first a u d it s tu d y to te st for h irin g d iscrim in atio n on the basis of p a re n ta l sta tu s, a n d therefore the first to p ro v id e cau sal evidence th a t m o th e rs experience h irin g dis­ crim in atio n . B y u sing ap p lic a tio n m a te ria ls a d a p te d from the la b o ra to ry ex p erim en t to ap p ly to over six h u n d re d jo b s, we fo u n d th a t pro sp ectiv e em ployers called m o th e rs b a c k a b o u t h a lf as often as no n m o th ers. F a th e rs, b y co n tra st, w ere n o t d is a d v a n ta g e d in the h irin g process. In general, the findings of the a u d it s tu d y co rresp o n d closely to those fro m the la b o ra to ry study, p ro v id in g co n verging evidence for the m o th e rh o o d p e n a lty across tw o stu d ies em ploying d ifferen t m e th o d s a n d sam ples.

W hile the d a ta s u p p o rt the m a in h y p othesis, the p ro je c t h as several lim itatio n s. F irst, the ex p e rim e n t only e v a lu a te d the sta tu s-b a se d d iscrim ­ in a tio n m e c h a n ism for a h ig h -sta tu s jo b th a t a p p e a re d to req u ire high levels of com m itm en t. W h e th e r m o th e rs w o u ld experience the sam e type a n d a m o u n t of d iscrim in atio n in lo w er-statu s jo b s or in jo b s th a t are m ore or less gen d er-ty p ed is an open q uestion. W hile we expect the m o th e rh o o d p e n a lty to a p p ly to a w ide ran g e of jo b s (as all jo b s req u ire som e degree of com petence a n d com m itm ent), the m a g n itu d e of the effect likely v aries w ith the jo b type. E x istin g su rv e y an alyses h a v e fo u n d a m o th e rh o o d p e n a lty across a w ide ran g e of o c cu p atio n s a n d jobs. A dditionally, one stu d y h as sh o w n th a t the m a g n itu d e of the w age gap is a c tu a lly la rg e st for those w ho h a v e only a h ig h sch o o l-lev el e d u catio n , suggesting th a t the p e n a lty is n o t lim ited to h ig h -sta tu s jo b s (A nderson et al. 2003). T h u s, alth o u g h th ere is reaso n to su sp ect th a t the m e c h a n ism d escrib ed here w o u ld ap p ly in a w id e r ran g e of jo b s, ex p erim en ts th a t v a ry the type of jo b are need ed to e v a lu ate this pred ictio n .

Second, th is stu d y exam ines d iscrim in atio n only a t the p o in t of hire. We p re d ic t th a t w o m en w ho give evidence of bein g a m o th e r w o u ld be h eld to a h a rs h e r s ta n d a rd a n d suffer d ecreased w o rk p lace ev a lu a tio n s a t o th er ju n c tu re s, such as w h en p ro m o tio n decisions or decisions to a w a rd raises are m ad e, b u t w h e th e r the m e c h a n ism holds a t these o th er crucial ju n c tu re s is also an em p irical q uestion. Finally, the s tu d y exam ines only one a v en u e for gettin g a job. I t is possible th a t e v a lu a to rs are m ore or less d iscrim in atin g , for exam ple, w h e n c a n d id a te s are reco m m en d ed th ro u g h social n etw o rk s.

T h e resu lts of th is s tu d y h a v e im p licatio n s for u n d e rs ta n d in g som e of the en d u rin g p a tte rn s of g en d er in e q u a lity in p a id w ork. T h e m o th e rh o o d

1333

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

p e n a lty a p p e a rs ro b u s t b o th in te rn a tio n a lly a n d historically. O ne analysis of incom e d a ta fo u n d a m o th e rh o o d p e n a lty in A u stralia, C a n a d a , the U n ite d K ingdom , the U n ite d S tates, G erm any, F in la n d , a n d S w eden (H a rk n e ss a n d W aldfogel 1999). A m ore re c e n t analysis (M isra, B udig, a n d M oller 2005) disco v ered a m o th e rh o o d p e n a lty in A u stria, G erm any, Italy, L u x em b o u rg , the N e th e rla n d s, C a n a d a , the U n ite d K ingdom , the U n ite d S tates, B elgium , F ra n c e , a n d Sw eden. F u rth e rm o re , the m o th e r­ h o o d p e n a lty a p p e a rs to h a v e re m a in e d stab le o v er tim e (A vellar a n d Sm ock 2003). T h is s tu d y offers a p a rtia l e x p lan atio n for the m ech an ism b e h in d a w id e sp re a d , d u ra b le p h en o m en o n w ith im p licatio n s for a b ro a d seg m en t of the p o p u latio n .

M o re generally, a g en d er gap in w ages h a s p e rsiste d over the v a s t m o v e m e n t of w o m en into p a id la b o r in the U n ite d S tates since the early 1970s. W hile the m a g n itu d e of the gap d ecreased over m u ch of this p erio d (C harles a n d G ru sk y 2004; B la u a n d K a h n 2004), a sizeable gap rem ains, a n d the gap h a s n o t n a rro w e d in re cen t y ears (In stitu te for W o m e n ’s Policy R esearch 2006). As G lass (2004) notes, em ployed m o th e rs are the gro u p of w o m en w ho ac c o u n t for m o st of this gap. W hile m a n y factors are certa in ly responsible for its p ersistence, this stu d y suggests th a t cul­ tu ra l beliefs a b o u t the tension b e tw een the m o th e rh o o d a n d “id eal w o rk e r” roles m a y p la y a p a r t in rep ro d u c in g th is p a tte rn of inequality. A second en d u rin g p a tte rn of g en d er in e q u a lity is the so-called “glass ceiling,” a m e ta p h o r for the b a rrie rs th a t re stric t w o m e n ’s m o v e m e n t up the career la d d e r to the h ig h e st positio n s in o rg an izatio n s a n d firm s. To the e x ten t th a t em ployers view m o th ers as less c o m m itted to th e ir jo b s a n d less “p ro m o ta b le ,” the glass ceiling w om en face could be, in p a rt, a m o th e rh o o d ceiling.

W ritin g for the N a tio n a l C e n te r for P olicy A nalysis, D enise V enable (2002) describes an analysis fro m the congressional b u d g e t office th a t fo u n d th a t am o n g people ages 2 7 to 33 w h o h a v e n e v e r h a d children, w o m e n ’s earn in g s a p p ro a c h 98% of m e n ’s. She concludes, “W h en w om en b eh av e in the w o rk p lace as m en do, the w age gap b e tw een th e m is sm a ll.” C laim s of u n e q u a l pay, she continues, “a lm o st alw ay s involve c o m p arin g ap p les a n d o ra n g e s.” H ow ever, since m o st em ployed m en a n d em ployed w o m en h a v e ch ild ren a t som e p o in t in th e ir lives, the m o st illu stra tiv e “w ith in f r u it” co m p ariso n is n o t the co m p ariso n of childless m en to ch ild ­ less w om en, b u t the c o m p ariso n of m en w ith ch ild ren to w o m en w ith children. As the tw o stu d ies re p o rte d here show, w h en w o m en “b eh av e as m en d o ,” giving evidence of bein g a p a re n t, th ey are d isc rim in a te d ag ain st, w hile th e ir m ale c o u n te rp a rts are often a d v a n ta g e d b y th eir p a ­ re n ta l status. F a r from bein g an “ap p les to o ra n g e s” co m parison, the m ale a n d fem ale a p p lic a n ts w ho w ere e v a lu a te d in these stu d ies w ere exactly

1334

This content downloaded from 131.94.16.10 on Fri. 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

eq u al b y e x p e rim e n ta l design. T h a t p a re n ta l s ta tu s d is a d v a n ta g e d only fem ale a p p lic a n ts is stro n g evidence of discrim in atio n .

REFER EN C ES A cker, Jo a n . 1990. “H ie ra rc h ie s, Jo b s , a n d B odies: A T h e o ry of G e n d e re d

O rg a n iz a tio n s .” G e n d er a n d S o c ie ty 4:139-58. A igner, D e n n is J., a n d G len G. C a in . 1977. “S ta tistic a l T h e o rie s o f D isc rim in a tio n in

L a b o r M a r k e ts .” I n d u s tr ia l a n d L a b o r R e la tio n s R e v ie w 30:175-87. A n d e rso n , D e b o ra h J., M e lissa B in d er, a n d K a te K ra u se . 2003. “T h e M o th e rh o o d W age

P e n a lty R e visited: E x p e rie n c e , H e tero g e n eity , W o rk E ffo rt a n d W o rk -S ch e d u le F le x ib ility .” I n d u s tr ia l a n d L a b o r R e la tio n s R e v ie w 56:273-94.

A ro n so n , E llio t, P h o e b e C. E llsw o rth , J. M e rril C a rls m ith , a n d M a rti H o p e G onzales. 1990. M e th o d s o f R e s e a r c h in S o c ia l P sy ch o lo g y. N e w Y ork: M c G ra w -H ill.

A rrow , K e n n e th J. 1973. “T h e T h e o ry o f D is c rim in a tio n .” P p. 3 -3 3 in D is c r im in a tio n in L a b o r M a r k e ts , e d ite d b y O. A s h e n fe lte r a n d A. Rees. P rin c e to n , N .J.: P rin c e to n U n iv e rsity Press.

A vellar, S a ra h , a n d P a m e la Sm ock. 2003. “H a s th e P rice o f M o th e rh o o d D e clin ed o v e r T im e ? A C ro ss -C o h o rt C o m p a riso n o f th e M o th e rh o o d W age P e n a lty ." J o u r n a l o f M a rria g e a n d th e F a m ily 65:597-607.

B ecker, G a ry S. 1985. “H u m a n C a p ita l, E ffo rt a n d th e S exual D iv isio n of L a b o r.” J o u r n a l o f L a b o r E c o n o m ic s 3:S 33-58.

B erger, Jo s e p h , B e r n a r d P. C o h e n , a n d M o rris Z e ld itc h , Jr. 1966. “S ta tu s C h a ra c te ristic s a n d E x p e c ta tio n S ta te s .” P p. 2 9 -4 6 in S o c io lo g ica l T h e o rie s in P rogress, e d ite d by J o s e p h B erger, M o rris Z eld itc h , Jr., a n d B o A n d e rso n . N e w Y ork: H o u g h to n M ifflin.

----------. 1972. “S ta tu s C h a ra c te ris tic s a n d Social In te r a c tio n .” A m e ric a n Soc io lo g ica l R e v ie w 37:241-55.

B erger, Jo s e p h , H a m it Fi§ek, R o b e rt N o rm a n , a n d M o rris Z eld itc h , Jr. 1977. S ta tu s C h a ra c teristic s a n d S o c ia l In te r a c tio n . N e w Y ork: E lsevier.

B e rtra n d , M a ria n n e , a n d S e n d h il M u lla in a th a n . 2003. “A re E m ily a n d G re g M ore E m p lo y a b le T h a n L a k is h a a n d J a m a l? A F ie ld E x p e rim e n t on L a b o r M a rk e t D is c rim in a tio n .” N a tio n a l B u re a u of E co n o m ic R e se a rc h W o rk in g P a p e r Series, no. 9873. A ccessed S e p te m b e r 2003. h ttp ://w w w .n b e r.o rg /p a p e rs/w 9 8 7 3 .

Bielby, D e n ise D ., a n d W illia m T. Bielby. 1984. “W o rk C o m m itm e n t, Sex-Role A ttitu d e s, a n d W o m e n ’s E m p lo y m e n t.” A m e ric a n S o c io lo g ica l R e v ie iv 49:234-47.

Bielby, W illia m , a n d Ja m e s B a ro n . 1986. “M e n a n d W om en a t W ork: Sex S eg re g atio n a n d S ta tistic a l D is c rim in a tio n .” A m e ric a n J o u r n a l o f S o c io lo g y 91:759-99.

B ie rn a t, M o n ic a, a n d D ia n e D . K o b ry n o w ic z. 1997. “G e n d e r a n d R a ce -B ased S ta n d a rd s o f C o m p ete n ce : L o w e r M in im u m S ta n d a rd s b u t H ig h e r A b ility S ta n d a rd s fo r D e v a lu e d G ro u p s .” J o u r n a l o f P e r so n a lity a n d S o c ia l P sy ch o lo g y 72:544-57.

B lair-L oy, M ary. 2003. C o m p e tin g D e v o tio n s: C areer a n d F a m ily a m o n g W om en E x e c u tiv e s . C a m b rid g e , M ass.: H a r v a r d U n iv e rsity Press.

B la u , F ra n c in e D ., a n d L a w re n c e M . K a h n . 2004. “T h e U.S. G e n d e r P a y G a p in th e 1990s: S low ing C o n v e rg e n c e .” W o rk in g P a p e r no. 10853. N a tio n a l B u r e a u of E c o n o m ic R esearch.

B u d ig , M ichelle, a n d P a u la E n g la n d . 2001. “T h e W age P e n a lty for M o th e rh o o d .” A m e ric a n S o c io lo g ica l R e v ie w 66:204-25.

C h a rle s, M a ria , a n d D a v id B. G rusky. 2004. O c cu p a tio n a l G hettos: T he W orld w id e S eg re g a tio n o f W om en a n d M e n . S ta n fo rd , C alif.: S ta n fo rd U n iv e rsity Press.

C le v e la n d , J e a n e tte N . 1991. “U sin g H y p o th e tic a l a n d A c tu a l A p p lic a n ts in A ssessing P e rso n -O rg a n iz a tio n F it: A M e th o d o lo g ic a l N o te .” J o u r n a l o f A p p lie d S o c ia l P sy ch o lo g y 21:1004-11.

C le v e la n d , J e a n e tte N ., a n d A n d re w H . B e rm a n . 1987. “Age P e rc e p tio n s o f Jobs:

1335

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

A g re em e n t b e tw e e n S a m p le s of S tu d e n ts a n d M a n a g e rs .” P sy ch o lo g ica l R e p o r ts 61: 565-66.

Clogg, C liffo rd C., E v a P e tk o v a , A d a m a n tio s H a rito u . 1995. “S ta tistic a l M e th o d s for C o m p a rin g R egression C oefficients b e tw e e n M o d e ls." A m e ric a n J o u r n a l o f Soc io lo g y 100:1261-93.

C o h e n , Ja c o b . 1988. S ta tis tic a l P o w e r A n a ly s is f o r th e B e h a v io r a l S c ie n c e s. H illsd ale , N .J.: E rlb a u m A ssociates.

C o rre ll, Shelley J. 2001. “G e n d e r a n d th e C a re e r C h oice Process: T h e Role o f B iased S elf-A ssessm en ts.” A m e ric a n J o u r n a l o f S o c io lo g y 106:1691-1730.

---------- . 2004. “C o n s tra in ts in to P references: G ender, S ta tu s a n d E m e rg in g C a re e r A s p ira tio n s .” A m e ric a n S o c io lo g ica l R e v ie iv 69:93-133.

C o rre ll, Shelley J., a n d S te p h e n B e n a rd . 2006. “B ia s e d E s tim a to rs ? C o m p a rin g S ta tu s a n d S ta tistic a l T h e o rie s o f G e n d e r D is c rim in a tio n .” P p. 8 9 -1 1 6 in S o c ia l P sy ch o lo g y o f th e W orkplace: A d v a n c e s in G roup P ro c ess, vol. 23. N e w Y ork: E lsevier.

C o rre ll, Shelley J., a n d C e cilia L. R idgew ay. 2003. “E x p e c ta tio n S ta te s T h e o ry .” Pp. 2 9-51 in H a n d b o o k o f S o c ia l P sy c h o lo g y , e d ite d by J o h n D e lam ate r. N e w Y ork: K lu w e r A cad e m ic Press.

C orse, S a ra J. 1990. “P re g n a n t M a n a g e rs a n d T h e ir S u b o rd in a te s: T h e E ffe c ts of G e n d e r E x p e c ta tio n s on H ie ra rc h ic a l R e la tio n s h ip s .” J o u r n a l o f A p p lie d B e h a v io r a l S c ie n c e 26:25-48.

C ritte n d e n , A nn. 2001. T he P ric e o f M o th e r h o o d : W h y th e M o s t Im p o r ta n t J o b in the W orld I s S t il l th e L e a st Valued. N e w Y ork: M e tro p o lita n B ooks.

C uddy, A m y J. C ., S u s a n T. F isk e , a n d P e te r C lic k . 2004. “W h e n P ro fe ssio n a ls B ecom e M o th e rs, W a rm th D o e s n ’t C u t th e Ic e .” J o u r n a l o f S o c ia l Iss u e s 60:701-18.

E p s te in , C y n th ia , C a rro ll S eron, B o n n ie O glensky, a n d R o b e rt S au te. 1999. T he P a rt T im e P a ra d o x : T im e N o rm s, P ro fessio n a l L iv e s, F a m ily a n d G ender. N e w Y ork: R outledge.

Fix, M ic h ae l, G eorge C. G alster, a n d R a y m o n d J. S tru y k . 1993. “A n O v e rv ie w of A u d itin g fo r D is c rim in a tio n .” Pp. 1-68 in Clear a n d C o n v in c in g E v id e n c e : M e a su re m e n t o f D is c r im in a tio n in A m e ric a , e d ite d b y M ic h a e l F ix a n d R a y m o n d J. S tru y k . W a sh in g to n , D .C .: U rb a n I n s titu te Press.

F o sch i, M a rth a . 1989. “S ta tu s C h a ra c te ristic s, S ta n d a rd s a n d A ttrib u tio n s .” Pp. 58-72 in S o c io lo g ica l T heories in P rogress: N e w F o r m u la tio n s , e d ite d by Jo s e p h B erger, M o rris Z e ld itc h , Jr., a n d B o A n d e rso n . B oston: H o u g h to n M ifflin.

---------- . 1996. “D o u b le S ta n d a rd s in th e E v a lu a tio n of M e n a n d W o m e n .” S o c ia l P sy ch o lo g y Q u a rte rly 59:237-54.

F u e g e n , K a th le e n , M o n ic a B ie rn a t, E liz a b e th H a in e s , a n d K a y D e au x . 2004. “M o th e rs a n d F a th e r s in th e W o rkplace: H o w G e n d e r a n d P a re n ta l S ta tu s In flu en c e J u d g m e n ts o f Jo b -R e la te d C o m p e te n c e .” J o u r n a l o f S o c ia l Is s u e s 60:737-54.

G lass, Je n n ifer. 2004. “B lessin g o r C u rse ? W o rk -F a m ily Policies a n d M o th e r s ’ W age G ro w th o v e r T im e .” W ork a n d O c cu p a tio n s 31:367-94.

H a lp e rt, J a n e A., M idge L. W ilson, a n d J u li a H ic k m a n . 1993. “P re g n a n c y a s a Source o f B ia s in P e rfo rm a n c e A p p ra is a ls .” J o u r n a l o f O rg a n iza tio n a l B e h a v io r 14:649-63.

H a rk n e s s , S u sa n , a n d J a n e W aldfogel. 1999. “T h e F a m ily G a p in Pay: E v id e n c e from S even In d u s tria lis e d C o u n trie s .” C A S E p a p e r 29: C e n tre fo r A n a ly sis o f Social E xclusion.

H a y s, S h a ro n . 1996. T he C u ltu ra l C o n tra d ic tio n s o f M o th e r h o o d . N e w H a v e n , C onn.: Y ale U n iv e rs ity Press.

H e id er, F ritz . 1958. T h e P sy ch o lo g y o f In te r p e r so n a l R e la tio n s . N e w Y ork: Wiley. H e rsc h , Jo n i, a n d L eslie S. S tra tto n . 2000. “H o u s e h o ld S p e c ia liz atio n a n d th e M ale

M a rria g e W age P re m iu m .” I n d u s tr ia l a n d L a b o r R e la tio n s R e v ie iv 54:78-94. H ill, M a rth a . 1979. “T h e W age E ffe c ts o f M a rita l S ta tu s a n d C h ild re n .” J o u r n a l o f

H u m a n R e so u rc e s 14:579-94. I n s titu te fo r W o m e n ’s Policy R e se arc h . 2006. “T h e G e n d e r W age R atio: W o m e n ’s a n d

1336

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

M o th e rh o o d P e n a lty

M e n ’s E a rn in g s .” h ttp ://w w w .iw p r.o rg /p d f/U p d a te d 2 0 0 6 _ C 3 S 0 .p d f. A ccessed F e b r u a ry 17, 2007.

K ennelly, Ivy. 1999. “T h a t Single M o th e r E le m en t: H o w W h ite E m p lo y e rs T ypify B la c k W o m e n .” G e n d er a n d S o c ie ty 13:168-92.

K o b ry n o w ic z, D ia n e D ., a n d M o n ic a B ie rn a t. 1997. “D e co d in g S u b je c tiv e E v a lu a tio n s: H o w S te re o ty p e s P ro v id e S h iftin g S ta n d a r d s .” J o u r n a l o f E x p e r im e n ta l S o c ia l P sy ch o lo g y 33:579-601.

K o re n m a n , S a n d e rs, a n d D a v id N e u m a rk . 1991. “D oes M a rria g e R eally M a k e M en M o re P ro d u c tiv e ? ” J o u r n a l o f H u m a n R e so u rc e s 26:282-307.

L o h , E n g Seng. 1996. “P ro d u c tiv ity a n d th e M a rria g e P re m iu m for W h ite M a le s .” J o u r n a l o f H u m a n R e so u rc e s 31:566-89.

L o v a g lia , M ic h ae l J., Je ffre y W. L u c a s, J e ffre y A. H o u ser, S h a n e R. T h y e , a n d B a rry M a rk o v sk y . 1998. “S ta tu s P rocesses a n d M e n ta l A b ility T est S cores.” A m e ric a n J o u r n a l o f S o c io lo g y 104:195-228.

M isra , Jo y a , M ichelle B u d ig , a n d S te p h a n ie M oller. 2005. “E m p lo y m e n t, W ages, a n d P o v e rty : R e co n c iliatio n Policies a n d G e n d e r E q u ity .” P a p e r p re s e n te d a t th e a n n u a l m e e tin g of th e A m e ric a n Sociological A ssociation, P h ila d e lp h ia , A ugust.

M o n in , B e n o it, a n d D a le T. M iller. 2001. “M o ra l C re d e n tia ls a n d th e E x p re s sio n of P re ju d ic e .” J o u r n a l o f P e r so n a lity a n d S o c ia l P sy ch o lo g y 81:33-43.

M u th é n , L in d a K ., a n d B e n g t M u th é n . 1998-2004. M p lu s U ser’s G u id e , 3d ed. L os Angeles: M u th é n a n d M u th é n .

O e ttin g e r, G e ra ld S. 1996 “S ta tistic a l D isc rim in a tio n a n d th e E a rly C a re e r E v o lu tio n o f th e B la c k -W h ite W age G a p .” J o u r n a l o f L a b o r E c o n o m ic s 14:52-78.

O lian , J u d y D ., a n d D o n a ld P. S c h w a b . 1988. “T h e Im p a c t o f A p p lic a n t G e n d e r C o m p a re d to Q u a lific a tio n s on H irin g R e co m m en d a tio n s: A M e ta -a n a ly sis of E x p e rim e n ta l S tu d ie s .” O rg a n iza tio n a l B e h a v io r a n d H u m a n D e c is io n P rocesses 41: 180-95.

O rloff, A nn. 1996. “G e n d e r a n d th e W elfare S ta te .” A n n u a l R e v ie iv o f S o c io lo g y 22: 5 1-78.

Pager, D e v a h . 2003. “T h e M a r k of a C rim in a l R e c o rd .” A m e ric a n J o u r n a l o f Soc io lo g y 108 (5): 9 37-75.

P h e lp s, E d m u n d S. 1972. “T h e S ta tistic a l T h e o ry of R a cism a n d S exism .” A m e ric a n E c o n o m ic R e v ie w 62:659-61.

R idgew ay, C e cilia L. 2001. “Social S ta tu s a n d G ro u p S tru c tu re .” Pp. 352-75 in B la c k w e ll H a n d b o o k o f S o c ia l P sychology: Group P ro cesses, e d ite d b y M ic h a e l A. H o g g a n d Scott T in d a le . M a ld e n , M ass.: B lac k w e ll.

R idgew ay, C ecilia, a n d Shelley J. C orrell. 2004. “M o th e rh o o d as a S ta tu s C h a ra c te ris tic .” J o u r n a l o f S o c ia l Iss u e s 60:683-700.

T o w n se n d , N ic h o la s W. 2002. T he Package Deal: M arriage, W ork a n d F a th e rh o o d in M e n ’s L iv e s . P h ila d e lp h ia : T em ple U n iv e rs ity Press.

TVoyer, L isa, a n d L esley C. Y ounts. 1997. “W h o se E x p e c ta tio n s M a tte r? T h e R e la tiv e P o w e r o f F ir s t- O r d e r a n d S e c o n d -O rd e r E x p e c ta tio n s in D e te rm in in g Social In flu e n c e .” A m e ric a n J o u r n a l o f S o c io lo g y 103:692-732.

V enable, D enise. 2002. “T h e W age G a p M y th .” N a tio n a l C e n te r fo r Policy A nalysis, A p ril 12. B rie f A n a ly sis no. 392. A ccessed D e c e m b e r 12, 2004. h ttp ://w w w .n c p a .o rg / p u b /b a /b a 3 9 2 /.

W aldfogel, J a n e , a n d S u s a n E. M eyer. 2000. “G e n d e r D iffere n c es in th e L ow -W age L a b o r M a r k e t.” Pp. 193-232 in F in d in g J obs: W ork a n d W elfare R e fo r m , e d ite d by D a v id C a rd a n d R e b e c c a B la n k . N e w Y ork: R ussell Sage.

W ebster, M u rra y , a n d M a r th a F oschi. 1988. “O v e rv ie w of S ta tu s G e n e ra liz a tio n .” Pp. 1-22 in S ta tu s G e n eralization: N e w T h e o ry a n d R e s e a r c h , e d ite d by M u rra y W eb ste r a n d M a r th a F oschi. S ta n fo rd , Calif.: S ta n fo rd U n iversity.

W eeden, K im A. 2005. “Is T h e re a F le x ig la ss C eiling? F le x ib le W o rk A rra n g e m e n ts a n d W ages in th e U n ite d S ta te s .” S o c ia l S c ie n c e R e s e a r c h 34 (2): 4 5 4 -8 2 .

1337

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions

A m erican J o u rn a l of Sociology

W illia m s, Jo a n . 2001. U n b e n d in g G ender: W h y W ork a n d F a m ily C o n flic t a n d W h a t to Do a b o u t I t . O xford: O x fo rd U n iv e rsity Press.

Y inger, Jo h n . 1986. “M e a su rin g D isc rim in a tio n w ith F a ir H o u s in g A u dits: C a u g h t in th e A c t.” A m e ric a n E c o n o m ic R e v ie iv 76:881-93.

1338

This content downloaded from 131.94.16.10 on Fri, 1 Aug 2014 16:08:30 PM All use subject to JSTO R Terms and Conditions