JOU

RN

AL

OF

HE

AL

TH

AN

D S

OC

IAL

BE

HA

VIO

RD

EC

EM

BE

R 2

00

9, V

OL

UM

E 50

, NU

MB

ER

4

V O LU M E 5 0 N U M B E R 4 D E C E M B E R 2 0 0 9

2008 LEO G. REEDER AWARD PAPERToward Explaining Mental Health DisparitiesCarol S. Aneshensel

SUBSTANCE USE IN THE TRANSITION TO ADULTHOODAge at First Birth and Alcohol UseJoseph D. Wolfe

The Gender Gap in Alcohol Consumption during Late Adolescence and Young Adulthood: Gendered Attitudes and Adult RolesC. André Christie-Mizell and Robert L. Peralta

UNRAVELING RACIAL AND ETHNIC HEALTH DISPARITIESResidential Segregation and Birth Weight among Racial and Ethnic Minorities in the United StatesEmily Walton

Perceived Ethnic Discrimination versus Acculturation Stress: Influences on Substance Use among Latino Youth in the SouthwestStephen Kulis, Flavio Francisco Marsiglia and Tanya Nieri

Living Arrangements, Social Integration, and Loneliness in Later Life:The Case of Physical DisabilityDavid Russell

Putting Work to Bed: Stressful Experiences on the Job and Sleep QualitySarah A. Burgard and Jennifer A. Ailshire

PM

S202

PM

S425

JOU

RN

AL

OF

HE

AL

TH

AN

D S

OC

IAL

BE

HA

VIO

RD

EC

EM

BE

R 2

00

9, V

OL

UM

E 50

, NU

MB

ER

4

V O LU M E 5 0 N U M B E R 4 D E C E M B E R 2 0 0 9

2008 LEO G. REEDER AWARD PAPERToward Explaining Mental Health DisparitiesCarol S. Aneshensel

SUBSTANCE USE IN THE TRANSITION TO ADULTHOODAge at First Birth and Alcohol UseJoseph D. Wolfe

The Gender Gap in Alcohol Consumption during Late Adolescence and Young Adulthood: Gendered Attitudes and Adult RolesC. André Christie-Mizell and Robert L. Peralta

UNRAVELING RACIAL AND ETHNIC HEALTH DISPARITIESResidential Segregation and Birth Weight among Racial and Ethnic Minorities in the United StatesEmily Walton

Perceived Ethnic Discrimination versus Acculturation Stress: Influences on Substance Use among Latino Youth in the SouthwestStephen Kulis, Flavio Francisco Marsiglia and Tanya Nieri

Living Arrangements, Social Integration, and Loneliness in Later Life:The Case of Physical DisabilityDavid Russell

Putting Work to Bed: Stressful Experiences on the Job and Sleep QualitySarah A. Burgard and Jennifer A. Ailshire

PM

S202

PM

S425

TH

E JO

UR

NA

L O

F H

EA

LTH

AN

DSO

CIA

L B

EH

AVIO

R(ISSN

00

22-1465)1430

K Street N

W, Suite 60

0W

ashington, DC

200

05

Periodicals postage paidat W

ashington, DC

andadditional m

ailing offices

Why Use EQS?EQS is the only Structural Equation Modeling (SEM) program to provide new accurate statistics for non-normal data.

EQS’ comprehensive data management capabilities allows users to work with data without having to use other programs such as SPSS® or PRELIS®. Imports SPSS files.

EQS tests the full range of SEM includ-ing latent regression, confirmatory factor analysis, LISREL®, structured means, multiple population, latent growth curve, multilevel analysis, correlation structure & categorical data models.

EQS’ improved DIAGRAMMER makes model setup easy. Use of command lan-guage and knowledge of matrix algebra are no longer necessary!

EQS is preferred by thousands of re-searchers worldwide providing them with a wide range of easy to use statistical and data exploration tools.

Multivariate Software, Inc.818-906-0740 or 800-301-4456FAX: 818-906-8205 Email: sales@mvsoft.com

New EQS 6.1 Features• Help & Setup Wizards, Graphs & Plots• Automatic Command Generator• Statistical and Data Exploration Tools• Missing Data: EM, Robust ML, MCAR Tests• Tests for Smaller Samples: F-tests, Residual-based, Mardia’s Kurtosis• Case Weighted Methods for Outliers, Influ-ential Cases, Complex Samples • Multilevel Models: ML, MUML, HLM-like• Heterogeneous Kurtosis Method & Test, Heteroskedastic Regression • Easy Bootstrapping & Simulation with Sum-maries• Maximal, Model-based, and Greatest Lower Bound Reliability• Polyserial/polychoric Correlations for IRT and SEM

Download a FREE Demo at:www.mvsoft.com

Structural Equation Modeling Has Never Been Easier!

New EQS Version 6.1

Job Name:95547 PDF Page:95547pbc.p1.pdfColor:

Date:09-11-19PANTONE 202 C PANTONE 425 C

EditorELIZA K. PAVALKO

Indiana University

Deputy EditorsPAMELA BRABOY JACKSON BERNICE PESCOSOLIDO

Indiana University ——Indiana UniversityBRIAN POWELL JILL QUADAGNO

Indiana University ——Florida State UniversitySCOTT SCHIEMAN——University of Toronto

Associate EditorsRONALD J. ANGEL

University of Texas-AustinCLIFFORD L. BROMAN

Michigan State UniversityCHRISTOPHER BROWNING

Ohio State UniversityRALPH CATALANO

University of California, BerkeleyBRIAN K. FINCH

San Diego State UniversityGILBERT C. GEE

University of California, Los AngelesSUSAN GORE

University of Massachusetts, BostonBRIDGET K. GORMAN

Rice UniversityJOSEPH GRZYWACZ

Wake Forest UniversitySTEVEN HAAS

Arizona State UniversityPAMELA HERD

University of Wisconsin–MadisonKARA JOYNER

Bowling Green State UniversityDONALD A. LLOYD

Florida State UniversityKAREN LUTFEY

New England Research InstitutesSCOTT LYNCH

Princeton UniversityFRED E. MARKOWITZ

Northern Illinois UniversityPEGGY MCDONOUGH

University of TorontoELIZABETH MENAGHAN

Ohio State University

ILAN H. MEYERColumbia University

RICHARD ALLEN MIECHUniversity of Colorado Denver

MELISSA A. MILKIEUniversity of Maryland

SAMUEL NOHUniversity of Toronto

DEBORA K. PADGETTNew York University

FRED C. PAMPELUniversity of Colorado at Boulder

JO C. PHELANColumbia University

JOHN R. REYNOLDSFlorida State University

CHRISTIAN RITTERNortheast Ohio Universities Colleges of Medicine andPharmacy

STEPHANIE A. ROBERTUniversity of Wisconsin–Madison

SUSAN ROXBURGHKent State University

TERESA L. SCHEIDUniversity of North Carolina at Charlotte

MICHAEL J. SHANAHANUniversity of North Carolina at Chapel Hill

MARK TAUSIGUniversity of Akron

STEFAN TIMMERMANSUniversity of California, Los Angeles

KRISTI L. WILLIAMSOhio State University

ANDREA E. WILLSONUniversity of Western Ontario

Managing EditorsINDERMOHAN VIRK

Indiana UniversityJOSEPH D. WOLFE

Indiana University

Copy EditorANDREW J. COGNARD-BLACK

St. Mary’s College of Maryland

American Sociological Association Executive OfficerSALLY T. HILLSMAN

Edited by Julia Adams, Jeffrey Alexander, Ron Eyerman and Philip Gorski

Published on behalf of the American Sociological Association

Published for the American Sociological Association, this important journal covers the full range of sociological theory—from ethnomethodology to world systems analysis, from commentaries on the classics to the latest cutting-edge ideas, and from re-examinations of

neglected theorists to metatheoretical inquiries. Its themes and contributions are interdisciplinary, its orientation pluralistic, its pages open to commentary and debate. Renowned for publishing the best international research and scholarship, Sociological Theory is essential reading for sociologists and social theorists alike.

For more information and to subscribe online visitwww.blackwellpublishing.com/st

SociologicalTheory

SOTH Filler Ad 32p_x_45p.indd 1 5/1/08 9:29:51 AM

Job Name:95547 PDF Page:95547pbc.p2.pdfColor:

Date:09-11-19Black

2008 LEO G. REEDER AWARD PAPERToward Explaining Mental Health Disparities 377Carol S. Aneshensel

SUBSTANCE USE IN THE TRANSITION TO ADULTHOODAge at First Birth and Alcohol Use 395Joseph D. Wolfe

The Gender Gap in Alcohol Consumption during Late Adolescence and Young Adulthood:Gendered Attitudes and Adult Roles 410C. André Christie-Mizell and Robert L. Peralta

UNRAVELING RACIAL AND ETHNIC HEALTH DISPARITIESResidential Segregation and Birth Weight among Racial and Ethnic Minorities in the United States 427Emily Walton

Perceived Ethnic Discrimination versus Acculturation Stress: Influences on Substance Use among Latino Youth in the Southwest 443Stephen Kulis, Flavio Francisco Marsiglia and Tanya Nieri

Living Arrangements, Social Integration, and Loneliness in Later Life: The Case of PhysicalDisability 460David Russell

Putting Work to Bed: Stressful Experiences on the Job and Sleep Quality 476Sarah A. Burgard and Jennifer A. Ailshire

Volume 50 Number 4 December 2009

The JOURNAL OF HEALTH AND SOCIAL BEHAVIOR (ISSN 0022-1465, formerly theJournal of Health and Human Behavior) is published quarterly in March, June, September, andDecember by the American Sociological Association, 1430 K Street NW, Suite 600,Washington, DC 20005. JHSB is printed by Edwards Brothers, Inc., Lillington, North Carolina.Periodicals postage paid at Washington, DC, and additional mailing offices. Postmaster: Sendaddress changes to the Journal of Health and Social Behavior, 1430 K Street NW, Suite 600,Washington, DC 20005.

Editorial Scope

The Journal of Health and Social Behavior publishes articles that apply sociological conceptsand methods to the understanding of health and illness and to the organization of medicine andhealth care. Its editorial policy favors those manuscripts that build and test knowledge in med-ical sociology, that show stimulating scholarship and clarity of expression, and that, takentogether, reflect the breadth of interests of its readership.

Manuscript Submission

Manuscript format: Manuscripts should meet the format guidelines specified in the Notice toContributors, which is published in the March and September issue of each volume. Electronicfiles (excluding artwork) must be in Microsoft Word or Excel format. All text must be printeddouble-spaced on 8-1/2-by-11-inch white paper. Use Times New Roman font, 12-point typesize. Margins must be at least 1 inch on all four sides. On the title page, note the manuscript’stotal word count (include all text, references, and notes; do not include word counts for tables orfigures). If you cite your own work, do not use wording that identifies you as the author.Submission requirements: Submit one (1) print copy of the original manuscript and one (1)print copy of a blinded manuscript from which all identifying information has been removed.Also submit one (1) electronic copy of the original manuscript and one (1) electronic copy ofthe blinded manuscript on CD or via e-mail. Enclose a $25 manuscript processing fee in theform of a check or money order payable to the American Sociological Association. The fee mustbe paid in order for the review process to begin. The fee is waived for student members of theASA. This reflects a policy of the ASA Council and Committee on Publications affecting allASA journals. Provide e-mail address, and JHSB will acknowledge the receipt of your manu-script. Manuscripts are not returned after review.Address for manuscript submission: Journal of Health and Social Behavior, IndianaUniversity, Karl F. Schuessler Institute for Social Research, 1022 East Third St., Bloomington,IN 47405-7103; phone (812) 856-6979; e-mail jhsb@indiana.edu.Editorial decisions: Median time between submission and decision is approximately 12 weeks.

Advertisements, Subscriptions, and Address Changes

Advertisements: Submit to Publications Department, American Sociological Association, 1430K Street NW, Suite 600, Washington, DC 20005; phone: (202) 383-9005 ext. 303; e-mail:publications@asanet.org.Subscriptions: ASA members, $35; ASA student members, $25; institutions, $155. Add $20 forpostage outside the United States and Canada. New subscriptions will be entered on a calendaryear basis only. To subscribe to JHSB or to request single issues, contact the ASA CustomerService Department; phone: (202) 383-9005 ext. 389; e-mail: subscriptions@asanet.org.Address Changes: Subscribers must notify the ASA Executive Office (e-mail:customer@asanet.org) six weeks in advance of an address change. Include both old and newaddresses. Claims for undelivered copies must be made within the month following the regularmonth of publication. When the reserve stock permits, the ASA will replace copies of JHSB thatare lost because of an address change.Copyright © 2009, American Sociological Association. Copying beyond fair use: Copies of arti-cles in this journal may be made for teaching and research purposes free of charge and withoutsecuring permission, as permitted by Sections 107 and 108 of the United States Copyright Law.For all other purposes, permission must be obtained from the publisher.The American Sociological Association acknowledges with appreciation the facilities and assis-tance provided by Indiana University.Cover design by Michael Mabry Design.

Putting Work to Bed: Stressful Experienceson the Job and Sleep Quality*

SARAH A. BURGARDJENNIFER A. AILSHIREUniversity of Michigan

Journal of Health and Social Behavior 2009, Vol 50 (December):476–492

Most adults spend one-third of every day sleeping and another third of mostdays at work. However, there is little analysis of the possible connections be-tween common workplace experiences and sleep quality. This study uses the lon-gitudinal and nationally-representative Americans’ Changing Lives study to ex-amine whether and how common conditions and experiences at work may “fol-low workers home” and impinge on their quality of sleep. We also explore howcompeting stressful experiences at home may influence sleep quality, andwhether these are more salient than work experiences. Results show that fre-quently being bothered or upset at work is associated with poorer sleep quali-ty, and the association is not explained by stressful experiences at home. Thesefindings are discussed in relation to the sociological literatures on work, stress,and emotion.

476

Most adults spend about one-third of most24-hour days in paid employment, and anotherthird sleeping, but our understanding of thelinks between experiences at work and sleepquality is limited. Biomedical studies havesuggested an association between workplaceconditions and sleep, but they have focused onparticular employee populations. In the socialscientific literature, there is substantial evi-dence that stressful working conditions are

linked to poorer health, while paid employmentinvolving positive aspects such as autonomyand creativity is associated with better healthand functioning (House 1987; Kohn andSchooler 1982; Kohn and Schooler 1983;Lennon 1994; Link, Lennon, and Dohrenwend1993; Mirowsky and Ross 2007). With only ahandful of exceptions (e.g., Arber et al. 2007;Hochschild and Machung [1989] 2003), how-ever, sociologists have all but ignored the con-tribution of experiences in the workplace tosleep quality. This is a major shortcoming be-cause poor sleep quality may act as a sensitivemarker of the consequences of stressful expe-riences in major macrosocial systems like theworkplace or at home. A better understandingof the work-sleep relationship in the generalpopulation is needed because sleep is a basichuman need and inadequate sleep has costs forindividuals, in terms of their health and safety,and for society, in the form of lost productivi-ty and medical care costs (Lamberg 2004).This study uses a nationally-representative,prospective sample of U.S. workers to examinewhether and how common conditions and ex-periences at work may “follow workers home”and impinge on their quality of sleep, and howthis may vary for those who are married or co-

* Versions of this article were presented at thePopulation Association of America and theAmerican Sociological Association meetings in2008. This study was supported by core fundingfrom Eunice Kennedy Shriver National Institute ofChild Health and Human Development grant R24HD041028 and National Institute on Aging grantP30 AG012846-14 to the Population Studies Center,University of Michigan. We would like to thankDavid Featherman, James House, YasaminKusunoki, Jason Schnittker, Pam Smock, theUniversity of Wisconsin Center for the Demographyof Health and Aging Seminar, and several anony-mous reviewers for helpful feedback. Address corre-spondence to Sarah Burgard, University ofMichigan, Department of Sociology, 500 SouthState Street, Ann Arbor, MI 48109-1382 (e-mail:burgards@umich.edu).

habiting or have children, compared to peoplewithout these family characteristics.

Extant research has generally neglected therange of stressors that are prevalent in the con-temporary workplace, focusing mainly on thesleep consequences of night shift and, in par-ticular, rotating shift work (for a review, seeAkerstedt 2003). Shift work makes it difficultto achieve a typical sleep schedule, disruptingsleep duration, timing, and the circadianrhythm. While understanding the conse-quences of shift work is important, a focussolely on this exposure limits our understand-ing of the range of mechanisms by which thework role may influence individuals’ lives out-side of work hours. Other, more common, oc-cupational stressors could exert their effects onsleep via very different means, and could affectsleep quality more than, or in addition to, itsduration. Perceived low control on the job(Karasek 1979), perceived job insecurity(Heaney, Israel, and House 1994), and negativeemotional experiences at work may create orindicate stress responses that raise hormonallevels and make it difficult for workers to “un-wind” at the end of the day. However, unlikerotating shift work, which is likely to presentsignificant objective obstacles to achieving ad-equate hours of sleep for most who perform it,common psychosocial stressors, such as lowcontrol or perceived job insecurity, may not af-fect all those who experience them, but onlyworkers who appraise them as threatening.

Another serious limitation of most existingstudies is that they rely on cross-sectional data,limiting researchers’ ability to understand howreverse causality, spurious association, or se-lection mechanisms may influence the rela-tionship. Sleepy workers may have a more neg-ative view of their working conditions than thewell-rested, for example, rather than, or in ad-dition to, troubles at work acting to reducesleep quality. Also, workers are not randomlyselected into jobs with negative working con-ditions, and the same characteristics that makethem more likely to face low control or othernegative experiences on the job could be theunderlying causes of poor sleep quality. For ex-ample, healthier people are more likely to beselected into employment and into particularkinds of jobs than their less healthy counter-parts (Pavalko, Gong, and Long 2007), andhealthier people may have better workplace ex-periences that could promote an existing ad-vantage in sleep quality. Moreover, when

studying self-reported occupational stressorssuch as perceived job insecurity, and also usingself-reported measures of sleep outcomes, as istypically done in survey-based studies, an un-derlying negative reporting style could lead toa spurious association that can best be ad-dressed if longitudinal data are available (Briefet al. 1988). Our study uses repeated measuresof working conditions and sleep to eliminatethe impact of stable individual characteristics,and we include baseline measures of respon-dents’ negative reporting style and health toprovide more robust estimates of the associa-tion.

This study thus has several strengths. First,we add to the very limited empirical analysis ofthe importance of common experiences atwork for sleep quality in the general popula-tion. We examine three stressful experiences atwork that have engaged sociologists and othersinterested in the ways social structure influ-ences individuals, and that are associated withother aspects of well-being. Importantly, weare able to address shortcomings of prior stud-ies of sleep quality by using nationally-repre-sentative, prospective data from a U.S. samplefollowed for about three years. This study ap-pears to be the first using U.S. data to do so, asexisting nationally-representative longitudinalstudies of sleep quality have been conductedon samples of European and Japanese workers,where working conditions and employmentcontexts may differ. Additionally, we explorepotentially competing stressors at home—in-cluding financial, spousal/partner, and child-related strains—to explore how importantthese experiences are, and how workers withdifferent family characteristics are influencedby their experiences in the workplace.Everybody sleeps, and most people will spendthe major part of their adult life working, soimproving understanding of the connection be-tween the two is important for an understand-ing of the way that social institutions and rolesstructure individual well-being.

THEORY AND EVIDENCE

Workplace Experiences and Sleep Quality

Inadequate sleep has serious consequencesranging from increased risk for traffic acci-dents (National Highway Traffic SafetyAdministration 2006), health problems (Mooreet al. 2002), chronic disease (Tasali et al.2008), and mortality (Ferrie et al. 2007).Moreover, while estimates vary considerably

PUTTING WORK TO BED 477

across studies and depend on the definition ofsleeping problems, they appear to be relativelycommon. A recent report suggests that 50 to 70million Americans suffer from a disorder ofsleep and wakefulness (Colten and Altevogt2006). The majority of research on the predic-tors of sleep quality has been biomedical orpsychological in nature and has focused onproximate risk factors, such as health condi-tions (Kutner, Bliwise, and Zhang 2004), per-sonality dispositions (Espie 2002), or other in-dividual or behavioral causes.

Psychological stress and reactivity to stressalso have been implicated in the developmentof insomnia, one of the major diagnosed con-ditions that indicates poor sleep quality (Espie2002; Morin, Rodrigue, and Ivers 2003). Thestress response increases neurological arousalthat involves the release of key neurotransmit-ters (such as adrenaline and noradrenaline) andneuron-effective hormones (such as cortisol).The presence of cortisol, in particular, can in-terfere with a worker’s ability to “switch off ”at the end of the work period and could alsolead to depressed mood or enduring agitationor anxiety about the day’s events, all of whichcould prevent adequate sleep (Linton 2004).While not intrinsically harmful, the stress re-sponses that lead to a poor night’s sleep couldbecome maladaptive if they occur chronically(House 2002; Pearlin et al. 1981). Thus, peoplewho are more likely to encounter psychologi-cally stressful experiences and those who aremore likely to appraise given conditions asthreatening may be at greater risk of poor sleepquality. This suggests that, beyond individual-level risk factors, social structure is also im-portant for sleep quality. Specifically, we arguethat social stratification across jobs and withinworkplaces leads to variation in the negativeexperiences individuals encounter at work, anddetermines exposure to the chronic psycholog-ical stressors that could lead to poor sleep qual-ity.

Work-related stress is frequently cited byworkers themselves as a cause of sleeping dif-ficulties (Henry et al. 2008; Linton 2004), butsince researchers have examined differentworking conditions and generally have notconsidered a variety of potentially stressful ex-periences in the same models, there is limitedunderstanding of which common working con-ditions have robust associations with sleepquality (but see, as exceptions, Knudson,Ducharme, and Roman 2007; Ribet and

Derriennic 1999; Sekine et al. 2006). We focuson three common workplace experiences—perceived low control, perceived job insecuri-ty, and feeling bothered or upset on the job—that are likely to be perceived as stressful by asubstantial fraction of individuals who experi-ence them.

Low control over tasks and decisions on thejob has received considerable attention fromsocial scientists, psychologists, and epidemiol-ogists. Longitudinal studies have shown thatoccupational self-direction enhances self-di-rected personality orientations, increasing theoverall sense of control (Kohn and Schooler1982) and lowering the risk for depression,psychological distress, and anxiety (Kohn andSchooler 1982; Kohn and Schooler 1983; Linket al. 1993). By contrast, low control preventsan individual from resolving problems on thejob or exercising autonomy or creativity, andthe stress and frustration of these experiencescould be carried home after work. A few stud-ies have shown that low control at work islinked with poor sleep quality, though priorstudies have examined workers outside theUnited States (Kalimo et al. 2000) and/or usedcross-sectional data (Knudson et al. 2007;Sekine et al. 2006), so further assessment ofthe association is needed.

Perceived job insecurity can involve antici-pating problems associated with a job loss, ex-periencing the mental strain of being in a pow-erless position, and feeling ambiguity aboutwhat the future might hold and what actionswould be most appropriate to reduce the strain(Heaney et al. 1994; Joelson and Wahlquist1987). We have found no studies that directlyexamined the association between perceivedjob insecurity and poor sleep quality as wemeasure it here, though prior studies havefound links between impending job loss andshort or long sleep duration among Britishmale civil servants (Ferrie et al. 1998a), andhave noted sleep disturbance among Swedishmale shipyard workers in the midst of major in-dustrial reorganization (Mattiasson et al.1990). Another study found that workers whoactually lost jobs during a major economic re-cession in Finland experienced increased in-somnia (Hyyppä, Kronholm, and Alanen1997). Perceived job insecurity also has beenlinked to depressive symptoms and physicalhealth indicators that reflect the impact ofstress (Burgard, Brand, and House 2009; Ferrieet al. 1995; Ferrie et al. 1998b), so a link be-

478 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

tween perceived job security and sleep qualityis plausible.

We have found no prior studies that directlyexamine the association between being both-ered or upset at work and sleep quality, thoughthere are several that deal with it indirectly. Across-sectional study of Australian nurses sug-gested that psychologically stressful experi-ences reflecting negative emotional load andpoor relations with coworkers, as well as otherpsychological demands, were much morestrongly related to poor sleep quality than thephysical demands of nursing (Winwood andLushington 2006). A longitudinal study of 47U.S. men and women also found that daytimeinterpersonal conflict was associated with poorsleep quality that night (Brissette and Cohen2002). Interpersonal conflict or negative emo-tional load could contribute to feeling botheredor upset at work, but without prior empiricalevidence for the measure we use in this study,we rely on related theoretical and empiricalfindings about emotion in the workplace.Sociologists have examined how workers ex-press emotions (Lively and Powell 2006) andface challenges in regulating their emotions inthe workplace (Hochschild 1983), and howstressful emotional experiences at work mayspill over into life at home, influencing familyinteractions (Menaghan 1991). These studieslead us to argue that being bothered or upset atwork indicates a negative emotional experiencelinked to psychological stress that could influ-ence sleep quality.

Importantly, unlike low control and per-ceived insecurity, feeling bothered or upset isan explicit measure of emotional reaction toconditions at work, rather than a measure ofsimply being exposed to specific working con-ditions. As such, it is a more direct measure ofstress and arousal, because all workers who re-port it have necessarily appraised their condi-tions as threatening or disturbing. This meansthat being bothered or upset at work may havea stronger or more consistent relationship withsleep quality than reports of low control or per-ceived job insecurity, which may or may not beviewed as threatening by a given individual. Onthe other hand, underlying stable personalitycharacteristics may more completely explainemotional reaction to working conditions, soour controls for those characteristics may ex-plain any link with sleep quality, leaving no re-maining association in longitudinal models.Based on prior theoretical and empirical re-

search on stressful experiences in the work-place, our first research question asks:

Question 1: Are perceived low control, per-ceived job insecurity, and feeling bothered orupset at work prospectively associated withpoor sleep quality?

Competing Stressful Experiences at Home

For many adults, the remaining third of theday when they are not at work or asleep is filledby family and home experiences and responsi-bilities. Home life could provide competingstressful experiences that may be as important,or more important, than working conditions forsleep quality. After all, individuals who experi-ence stress in interactions with a spouse orpartner often share a bed with that person,which could make such experiences particular-ly salient for sleep quality. Additionally, timespent dealing with bills and financial issues ordisciplining children may occur closer to bed-time than problems arising at work. However,there is very little literature that examines howthese various realms of stressors intersect to in-fluence sleep.

Individuals with a spouse or children may bemore heavily influenced by home and familyexperiences than by experiences at work, com-pared to individuals who do not hold the com-peting roles of parent or spouse. For example,the stresses associated with having children inthe home, particularly if they do not sleep con-sistently or are out late, could affect parents’sleep (Meltzer and Mindell 2007). Moreover,negative experiences in the workplace couldinfluence the way individuals interact withtheir partners or children (Menaghan 1991),and this spillover could create interpersonalproblems at home that more proximally influ-ence sleep. In this study we examine how com-peting stressful experiences in the home spheremay overshadow or explain the association be-tween negative experiences at work and sleepquality for working-aged individuals who havea spouse/partner or who live with their chil-dren, compared to the sample of working indi-viduals overall.

Focusing on finances, a spouse or partner,and children as key sources of potential sleep-disrupting stress, we explore self-reports of rel-atively objective conditions as well as directmeasures of emotional response to negative ex-periences, to parallel our measures of workingconditions and experiences. For example, weassess the association between poor sleep qual-

PUTTING WORK TO BED 479

ity and reported difficulty paying bills (a moreobjective measure) as well as dissatisfactionwith finances (better reflecting the appraisedthreat of one’s financial situation). Among in-dividuals living with a spouse or partner, weexamine the association between poor sleepquality and the degree of negative hassles fromthe spouse/partner, as well as feeling botheredor upset by one’s marriage or relationship.Among those with children in the home, we ex-amine the importance of feeling bothered orupset as a parent. We examine three explorato-ry research questions to assess negative expe-riences at home as competing risks for poorsleep quality, compared to those at work.

Question 2: Are financial, spousal/partner,and/or child-related negative experiencesprospectively associated with poor sleep quali-ty?

Question 3: Are financial, spousal, and/orchild-related negative experiences morestrongly associated with poor sleep qualitythan workplace experiences?

Question 4: Do financial, spousal, and/orchild-related negative experiences explain theassociation between workplace experiencesand poor sleep quality?

DATA AND METHODS

Data

We use the Americans’ Changing Lives(ACL) study and focus on respondents work-ing at least 20 hours per week at baseline. Welimit the sample in this way because only thosewith at least 20 hours of work per week havedata on low workplace control. This means thatwe study respondents whose exposure to nega-tive working conditions is likely to be substan-tial, but when we re-estimated simpler modelsusing all employed respondents (not shown),our conclusions did not change. The ACL is astratified, multi-stage area probability sampleof 3,617 non-institutionalized adults 25 yearsand older living in the United States in 1986,with oversampling of adults 60 and older andof African Americans. Follow-up interviewswere conducted in 1989, 1994, and 2001/2002,but our analysis uses data from 1986 and 1989because some necessary questions were omit-ted in later waves. Sample weights designed toadjust for oversampling of special populationsand sample nonresponse or noncoverage atbaseline, as well as loss to follow-up due to at-trition or death, are used in all appropriate de-scriptive statistics and multivariate models.

Excluding ACL respondents who did not workat least 20 hours per week in 1986 (N = 1,930),the vast majority of whom were already retiredor not working for pay; those who were not pre-sent for the 1989 interview (N = 297); and cas-es missing on covariates (N = 60), 1,330 indi-viduals are eligible for inclusion in analysesusing information on working conditions in1986. In most of the multivariate analyses, wefocus on those who were working for pay for atleast 20 hours per week in both 1986 and 1989(N = 1,101). In some analyses we use subsam-ples of respondents who were married andworking in 1986 (N = 869) or who were bothmarried and working in both 1986 and 1989 (N= 670); and subsamples of those who wereworking and had children 18 years old oryounger living in the home in 1986 (N = 596)and those who were working and had childrenin the home both in 1986 and 1989 (N = 435).

Measures

Sleep quality. Poor sleep quality is typicallymeasured in surveys with indicators of delayed,disrupted, or nonrestorative sleep. We measurepoor sleep quality with a global item obtainedfrom the Center for Epidemiologic StudiesDepression Scale, or CES-D (Radloff 1977):“During the past week my sleep was restless:most of the time, some of the time, or hardlyever.” We dichotomize the responses so that 0= hardly ever and 1 = some or most of the time,focusing on respondents who reported troubledsleep for at least some meaningful fraction ofthe last week. Additionally, we collapsed“some” and “most” of the time because only asmall percentage of respondents reported theresponse “most of the time” (about 10% in1986 and 7% in 1989). This item was di-chotomized similarly in a prior study of socialfactors and sleep quality (Kutner et al. 2004);results using an alternative coding are dis-cussed below.1

Working conditions. Perceived low control isderived from three items based on Karasek’s(1979) measure of decision latitude, including:“I get to do a variety of different things in mywork,” “I have a lot to say about what happensin my work,” and “I have very little chance todecide how I do my work” (reverse-coded).Response categories are strongly agree = 1,agree somewhat = 2, disagree somewhat = 3,and strongly disagree = 4; we create a measureof low control by summing all items (ranging3–12); these items have an alpha of .61. To

480 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

measure perceived job insecurity, respondentswere asked the following question: “How like-ly is it that during the next couple of years youwill involuntarily lose your main job—not atall likely = 1, not too likely = 2, somewhat like-ly = 3, or very likely = 4?” To capture negativeemotional experiences at work we use a singleitem: “In general, how often do you feel both-ered or upset in your work—almost always = 4,often = 3, sometimes = 2, rarely = 1, or never= 0?” For each working condition measure, wealso created an indicator of change over fol-low-up by subtracting the value for 1986 fromthe value for 1989. A positive value on thechange score means that the negative exposureworsened over time, while a negative value in-dicates that it lessened.

Home conditions. Measures of financialstrain include reported difficulty paying bills(not difficult = 1, slightly difficult = 2, some-what difficult = 3, very difficult = 4, extreme-ly difficult = 5) and an indicator of dissatisfac-tion with the respondent’s present financial sit-uation (completely satisfied = 1, very satisfied= 2, somewhat satisfied = 3, not very satisfied= 4, not at all satisfied = 5). A negative hasslesindex referring to the respondent’s spouse orlive-in partner uses two items: (1) “How muchdo you feel (he/she) makes too many demandson you?” and (2) “How much is (he/she) criti-cal of you or what you do?” Response cate-gories for each were “a great deal, quite a bit,some, a little, or not at all,” and reverse-codedvalues for the two items were averaged and theindex standardized with a mean of 0 and a stan-dard deviation of 1. The spouse/partner nega-tive hassles index ranges from –1.3 to 2.9. Tocapture negative emotional experiences withfamily members, respondents were asked howoften they felt bothered or upset (1) by theirmarriage/relationship or (2) as a parent, withresponse categories coded so that almost al-ways = 4, often = 3, sometimes = 2, rarely = 1,and never = 0. Change scores were created foreach of these measures of negative experiencesat home by subtracting the 1986 value from the1989 value, with positive values on the changemeasure indicating conditions that worsenedover time.

Other predictors. To explore whether the as-sociation between working conditions andsleep quality is spurious, we adjust for neuroti-cism and health at baseline, as well as adjust-ing for prior poor sleep quality. Neuroticism isa relatively stable underlying personality trait

that may mark a negative reporting style, andwe use a neuroticism index based on four ques-tions from the Eysenck Personality Inventory(Eysenck and Eysenck 1975), such as “Are youa worrier?” The standardized scale ranges from–1.2 to 2.2 (most neurotic). Self-rated health, ageneral indicator used to distinguish respon-dents who may have health conditions that in-fluence their ability to sleep, is measured witha single item: “How would you rate your healthat the present time— poor = 1, fair = 2, good =3, very good = 4, and excellent = 5?” We alsocontrol for obesity, a risk factor for sleep ap-nea, which could negatively impact sleep qual-ity. Using self-reported weight and height, obe-sity is coded so that 0 = body mass index lessthan 30, while 1 = body mass index 30 orabove.

In multivariate analyses we also adjust forbaseline sociodemographic characteristics thatare predictive of sleep quality, working condi-tions, or both. Age is measured in years, and asquared term for age is included in multivari-ate models to adjust for nonlinearities in the as-sociation between age and sleep. Respondent’srace is coded into a series of dichotomous in-dicator variables denoting whites, AfricanAmericans, and “other” race-ethnicity.2 Sex iscoded so that 0 = female and 1 = male; maritalstatus is coded so that 0 = married or livingwith a partner and 1 = unmarried/not livingwith a partner; and parental status is coded sothat respondents without children under 18 liv-ing with them = 0 and those with children = 1.Educational attainment at baseline is coded as0 = some college or more and 1 = high schoolgraduate or less. We also include a measure ofhousehold income, reported in Table 1 in 2008dollars, but transformed for multivariate analy-sis by adding $500 before taking the log so thatindividuals with a score of zero on the measureare retained. Work hours at the main job aremeasured as average hours per week.Employment status in 1989 (0 = not employed,1 = employed) is included in Table 1 to indicatethe loss of respondents from the paid laborforce between 1986 and 1989.

Analytic Strategy

We first examine simple associations be-tween negative experiences at work and homeand poor sleep quality. We then estimate logis-tic regression models to explore the associationbetween negative working conditions and sleepquality in 1986. Next, in longitudinal models,

PUTTING WORK TO BED 481

482 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 1. Descriptive Statistics for Dependent and Independent Variables by Analytic Sample, ACLRespondents

Working in 1986 Working in 1986 and 1989

Mean/% S.D. Mean/% S.D.

Poor sleep quality 1986 49.5% 49.0%

Poor sleep quality 1989 48.1% 48.0%

Perceived Low Control 1986 5.10 (2.00) 5.06 (2.01)—Change 1986–1989 .— –.08 (1.95)

Perceived Job Insecurity 1986 1.73 (.859) 1.73 (.852)—Change 1986–1989 .— –.03 (.967)

Bothered/Upset at Work 1986 1.62 (.829) 1.62 (.812)—Change 1986–1989 .— .06 (.950)

Difficult to Pay Bills 1986 1.93 (1.012) 1.93 (.990)—Change 1986–1989 .— –.12 (.973)

Dissatisfaction with Finances 1986 2.79 (.993) 2.79 (.962)—Change 1986–1989 .— –.05 (.963)

Spouse/Partner Negative Hassles 1986a .05 (.954) .04 (.949)—Change 1986–1989 .— .10 (.869)

Bothered/Upset by Marriage/Relationship 1986a 1.03 (.819) 1.02 (.797)—Change 1986–1989 .— .02 (.792)

Bothered/Upset as Parent 1986b 1.80 (.833) 1.79 (.834)—Change 1986–1989 .— –.06 (.833)

Neuroticism Score 1986 –.093 (.947) –.116 (.933)

Self-rated Health 1986 4.01 (.871) 4.04 (.846)

% Obese (BMI 30 or higher) 1986 13.5% 13.6%

Age (years) 1986 40.5 (11.3) 39.6 (10.4)

% Male 56.4% 58.4%Race—% White 85.1% 84.7%—% African American 9.7% 9.6%—% Other 5.2% 5.6%

% Unmarried/No Partner 1986 22.9% 21.8%

% Has Children under 18 years in 1986 52.6% 53.7%

% High School or less Education 1986 45.8% 43.9%

Household Income 1986 in 2007 dollars 74,845 (47,350) 76,390 (47,109)

Work Hours per Week 1986 44.0 (11.7) 44.7 (11.7)

% Employed in 1989 91.9% 100.0%N 1,330 1,101

Note: Figures are weighted using 1986 sampling weight, column totals unweighted.a Reports about spouses only collected from those who are married; for those working in 1986 and married in 1986, N= 869; for those working in 1986 and 1989 and married in both years, N = 670.b Reports about children only collected from those who have children; for those working in 1986 and with children un-der 18 years in 1986, N = 596; for those working in 1986 and 1989 and with children under 18 in both years, N = 435.

we consider the association between workingconditions in 1986 and change in working con-ditions between 1986 and 1989 with changes insleep quality between 1986 and 1989. Using aparallel set of models, we explore the compet-ing risks of negative experiences at home forsleep quality in 1989. We use different sub-samples to target respondents who were at riskof particular exposures; the importance of fi-nancial strain is examined for all respondents,while models that examine spousal/partnerstrain include only respondents who were mar-ried or lived with a partner at both waves, andthose that examine child-related strain are re-stricted to respondents living with their chil-dren under age 18 at both waves. Attrition ofrespondents is always a concern when usinglongitudinal samples. All longitudinal modelsuse wave two survey weights, which adjust forsurvey attrition, while cross-sectional figuresuse baseline sampling weights. All analyses areconducted using Stata 10SE software.

RESULTS

Descriptive Results

Table 1 presents means and standard devia-tions or percentages for all variables used in theanalysis separately for the two main analyticsamples described above. The first column pre-sents characteristics for all respondents work-ing at baseline, and the second column presentscharacteristics for the sample working in 1986and 1989. Comparison across columns showsthat sample means are very similar on mostcharacteristics. Across samples, respondentswere about 40 years old at baseline, on average,with a higher fraction of males than females inthis sample of individuals working more than20 hours per week (56–58%). Most are white,almost four out of five were married at base-line, and close to half had a high school educa-tion or less.

As shown in Table 1, about 49 percent of therespondents reported poor sleep quality atbaseline in 1986, and about 48 percent did soin 1989. By comparison, a study of U.S. work-ers using data from the 2002–2003 NationalEmployee Survey showed that about 58 percentreported at least some trouble falling asleep inthe past month and about 56 percent reportedat least some trouble staying asleep (Knudsonet al. 2007), suggesting that our figures are rea-sonable. Turning to negative working condi-tions at baseline, respondents average a lowcontrol score of 5.1, close to the bottom of the

possible range. The average response on per-ceived job insecurity is about 1.7, closer to “nottoo likely” than to “not at all likely,” and the av-erage score for being bothered or upset at workis 1.6, which is closer to “sometimes” than“rarely” on this measure. The average amountof change was very close to zero on all threeworking conditions, but tabulations not shownindicate that only about one-third of respon-dents had the same low control score in 1986and 1989, while about half reported no changein job insecurity or being bothered or upset atwork. About one-quarter to one-third of re-spondents showed improvement over this peri-od, while the remainder reported worse work-ing conditions.

Table 2 presents the percentages reportingpoor sleep quality across categories of stressfulexperiences at work or at home. The first col-umn shows the percentage of respondents re-porting poor sleep quality in 1986 for each cat-egory of the exposure variables in 1986 (lowcontrol at work and spouse negative hassles arepresented categorically here for ease of inter-pretation, but used as linear terms in multivari-ate models), while the second column showsthe percentages reporting poor sleep quality in1989. The p-values for chi-square tests of dif-ference are presented for the low control andspousal hassles comparison groups, and fornonparametric tests of trend for comparisonsacross categories of the other measures.Results in Table 2 suggest that before adjustingfor any individual characteristics, respondentsreporting negative experiences at work or athome were significantly more likely to reportpoor sleep at all survey waves. The only ex-ceptions were for the comparison of sleep qual-ity in 1989 across categories of low control andspousal/partner hassles.

Multivariate Results

Table 3 presents odds ratios and 95 percentconfidence intervals from logistic regressionmodels predicting poor sleep quality, withsample sizes and tests of model fit presented atthe bottom of the table. Model 1 examines theassociation between low control and perceivedjob insecurity and poor sleep quality in 1986,adjusting only for age, age-squared, and sex.Model 2 adds the measure of feeling both-ered/upset at work in 1986 to test whether thisindicator of emotional response mediates theimpact of the other working conditions, andmodel 3 adds controls for all other baseline

PUTTING WORK TO BED 483

484 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

TABLE 2. Percentage of Respondents Reporting Poor Sleep Quality in 1986 or 1989 by Categories ofStressful Work or Home Conditions, ACL Respondents Working in 1986

% Poor Sleep Quality % Poor Sleep Quality1986 1989

Low Control 1986—Control at or above median 45.9% 48.1%—Control below median 55.0% 48.2%

——p-value for difference .001 .254

Perceived Job Insecurity 1986—Job loss not at all likely 45.3% 44.4%—Not too likely 48.9% 48.9%—Somewhat likely 61.2% 60.8%—Very likely 65.0% 44.3%

——p-value for trend < .001 .002

Bothered/Upset at Work 1986—Never 37.1% 39.8%—Rarely 40.2% 41.7%—Sometimes 55.6% 51.4%—Often 62.8% 63.4%—Almost always 77.0% 61.8%

——p-value for trend < .001 < .001

How Difficult to Pay Bills 1986—Not difficult 47.1% 44.0%—Slightly difficult 48.9% 47.9%—Somewhat difficult 50.5% 53.2%—Very difficult 62.1% 53.0%—Extremely difficult 68.2% 73.5%

——p-value for trend < .001 < .001

Dissatisfaction with Finances 1986—Completely satisfied 37.4% 34.8%—Very satisfied 44.9% 42.8%—Somewhat satisfied 50.6% 50.5%—Not very satisfied 56.9% 49.2%—Not at all satisfied 65.3% 71.7%

——p-value for trend < .001 < .001

Spouse/Partner Negative Hassles 1986a

—Hassles below median 46.5% 45.2%—Hassles at or above median 51.4% 51.3%

——p-value for difference .073 .212

Bothered/Upset by Marriage/Relationship 1986a

Never 44.6% 40.3%—Rarely 46.3% 48.9%—Sometimes 57.7% 52.8%—Often 65.3% 70.2%—Almost always 60.8% 68.0%

——p-value for trend .001 < .001

Bothered/Upset as Parent 1986b

—Never 37.1% 33.3%—Rarely 41.2% 42.9%—Sometimes 51.2% 44.7%—Often 59.3% 52.0%—Almost always 92.1% 80.8%

——p-value for trend < .001 .006

Note: p-values obtained from chi-square or nonparametric tests for trend across ordered categories.a Only respondents who were married/living with a partner reported on that person.b Only respondents with children under 18 years reported on them.

PUTTING WORK TO BED 485

TA

BL

E 3

.O

dds

Rat

ios

and

95%

Con

fide

nce

Inte

rval

s fr

om L

ogis

tic

Reg

ress

ion

Mod

els

of P

oor

Slee

p Q

ualit

y in

198

6 or

198

9,A

CL

res

pon

dent

s

Mod

el 1

: Po

or S

leep

Mod

el 2

: Po

or S

leep

Mod

el 3

: Po

or S

leep

Mod

el 4

: Po

or S

leep

Mod

el 5

: Po

or S

leep

Qua

lity

198

6Q

uali

ty 1

986

Qua

lity

198

6Q

uali

ty 1

989

Qua

lity

198

9

O.R

.(9

5% C

.I.)

O.R

.(9

5% C

.I.)

O.R

.(9

5% C

.I.)

O.R

.(9

5% C

.I.)

O.R

.(9

5% C

.I.)

Low

Con

trol

1.07

*(1

.008

–1.1

27)

1.05

(.99

0–1.

109)

1.03

(.96

8–1.

094)

1.01

(.93

5–1.

094)

.99

(.91

7–1.

077)

—C

hang

e 19

86–8

9.—

.—.—

1.05

(.96

9–1.

133)

1.03

(.95

2–1.

116)

Job

Inse

curi

ty1.

29**

*(1

.135

–1.4

72)

1.27

***

(1.1

10–1

.445

)1.

21**

(1.0

51–1

.385

)1.

11(.

911–

1.34

1)1.

09(.

895–

1.32

2)—

Cha

nge

1986

–89

.—.—

.—1.

04(.

879–

1.23

2)1.

04(.

873–

1.22

8)B

othe

red/

Ups

et a

t Wor

k.—

1.56

***

(1.3

59–1

.796

)1.

40**

*(1

.201

–1.6

23)

.—1.

35**

(1.0

89–1

.676

)—

Cha

nge

1986

–89

.—.—

.—.—

1.27

**(1

.067

–1.5

02)

Age

(ye

ars)

.98

(.92

0–1.

054)

.98

(.91

5–1.

052)

.98

(.90

6–1.

056)

.89*

(.81

0–.9

79)

.88*

(.80

5–.9

72)

Age

2(y

ears

)1.

00(.

999–

1.00

1)1.

00(.

999–

1.00

1)1.

00(.

999–

1.00

1)1.

00*

(1.0

00–1

.002

)1.

00*

(1.0

00–1

.002

)M

ale

.95

(.76

3–1.

183)

.92

(.74

0–1.

156)

.98

(.77

3–1.

255)

1.03

(.78

3–1.

354)

1.01

(.76

8–1.

332)

Afr

ican

Am

eric

an.—

.—.9

8(.

662–

1.45

3)1.

30(.

843–

2.01

9)1.

33(.

855–

2.05

8)O

ther

Rac

e.—

.—1.

88*

(1.1

04–3

.211

)1.

36(.

805–

2.29

3)1.

38(.

813–

2.33

8)U

nmar

ried

/No

Part

ner

.—.—

.98

(.71

3–1.

344)

.80

(.56

0–1.

152)

.81

(.56

2–1.

158)

Has

chi

ldre

n <

18

in h

ome

.—.—

.96

(.73

3–1.

251)

.59*

*(.

439–

.800

).6

2**

(.45

6–.8

35)

Hig

h S

choo

l or

les

s.—

.—1.

32*

(1.0

29–1

.698

).7

9(.

592–

1.04

2).8

3(.

621–

1.10

0)H

ouse

hold

Inc

ome

.—.—

.97

(.79

6–1.

190)

1.08

(.86

0–1.

360)

1.06

(.84

5–1.

339)

Hou

rs/W

eek

.—.—

1.00

(.99

3–1.

014)

.99

(.98

1–1.

005)

.99

(.98

0–1.

004)

Neu

roti

cism

Sco

re.—

.—1.

45**

*(1

.269

–1.6

60)

1.40

***

(1.2

10–1

.624

)1.

33**

*(1

.138

–1.5

53)

Sel

f-ra

ted

Hea

lth

.—.—

.74*

**(.

645–

.856

).8

5*(.

723–

.992

).8

5*(.

721–

.991

)O

bese

.—.—

1.44

*(1

.027

–2.0

32)

.88

(.60

3–1.

293)

.90

(.61

7–1.

323)

Poor

Sle

ep 1

986

.—.—

.—3.

03**

*(2

.332

–3.9

49)

2.89

***

(2.2

12–3

.774

)

N1,

330

1,33

01,

330

1,10

11,

101

LR

Chi

225

.3**

*66

.6**

*15

1.9*

**15

3.7*

**16

3.1*

**

Not

e:**

* p

< .0

01;

** p

< .0

1; *

p<

.05.

predictors. Taken together, the cross-sectionalmodels 1 through 3 suggest that perceived jobinsecurity and feeling bothered or upset atwork are most strongly linked to poor sleepquality at baseline, and that feeling both-ered/upset may mediate a limited amount ofthe impact of low control and job insecurity.

Turning to longitudinal models that examinechange in sleep quality, model 4 examines theimpact of 1986 low control and job insecurity,and change in these working conditions be-tween 1986 and 1989 on poor sleep quality in1989, controlling for all predictors used inmodel 3 and adding a measure of poor sleepquality in 1986. Model 5 adds measures offeeling bothered/upset at work in 1986 andchange between 1986 and 1989 in feeling both-ered/upset. The results for models 4 and 5 showthat only being bothered or upset at work in1986 (OR = 1.35) and increases in being both-ered or upset by 1989 (OR = 1.27) are inde-pendently associated with subsequent poorsleep quality, net of sleep quality in 1986.

Table 4 presents models that assess the asso-ciations between negative experiences at workand poor quality sleep when competing stress-ful experiences at home are added. The toppanel presents results with controls for finan-cial stress, the middle panel considersspousal/partner-related stress, and the bottompanel considers the stress of parenting. Model1 for each panel considers cross-sectional rela-tionships and includes all predictors from mod-el 3 in Table 3, though we present only the fo-cal odds ratios. Models 2 and 3 for each panelconsider models of change in sleep quality.Model 2 adds baseline and change values ofmore “objective” indicators of working condi-tions (low control and job insecurity), financialstrain (difficulty paying bills), or spousal/part-ner stress (negative hassles), while model 3adds indicators of emotional responses to work(bothered/upset at work), finances (dissatisfac-tion with finances), spouse/partner (both-ered/upset with spouse/partner), and children(bothered/upset as a parent).

Results presented in Table 4 show that whiledissatisfaction with finances in 1986 (OR =1.40) and an increase in dissatisfaction be-tween 1986 and 1989 (OR = 1.29) are inde-pendently and significantly associated withpoor sleep quality in 1989 (model 3, panel 1),these and the other additional predictors do notsubstantially alter the main findings from Table3. Being bothered/upset at work is still

prospectively associated with change towardpoorer sleep quality in all models except mod-el 3 in panel 3. In that model for ACL respon-dents living with children under 18 years inboth 1986 and 1989, perceived job insecurityin 1986 is significantly associated withchanges in sleep quality (OR = 1.49), while be-ing bothered or upset at work in 1986 is nolonger a significant predictor, though an in-crease in feeling bothered or upset at work re-mains a significant predictor (OR = 1.38).Being bothered or upset as a parent is not as-sociated with sleep quality in panel 3, so thedifferences in the results observed amongthose living with their children compared to thesample overall are probably partially due to thereduction in sample size and to factors notmeasured here.

Sensitivity Analyses

Given the exploratory nature of our study,we conducted additional analyses to verify therobustness of our results; all are available onrequest. First, our classification of poor sleepquality may be too generous because we in-clude respondents who report “sometimes” ex-periencing troubled sleep. In models not shownhere we re-estimated the models in Table 3 us-ing an indicator that distinguished those re-spondents reporting troubled sleep “most ofthe time” (coded 1; about 10% in 1986, about7% in 1989) from all others (coded 0). The as-sociations between feeling bothered or upset atwork and poor sleep quality in 1986 and 1989were very similar to those reported in Table 3,though change in feeling bothered/upset atwork between 1986 and 1989 was no longersignificantly associated with worsening sleepquality over follow-up. With this more conser-vative coding of poor sleep quality, low con-trol, and perceived job insecurity were neversignificant predictors. We also explored theways that selection on measured or unmea-sured characteristics may influence our results.To further clarify the temporal ordering ofevents and address concerns about selection,we predicted change in sleep quality between1989 and 1994 as a function of 1986 workingcharacteristics and changes in these workingcharacteristics between 1986 and 1989. Resultswere substantively identical to those presentedin Table 3.

Because work and home roles are stronglygendered in the United States, and thus couldaffect sleep differently for men and women, we

486 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

also re-estimated all the models presented inTables 3 and 4 with interactions by respon-dent’s sex. While the association between feel-ing bothered/upset at work and poor sleepquality is stronger for men in cross-sectionalmodels, there are not significant differences by

sex in the association between working condi-tions and sleep quality in longitudinal models.Moreover, there is little indication that the as-sociations between spousal/partner or parentalstressors and sleep quality varies substantiallyby sex, though our sample sizes are not optimal

PUTTING WORK TO BED 487

TABLE 4. Odds Ratios and 95% Confidence Intervals from Logistic Regression Models of PoorSleep Quality in 1986 or 1989 with Adjustment for Other Stressors, ACL respondentsWorking in 1986 (Model 1) or in 1986 and 1989 (Models 2 and 3)

Financial Stress

M 1: Poor Sleep 1986 M 2: Poor Sleep 1989 M 3: Poor Sleep 1989

O.R. (95% C.I.) O.R. (95% C.I.) O.R. (95% C.I.)

Low Control 1.02 (.959–1.085) 1.00 (.928–1.088) .98 (.907–1.067)—Change 1986–89 .— 1.04 (.965–1.130) 1.03 (.948–1.116)Job Insecurity 1.19* (1.037–1.368) 1.09 (.893–1.319) 1.05 (.858–1.276)—Change 1986–89 .— 1.03 (.869–1.221) 1.01 (.850–1.199)Bothered/Upset at Work 1.39*** (1.194–1.614) .— 1.30* (1.044–1.614)—Change 1986–89 .— .— 1.22* (1.029–1.457)Difficult to Pay Bills .90 (.776–1.035) 1.17 (.992–1.387) 1.00 (.822–1.224)—Change 1986–89 .— 1.15 (.982–1.344) 1.01 (.851–1.210)Dissatisfaction with Finances 1.24** (1.069–1.431) .— 1.40** (1.126–1.750)—Change 1986–89 .— .— | 1.29** (1.067–1.556)

N 1,330 1,101 1,101LR Chi2 160.1*** 158.0*** 177.6***

Spousal/Partner Stress

M 1: Poor Sleep 1986 M 2: Poor Sleep 1989 M 3: Poor Sleep 1989

O.R. (95% C.I.) O.R.| (95% C.I.) O.R. (95% C.I.)

Low Control 1.02 (.939–1.098) 1.03 (.930–1.146) 1.00 (.901–1.116)—Change 1986–89 .— 1.04 (.942–1.154) 1.02 (.920–1.135)Job Insecurity 1.21* (1.012–1.442) 1.07 (.831–1.381) 1.07 (.822–1.383)—Change 1986–89 .— .95 (.765–1.176) .96 (.773–1.197)Bothered/Upset at Work 1.43** (1.174–1.732) .— 1.42* (1.063–1.896)—Change 1986–89 .— .— 1.38** (1.093–1.742)Spouse/Partner Negative Hassles .92 (.786–1.087) 1.16 (.947–1.433) 1.10 (.877–1.383)—Change 1986–89 .— 1.21 (.968–1.503) 1.12 (.891–1.419)Bothered/Upset by Marriage/ 1.17 (.964–1.427) .— 1.14 (.859–1.503)—Relationship—Change 1986–89 .—| .—| .—| | 1.28| (.986–1.657)

N 869 670 670LR Chi2 117.2*** 98.7*** 110.1***

Parenting Stress

M 1: Poor Sleep 1986 M 2: Poor Sleep 1989 M 3: Poor Sleep 1989

O.R.| (95% C.I.) O.R.|(95% C.I.) O.R.| (95% C.I.)

Low Control 1.04 (.939–1.143) .— 1.12 (.971–1.281)—Change 1986–89 .— .— 1.12 (.985–1.280)Job Insecurity 1.23 (.980–1.533) .— 1.49* (1.055–2.112)—Change 1986–89 .— .— 1.12 (.830–1.500)Bothered/Upset at Work 1.29* (1.016–1.645) .— 1.10 (.738–1.645)—Change 1986–89 .— .— 1.38* (1.020–1.878)Bothered/Upset as Parent 1.20 (.952–1.519) .— 1.08 (.766–1.530)—Change 1986–89 .—| | .—| | 1.19| (.856–1.643)N 596 — 435LR Chi2 112.6*** — 92.7***

Note: *** p < .001; ** p < .01; * p < .05; † p < .10. Model 1 adjusts for all covariates included in Model 3, Table 3;Models 2 and 3 adjust for all covariates included in Model 5, Table 3.

for testing these interactions and these findingsshould be verified in future research.

We also tested potential confounders of theassociations shown here. First, pre-existing de-pression could influence respondents’ reportsof their working conditions and sleep quality,or, alternatively, repeated exposures to workrole-related stressors could increase depressivesymptoms that account for sleep difficulties(Link et al. 1993).3 While we already controlfor neuroticism, a characteristic strongly asso-ciated with depressive symptoms, we also re-estimated our models after eliminating respon-dents in the top fourth of depressive symptomsat baseline. The results were substantivelyidentical to those presented here. Second, weexamined other relevant health behaviors in-cluding measures of smoking status, alcoholuse, and an index of physical activity.Measures of these behaviors in 1986 and ofchanges in behaviors between 1986 and 1989were not associated with sleep quality, and didnot change the results of the main analysis.Finally, negative experiences at work maymake it difficult for individuals to get enoughsleep, thereby leading them to report negative-ly about both work and sleep quality. However,the magnitude of the correlation between sleepduration and quality measures is –.2 or lessamong ACL respondents. Sleep duration in1986 is negatively associated with poor sleepquality in 1986, but it is not a significant pre-dictor in longitudinal models and did notchange the associations between working con-ditions and sleep quality.

DISCUSSION

For past generations of workers, the strain ofphysical effort on the job tended to push themtoward physical fatigue and restorative sleep,but emerging research shows that commonpsychosocial stressors at work seem to exertthe opposite effect, making it more difficult forindividuals to achieve restful sleep (Linton2004; Ota et al. 2005; Winwood andLushington 2006). Most of the prior evidence,however, is based on cross-sectional data orsamples of workers who have unusually diffi-cult work conditions (such as rotating shiftwork). Instead, we prospectively examine theway that common negative experiences in theU.S. workplace may “follow workers home”and impinge on sleep quality.

We explored four research questions. First,we asked whether perceived low control, per-

ceived job insecurity, or being bothered or up-set at work are prospectively associated withpoor sleep quality. Second, we asked if finan-cial, spousal/partner, or child-related negativeexperiences are independently associated withsubsequent poor sleep quality. Our third andfourth research questions asked if negative ex-periences at home were stronger predictorsthan negative experiences at work, and if theyexplained the impact of workplace experienceson sleep quality. The results showed mixedsupport for our first hypothesis, some supportfor the second, and little support for the re-maining two; below, we discuss each in turn.Most centrally, our results show that beingbothered or upset frequently at work predictschanges toward poorer sleep quality, an associ-ation robust to multiple alternative specifica-tions. By contrast, perceived low control wasnot significantly associated with change insleep quality in longitudinal models, whilethere was limited and inconsistent evidencethat job insecurity is associated with poorersleep quality.

Why does being bothered or upset at workshow the most robust prospective associationwith poor sleep quality when compared to theother measures of stressful experiences at workthat we explored here? One possibility is thatlow control and job insecurity as measuredhere reflect individuals’ perceptions of theirobjective working conditions, but do not nec-essarily capture their appraisal of how threat-ening or disturbing these conditions may be.While perceived job control or job insecuritymay lead to a negative stress response for someworkers, being bothered or upset at work is adirect measure of emotional response.Moreover, frequently being bothered or upsetmay indicate that an individual has to take di-rect action to remediate the bothersome or up-setting experiences, while direct action is notrequired by the other working conditions ex-amined here. To better understand this newfinding, future research could more carefullyexplore what kinds of negative experienceslead workers to report being bothered or upsetat work. The sociological literature on emo-tions in the workplace suggests that power andstatus differentials in the workplace (Livelyand Powell 2006; Lovaglia and Houser 1996),and the difficulties of managing emotions, par-ticularly for those working in the service econ-omy (Hochschild 1983), would be fertile direc-tions to explore.

488 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Among the indicators of negative experi-ences at home studied here, we found that on-ly dissatisfaction with one’s financial situationhad a significant independent association withsubsequent poor sleep quality in 1989, an as-sociation suggested by prior studies (Steptoe etal. 2008). Additionally, the magnitude of theassociations showed that negative experiencesat home are not more strongly associated withsleep quality than workplace experiences, anddo not explain the impact of negative emotion-al experiences at work. Moreover, by includingmultiple measures of feeling bothered or upset(i.e., at work, with one’s marriage/relationship,as a parent) we show that feeling bothered orupset at work does not simply reflect a generalreporting tendency, but appears to have a do-main-specific association with sleep quality.The only subsample that showed distinct pat-terns was workers living with children under18 years of age, and this group deserves furtherstudy with larger samples. However, we maynot have adequately specified stressful experi-ences associated with the spousal/partner orparental roles; other measures and mechanismsshould be proposed and explored in futurework. Additionally, these findings should beinterpreted in light of the analytic sample used:All respondents were working at least 20 hoursper week at baseline and at follow-up, andfindings may vary for persons with greatercare-giving responsibilities that limit their par-ticipation in paid work, for example.

Several other limitations should be consid-ered when assessing these results. First, themeasure of poor sleep quality used here isbased on a single item drawn from a scale ofdepression items; a more detailed set of ques-tions designed specifically to measure dis-turbed sleep or insomnia symptoms couldmore accurately reveal the association betweenworking conditions and sleep quality, but sur-veys that include higher quality measures, suchas the Pittsburgh Sleep Quality Index (Buysseet al. 1989), generally do not observe national-ly-representative samples of U.S. workers orfollow workers over time. Compared to survey-based measures, laboratory polysomnographyis the gold standard for measuring sleep distur-bance, but it is typically not applied to largepopulation-based samples. Fortunately, self-re-ports have been found to provide reasonableestimates of sleep quality (Karacan et al.1976).

Second, the ACL data capture job qualityand experiences from the late 1980s, and theworld of work has changed in the interim. Asperceived job security has declined further indeclining industries, such as manufacturing,and even in the growing service sector, the im-portance of this stressor for sleep quality mayhave grown, and associations with insecurityand job control should be explored using morerecent data. Our results suggest that emotionalresponses to experiences at work may be mostsalient for sleep quality, however, so it is notclear how changes in more objective workingconditions since the late 1980s may have af-fected the nature or prevalence of these emo-tional experiences. In future studies, otherworking conditions deserve attention, as domore direct measures of their appraised threatto the individual. For example, we have nomeasure of shift work in our study, though ithas been shown to strongly influence sleep pat-terns in other research. Larger samples ofworkers and additional waves of survey data al-so would be helpful for isolating associationswithin subgroups and more carefully investi-gating the temporal ordering of changes in ex-posures and outcomes.

Despite these limitations, the ACL sampleprovides the best data currently available togain insight into the links between stressfulworking conditions and poor sleep qualityamong U.S. workers. Our conclusions arestrengthened by our access to longitudinal da-ta on workers from across the occupationalspectrum, and results are consistent acrossmultiple alternative specifications. Controlsfor neuroticism and baseline health in our lon-gitudinal models mean that our findings arenot likely unduly affected by negative reportingstyles or selection into particular jobs on thebasis of health. Future research is needed, how-ever, to substantiate these results and furtherexplore the factors that could buffer workersfrom these negative conditions or interventionsthat could break the link between conditions inthe workplace and maintenance of healthysleep patterns.

NOTES

1. Supporting the validity of this measure ofsleep quality, we found significant associa-tions between this CES-D item for restlesssleep and other, more detailed items, specif-ically designed to measure sleep quality inthe Chicago Community Adult Health

PUTTING WORK TO BED 489

Study in 2001 (Morenoff et al. 2007) and inthe NLSY 79 cohort (among 40 to 50 yearolds) in 1998–2006. Future studies wouldbenefit from more detailed measures ofsleep quality to verify our results.

2. The racial-ethnic backgrounds of ACL re-spondents reflects the population composi-tion of the United States in 1986 when thefraction of groups other than non-Hispanicwhites and blacks was smaller than it is to-day. The indicator for “other” race-ethnicityincludes the small number of Hispanic,Asian, and other respondents.

3. The reciprocal associations between depres-sion and poor sleep quality have been notedin the biomedical and psychological litera-tures, and the relationship may be self-rein-forcing. For example, if workers with lowcontrol develop depressive symptoms thatmake it more difficult to achieve high qual-ity sleep, they may come to work fatiguedand subsequently have more difficulty onthe job, reinforcing their depressive symp-toms and creating a vicious cycle leading tolonger-term insomnia (Espie 2002).

REFERENCESAkerstedt, T. 2003. “Shift Work and Disturbed

Sleep/Wakefulness.” Occupational Medicine53:89–94.

Arber, S., J. Hislop, M. Bote, and R. Meadows.2007. “Gender Roles and Women’s Sleep in Midand Later Life: A Quantitative Approach.” Soci-ological Research Online 12. Retreived July 14,2008 (http://www.socresonline.org.uk/index.html).

Brief, A. P., M. J. Burke, J. M. George, B. S. Robin-son, and J. Webster. 1988. “Should NegativeAffectivity Remain an Unmeasured Variable inthe Study of Job Stress?” Journal of AppliedPsychology 71:193–98.

Brissette, I. and S. Cohen. 2002. “The Contributionof Individual Differences in Hostility to theAssociations between Daily Interpersonal Con-flict, Affect, and Sleep.” Personality and SocialPsychology Bulletin 28:1265–74.

Burgard, Sarah A., Jennie E. Brand, and James S.House. 2009. “Perceived Job Insecurity andWorker Health in the United States.” Social Sci-ence and Medicine 69(5):777–785.

Buysse, D. J., C. F. Reynolds, T. H. Monk, S.R.Berman, and D.J. Kupfer. 1989. “The PittsburghSleep Quality Index: A New Instrument for Psy-chiatric Practice and Research.” PsychiatryResearch 28:193–213.

Colten, Harvey R. and Harvey R. Altevogt. 2006.Sleep Disorders and Sleep Deprivation: An

Unmet Public Health Problem. Washington, DC:The National Academies Press.

Espie, Colin A. 2002. “INSOMNIA: ConceptualIssues in the Development, Persistence, andTreatment of Sleep Disorder in Adults.” AnnualReview of Psychology 53:215–43.

Eysenck, H. J. and S. B. G. Eysenck. 1975. TheEysenck Personality Questionnaire. London:Hodder & Stoughton.

Ferrie, Jane E., Martin J. Shipley, Francesco P. Cap-puccio, Eric Brunner, Michelle A. Miller, MeenaKumari, and Michael G. Marmot. 2007. “AProspective Study of Change in Sleep Duration;Associations with Mortality in the Whitehall IICohort.” Sleep 30: 1659–1666.

Ferrie, Jane Elizabeth, Martin John Shipley,Michael Gideon Marmot, Stephen Alfred Stans-feld, and George Davey Smith. 1995. “HealthEffects of Anticipation of Job Change and Non-employment: Longitudinal Data from the White-hall II Study.” British Medical Journal 311:1264–69.

———. 1998a. “The Health Effects of MajorOrganizational Change and Job Insecurity.”Social Science and Medicine 46:243–54.

———. 1998b. “An Uncertain Future: The HealthEffects of Threats to Employment Security inWhite-Collar Men and Women.” American Jour-nal of Public Health 88:1030–36.

Heaney, Catherine A., Barbara A. Israel, and JamesS. House. 1994. “Chronic Job Insecurity amongAutomobile Workers: Effects on Job Satisfactionand Health.” Social Science and Medicine38:1431–37.

Henry, Doug, Dana McClellen, Leon Rosenthal,David Dedrick, and Melissa Gosdin. 2008. “IsSleep Really for Sissies? Understanding theRole of Work in Insomnia in the US.” Social Sci-ence and Medicine 66:715–26.

Hochschild, Arlie. 1983. The Managed Heart.Berkeley: University of California Press.

Hochschild, Arlie and Anne Machung. [1989] 2003.The Second Shift: Working Parents and the Rev-olution at Home. New York: Penguin Books.

House, James S. 1987. “Chronic Stress and Chron-ic Disease in Life and Work: Conceptual andMethodological Issues.” Work and Stress1:129–34.

———. 2002. “Understanding Social Factors andInequalities in Health: 20th Century Progressand 21st Century Prospects.” Journal of Healthand Social Behavior 43:125–42.

Hyyppä, Markku T., Erkki Kronholm, and ErkkiAlanen. 1997. “Quality of Sleep during Eco-nomic Recession in Finland: A LongitudinalCohort Study.” Social Science and Medicine45:731–38.

Joelson, L. and L. Wahlquist. 1987. “The Psycho-logical Meaning of Job Insecurity and Job Loss:Results of a Longitudinal Study.” Social Scienceand Medicine 25:179–82.

490 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Kalimo, Raija, Leena Tenkanen, Mikko Harma,Esko Poppius, and Pertti Heinsalmi. 2000. “JobStress and Sleep Disorders: Findings from theHelsinki Heart Study.” Stress Medicine16:65–75.

Karacan, Ismet, John I. Thornby, Michael Anch,Charles E. Holzer, George J. Warheit, John J.Schwab, and Robert L. Williams. 1976. “Preva-lence of Sleep Disturbance in a Primarily UrbanFlorida County.” Social Science and Medicine10:239–44.

Karasek, R. A. 1979. “Job Demands, Job Control,and Mental Strain: Implications for JobRedesign.” Administrative Science Quarterly24:285–308.

Knudson, Hannah K., Lori J. Ducharme, and PaulM. Roman. 2007. “Job Stress and Poor SleepQuality: Data from an American Sample of Full-Time Workers.” Social Science and Medicine64:1997–2007.

Kohn, Melvin L. and Carmi Schooler. 1982. “JobConditions and Personality: A LongitudinalAssessment of their Reciprocal Effects.” Ameri-can Journal of Sociology 87:1257–86.

———. 1983. Work and Personality: An Inquiry inthe Impact of Social Stratification. Norwood, NJ:Ablex Publishing.

Kutner, Nancy G., Donald L. Bliwise, and RebeccaZhang. 2004. “Linking Race and Well-Beingwithin a Biopsychosocial Framework: Variationin Subjective Sleep Quality in Two RaciallyDiverse Older Adult Samples.” Journal ofHealth and Social Behavior 45:99–113.

Lamberg, Lynne. 2004. “Promoting Adequate SleepFinds a Place on the Public Health Agenda.”Journal of the American Medical Association291: 2415–17.

Lennon, Mary Clare. 1994. “Women, Work, andWell-Being: The Importance of Work Condi-tions.” Journal of Health and Social Behavior35:235–47.

Link, Bruce G., M. C. Lennon, and B. P. Dohren-wend. 1993. “Socioeconomic Status and Depres-sion: The Role of Occupations Involving Direc-tion, Control, and Planning.” American Journalof Sociology 98:1351–87.

Linton, Steven J. 2004. “Does Work Stress PredictInsomnia? A Prospective Study.” British Journalof Health Psychology 9:127–36.

Lively, Kathryn J. and Brian Powell. 2006. “Emo-tional Expression at Work and at Home:Domain, Status, or Individual Characteristics?”Social Psychology Quarterly 69:17–38.

Lovaglia, Michael J. and Jeffrey A. Houser. 1996.“Emotional Reactions to Status in Groups.”American Sociological Review 61:867–83.

Mattiasson, I., F. Lindgarde, J. A. Nilsson, and T.Theorell. 1990. “Threat of Unemployment andCardiovascular Risk Factors: Longitudinal Studyof Quality of Sleep and Serum Cholesterol Con-

centrations in Men Threatened with Redundan-cy.” British Medical Journal 301:461–66.

Meltzer, L. J. and J. A. Mindell. 2007. “Relationshipbetween Child Sleep Disturbances and MaternalSleep, Mood, and Parenting Stress: A PilotStudy.” Journal of Family Psychology 21:67–73.

Menaghan, Elizabeth A. 1991. “Work Experienceand Family Interaction Processes: The LongReach of the Job?” Annual Review of Sociology17:419–44.

Mirowsky, John and Catherine E. Ross. 2007. “Cre-ative Work and Health.” Journal of Health andSocial Behavior 48:385–403.

Moore, Philip J., Nancy E. Adler, David R.Williams, and James S. Jackson. 2002. “Socioe-conomic Status and Health: The Role of Sleep.”Psychosomatic Medicine 64:337–44.

Morenoff, Jeffrey D., James S. House, Ben B.Hansen, David R. Williams, George A. Kaplan,and Haslyn E. Hunte. 2007. “UnderstandingSocial Disparities in Hypertension Prevalence,Awareness, Treatment, and Control: The Role ofNeighborhood Context.” Social Science andMedicine 65:1853–66.

Morin, Charles M., Sylvie Rodrigue, and HansIvers. 2003. “Role of Stress, Arousal, and Cop-ing Skills in Primary Insomnia.” PsychosomaticMedicine 65:259–67.

National Highway Traffic Safety Administration.2006. “The Impact of Driver Inattention onNear-Crash/Crash Risk.” Washington, DC: U.S.Department of Transportation.

Ota, Atsuhiko, Takeshi Masue, Nobufumi Yasuda,Akizumi Tsutsumi, Yoshio Mino, and HiroshiOhara. 2005. “Association between PsychosocialJob Characteristics and Insomnia: An Investiga-tion Using Two Relevant Job Stress Models—The Demand-Control-Support (DCS) Model andthe Effort-Reward Imbalance (ERI) Model.”Sleep Medicine 6:353–58.

Pavalko, Eliza, Fang Gong, and J. Scott Long. 2007.“Women’s Work, Cohort Change, and Health.”Journal of Health and Social Behavior48:352–68.

Pearlin, Leonard I., Elizabeth G. Menaghan, A.Lieberman Morton, and Joseph T. Mullan. 1981.“The Stress Process.” Journal of Health andSocial Behavior 22:337–56.

Radloff, Lenore S. 1977. “The CES-D Scale: ASelf-Report Depression Scale for Research inthe General Population.” Applied PsychologicalMeasurement 1:385–401.

Ribet, C. and F. Derriennic. 1999. “Age, WorkingConditions, and Sleep Disorders: A LongitudinalAnalysis in the French Cohort ESTEV.” Sleep22:491–504.

Sekine, M., T. Chandola, P. Martikainen, M. Mar-mot, and S. Kagamimori. 2006. “Work and Fam-ily Characteristics as Determinants of Socioeco-nomic and Sex Inequalities in Sleep: The Japan-ese Civil Servants Study.” Sleep 29:206–16.

PUTTING WORK TO BED 491

Sarah A. Burgard is Assistant Professor of Sociology, Assistant Research Scientist at the Population StudiesCenter, and Joint Assistant Professor of Epidemiology at the University of Michigan. Her work examinesthe consequences of social stratification for population health and health disparities. She studies the linksbetween working conditions and adult health over the career as well as the multilevel social determinants ofthe health of children and adolescents in multiple international contexts.

Jennifer A. Ailshire is a doctoral candidate in the Department of Sociology and a National Institute onAging Trainee in the Population Studies Center at the University of Michigan. Her research focuses on un-derstanding how social and economic inequality over the life course produces differences in health trajec-tories and on determining the extent to which the social and physical conditions in which people live ex-plain social disparities in health and health-related behaviors.

492 JOURNAL OF HEALTH AND SOCIAL BEHAVIOR

Steptoe, A., K. O’Donnell, M. Marmot, and J. War-dle. 2008. “Positive Affect, Psychological Well-Being, and Good Sleep.” Journal of Psychoso-matic Research 64:409–15.

Tasali, E., R. Leproul, D. A. Ehrmann, and E. VanCauter. 2008. “Slow-Wave Sleep and the Risk ofType 2 Diabetes in Humans.” Proceedings of the

National Academy of Sciences of the UnitedStates of America 105:1044–49.

Winwood, Peter C. and Kurt Lushington. 2006.“Disentangling the Effects of Psychological andPhysical Work Demands on Sleep, Recovery andMaladaptive Chronic Stress Outcomes within aLarge Sample of Australian Nurses.” Journal ofAdvanced Nursing 56:679–89.