An End to Gender Display Through the Performance of ... to gender display... · article, I review...

ORIEL SULLIVAN Oxford University

An End to Gender Display Through the Performance

of Housework? A Review and Reassessment

of the Quantitative Literature Using Insights

From the Qualitative Literature

According to the gender-deviance neutraliza-tion hypothesis, men and women in house-hold circumstances that contradict the normalexpectations of gender display their genderaccordingly, by either increasing or decreas-ing their contribution to household tasks. In thisarticle, I review and reassess the large-scalequantitative evidence, concluding that consider-able doubt has subsequently been cast on thishypothesis. For women, research shows thatthe original identification of gender-devianceneutralization behavior was questionable, as itfailed to take into account women’s absolutelevels of income. For men, both more recentquantitative and indicative qualitative researchsuggests that such behavior was always limitedto a very small group. Subsequent changes in thecontributions to housework of men from lowersocioeconomic groups suggest that such displaymay no longer be evident.

In this article, I review and reassess thehypothesis of gender display in relation to house-work, as originally discussed in a series of arti-cles based on the analysis of large-scale survey

Centre for Time Use Research, Department of Sociology,Manor Road Building, Manor Road, Oxford OX1 3UQ,United Kingdom ([email protected]).

Key Words: gender relations, household labor.

data (Bittman, England, Folbre, Sayer, & Math-eson, 2003; Brines, 1994; Greenstein, 2000).According to this hypothesis, heterosexual mar-ried men and women in household circumstancesthat contradict the normal expectations of gen-der display their gender accordingly, by eitherincreasing or decreasing their contribution tohousehold tasks.1 So, for example, a womanwho earns more than her husband might com-pensate for this reversal of gender expectationsby doing most of the housework. In this case,the effect of gender can be said to overridethat of the marital bargaining power that herhigher income brings her, a bargaining powerthat can (according to household economicexchange theory and the economic dependencyand relative resources perspectives) be used tonegotiate a lesser contribution to housework. Inother words, in such cases, the effect of genderoverrides the power of money—a forceful argu-ment against explanations based on economicexchange theory. According to the same logic,men’s gender display in relation to houseworkmay be indicated in situations where the manof a couple cannot be classified as the mainbreadwinner of the family. In this situation of

1The articles based on large-scale quantitative data thatoriginally identified a gender display effect in the contextof housework excluded same-sex couples from the analyses.Therefore, the interesting issue of gender display in same-sexhouseholds has not been addressed to date in this literature.

Journal of Family Theory & Review 3 (March 2011): 1–13 1DOI:10.1111/j.1756-2589.2010.00074.x

2 Journal of Family Theory & Review

threat to his gender identity, he may display hisgender by minimizing his contribution to house-work (which is, of course, feminine defined).

Gender display in these situations thereforerefers to the performance of specific behaviors(in this case, housework) designed to emphasizetraditional masculinities and femininities. Assuch, it provides a sort of litmus test: If we canfind less evidence over time for gender displayin relation to housework or, indeed, evidencecontradicting it, this would indicate a change notjust in gendered time use but also in the genderideologies supporting gendered time use.

The above-mentioned series of articles werehighly influential in apparently showing thatgender can override the power of money. Inthis article, through a review of the large-scalequantitative research supported with indicativefindings from the qualitative literature, I arguethat there is now enough evidence to indicate thatthe original identification of gender display inrelation to housework among women may havebeen the result of a misinterpretation of data.Among men, more recent large-scale quantita-tive research points to the fact that such behaviorwas always limited to a very small group. Subse-quent changes in the contributions to houseworkof men from lower socioeconomic groups sug-gest that such display may no longer be evident.Again, qualitative studies aid in understandingthe processes involved.2

DEVELOPMENT OF THE GENDER DISPLAYPERSPECTIVE IN RELATION TO HOUSEWORK

The concept of gender display was introducedinto the quantitative literature on the domesticdivision of labor by Brines (1994). It forms partof the theoretical framework of ‘‘doing’’ gen-der (Berk, 1985; West & Zimmerman, 1987),in which gender is regarded as actively accom-plished and negotiated in interaction. Interactionwith a partner in a heterosexual relationshipcan be considered among the most significantlocations for this enactment of gender, and theroutine performance of housework is regardedas a key indicator of such enactments (West

2The review of literature is based on electronic searchesfor gender display and gender deviance neutralization inboth EBSCO and Sociological Abstracts. Tables listing thelarge-scale quantitative studies are found in Gupta (2007)and Gupta and Ash (2008).

& Zimmerman, 1987). In this sense, doing thehousework may be regarded, in the words ofButler (2006), as a performative act. In theearly quantitative literature, the identification ofgendered behaviors around the performance ofroutine housework that seemed to be compen-sating for deviations from the expected genderstructure resonated with the ideas of emphasizedfemininity and emphasized masculinity in thedoing gender literature.

Qualitative research conducted in Britain andthe United States during the 1980s had alreadybegun to document behaviors among men thatwould later be identified as gender display. TheBritish studies were conducted among groupsof working-class couples, with a specific focuson communities in which the dramatic declineof traditional sources of employment for men(e.g., coal mining, steel working) meant thatlarge numbers of men had lost their jobs, alongwith their normative role of family breadwinner(Morris, 1985, 1987; Wheelock, 1990). Morris’sconclusions indicated that male redundancy didnot, in general, lead to any renegotiation of thedomestic division of labor. Any renegotiationthat did take place was conducted in the strongshadow of traditional normative expectationsof gender. Crucial to the (re)negotiation ofgender roles in these situations, the choicefor women was to remain at home with noemployment or, where full-time employmentwas available, to become the sole provider ofincome to the family.3 Morris (1987) foundthat women in the first situation were likelyto withdraw into the domestic role, notablymotherhood, as ‘‘providing the most secure basisfor identity’’ (p. 202). Men in this situation,having been deprived through unemploymentboth of their identity in the public sphere and oftheir normative position in the domestic sphere,emphasized their normative role at home bynot contributing much to the household chores.Although younger men tended to participate inparticular domestic tasks on a regular basis, thiswas not the case among families with children,in which a strong emphasis on a traditionalmodel of motherhood encouraged the husband towithdraw from the domestic sphere, even whentheir wives were employed outside the home.

3Because of the intricacies of the tax laws, while drawingunemployment benefits it was not financially worthwhile fora woman in this situation to take on part-time work.

An End to Gender Display Through Housework? 3

At around the same time, Hochschild andMachung (1989) reported similar findings fromtheir qualitative study of 50 dual-earner couplesin the United States. They found differencesin the sharing of housework between couplesin which the husbands earned less than theirwives, none of whom shared the housework,and those in which the husbands earned thesame as or more than their wives, among whombetween one fifth to one third shared housework.Their interpretation from intensive interviewswas founded on the threat to male identity fromfinancial dependence in situations where thehusband earned less than his wife, and the cor-responding reluctance to threaten it further bytaking on feminine-associated housework.

Following these earlier qualitative studies,it was Brines’s influential 1994 article anda follow-up by Greenstein (2000) that firmlyestablished the gender-display thesis in the lit-erature on the domestic division of labor. Thesestudies, based on large-scale national data fromthe United States, provided support for the ideathat men and women in situations that deviatedfrom the traditional norms of gender react byemphasizing their normative identities (a) bycontributing less to housework (in the caseof men who are economically dependent) or(b) through the overaccomplishment of house-work (in the case of breadwinner wives).4 Brines(1994) analyzed Panel Study of Income Dynam-ics (PSID) data from 1985 to show that, althoughthe housework hours of wives largely conformedto the assumptions of conventional householdeconomic exchange theory (the more econom-ically dependent, the more housework the wifeperformed), husbands who were economicallydependent (in particular those in low-incomehouseholds and the long-term jobless) performedless housework than others. Brines (1994) noted:‘‘It appears, then, that lower-income husbandsdo less housework because they are more likelyto depend on their wives for economic supportand respond by avoiding housework to reclaimtheir constitutive masculinity’’ (p. 667).

The significance of Brines’s (1994) arti-cle was that it provided for the first timean indication from large-scale data that, formen in specific structural situations, the impor-tance of doing gender, of emphasizing their

4In this literature, the term housework includes routineand nonroutine housework but not child care.

normative masculinity, overrides the predictionsof economic dependency theory, which wouldforecast that those with less income contributeto the household economy by doing more house-work. It seemed to demonstrate the importanceof gender as a mediator of economic dependencytheory, and it provided one possible answer to thequestion of why it appeared that men were failingto take up the slack in the routine performanceof housework in a period when women wereincreasingly entering the primary labor force.

Brines (1994) did not find any statistically sig-nificant support in her results for the existenceof equivalent gender display behavior by bread-winner wives. The relationship between incomeand housework hours for women was negativeand linear. However, she indicated that therewas a modicum of evidence to suggest that theremight be some effect, which did not reach statis-tical significance in analysis. This idea was takenup and developed further by Greenstein (2000),from whom the idea of complementary male andfemale gender display effects in the performanceof housework became established in the litera-ture. Using large-scale nationally representativeNational Survey of Families and Households(NSFH) data from 1987–1988, Greenstein mod-eled the relative proportion of total houseworkcontributed by men and women in partnershipsinstead of the absolute housework hours ashis dependent variable, to capture the distri-butional aspects of the division of housework.He also added a control for the gender ideologyof partners. His main conclusion is that botheconomically dependent men and breadwinnerwives tended to neutralize the gender deviancein their economic performance by undertakingless and more housework, respectively. That is,Greenstein (2000) reported a curvilinear rela-tionship between housework participation andeconomic dependence for both men and women,whereas Brines (1994) had only shown this effectfor men.

Because Brines’s (1994) most striking find-ings related to low-income and long-term joblessmen, and because the findings of the Britishqualitative studies of the 1980s were conductedamong traditional working-class communities,Greenstein (2000) had expected to find an asso-ciation with a gender ideology variable. In fact,he detected no differences between those withtraditional and nontraditional gender ideologies.Irrespective of their gender ideology, men and


women performed a lesser or greater propor-tion of housework according to extent to whichthey fulfilled normatively defined gender expec-tations—breadwinner wives performed a greaterproportion of housework than other wives, anddependent husbands performed a smaller propor-tion than other husbands. Greenstein thereforechose to refer to his findings as gender-devianceneutralization (Atkinson & Boles, 1984) ratherthan the more general gender display. In a studyof couples organized around the career of thewomen, Atkinson and Boles (1984) documentedbehaviors for both partners that they defined(following Goffman) as covering or concealingto neutralize as much as possible the stigmaattached to their deviant structural positions.From this point on, the two terms gender dis-play and gender-deviance neutralization havebeen used more or less interchangeably in theliterature on the domestic division of house-work to describe the same behaviors. I usegender-deviance neutralization here, even whenreferring to results originally presented as a gen-der display effect, as its meaning in the contextof the division of housework is more specific.Indeed, gender-deviance neutralization can beregarded as an instance of gender display, andthe case of housework as an instance of gender-deviance neutralization.

Further support for a gender-deviance neu-tralization effect among breadwinner womenappeared in Bittman et al. (2003). In an anal-ysis of large-scale quantitative time-use datafrom the Australian 1992 Time Use Survey, andincluding information on hours spent in paidemployment per week for both spouses to takeinto account the wider question of time avail-ability (see Presser, 1994), Bittman et al. (2003)found curvilinear relationships between relativeearnings and housework hours in the case ofmarried women but not in the case of marriedmen. In other words, only among couples withbreadwinner wives was the effect of gender-deviance neutralization evident. They concludedthat gender trumps money in Australia in caseswhere the wife earned more than half thetotal income of both partners. An additionalarticle providing support for a gender-deviance-neutralization effect comes from Evertsson andNermo (2004), who compared Swedish andU.S. couples between 1973 and 2000 usingPSID and the Swedish Level of Living Surveydata. As had Bittman et al. (2003), Evertssonand Nermo (2004) found persistent evidence

of doing gender among U.S. married women,though not among Swedish women, whereresults supported a relative resources (economicdependency) hypothesis, or among men, whereresults were inconclusive (see below). They sug-gested that women in the United States dependedon their husbands to a greater extent than didSwedish women, which led to a stronger gender-deviance-neutralization effect.

The analyses described thus far suggest quitestrong empirical support in large-scale quanti-tative data for a gender-deviance-neutralizationeffect in the performance of housework for bothmen and women. This effect is still frequentlyreferred to in both academic and other media as awinning argument for the significance of genderas measured against more conventional materialindicators of power, such as money (see, e.g.,Tichenor, 2005).

GENDER DEVIANCE NEUTRALIZATION:A REASSESSMENT OF THE EVIDENCE

Although gender-deviance-neutralization behav-ior in the division of housework has been widelyreferred to in the literature on families andhouseholds, doubt has subsequently been caston certain aspects of the original identificationof the phenomenon, both in quantitative and inqualitative research. Recent reassessments of theoriginal articles based on large-scale quantitativeanalysis do not succeed in showing a gender-deviance-neutralization effect. For example,Kan’s (2008a) analysis of pooled British House-hold Panel Data over the period 1993–2003found no conclusive evidence that either highlyeconomically dependent men or highly economi-cally independent women do gender by resortingto a traditional-normative division of domesticlabor. Kan (2008a) directly compared the effectof a relative income measure with a measureof gender attitudes on the absolute number ofhours contributed to housework and found sub-stantive support for a negative, linear effect ofrelative income on housework hours. Both men’sand women’s housework hours decrease signif-icantly as their relative income increases (net ofgender attitudes). There was more limited sup-port, also, for an effect of gender attitudes amongwomen once relative income was controlled for;women with more traditional attitudes did morehousework. However, there was no evidencefor a gender-deviance-neutralization effect,whereby women with high relative incomes or


men with low relative incomes compensate fortheir deviant normative situation through theover- or underperformance of housework.

In a series of articles, Gupta (2006, 2007;Gupta & Ash, 2008) has recently reassessedthe whole basis of the economic depen-dency and gender-deviance-neutralization per-spectives, arguing that it is crucial to take intoaccount women’s autonomous agency and thatprevious findings in relation to housework canbe explained more simply in terms of women’sabsolute rather than relative earnings (which heterms the autonomy hypothesis). Focusing onan empirical test of the economic dependence,gender-display, and autonomy hypotheses basedon NSFH data from 1992 to 1994, Gupta hasquestioned the validity of both the dependenceand the gender-deviance-neutralization thesesby demonstrating that women’s relative earningscontribute little to the explanation of houseworkhours when the analytic model also includes theirabsolute hours. He concluded that the findingthat women who earn more than their husbandscompensate by performing more housework isa statistical artifact and results from the rela-tionship between women’s absolute and relativeearnings. Women at the high end of the relativeearnings distribution are in fact generally poorer,in terms of their own earnings, than are womenin the equality range. This is because pay scalesfor blue-collar jobs are narrower than for white-collar jobs, so that when the wives of blue-collarmen get paid jobs, the ratio of their income totheir husbands’ is often better than the wife-to-husband pay ratio among wealthier families (seeBlumberg & Coleman, 1989). Therefore, theargument that women with high earnings (theusual misinterpretation of the term breadwinnerwife) deliberately do more housework is a spu-rious one, as the findings commonly referred toare in fact largely based on a group of womenwith rather low incomes (Gupta, 2007). In a fur-ther reanalysis of NSFH based on nonparametricstatistics (which do not make assumptions aboutthe underlying pattern of the relationships in thedata), Gupta and Ash (2008) demonstrated thatestimates based on women’s earnings of around$20,000 per year and greater were unreliablebecause they were based on sparse data. Theyshowed again that, overall, women’s earningsare negatively associated with their houseworkhours, independent of their partners’ earningsand their shares of couples’ total earnings. Theyconclude that an alternative model—her money,

her time—is more appropriate to the data thaneither the economic dependency or the gender-display models.

The autonomy model finds some support ina recent mixed-methods study from Britain.Crompton and Lyonette (2008), using the BritishSocial Attitudes Survey of 2002–2006, foundthat higher-earning women reported a signif-icantly less traditional division of domesticlabor than did other women. Semistructuredinterviews with some of the professional andmanagerial class women who earned more thantheir male partners revealed that, although mostdescribed themselves as more likely to takeresponsibility for housework, they gave no indi-cation that this responsibility was part of an effortto neutralize gender deviance. On the contrary,they appeared, according to the authors, to besimultaneously challenging (or ‘‘undoing’’ gen-der) (Deutsch, 2007; Risman, 2009) and doinggender. Their high levels of gender conscious-ness led them to contest their male partners’performance of housework where they thoughtthe division of labor was unfair (Crompton &Lyonette, 2009).

In summary, if we accept the argumentthat earlier findings suggesting the existence ofa gender-deviance-neutralization effect amongwomen were due to a confusion betweenwomen’s relative and absolute earnings, andcombine this with both the quantitative (Gupta,2008) and qualitative (Crompton & Lyonette,2009) evidence that women with higher earn-ing power do less housework than those withlower earning power (perhaps partly becausethey are buying domestic services from others5),we have quite a powerful argument for reject-ing the gender-deviance-neutralization modelfor high-earning women.

This still leaves outstanding the question ofwhether men (or some men) do gender devianceneutralization. The earliest qualitative papersand subsequent quantitative studies based onanalyses using data primarily from the 1970sand 1980s (Akerlof & Kranton, 2000; Bittmanet al., 2003; Brines, 1994; Greenstein, 2000;Morris, 1985, 1987; Wheelock, 1990) suggested

5It should be noted that the purchase of suchservices would be counterproductive for a woman earningsubstantially more than her male partner and attempting toneutralize the gender deviance of her situation through the(over)performance of housework!


that working-class men or men with the low-est levels of economic resources display theirgender by reasserting their traditional masculinegender identity through the underperformanceof housework. However, large-scale quantita-tive research based on more recent data (fromthe 1990s and 2000s) has failed to show thisassociation (Kan, 2008b). In addition, Evertssonand Nermo’s (2004) conclusion that the evi-dence from the United States for men’s gender-deviance-neutralization behavior in relation tohousework is inconclusive was based on anal-yses of PSID data from the period 1973–1999.They found evidence for men’s gender displayfor the 1973 data but not for the 1999 data. Thissynopsis raises two questions on the standing ofthe issue of men’s gender-deviance neutraliza-tion: First, what are (or were) the characteris-tics of the men who were originally identifiedas displaying gender through nonperformanceof housework? Second, could men’s gender-deviance neutralization be a thing of the past?

The answer to the first question appears tocome both from a consideration of the earli-est qualitative articles and from some recentreassessments of Brines’s (1994) and Green-stein’s (2000) analyses based on large-scale data.The earliest qualitative research that identifiedgender-deviance neutralization among men wasconducted among British traditional working-class communities, focusing particularly onthose households in which men had been maderedundant or occupied marginal labor marketpositions—in other words, among groups likelyto have the lowest incomes and the most tra-ditional gender attitudes. This identification issupported by Gupta’s (1999) reassessment ofBrines’s (1994) original analysis of 1984 PSIDdata in which he refers to the extreme tail ofthe distribution of male respondents. He showsthat once this tail—the small proportion of long-term jobless men—is removed from the analysis,the relationship between income and house-work hours becomes linear and negative, inconformity with economic dependency theory.The idea that this extreme tail was responsi-ble for the identification of the male gender-deviance-neutralization effect is also supportedin the work of Bittman et al. (2003), whofound that when they deleted the 2–3% ofcases in which men contribute no earningsto the household from the 1987–1988 NSFHdata, the male gender-deviance-neutralizationeffect disappeared. They conclude that ‘‘the U.S.

phenomenon of men’s responding to economicdependence on women by decreasing houseworkis real but that it comes only from the extremetail of the men’s earnings distribution’’ (p. 207).

These recent reassessments of earlier analysesboth conclude that the effect found for men in theinitial quantitative research was real but confinedto a small and specific group of men. However,an alternative explanation, which adds anotherelement of doubt, can be found in the literature onreporting differences in questionnaire-based andtime-use-diary survey instruments. These report-ing differences could explain why, for example,neither Kan’s (2008a) analysis of British time-use data nor Bittman et al.’s (2003) analysis ofAustralian time-use data identified any gender-deviance-neutralization effect for men. Thisliterature is generally treated as technical in con-tent, so its implications for assessments of thegender-deviance-neutralization thesis have notbeen widely recognized. It is paralleled by someindicative findings from the qualitative literatureon husband’s performance of housework.

In a comparison of time-use-diary data withquestionnaire responses from the same respon-dents to a nationally representative British sur-vey (Home On-Line, 1999–2001), Kan (2008b)found systematic differences between the twotypes of responses, which varied by gen-der. In general, women were more accuratein their reporting of housework hours. Inother words, there was greater correspondencebetween women’s time-use-diary reports andtheir questionnaire reports than in the case ofmen, who in general overestimated their house-work contributions in questionnaire reports bycomparison with their diary reports. It is knownthat time-use-diary reports of time spent inparticular activities are more reliable than retro-spective recall questions asked in questionnaires(Bianchi, Robinson, & Milkie, 2006). This isso because time-use diaries are in general bothmore immediate than responses to such ques-tions (suffering less from the problem of recallbias), and more important, they are less subjectto normative response errors (in other words, theperceived pressure to respond to a question in asocially approved way). Tellingly for the currentdiscussion, though, while men with traditionalgender attitudes reported more housework timein the questionnaire than in the time-use diarythan was the general case for men, those menwith traditional gender attitudes who contributeda relatively high number of hours to housework


reported fewer hours in the questionnaire thanthey recorded in the time-use diary. The find-ing that men with traditional gender attitudeswho also make a relatively large contribution tohousework in particular tend to underreport theirhours of housework therefore calls into questionresearch based on questionnaire responses (e.g.,PSID, NSFH) that suggested that men with thelowest relative incomes do less housework ina process of gender-deviance neutralization. Itcould simply be that they underreport their timespent in housework, and though such under-reporting could also be viewed as a form ofgender-deviance neutralization, it is of a quitedifferent type than the literature suggests.

Because men, in general, tend to overreporttheir housework, it is an interesting questionwhy a particular group of men (those with tra-ditional gender attitudes who also do a lot ofhousework) should be underreporting, therebyperhaps contributing to a spurious finding ofa male gender-deviance-neutralization effect inthe actual performance of housework. Thereare hints of this phenomenon in the qualitativeliterature on the relationship between gender,family work, and gender attitudes among sharingcouples. For instance, various suggestive indi-cations about the underreporting of men’s con-tributions among working-class couples appearin Deutsch’s (1999) work on equally sharingcouples, based on interviews conducted with150 dual-earning couples. As an example, oneworking-class couple (Paul and Mary) appearedembarrassed by their equality, especially Mary,who reported that she doesn’t always tell peo-ple what her husband does at home. They areportrayed in contrast to Janet and Daniel, col-lege professors, who are proud of their degreeof equality and eager to report it. Deutsch con-firmed (by personal correspondence) that, in theinitial telephone screening questions she used todetermine equality of sharing, men from blue-collar couples tended to report less houseworkthan they actually contributed according to thesubsequent depth interviews with them and theirwives. In other words, their responses to the ini-tial, more superficial questions conform more totraditional gender expectations than does theirsubsequent discourse in the intensive interviews.The implication is that the reporting issue, too,may represent a form of gender display. Whatcould be happening is that particular groupsof working-class men are only reporting fewerhours of housework, which also suggests a kind

of gender-deviance-neutralization response butof a different kind than the literature has sug-gested to date.

AN END TO GENDER DEVIANCENEUTRALIZATION THROUGH HOUSEWORK?

The second question that I raised above inconnection to men’s gender-deviance neutral-ization through housework was, has there beenany change over time? The coincidence ofthe findings of the earliest quantitative stud-ies and the early qualitative findings identifyingmen’s gender-deviance neutralization throughhousework among low-income, working-classhouseholds in the 1970s and 1980s suggest thatsome real phenomenon was recognized. If weaccept that this is so but that it does not appear inlater data (see, e.g., Evertsson & Nermo, 2004),what additional evidence is there that might sup-port the idea of change among this particulargroup of men? Here again, a mix of sources cancontribute to our understanding. I present a rangeof indicative evidence, dating from the 1980s tothe first decade of the 21st century, which sug-gests that there was always at least some degreeof flexibility in men’s performance of familywork according to employment and sociodemo-graphic characteristics of the couple, even (orindeed, perhaps, especially) among those withtraditional gender ideologies and low-incomeoccupations.

Few studies have attempted large-scale longi-tudinal analyses of what happens to the domesticdivision of labor in couples in which the womanenters paid employment. The most comprehen-sive study used data from three longitudinalnational household panel surveys—the BritishHousehold Panel Survey, the German Socioe-conomic Panel, and the U.S. Panel Study ofIncome Dynamics (Gershuny, Bittman, & Brice,2005) to show direct longitudinal evidence ofchange in the balance of domestic labor incouples during the 1980s and 1990s. Womenentering full-time paid work made immediatelarge downward adjustments in their domesticwork time, but men also made changes—theircontribution to domestic work increased suc-cessively in the years following their partner’sentry into full-time employment (a process thatGershuny et al., 2005, referred to as lagged adap-tation). Similar results appear in the work ofCunningham (2007), who analyzed a 30-yearpanel data sample of parents and children in


Detroit from the 1960s to the 1990s, showingthat women’s previous employment history hadsignificant positive effects on husbands’ contri-butions to routine housework, an effect mediatedby wives’ attitudes to gender egalitarianism.

These findings demonstrating flexibility inmale partners’ responses to the woman’semployment status, however, still leave openthe question of whether there has been large-scale change observed among men belonging tothose socioeconomic groups with low-incomeemployment (commonly associated with themost traditional gender ideologies). In fact, thereis some strong evidence for this. Data from theUnited States suggest that, whereas managerialand professional couples were the most likely toshare family work in the 1970s and 1980s, by the1990s and 2000s, the most change (and some-times the most sharing) had occurred in couplesin blue- and pink-collar professions (Coltrane,2000, 2004). In an analysis of dual-earner house-holds using British and U.S. time-use datafrom the 1970s to the turn of the 21st cen-tury (Sullivan, 2006, forthcoming) demonstratedsome similar results. Over the period, men indual-earner couples with the lowest levels ofeducational attainment from both Britain andthe United States had increased the time theyspent on housework to a far greater extent thanequivalent men with higher levels of educa-tional attainment. Starting from a lower base inthe 1970s, by the year 2000–2001, men in dual-earner couples who did not complete secondaryeducation had assumed as much responsibilityfor household tasks as men who had some highereducation. So whereas in the earlier periods therewas a clear gradient in such men’s houseworktime according to educational attainment (withthose attaining higher levels of education con-tributing more time), by the later surveys, thisgradient had disappeared. For example, in theUnited States, although the most highly edu-cated group (college educated) increased thetime they spent in housework by 33% over theperiod 1965–2003, those who graduated fromhigh school increased theirs by 52%, and thosewith the lowest level of educational attainment(0–11th grade only) doubled theirs with a 100%increase in contributions. So over the 25-yearperiod between the surveys, men in dual-earnercouples with the lowest levels of educationalattainment had at least caught up with thecontributions of men with higher levels. Withrespect to country differences, the percentage

increases over the period 1975–2000 for thegroups with lower levels of educational attain-ment (incomplete secondary and secondary)from Britain were greater than those over theperiod 1975–2003 in the United States, from alower base in 1975. However, by 2000, Britishmen’s contributions equaled or even exceededthose of American men in 2003. For both coun-tries, then, these findings demonstrate a clearcatch-up effect, whereby men in dual-earnercouples with lower levels of educational attain-ment, from a lower starting point 25–35 yearsago, equaled or even exceeded the contributionsto domestic work of the most highly educatedgroup of men by the start of the 21st century.

In support of these recent quantitative find-ings, there are several indicative suggestionsfrom the qualitative literature to suggest thatsome men in traditional working-class occu-pations were in fact helping out around thehome even in the 1980s and 1990s. Severalauthors at that time pointed out that working-class men, the least likely to endorse genderequality, frequently engaged in behavior thatwas more egalitarian than the traditional gen-der ideologies they espoused (Deutsch, 1999;Hochschild, 1989; Pyke, 1996). I examine thisfirst in the early qualitative studies referred toabove, based on research conducted in the 1980s.Morris (1987), for example, found that amongher sample of couples from the old industrialnortheastern England, the young and childlesswhere both partners were employed full-time orwhere the male partner was unemployed hada remarkably more flexible domestic divisionof labor: ‘‘Although the traditional ideas aboutmale and female responsibility remain, there isalso a fairly high level of male participation indomestic tasks, notably washing dishes, cook-ing and cleaning through the house on a fairlyregular basis’’ (p. 203).

In such cases, the notion of justice is oftenused to explain male participation in domestictasks. For most cases in Wheelock’s (1990)study of low-income couples from in theNortheast, she, too, concluded that traditionalvalues on gender roles presented no barriers tothe negotiation and adjustment of the division oflabor where practical circumstances suggested it.Flexibility in the face of practical circumstancesand the appeal to notions of fairness (Thompson,1991) are also themes that run through a bodyof research on the division of household tasksamong couples working unsocial shift hours


(who are more likely to be working classthan middle class). In a comprehensive reviewof the findings of research on the domesticdivision of labor, Coltrane (2000) argued thatwives’ and husbands’ employment schedulesare perhaps the most consistent and importantpredictors of domestic sharing that researchershad documented to date. In an analysis of1980s NSFH data, Presser (1994), for example,found that spouses’ employment schedules wereamong the most important factors influencinghusbands’ performance of tasks traditionallydone by women. The more hours husbandswere not doing paid work during times whentheir wives were, the more likely they wereto do feminine-associated housework, therebybreaking traditional gender expectations.

Although the central quantitative literatureand the discussion so far has addressed theperformance of housework without reference tocontributions to child care, further backgroundon motivations and justifications may perhapsbe gleaned from some qualitative reports ondifferences in child care between working-class and middle-class families. For example,Deutsch and Saxon (1998) were interested ina group of U.S. dual-earner couples defyingtraditional gender stereotypes while maintainingtraditional gender ideologies. In an articletitled ‘‘Traditional Ideologies, Non-TraditionalLives,’’ they interviewed couples employed onalternating shifts who had dependent childrenand who were therefore, by necessity, alternatingchild care. The outcome of this employmentpattern was that fathers spent more timewith their children than did other fathers andfrequently performed the feminine-defined tasksassociated with the physical care of children.They point out that the occupations relyingheavily on shift work are traditional working-class occupations and that the expressed stancesof the fathers of their sample were generallyconservative in relation to gender. However, inintensive interviews, it became clear that one ofthe primary motivations for these couples wasthe idea of fairness (see Thompson, 1991).

A decade later, Gerstel, Clawson, andHuyser (2007) compared physicians (high-earning professionals) and emergency medicaltechnicians (EMTs, low-income employmentfrequently involving shift work). They found thatthe doctors do what they call public fatherhoodbut don’t put in the time. They were muchmore likely to emphasize their involvement

in the special events in their children’s lives(sports and performances) over the tasks ofdaily care. However, for the working-class EMTfathers, there was a very different pattern andmotivation. They put the emphasis on privatefathering, talking about their involvement inthe daily routines of their children’s lives—theconventionally feminized tasks of picking themup from day care or school, feeding themdinner, or staying at home with them when theygot sick. Many reported that they were happywith their often-antisocial working schedulesbecause it allowed them to participate fully inchild care and to spend a long time at homewith their children. Gerstel et al. (2007) wrote,‘‘These working class men exhibit a model ofmasculinity—based both at home and on thejob—that provides valued involvement in theirchildren’s daily lives’’ (p. 25).

They locate their findings in an existingbody of literature that points to a disjuncturebetween traditional masculine gender attitudesemphasizing a breadwinner or provider roleand the actual practice of fatherhood amongworking-class men (e.g., Pyke, 1996; Williams,2000). This is contrasted with studies conductedamong highly educated professional men, whichsuggest that, though they espouse egalitarianideals, in practice, they often do not live up tothose ideals given constraints of employmentthat emphasize a culture of long hours andput a status premium on busyness (Coltrane,2004; Pyke, 1996; Williams, 2000). Althoughthese studies relate to fatherhood (rather thanto housework), it is clear that there is a basisfor generalization in terms of the notions offairness and practicality in the division of laborexpressed by the interviewees in the face ofparticular employment conditions. For example,Deutsch and Saxon (1998) wrote, ‘‘They adoptthe non-traditional roles they do not becausethey question the gendered basis of those roles,but because with women’s entry into the paidlabor force, the exchange underlying those roleshas been undermined’’ (p. 346).

SUMMARY AND CONCLUSIONS

The idea of gender-deviance neutralization inrelation to housework has been so highlyinfluential because it lent support to developingfeminist theories concerning the relationshipsamong gender, relative income, and domesticlabor. In the case of men’s gender-deviance


neutralization, it contributed to the explanationof why men had not increased their houseworkcommensurately as women moved into paidemployment; in the case of women’s gender-deviance neutralization, it provided strongsupport for the idea that gender could have amore powerful influence on domestic relationsthan relative income (gender trumps money),thus casting doubt on conventional quasi-economic explanations that tended to ignoregender.

In this article, I have reviewed and reassessedthe quantitative studies that firmly establishedthe idea of gender-deviance neutralization inthe area of housework through to more recentresearch that has raised substantial questionsabout it. I have traced the origins of theidea of men’s gender-deviance neutralizationin some early qualitative research from the1980s and 1990s in both the United Statesand Britain, and I have deployed indicativefindings from the qualitative literature to exploresome of the possible motivations of thoseprofessional women and working-class menwhose performance of housework is notablydifferent from that which the gender-deviance-neutralization hypothesis suggests.

I conclude that considerable doubt is caston the gender-deviance-neutralization thesis inthe sense that it is commonly understood, inrelation to both women and men. For women,the common interpretation is that a group ofhigh-earning women married to men earning lessthan them deliberately do gender by contributingmore to housework than those with more equalrelative incomes in a kind of compensationfor their deviant, more masculine, structuralsituation. Subsequent large-scale research hasshown that the original identification of thiseffect was the outcome of a misspecificationin analysis. In the case of men, subsequentresearch both calls into question the originalidentification in large-scale quantitative analysisof gender-deviance neutralization and raisesthe question of whether subsequent dramaticchanges in the contributions to housework ofmen from lower socioeconomic levels have notput an end to gender-deviance-neutralizationbehaviors.

For women, Gupta (2006, 2007) and Guptaand Ash (2008) have demonstrated convinc-ingly that the original quantitative studiesidentifying gender-deviance neutralization asit is commonly understood resulted from the

misspecification of the analytic models, whichrelied on measures of relative income betweenspouses and did not include absolute income.Indeed, the idea that high-earning women withstrong positions in the primary labor market wereat the same time compensating by undertakingmore housework than those with more equalincome distributions is counterintuitive and fliesin the face of contrary evidence from the relativeresources perspective that documents empower-ment deriving from labor market position. Theapparent identification of gender-deviance neu-tralization behavior in analysis can be explainedby the simple fact that those women withhigher earnings than their husbands are morelikely to be in lower income households withmore traditional gender ideologies. To postulatea gender-deviance-neutralization effect that islogically equivalent to men’s gender-devianceneutralization (i.e., women in gender-deviantstructural locations display a more traditional-normative feminine gender through the over-compensatory performance of housework), highrelative income has been perceived as meaninghigh absolute income, which is not the case.When absolute income is added to the mod-els, the gender-deviance-neutralization effectdisappears.

Support for the autonomy model is alsofound in other types of analysis. For example,according to the research of Crompton andLyonette (2009), support for the idea ‘‘hermoney, her time’’ is found in the discourseof professional and managerial women whooutearn their husbands. I do not claim thatwomen in this situation do not do gender invarious ways in an effort to minimize the effectson their male partners of their gender-deviantstructural situation. It is also clear that hermoney, her time assumes that women still takeoverall responsibility for the management ofhousehold labor if they chose to use their moneyto employ domestic help. What I do question iswhether this doing of gender any longer (if it everdid) takes the form of the overperformance ofroutine housework. There are many other ways,perhaps more subtle and telling, as describedby Tichenor (2005), among others, in which amale partner may be deferred to, placated, ormade to feel more in control without resortingto the overperformance of housework. Giventhe high levels of gender consciousness amongsuch women (Crompton & Lyonette, 2009), it isquite difficult to imagine them hovering under


the furniture nightly in an effort to soothe thefeelings of their male partners.

In the case of men’s gender-deviance-neutralization behavior, we need to take threerelated strands of argument into account. First,although such behavior in the performance ofhousework was first identified in qualitativeresearch among groups of working-class menwith traditional gender ideologies, other evi-dence suggests that it is possible that therewas always more flexibility in practice amongthese groups. In fact, male gender-deviance-neutralization behavior in the major quantitativestudies that first identified it has been foundin subsequent reanalyses to be restricted to themost extreme 2–3% of men in terms of incomeand labor market position. Support for a greaterdegree of flexibility in such families in the faceof practical circumstances comes both from theoriginal studies of Morris (1985, 1987) andWheelock (1990) and from longitudinal quan-titative research showing clear adaptation overtime by men to their wives’ changing employ-ment situations. Second, there is recent evidencefrom the methodological literature comparingtime-use and questionnaire reporting, supportedby references in some smaller-scale qualita-tive interview studies, that men with moretraditional gender ideologies may disproportion-ately underreport their contributions to domesticwork in responses to survey-type questions.This may have led to a mistaken identificationof male gender-deviance-neutralization behav-ior in quantitative papers, which was in factdue to reporting bias. Bittman et al. (2003),for example, found evidence for male gender-deviance neutralization from questionnaire datafrom the United States but not in time-use datafrom the same period in Australia. The method-ological literature suggests that this may reflectreporting differences for the different surveyinstruments rather than national level differencesin men’s actual performance of housework.Whether or not such underreporting itself consti-tutes a kind of gender-deviance neutralization, itis not the same behavioral effect as is frequentlyquoted and referred to.

Third, there is some strong evidence forchange in the performance of houseworkamong men since the period in the 1980swhen male gender deviance neutralization inthis area was first identified, in particularamong those from lower socioeconomic levels,which are generally associated with more

traditional gender ideologies (Sullivan, 2006,forthcoming). Men in dual-earner couples withthe lowest levels of educational attainment havedisproportionately increased their contributionsto housework over the past 20 years. Theimplication is that, if male gender-devianceneutralization in the performance of houseworkdid exist 20–30 years ago, it probably did soamong a relatively small group in specificstructural locations, and recent changes in bothattitudes and practice among men have meantthat it is probably no longer a significantphenomenon.

Finally, it should be noted that all the studiesreferred to in this article are of heterosexualcouples. Indeed, same-sex couples are excludedfrom the large-scale quantitative analyses, andrace enters the analyses only as a controlvariable (so that it is not possible to examinedifferences in gender-deviance-neutralizationbehavior between races or between couples ofdifferent sexualities). It is true that the mainreferential theoretical perspectives of the gender-deviance-neutralization hypothesis (economicexchange, economic dependency, and relativeresource theories) and even gender displayitself have primarily been formulated in relationto heterosexual couples. However, althoughit is correct not to assume that the samegender display behaviors exist in same-sexcouples (the reason for the exclusion of same-sex couples from the large-scale quantitativeanalyses), the intriguing question of genderdeviance neutralization in the performance ofhousework among same-sex couples remains tobe addressed.

NOTE

Oriel Sullivan is Research Reader at the Departmentof Sociology, University of Oxford. She has publishedwidely in the areas of time use and gender relations.Recent publications include Changing Gender Relations,Changing Families (Rowman and Littlefield Gender LensSeries); ‘‘Father-Friendly Polices and Time Use Data in aCross-National Context,’’ Annals of the American Academyfor Political and Social Science (with S. Coltrane, L.McAnnally, and E. Altintas).

The writing of this article was supported by Economicand Social Research Council Grant No. RES 060 25 0037.

REFERENCES

Akerlof, G. A., & Kranton, R. E. (2000). Economicsand identity. Quarterly Journal of Economics, 115,715 – 753.


Atkinson, M. P., & Boles, J. (1984). WASP (wivesas senior partners). Journal of Marriage and theFamily, 46, 861 – 870.

Berk, S. (1985). The gender factory: The apportion-ment of work in American households. New York:Plenum Press.

Bianchi, S. M., Robinson, J. P., & Milkie, M. A.(2006). Changing rhythms of American family life.New York: Russell Sage Foundation.

Bittman, M., England, P., Folbre, N., Sayer, L., &Matheson, G. (2003). When does gender trumpmoney? Bargaining and time in household work.American Journal of Sociology, 109, 186 – 214.

Blumberg, R. L., & Coleman, M. T. (1989). A theo-retical look at the gender balance of power in theAmerican couple. Journal of Family Issues, 10,225 – 250.

Brines, J. (1994). Economic dependency, gender, andthe division of labor at home. American Journal ofSociology, 100, 652 – 688.

Butler, J. (2006). Gender trouble: Feminism and thesubversion of identity. London: Routledge.

Coltrane, S. (2000). Research on household labor.Journal of Marriage and the Family, 62,1209 – 1233.

Coltrane, S. (2004). Elite careers and family com-mitment: It’s (still) about gender. Annals of theAmerican Academy of Political and Social Science,596, 214 – 220.

Crompton, R., & Lyonette, C. (2008). Who doesthe housework? The division of labour withinthe home. In A. Park, J. Curtice, K. Thomson,M. Phillips, & M. Johnson (Eds.), British socialattitudes: The 24th report (pp. 53 – 80). London:Sage.

Crompton, R., & Lyonette, C. (2009). Partners’relative earnings and the domestic division oflabour. Manuscript submitted for publication.

Cunningham, M. (2007). Influences of women’semployment on the gendered division of householdlabor over the life course: Evidence from a 31-year panel study. Journal of Family Issues, 28,422 – 444.

Deutsch, F. M. (2002). Halving it all: How equallyshared parenting works. Cambridge, MA: HarvardUniversity Press.

Deutsch, F. M. (2007). Undoing gender. Gender andSociety, 21, 106 – 127.

Deutsch, F. M., & Saxon, S. E. (1998). Traditionalideologies, nontraditional lives. Sex Roles, 38,331 – 362.

Evertsson, M., & Nermo, M. (2004). Dependencewithin families and the division of labor:Comparing Sweden and the United States. Journalof Marriage and Family, 66, 1272 – 1286.

Gershuny, J., Bittman, M., & Brice, J. (2005). Exit,voice, and suffering: Do couples adapt to changingemployment patterns? Journal of Marriage andFamily, 67, 656 – 665.

Gerstel, N., Clawson, D., & Huyser, D. (2007).Explaining job hours of physicians, nurses, EMTs,and nursing assistants: Gender, class, jobs andfamily. In B. Rubin (Ed.), Workplace temporalities(pp. 369 – 402). Amsterdam: JAI, Elsevier.

Greenstein, T. N. (2000). Economic dependence,gender, and the division of labor in the home:A replication and extension. Journal of Marriageand the Family, 62, 322 – 335.

Gupta, S. (1999). Gender display? A reassessmentof the relationship between men’s economicdependence and their housework hours. Paperpresented at the annual meeting of the AmericanSociological Association, Chicago.

Gupta, S. (2006). Her money, her time: Women’searnings and their housework hours. Social ScienceResearch, 35, 975 – 999.

Gupta, S. (2007). Autonomy, dependence or display?The relationship between married women’s earn-ings and housework. Journal of Marriage andFamily, 69, 399 – 417.

Gupta, S., & Ash, M. (2008). Whose money, whosetime? A nonparametric approach to modeling timespent on housework in the United States. FeministEconomics, 14, 93 – 120.

Hochschild, A. R., with Machung, A. (1989). Thesecond shift: Working parents and the revolutionat home. Berkeley: University of CaliforniaPress.

Kan, M. Y. (2008a). Does gender trump money?Housework hours of husbands and wives in Britain.Work, Employment and Society, 22, 45 – 66.

Kan, M. Y. (2008b). Measuring housework participa-tion: The gap between ‘‘stylised’’ questionnaireestimates and diary-based estimates. Social Indi-cators Research, 86, 381 – 400.

Morris, L. (1985). Renegotiation of the domesticdivision of labour in the context of redundancy.In B. Roberts, R. Finnegan, & D. Gallie (Eds.),New approaches to economic life (pp. 400 – 416).Manchester, UK: Manchester University Press.

Morris, L. (1987). The life cycle and the labourmarket in Hartlepool. In A. Bryman, B. Bytheway,P. Allatt, & K. Teresa (Eds.), Rethinking the lifecycle (pp. 192 – 206). Dobbs Ferry, NY: SheridanHouse.

Presser, H. B. (1994). Employment schedules amongdual-earner spouses and the division of labor bygender. American Sociological Review, 59,48 – 69.

Pyke, K. D. (1996). Class-based masculinities: Theinterdependence of gender, class, and interpersonalpower. Gender and Society, 10, 527 – 549.

Risman, B. (2009). From doing to undoing: Genderas we know it. Gender and Society, 23, 81 – 84.

Sullivan, O. (2006). Changing gender relations,changing families: Tracing the pace of change.New York: Rowman and Littlefield.


Sullivan, O. (in press). Changing differences byeducational attainment in fathers’ domestic labourand child care. Sociology, 44.

Thompson, L. (1991). Family work: Women’s senseof fairness. Journal of Family Issues, 12, 181 – 196.

Tichenor, V. (2005). Maintaining men’s dominance:Negotiating identity and power when she earnsmore. Sex Roles, 53, 191 – 205.

West, C., & Zimmerman, D. H. (1987). Doing gender.Gender and Society, 1, 125 – 151.

Wheelock, J. (1990). Husbands at home: Thedomestic economy in a post-industrial society.London: Routledge.

Williams, J. (2000). Unbending gender: Why familyand work conflict and what to do about it. NewYork: Oxford University Press.

Date post:	26-Jun-2020
Category:	Documents
Upload:	others
View:	2 times
Download:	0 times

An End to Gender Display Through the Performance of ... to gender display... · article, I review...

Documents