the sociology of suicide · 2018-07-31 · (fatalism) toa now infamous footnote wherehesuggeststhat...

SO37CH24-Wray ARI 8 June 2011 21:37

The Sociology of SuicideMatt Wray,1 Cynthia Colen,2

and Bernice Pescosolido3

1Department of Sociology, Temple University, Philadelphia, Pennsylvania 19122;email: [email protected] of Sociology, Ohio State University, Columbus, Ohio 43210;email: [email protected] of Sociology, Indiana University, Bloomington, Indiana 47405;email: [email protected]

Annu. Rev. Sociol. 2011. 37:505–28

First published online as a Review in Advance onApril 26, 2011

The Annual Review of Sociology is online atsoc.annualreviews.org

This article’s doi:10.1146/annurev-soc-081309-150058

Copyright c© 2011 by Annual Reviews.All rights reserved

0360-0572/11/0811-0505$20.00

Keywords

Durkheim, social integration, social influence, social networks,mortality, health and illness

Abstract

Since Durkheim’s classic work on suicide, sociological attention tounderstanding the roots of self-destruction has been inconsistent. Inthis review, we use three historical periods of interest (pre-Durkheim,Durkheim, post-Durkheim) to organize basic findings in the body of so-ciological knowledge regarding suicide. Much of the twentieth-centuryresearch focused on issues of integration and regulation, imitation, andthe social construction of suicide rates. Innovations in the twenty-first-century resurgence of sociological research on suicide are describedin detail. These newer studies begin to redirect theory and analysistoward a focus on ethnoracial subgroups, individual-level phenomena(e.g., ideation), and age-period-cohort effects. Our analysis of sociol-ogy’s contributions, limits, and possibilities leads to a recognition of theneed to break through bifurcations in individual- and aggregate-levelstudies, to pursue the translation of Durkheim’s original theory into anetwork perspective as one avenue of guiding micro-macro research,and to attend to the complexity in both multidisciplinary explanationsand pragmatic interventions.

505

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


INTRODUCTIONAs an object of sociological inquiry, suicide hashad a long and strange career. At different mo-ments since the birth of the discipline, it hasvariously served as a foundational subject forthe establishment of the discipline, as a provingground for methodological debates, and as anindex of social integration nonpareil. Althoughsociological investigations of suicide flourishedthrough the 1970s, the closing decades of thetwentieth century saw sociological interest inthe topic wane. This same period, however, wit-nessed increased attention to suicide from med-ical professionals, public health researchers, ex-pert policy makers, and legislators, culminatingin The Surgeon General’s Call to Action to PreventSuicide (U.S. Public Health Service 1999), theInstitute of Medicine’s Reducing Suicide: A Na-tional Imperative (Goldsmith et al. 2002), theCenters for Disease Control’s (2008) reporton preventing suicide through social connect-edness, the Garrett Lee Smith Memorial Actauthorizing millions in federal funds for sui-cide prevention (2004), and the implementa-tion of comprehensive suicide prevention plansin dozens of states.

All these policy reports have called forsocial scientists to engage in a national researchagenda designed to reduce rates of suicide, butthe response from sociologists has been weakat best. As a result, sociological insights are sel-dom incorporated into trends and priorities insuicide research. Suicide continues to be framedand understood as a problem faced by individ-uals, even when social and contextual factorsare acknowledged. Given this individualisticframe, biomedical and psychiatric perspectiveshave become paradigmatic, with research onindividual-level risk factors generating preven-tion strategies aimed at high-risk individuals(usually understood as those with a history ofmental illness, suicidal ideation, or previousattempts). The models for suicide preventionthat flow from this framework tend to neglectsocial and ecological determinants or includethem only in superficial and cursory ways.

We argue that the present situation is prob-lematic for sociology, for the scientific research

agenda on suicide, and for the creation of solu-tions to this pressing social problem. In an agewhen biomedicine and genomics tend to domi-nate scientific and public policy debates, includ-ing the understanding of health and mortalityoutcomes, sociological research that strictly fol-lows its own disciplinary boundaries is out ofstep with current emphases on complexity andmultidisciplinarity. Equally undesirable is thedecision to abandon suicide research to otherdisciplines and perspectives. In the followingfour-part review, we briefly place sociologicalresearch on suicide in historical context; sketchthe outlines of suicide research from psychol-ogy, biomedicine, and public health; describecontemporary perspectives on suicide withinsociology, identifying current theoretical andmethodological challenges; and offer sugges-tions for future sociological research, argu-ing for a more multidisciplinary and problem-solving approach to the sociology of suicide.

THE SOCIOLOGY OF SUICIDE INHISTORICAL CONTEXT

Historians (e.g., Fedden 1938, Minois 1999)have provided in-depth intellectual historiesof social commentary and research on self-destruction; there is neither need nor spacefor us to do this here. Instead, we organizesociological writing on suicide into three erascentered on sociology’s fundamental, landmarkcontribution, Durkheim’s (2006 [1897]) OnSuicide.

Pre-Durkheimian Approaches

If, as Richard Sennett (2006, p. xi) has writ-ten, Durkheim “taught the modern world howto think about suicide,” there were nonethe-less several modern thinkers on whose workDurkheim built. From a sociology of knowl-edge perspective, it is of little surprise thathe relied heavily on “general frames of refer-ence” and “specific knowledge and forms ofthought” (Douglas 1967, p. 13) current amonglate-nineteenth-century European intellectu-als. These thinkers were engaged in trying to

506 Wray · Colen · Pescosolido

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


identify the forces behind the apparent increasein national suicide rates coincident with the riseof modernity, which itself was presumed to beloosening the chains of agrarian society andproviding greater freedoms to the individual.Most influential among these thinkers were themoral statisticians, including Quetelet (1842)and Morselli (1882), who collected, calculated,and attempted to inductively analyze a largebody of suicide statistics. What impressed thesemoral statisticians most was the regularity andstability of national suicide rates year to year,as well as the overall rise in suicide rates in themodern era.

Against prevailing views that conceptualizedsuicide as a matter of free-willed individuals act-ing out of private despair, the moral statisticiansreasoned that such systematic, geographic, andtemporal patterns pointed to forces larger thanindividuals. In addition to this core belief, thesethinkers shared assumptions that suicide wasmorally problematic and that a persuasive ex-planation or theory of suicide would have to dis-cover and empirically document whatever forceor forces determined and regulated suicide ratesin the modern world. Durkheim shared mostof these assumptions with his contemporaries(Giddens 1965).

Masaryk (1970 [1881]) preceded Durkheimin looking to the forces of modernization forexplanations about rising rates. He focusedon suicide as an unintended consequence ofincreasing levels of education, arguing that theabandonment of traditional ways of thinkingand a preference for rational reflection openedup the idea of suicide as a solution to individualproblems. Tarde (1903 [1895]) counteredthe theories of moral statisticians by positingthat geographic and temporal clusteringof suicides could also be caused by variouskinds of imitative behavior, which, alongwith innovation, he argued, constituted thefundamental drivers of all social interaction.In this view, which resonates strongly withsocial learning theory of today, humans oftenimitate the actions and behaviors of otherpeople in their proximate social environments,especially when they regard those actions as

meaningful and significant and hold those theyare imitating in high esteem. Thus, suicidesin a given time and place can produce moresuicides in chain-reaction fashion, leading tothe geographical pockets of high suicide ratesthat the moral statisticians had documented.

Durkheim’s Great Leap

While agreeing that modernity was ultimatelyto blame for rising suicide rates, Durkheim tooka more analytical view than did Masaryk andthe moral statisticians, while explicitly rejectingTarde’s imitation theory. Instead, he sought tounderstand how negative meanings and emo-tions were produced in individuals and groupsduring times of dramatic social change and howsuch changes made some groups more vulner-able than others to self-destruction. He pro-vided a multifaceted theoretical scheme thatprivileged social explanations and dismissed,in strongly polemical terms, other popular layand scientific explanations, such as mental ill-ness, imitation, climate, and temperature (Pope1976).

The result of his efforts was a fourfold ty-pology of suicide derived from the intersectionof two major axes, which he termed integrationand regulation.1 By integration, he targetedthe sense of social belonging and inclusion,the love, care, and concern that can flow (ornot flow) from social ties. Well-integratedgroups, he argued, enjoy stable, durable, andcohesive social ties. Individuals in such groupsare supported in their lives, particularly duringtimes of personal crisis, thereby reducing theirvulnerability to suicide.

Durkheim described the nature and influ-ence of social integration in family, religion,and political domains and famously concludedfrom his empirical analyses that “suicide ratesvary inversely with the degree of integration of

1Since Durkheim’s On Suicide has been the subject of severalgenerations of sociological commentary (see Halbwachs 1978[1930], Douglas 1967, Pope 1976), we offer only a brief sum-mary here. For insightful overviews, see Pickering & Walford(2000).

www.annualreviews.org • The Sociology of Suicide 507

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


the social groups to which the individual be-longs” (Durkheim 2006 [1897], p. 224). How-ever, this finding has often been taken in iso-lation from his other major conclusion: thatgroups that tie together individuals in loyalty orlove too closely also contribute to group suiciderates. Individuals who are overly integrated intosocial structures (for example, martyrs or warheroes) are also at risk of suicide precisely be-cause these groups (and the individuals withinthem) value the needs of the group over the in-dividual’s need to survive.

Although Durkheim never explicitly definedsocial integration, it serves as the core insightthat has dominated, and continues to dominate,sociological thinking and analysis (Pope 1976,Stack 1994, Baudelot & Establet 2008 [2006]).Yet Durkheim was quite clear that he was alsoconcerned with what he called regulation, themonitoring, oversight, and guidance that comefrom social ties. For Durkheim, individuals re-quire moral guidance and external restraint be-cause without them their desires and expecta-tions will exceed their grasp, with the resultingfailures and frustrations leading to continuousstates of despair.

Thus, regulation, as well as integration, iscentral to his theory, with both over- andunderregulation producing suicide.2 The so-cial forces of integration and regulation inter-act in U-shaped fashion to form the basis forDurkheim’s fourfold typology of suicide thatarises in social structures characterized by ex-tremes. Egoistic suicide, occurring when inte-gration is low, and altruism, occurring whenintegration is high, both represent situationswhere care and concern deviate from moderate

2As Pescosolido (1994) points out, this duality and symmetryare often missed because Durkheim relegated his fourth type(fatalism) to a now infamous footnote where he suggests thatoverregulated social structure only characterizes premodernsocieties. As Coser & Coser (1979) suggest in their analysis ofthe Jonestown mass suicide in 1978, however, Durkheim mayhave been incorrect in this assertion. The present-day ten-dency to disregard the regulation side of the theory may alsobe a result of Johnson’s (1965) influential and more parsimo-nious contention that Durkheim’s argument could be boileddown to positing integration as the “one cause of suicide.”

levels. Similarly, anomic and fatalistic suicidesrepresent the increased risk that comes from so-cial structures with under- or overregulated sys-tems, respectively (Bearman 1991, Pescosolido1994, Pescosolido & Georgianna 1989, Smelser& Warner 1976). When and where forces ofintegration and regulation are extremely lowor extremely high, more people become morevulnerable to suicide. Only when these forcesare balanced, when individuals feel in harmonywith their own needs and the demands of thegroup, does the suicide rate diminish.

Post-Durkheim: Testing, Elaborating,Rejecting, and Transforming

Durkheim’s On Suicide (2006 [1897]) spurredAmerican suicide research in the first half ofthe twentieth century. Throughout the post-Durkheim era, sociologists have kept social in-tegration as a touchstone theme, although ithas often been called by other names, such associal isolation (Trout 1980), social cohesion(Kawachi & Kennedy 1997), or social support(Berkman et al. 2000).3 Ecological modeling ofurban suicide, refracted through the prism ofthe urban sociology of the Chicago School, be-came a sociological mainstay during the 1920s–1950s (Cavan 1928, Schmid 1928, Porterfield1949, Sainsbury 1955; see Stack 1994 for anoverview). These studies blended statistical dataon suicide rates with social observation of thegeography of communities and neighborhoodsto assess how individuals and groups were in-fluenced by their urban environments.

With community-based ecological studiesfalling out of fashion by the late 1950s,4

3Interactionist studies of suicide have occasionally chal-lenged the Durkheimian frame. Goffman (1969) offered hisoblique reflections on suicide in the wake of a personal loss,and Garfinkel (1967) and Sacks (1995) both produced eth-nomethodological accounts of suicide prevention centers andhotlines. Yet these remarkable studies have not had lastinginfluence within the discipline.4This may be the result of Robinson’s (1950) influentialarguments about the ecological fallacy, or a reflection ofthe growing challenges to the functionalist paradigm withinthe discipline, or, as Coleman (1986) argued, a shift in


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


three major sociological statements on suicideemerged in the decades that followed. First,drawing insights from psychoanalytic theory,Henry & Short (1954) offered their aggression-frustration model, hypothesizing that suicideand homicide are opposite sides of a singlephenomenon—violent aggression—which canonly be understood by taking into account soci-ological and psychological variables. In their ac-count, increases in frustration lead to increasesin aggression. In turn, increases in aggressionlead to increases in lethal violence. Violence di-rected outward against others produces homi-cide; violence directed inward against the selfproduces suicide. Whether individuals perceiveothers or themselves as the source of the frus-tration, they argued, varies predictably by classand status. Low-status groups blame othersand have low suicide rates; high-status groupsblame themselves and have high suicide rates.Although the theory was taken up by sociol-ogists and criminologists, subsequent studiesproduced inconsistent findings on status andsuicide.5

A decade later, Gibbs & Martin (1964)issued the second major statement of the eraby developing a status integration theory thatdismissed the distinction Durkheim madebetween integration and regulation. They pos-tulated that conflict between social roles relatedto age, sex, occupation, and marital status (vari-ables from which they constructed an index ofstatus integration) was the crucial determiningfactor in suicide. Their approach was more for-mally theoretical, more accurately observable,and therefore more testable than Durkheim’sconcepts of social integration and regulation.

sociological focus away from community structures to na-tional ones. We remain uncertain about the causes of theshift.5Gold (1958, p. 652) criticized Henry & Short’s (1954) as-sumption “that members of higher-status categories are lessrestrained externally than their lower-status counterparts”.Gold’s revision of Henry & Short’s theory placed a strongemphasis on class as the crucial variable determining an in-dividual’s choice between homicide and suicide. For recentevidence that poor, low-status groups can have high rates ofsuicide, see Rehkopf & Buka (2006).

Specifically, in social circumstances in whichindividuals experienced a high degree of irre-solvable role conflict, they reasoned, suiciderates increased. They found solid empiricalsupport for their theory. However, despite thepromise of this approach, it was not widelyadopted either within or outside the discipline.

In the third major statement from thisera, Douglas (1967) offered a tightly reasonedcritique not only of Durkheim, but also ofall sociological research on suicide. His ma-jor complaint was essentially Weberian—theDurkheimian perspective on suicide was flawedbecause it failed to take into account the socialand cultural meanings of suicide. Moreover, itfailed to consider suicide as a meaningful so-cial action undertaken by individuals and failedlikewise to attend to the variation in definitionsof suicide observable among those who wereofficially charged to determine if a suicide hadoccurred (i.e., coroners and medical examiners).From this last point, Douglas developed a skep-tical argument about the validity and reliabilityof official suicide statistics, leading him to re-ject quantitative approaches to suicide researchaltogether. Rather than a Durkheimian typol-ogy of suicides, what sociology needed most,he argued, was a typology of “suicidal actionsas socially meaningful actions” (Douglas 1967,p. 255). In the years that followed, sociologiststended to either ignore or reluctantly acknowl-edge the importance of meaning and the issuesof unreliable statistics that Douglas raised.

Other sociologists of this era battledDurkheim on a different front. Phillips, re-visiting Tarde, suggested in a series of arti-cles (Phillips 1974, 1979; Phillips & Carstensen1986; Phillips & Paight 1987) that imitation—particularly imitation of celebrity suicide—plays a bigger role in shaping suicidal behav-ior than Durkheim allowed. Phillips dubbedthis phenomenon “the Werther effect,” afterGoethe’s story of a young artist’s suicide thatspawned multiple “copycat” suicides (see Baron& Reiss 1985 for an important methodologi-cal critique). Other sociologists continued bothU.S. (e.g., Breault 1986) and cross-national(e.g., Stack 1983) work in a Durkheimian frame,


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


with Maris (1981) adding a life course perspec-tive on “suicidal careers.”

As the preceding paragraphs suggest, thepost-Durkheim era of sociological research onsuicide is quite difficult to summarize. AlthoughDurkheim continued to exert paradigmatic in-fluence on the ways that sociologists concep-tualized and carried out research on suicide,voices of caution and dissent emerged, offer-ing a confusing series of friendly amendments,skeptical arguments, and plausible alternativehypotheses. Simply put, sociology entered intoa period of disciplinary confusion and uncer-tainty regarding one of its canonical subjects,and the Durkheimian consensus began to fray.

PARALLEL RESEARCH:PSYCHOLOGICAL, BIOMEDICAL,AND PUBLIC HEALTHPERSPECTIVES

The ways that sociology as a discipline has con-ceived of suicide as an object of scientific inquirydiverge sharply from the ways that other disci-plines have done so. Although it is outside thescope of this article to elaborate this point, a fewof the contours and consequences of this disci-plinary divergence are critical to future researchin sociology.

Historically, psychological explanations ofsuicide have focused squarely on individualsand their conflicting internal emotional andprimal drives as well as their struggles overlife’s meaning. These themes stand at the coreof psychoanalytic thought (Zilboorg 1936,Menninger 1938). Research into the psycho-dynamics of suicide flourished from the 1930sto the 1980s, when it was challenged withinpsychology as lacking empirical rigor (seeMaltsberger & Goldblatt 1996 and Jamison1999 for overviews). However, the use of thepsychological autopsy survived. This retrospec-tive case study approach, pioneered by Zilboorgin the 1930s, combines medical histories, policereports, medico-legal investigations, and in-depth interviews with friends and relatives ofpeople who committed suicide, with the goal ofconstructing a more detailed postmortem than

that offered by forensic autopsy alone. Studiesusing psychological autopsy data are the sourceof the widely reported statistic that 90% ormore of individuals who kill themselves have ahistory of mental illness (Conwell et al. 1996).Although the method has been embraced bysome sociologists—notably Maris (1981)—ithas not entered the mainstream of sociologicalresearch on suicide, even as the development ofcase-control psychological autopsies increasedmethodological rigor (Cavanagh et al. 2003).

Following theoretical shifts in other scien-tific disciplines, psychodynamic models of sui-cide have been displaced by psychiatric andmedical models that look to discover biolog-ical and genetic causes. This research inves-tigates physiological responses to stressors—particularly the role that neurotransmitterssuch as serotonin and norepinephrine play inthis system—as well as the potential influenceof genetics in the etiology of suicide (Goldsmithet al. 2002). Findings highlight the associationof suicidality (i.e., suicidal ideation or attempts)with (a) dysregulation of the hypothalamic-pituitary-adrenal axis, the primary stress re-sponse system (Brunner et al. 2001); (b) vari-ation in the functioning of the serotonergicand noradrenic systems, which help regulateaggression and impulsiveness, as revealed bybrain-mapping techniques and postmortems(Arango & Mann 1992); and (c) genetic fac-tors assessed through family studies (Stathamet al. 1998), particularly studies of adopteesfrom families with multiple suicides (Wenderet al. 1986) and twin studies (Roy et al. 1991).The search for candidate genes has thus far beeninconclusive.

The 1990s also witnessed growing attentionto suicide within the public health researchagenda (U.S. Public Health Service 1999).Research focused largely on identifying popula-tions at greatest risk for suicide and on develop-ing prevention and intervention strategies to re-duce suicide rates and risk of suicide (Knox et al.2004). In the Durkheimian tradition, researchon the spatial distribution of suicide found it wasconcentrated more heavily in rural than urbanareas in the United States (Branas et al. 2004).


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Public health researchers have also looked tothe significant role played by access to lethalmeans (e.g., guns, toxic gas, bridges, and evenacetaminophen) and on efforts to reduce access,which were shown to have both temporary andpermanent effects on lowering suicide rates(Mann et al. 2005, Miller & Hemenway 2008).

There are, of course, the usual reasons forsociologists to be skeptical regarding the the-ories and evidence about suicide put forthby other disciplines. In most cases, the socialand ecological factors surrounding suicide arepoorly considered, if at all; the causal path-ways considered important for suicide are farnarrower than sociological thinking requires;and the methods and logics of inquiry do notalways match sociological standards. Yet thisskepticism toward other disciplinary perspec-tives, however warranted, does not appear to beserving the discipline well. While there is evi-dence that psychology, biomedicine, and pub-lic health have adopted an open and multidis-ciplinary stance toward understanding suicide(e.g., Duberstein et al. 2004), sociology has yetto do the same.

The dangers of this insularity are threefold.First, sociologists may fail to identify innova-tive and sociologically interesting approachesto suicide that arise in other disciplines.Second, other disciplines will continue toregard sociology as having produced fewnew insights into the topic since 1897. Thisperception, although inaccurate, is not wildlyso given that sociological research on suicidehas been dwarfed by research on suicide inother disciplines. According to a recent surveyof over 30,000 academic articles on suicidepublished since 1980, only about 400 (1.3%)could be categorized as sociological (Agerboet al. 2009, reported in Scourfield et al. 2010).Third, sociologists may miss opportunities todemonstrate our theories and evidence of howthe biological processes presumed to be fun-damental by biomedical researchers are in factcausally impacted by social and environmentalcontexts (House et al. 1988). We return to thesepoints in our discussion below and offer more

explicit suggestions about the contributionssociology can offer these other disciplines.

CONTEMPORARYSOCIOLOGICAL PERSPECTIVES

In order to assess major sociological research onsuicide over the past two decades, we searchedelectronic databases for articles with “suicide”in the title or abstract in four of the top-tier sociological journals, American Journal ofSociology, American Sociological Review, Journalof Health and Social Behavior, and Social Forces,for the period 1990–2009. We excluded spe-cialized journals on suicide, such as Suicide andLife-Threatening Behavior and Archives of SuicideResearch, as well as journals with a high degree ofmultidisciplinary content, such as Social Scienceand Medicine, in order to keep the articles underreview to a manageable number and because de-ciding which articles in multidisciplinary jour-nals were clearly sociological in nature provedto be an exceedingly subjective process. More-over, limiting our sample to four top sociol-ogy journals met our goal of focusing on andassessing high-quality sociological research onsuicide.

Our search netted 41 articles, each of whichwas read by at least two of us. Figure 1 illus-trates the percentage of total articles in eachof the four journals for which the major substan-tive focus is suicide and documents changingtrends over the period of interest. Althoughthe absolute number of articles on suicidefluctuates over time, percentages remainedexceedingly small and failed to rise above 3%.In what follows, we highlight recurring andoverlapping themes, focusing on articles thatrepresent each theme particularly well.

Theoretical Themes

The theoretical themes in our sample tappedinto classic sociological foci (e.g., Durkheimiantheory and imitation) as well as new ones (e.g.,previously understudied groups, such as womenand racial minorities).


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


0

0.5

1.0

1.5

2.0

2.5

1990–1994 1995–1999 2000–2004 2005–2009

Per

cen

t

Year

SF

ASR

JHSB

AJS

Figure 1Articles on suicide as a percentage of all published articles, by journal. Journals: ASR, American SociologicalReview; AJS, American Journal of Sociology; JHSB, Journal of Health and Social Behavior; and SF, Social Forces.

Social stratification and suicide. One of themore novel and promising themes that emergesfrom recent empirical efforts concerning the so-cial patterning of suicide is the emphasis on howmechanisms of stratification may interact eitherto promote or to prevent suicide. Although sev-eral studies address how rates of suicide vary ac-cording to existing social cleavages such as race,class, and gender (Almgren et al. 1998, Burret al. 1999, Campbell & Troyer 2007, Ellisonet al. 1997, Girard 1993, Krull & Trovato 1994,Kubrin et al. 2006, Lehmann 1995, Pampel1998, Wadsworth & Kubrin 2007), a notablefew integrate theories of stratification in novelways to deepen our understanding of the com-plex web of social forces that drive suicidal acts.

Racial disparities. Research on racial/ethnicdisparities in suicide incorporates perspectivesfrom criminology, population health, andsocial psychology to gain a more nuancedunderstanding of why racial minorities faceeither reduced or heightened risk of suicide.These efforts extend beyond Black/White dif-ferentials to include Hispanics, for whom therisk for suicide is low, and Native Americans,for whom the risk is high.

Kubrin and colleagues (2006) employWilliam Julius Wilson’s deindustrialization

hypothesis to explain the historical increasein suicide among Black men in the UnitedStates. Incorporating measures specificallyconstructed to capture Black structural dis-advantage at the metropolitan level (e.g.,Black/White inequality), the authors find thatBlack disadvantage is a significant predictor ofyoung (15–34) Black male suicide rates. More-over, absolute levels and recent fluctuations inindustrial sectors of the local labor market arepositively associated with increases in suiciderates for this subpopulation.

Similarly, Wadsworth & Kubrin (2007) ex-amine structural correlates and cultural factorsto explain relatively low levels of suicide amongU.S. Hispanics. Asking whether the ethnoracialand socioeconomic composition of an areamoderates the association between individualrace/ethnicity and suicide, they find importantdifferences by nativity. For foreign-bornHispanics, higher levels of economic inequalitybetween Blacks and Hispanics are significantpredictors of lower suicide rates. For native-born Hispanics, higher levels of economicinequality between Whites and Hispanics aresignificant predictors of higher suicide rates.These somewhat surprising findings suggestthat foreign-born Hispanics are likely to useBlacks as a comparison group, whereas their


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


native-born counterparts appear to look towardWhites to provide a source of social assessment.

Focusing on individual suicidal ideation,rather than aggregate suicide rates, Campbell& Troyer (2007) investigate the extent to whichracial misclassification negatively impacts arange of psychological outcomes. Using datafrom the National Longitudinal Survey ofAdolescent Health (Add Health) to identifyNative Americans whose racial identity wasincorrectly identified by interviewers, they findracial misclassification is positively associatedwith both suicidal thoughts and attempts.Additionally, for Native Americans, residingin a neighborhood with a significant NativeAmerican population seems to be protectiveagainst suicidal ideation. This finding suggeststhat the presence of a critical mass of same-race residents may balance levels of socialintegration and regulation.

Gender disparities. One of the most classicfindings in suicide lies in sex and gender dif-ferences, with men being three to four timesmore likely to commit suicide than women(Center for Disease Control 2008). AlthoughDurkheim argued that women have internal, bi-ological protections from suicide, sociologicallythe protections for women have traditionallybeen linked to closer ties to friends and family aswell as to a lower emphasis on occupational at-tainment (Bielby & Bielby 1984, Gerson 1986,Eccles 1987, Maume 2006).

Using aggregate data from 18 developednations, Pampel (1998) examines whethermovements toward social equality for womenreduce disparities in suicide by sex. His findingsreveal a curvilinear trend in female suicide ratesover time. As women’s labor force participationincreases and marital stability decreases, thedifferential in suicide rates between menand women is reduced, with women’s ratesrising relative to men’s. However, as femaleworkforce participation continues to escalate,women’s rates begin to fall. Notably, the returnof the female suicide advantage occurs morequickly in nations that can be characterizedas collectivist in nature. Pampel theorizes that

such societies can adapt quickly and moreappropriately to women’s changing needs.Similarly, Krull & Trovato (1994) focus ontemporal changes in suicide in Quebec, findingthat numerous indicators of decreased integra-tion (e.g., divorce, childlessness, irreligiosity,and unemployment) impact suicide rates formen. However, only marital dissolution has asimilar effect on female suicide rates.

Religious disparities. Given the centrality ofreligion in Durkheim’s On Suicide, both clas-sic and recent work in sociology examines theextent to which suicide rates are associated withreligious affiliation, attendance, and beliefs. Bythe late 1980s and early 1990s, research sug-gested that the unique denominational historyof the United States as well as distinct reli-gious hubs required a rethinking and adapta-tion of Durkheim’s original hypotheses. Thesestudies found a continued protection from sui-cide in areas with a greater proportion ofCatholics. However, the effect of Protestantismwas now clearly bifurcated. The presence ofliberal groups (e.g., Episcopalians) predictedhigher rates, whereas the presence of evangel-ical groups predicted rates similar to those inareas with substantial Catholic profiles. Evenmore surprising, religious effects were regional,perhaps reflecting a web of support organiza-tions as well as coreligionists (Pescosolido &Georgianna 1989, Pescosolido 1990).

In our sample of recent work, Ellison andcolleagues (1997), constructing a unique mea-sure of religious homogeneity (the Herfindahlindex), follow in this tradition by examiningaggregate suicide rates from 1979–1981 acrossnearly 300 standard metropolitan statisticalareas. They find that high levels of religioushomogeneity are associated with lower suiciderates. Moreover, there appears to be significantgeographic variation in the strength of thisrelationship, with the strongest effect on ratesin the Northeast and the South.

The aforementioned studies indicate thatthe social processes involved in shapingsuicide risk and completion vary according tosocial cleavages, namely race/ethnicity, gender,


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


and religion. For example, it appears that aswomen’s roles in society move away from thefamily and toward the workplace, gender dis-parities in suicide contract and then expand.This type of curvilinear pattern makes sensegiven that, as a critical mass of women enterthe labor force, such a seismic demographicshift becomes less threatening, both to the ex-isting male-dominated social structure and tothe identities of individual women. Althoughfindings regarding gender and religious dis-parities in suicide add an important layer ofnuance to the body of research, the strengthof recent empirical work emerges from stud-ies that examine variations in suicide by race.Notably, these efforts have not been ham-pered by a heavy reliance on a Black/Whitedichotomy that typically dominates stratifica-tion research; rather, suicide risk and comple-tion among groups as disparate as young AfricanAmerican men, foreign- and native-bornHispanics, and Native American youth havebeen carefully explored. Moreover, differenti-ations within groups have been conscientiouslyexamined, the unique structures of national datasets have been exploited, and predictive mea-sures have been skillfully constructed by draw-ing upon established stratification theory.

Integration and regulation. Not surpris-ingly, much new research continued torely on Durkheimian theory, placing socialintegration—the presence of stable and durablerelationships—squarely in the center of theiranalyses (Gibbs 2000; Kposowa et al. 1995;Maimon & Kuhl 2008; Stockard & O’Brien2002a,b; Thorlindsson & Bjarnason 1998; vanTubergen et al. 2005). Most noteworthy are ef-forts to clarify, update, and challenge Durkheimby (a) modeling the effects of integration on theindividual rather than aggregate level, (b) reex-amining the neglected theme of regulation, and(c) revisiting the theory of status integration.

Thorlindsson & Bjarnason’s (1998) study ofsuicide attempts among Icelandic youth mod-els the relationship between social integra-tion and suicide on the interpersonal level.They find that familial integration, measured by

levels of emotional and material support withinthe family, protects youth against suicide at-tempts. Furthermore, both familial integrationand regulation, measured by parental rule set-ting and monitoring, protects against imitationeffects.

Gibbs (2000), following up on earlier work(Gibbs & Martin 1964, 1974), examines U.S.county-level data from 1980 by asking whetherstatus integration is inversely related to suiciderates because of the level of marital integration,occupational integration, or both. Earlier re-search (Stafford & Gibbs 1985) had indicatedthat the protective effects of occupational in-tegration were stronger than marital integra-tion and that this varied across age, gender, andracial groups. Gibbs (2000) finds that, althoughboth types of integration are important, the ef-fect of marital integration is stronger and linear.

Focusing on White males, Kposowa et al.(1995) use individual-level data from the 1979–1985 National Longitudinal Mortality Study totest hypotheses about social integration. Com-paring death from suicide to death from othercauses, they find only mixed support for theprotective effects of marriage. Men who are di-vorced, are widowed, or live alone are at greaterrisk for suicide; however, after controlling forsocioeconomic status, single and widowed menare not. They also report different protectiveeffects for urban versus rural residents andimmigrants versus native-born, findings thatboth support and contradict Durkheim’s origi-nal ideas.

Social imitation and cultural influence. Re-newed interest in the dynamics of microinter-actions and cultural arguments that emphasizeshared meanings and habits have led some so-ciologists to reconsider Durkheim’s claims ofthe irrelevance of timing on suicide. For exam-ple, Baller & Richardson (2002) take up Tarde’sargument that spatial clustering of suicidescould be produced by social imitation, test-ing these competing theories using data on ge-ographic patterning of suicides in France inthe 1870s and in the United States in 1990.Controlling for the geographic clustering of


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


multiple dimensions of social integration, theyfind evidence for an imitation effect in Franceand in most U.S. counties, except those in theAmerican West (see also Thorlindsson &Bjarnason 1998 on Iceland). Interestingly, theyargue that the absence of imitative effects in theWest—a region of very low social integration—may suggest that as social integration decreasesand isolation from community life becomesmore prevalent, susceptibility to imitative sui-cide decreases as well.

Returning to the frustration-aggressionmodel developed by Henry & Short (1954),Huff-Corzine et al. (1991) compare structuralinfluence (i.e., poverty rates) with cultural influ-ence (i.e., Southerness measured rather weaklyas percent born in the South) to estimate theireffects on state-level lethal violence rates ofWhites and Blacks from 1969–1971. They findthat state levels of severe poverty correlate posi-tively with violence rates for both racial groups.However, cultural influence does not, althoughit does affect the ratio of homicides to suicides.Poverty rates increase the proportion of suicideto homicide in the rates of White suicide, butnot Black suicide. The authors conclude that amerger of structural and cultural explanationsfor violence is long overdue and that greaterspecificity regarding how cultural mechanismsinfluence regional suicide and homicide rates isrequired.

Using a unique Swedish data set of 1.2million adults who lived in greater Stockholmin the 1990s, Hedstrom et al. (2008) pursue thequestion of social influence by asking whetherpersons who are exposed to suicides in theirfamilies or workplaces are at greater risk ofsuicide themselves. Controlling for other riskfactors, men exposed to a suicide in the familyare 8.3 times more likely to kill themselvesthan unexposed men. Workplace exposureshave a smaller but still impressively large effect(3.5) on men. The authors note, somewhatcounterintuitively, that from a suicide pre-vention standpoint workplace suicides actuallypose the greater risk because workplaces aretypically much larger than families, and as a re-sult greater numbers of individuals are exposed.

This is the first study to document exposureeffects for adults; previous work targetedexposure to suicide among adolescents.

Taken together, these articles, along withothers not discussed here (Pridemore 2006,Stack & Gundlach 1992, Trovato 1998), high-light the influence of imitation and culturalnorms on the patterning of suicide rates andsuggest that a return to the Tardean themes dis-missed by Durkheim is warranted. Cultural so-ciologists and others interested in mechanismsof social influence should take special note.

Methodological Themes

In recent work, suicide researchers tackledmethodological issues related to age, period,and cohort (APC) effects; multilevel analysis;and the construction and use of official suicidestatistics.

Age, period, and cohort effects. Disen-tangling concurrent and longer-term conse-quences of APC characteristics is an issuefraught with methodological challenges. Ageeffects refer to specific conditions that uniquelyshape the mortality profiles of individualswithin a given age range. Period effects—historical fluctuations in mortality that impactall age groups—are specific to the period im-mediately preceding or during which suiciderates are tallied. These may provide clues towhich macro-social and -economic trends maybe influencing population susceptibility to sui-cide within a given time frame. Cohort effectsrefer to historical fluctuations in mortality ratesacross groups of individuals who share a unify-ing event (e.g., all those born in the same yearor birth cohort). When an underlying cause ofdeath is thought to be significantly related toearly life exposures, suicide risk may be subjectto unexamined cohort effects, as well as the pe-riod and age effects mentioned above.

Arguing that there is a link between differen-tial opportunity structures experienced by birthcohorts and varying degrees of integration aswell as regulation, Stockard & O’Brien (2002b)investigate whether shifts in age-specific


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


suicide rates in the United States between1930 and 1995 were influenced by cohort char-acteristics such as family size and structure.Drawing on Elder (1994) and Easterlin (1987),they underscore the importance of consideringthe context within which members of a cohortundergo specific life course transitions. Theyfind that, independent of age and period effects,cohort variation—especially the relative size ofthe cohort and the predominance of nonmar-ital childbearing—is predictive of higher ratesof suicide.

In a similar vein, Pampel & Williamson(2001) further our understanding of how socialconditions, such as cohort size, family charac-teristics, and sociopolitical equality, impact age-related trajectories in suicide and homicide inhigh-income nations. Employing cause-specificmortality data from 18 countries for the periodbetween 1955 and 1994, the authors demon-strate that increases in the size of the youngest(15–24) or oldest (65–74) cohorts as well as re-cent changes in traditional family roles and fam-ily stability (measured by an index composedof rates of fertility, marriage, divorce, and fe-male labor force participation) are associatedwith increases in youth suicide rates relativeto elder suicide rates. These results are morerobust for rates of male suicide than for ratesof female suicide. However, the findings con-cerning the effect of country-specific sociopo-litical equality on suicide rates have little to dowith cohort variation. Mirroring the results ofPampel’s previous study of gender disparities insuicide discussed above, this study suggests thata nation’s degree of collectivism significantly af-fects group suicide rates.

Multilevel data and modeling approaches.Reflecting recent advancements in data col-lection, quality, and analytic tools, suicideresearchers have capitalized on the prolif-eration of detailed multilevel data sets andanalogous methodological strategies. Follow-ing a Durkheimian approach, Maimon & Kuhl(2008) employ data from Add Health to inves-tigate possible reasons for the recent increasein suicide attempts among American youth in

their teens and early twenties. Because of themultilevel structure and purposeful samplingstrategy of these data, the authors can examineintegrative social forces within and between theindividual, familial, school, and neighborhoodcontexts. Empirical findings suggest that theproportion of conservative religious adherentsin a neighborhood is negatively associated withindividual-level suicide attempts, even aftercontrolling for additional covariates. More-over, depression appears to be a significantlystronger predictor of suicide attempts withinsecular as opposed to religious neighborhoods.

To refine the ongoing debate regardingreligion and suicide further, Van Tubergen andcolleagues (2005) combine both individual- andaggregate-level information from all suicides inthe Netherlands from 1936–1973. They seek todifferentiate between the effects of communitynorms and support on suicide rates amongvarious religious denominations. Based on thefinding that regardless of religious affiliationor attendance, suicide rates decrease as theproportion of church members within a givenmunicipality increases, the authors concludethat community norms are a more salientfactor in the prevention of suicide. Theyalso note that as the Netherlands grew moresecular over time, the influence of municipalreligious composition had a weaker impact onsuicide rates. Unfortunately, this study suffersfrom a lack of basic individual-level statisticalcontrols, including age and sex, as well as a lackof key measures that could adequately capturecomplex social phenomena such as communitynorms and support.

Although not multilevel in the strictestsense, Norstrom’s (1995) study triangulatesecological and individual data from multiplesources to estimate the extent to which alcoholabuse increases suicide risk among Swedishmen. A heavy reliance on alcohol is cast notsimply as a psychological or health-relatedrisk factor for suicidal outcomes; rather, it isconceived of as an indicator of a serious break-down in the social system. Employing standardepidemiologic measures of attributable risk,Norstrom calculates the proportion of the


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


age-adjusted male suicide rate that can be tiedto alcohol abuse, divorce, and unemployment.Alcohol abuse and unemployment account forapproximately 37% and 10%, respectively, ofmale suicides, whereas, surprisingly, divorcedoes not appear to be a significant factor.

Although the incorporation of multileveldata and modeling strategies into the study ofsuicide stands to advance our understanding,we caution against viewing these developmentsas a panacea to the methodological challengesthat continue to confront researchers in thisarea. We return to this seemingly minor pointbelow.

The construction and (mis)use of officialsuicide statistics. As noted above, Douglas(1967), among others, has raised serious con-cerns about the reliability of official suicidestatistics and the degree of error they contain.Making a medico-legal determination that asudden death is, in fact, a suicide is generallydifficult, primarily because it requires strong ev-idence regarding the intent of the decedent atthe time of death. Absent a note or other com-pelling information, evidence of intent is oftenscant, opening up the possibility of misclassifi-cation into accidental or undetermined mortal-ity categories. Although the complexity of thedeath classification process was well known andwell documented by sociologists early on (e.g.,Atkinson 1978), two crucial questions persistfor sociologists. First, is there error in the sui-cide rates, and does it interfere with the abilityto determine the etiological factors underlyingsuicide? Second, what are the systematic mech-anisms that drive the medico-legal classificationprocess?

Twenty-five years ago, Pescosolido &Mendelsohn (1986) combined sociologicaltheory on discretionary subsystems and econo-metric modeling to examine the questionregarding error and etiology. For the specificcase of ecological and cross-sectional data in theUnited States, they found that error does existin rates, that the sources of error were not inline with predictions with many social construc-tionist hypotheses, and that the ability to detect

sociological correlates was not significantlyaffected by existing error (see also Sainsbury& Jenkins 1982). Despite these findings,sociologists continue to look for and documentclear-cut cases of the systematic misclassifi-cation of suicides (Whitt 2006). Sociologistshoping to settle the matter face the same prob-lem faced by medico-legal experts: The peopleyou really want to talk to are dead (Wray 2007).

Returning to Durkheim’s era, van Poppel& Day (1996) consider whether, after con-trolling for differences in patterns of Catholicand Protestant death classification, group dif-ferences in suicide rates remained. They arguethat group differences in suicide rates reportedin the early twentieth century were artifacts ofdifferential death recording. Using data froma Netherlands death registry from 1905–1910,which included each decedent’s religious affil-iation, van Poppel & Day observe that a muchhigher proportion of Catholics than Protes-tants died from “sudden,” “accidental,” or“unknown” causes of death, leading them toconclude that Catholic stigma surrounding sui-cide resulted in intentional misclassification(see Simpson’s 1998 critique; van Poppel &Day’s 1998 reply).

Timmermans (2005) takes up the secondpersistent question, the one regarding mecha-nisms of medico-legal classification. His ethno-graphic account of three years of fieldworkin a medical examiner’s office highlights howthe demands of professional authority facedby forensic death investigators cause them toadopt strict death investigation procedures thatmake family influence (pressure, for instance,to change a death provisionally ruled as suicideto an accident) nearly impossible. Following onPescosolido & Mendelsohn’s (1986) findingson lower suicide rates in areas with a medicalexaminer system, he also concludes that avariety of reasons lead medical examiners tounderreport suicides more frequently than docoroners.

Both methodological challenges and recentadvancements in suicide research suggest theneed to take a more integrated approach tothe study of this sociological phenomenon. It


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


is not enough simply to focus on the longitu-dinal demographic processes, such as APC ef-fects, that are likely to influence suicide rateswithout simultaneously considering how mul-tilevel conditions may impact this outcome ofinterest. Similarly, we should not be attempt-ing to estimate the effects of predictive factorsof suicide without better understanding howsuicide statistics are produced (Claassen et al.2010), including how work processes, organi-zational structures, and familial relations in-teract to socially construct mortality statistics(Timmermans 2006).

NEW DIRECTIONS IN THESOCIOLOGY OF SUICIDE

If sociology is going to remain relevant to theevolving multidisciplinary investigation of sui-cide as a social problem, the discipline needs topursue three avenues simultaneously. First, weneed to stand back and reconsider the macro-micro dilemma theoretically and methodolog-ically, including how to assemble a data setcomplex enough to provide rigorous empiri-cal examination; gain new insights into the so-cial and cultural mechanisms underlying suiciderisk; and use our strengths in qualitative andmultimethod approaches to push our under-standings further. Second, we need to take se-riously research findings on suicide from otherdisciplines, incorporating their insights into themultiple factors that affect suicide in individualsand societies. Third, we need to move forwardby demonstrating and evaluating the utility ofwhat we consider the robust conclusions of so-ciological research in pragmatic, real-world ef-forts to reduce suicide.

In what follows, we identify and discuss fourissues that stand as obstacles to forward move-ment along these avenues.1. Reconsidering the macro-micro insightsof sociology. Our review of recent suicideresearch in conjunction with a more com-prehensive historical overview of sociologicalcontributions reveals a trend toward studiesthat examine suicide as something that happensnot just within groups or societies, but also

to individuals who inhabit those groups andsocieties. By this we do not mean to suggestthat sociologists either adopt an overly indi-vidualistic approach to the study of suicideor abandon the study of aggregate measures;rather, this suggestion calls both for incorpo-rating our own insights on the individual leveland for pushing past the current insularity ofthe sociology of suicide to consider researchfrom other disciplines.

By the 1980s, sociologists such as Giddens,Coleman, and Stryker brought an explicit focuson the macro-micro link to the discipline’scentral research agenda (Pescosolido 1992).Furthermore, over the past three decades,scientific research and conceptual thinkinghave converged to suggest that suicide comesfrom a combination of genetic, developmental,environmental, physiological, social, andcultural factors operating through diverse,complex pathways (Goldsmith et al. 2002).These two avenues converge to suggest aneed for more complexity in sociologicaltheorizing on suicide. For example, the debateover sociology’s “One Law” ( Johnson 1965)about the role of religion in suicide can bereconsidered. Taking a simple example, ifareas with a higher share of Catholics havelower suicide rates, the next question becomes:Who in those areas commits suicide? Is it thenon-Catholics who are not members of a moreintegrative, and therefore protective, religiouscommunity? That has always been the implicitassumption. However, it may well be the casethat areas with high percentages of Catholicshave a more integrated set of social agenciesthat provide relief and support during timesof crisis (Pescosolido 1990). If this is the case,it may be that religious affiliation has no realimpact on the individual level given that theseagencies are open to all, not just to Catholics.In any case, there are many such parallels thatcan draw from past sociological research.

We have suitable analytic tools to in-corporate individual- and aggregate-levelconsiderations. However, other tools yet to beused in suicide research are promising and allowfor additional complexities with respect to how


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


individuals are embedded in societies. For ex-ample, individuals who live in areas with similarsocial profiles do so in different historical peri-ods and at different points in the life course. So,while the identification problem renders it im-possible to completely disentangle the effects ofAPC characteristics on suicide probabilities, arecent development in APC analysis may offer away to overcome this obstacle. The intrinsic es-timator approach appears to produce uniquelydetermined parameter estimates in both linearand nonlinear models (Yang et al. 2008, Yang2008). Given that much of the available infor-mation on complete suicides is obtained fromtime-series data composed primarily of aggre-gate mortality measures, disentangling whichpredictors of suicide are due to APC com-position could enable us to generate researchresults with utilitarian functions outside thewalls of academia—for example, results that areuseful to mental health practitioners or policymakers.

Yet another promising methodologicalapproach to incorporating multilevel pre-dictors and changes over time, as well asmodel interactions between the two, is com-plex systems analysis (CSA). Simply stated,CSA seeks to examine the dynamic interplayamong integrative systems without adheringto reductionist assumptions. Rather than tryto reduce complex social phenomena to themost parsimonious explanation, this analyticapproach seeks to obtain results that are simpleenough to further our understanding of theunderlying social processes of interest, yetintricate enough to retain sufficient contextualcontent to ensure that important interactions—both between levels and over time, as well asfeedback loops—are not overlooked. CSA mayenable future suicide researchers to capturemore accurately how individuals shape and areshaped by environmental forces that influencewhether or not they take their own lives.2. The intractable problem of macro-microdata on suicide? While the sociological imag-ination holds the potential to rethink the com-plexity of individual-aggregate influences, andwhile today’s sociologists have the analytic tools

to do this, ideas about how to construct suitabledata sets are in short supply. To be sure, high-quality, multilevel data regarding suicides arehard to come by, but a few studies discussedabove demonstrate that they do exist or canbe created for at least some outcomes relatedto suicide (e.g., suicidal ideation, attempts orbeliefs/attitudes) and for some countries out-side of the United States (Campbell & Troyer2007, Maimon & Kuhl 2008, Norstrom 1995,van Tubergen et al. 2005).

However, at least in the United States, an in-tegrated data set on completed suicide remainselusive. Four conditions stand as fundamentalbarriers to the scientific integration of etiolog-ical streams at micro and macro levels: (a) afundamentally different etiology exists betweencompleted suicide and attempts or ideation;(b) there are low base rates and high geograph-ical dispersion of completed suicide; (c) by thevery act, individuals who commit suicide are notavailable for research; and (d ) the comparisongroup is problematic. These problems go handin hand. Attempts and completions representdistinctly different social phenomena for whichdirect causes, mediating mechanisms, and mod-erating influences vary (Beck et al. 1985, Mariset al. 1992, Safer 1997). In the most standarddifference, women are more likely to attemptsuicide, whereas men are more likely to commitsuicide. Investigators who have explored mul-tilevel effects, even using large-scale, nation-ally representative, and contextualized data sets(e.g., Add Health, Bearman & Moody 2004),have focused on suicidal ideation or attempts—rather than completed acts—as their endpointsof interest.

Acts of completed suicide remain a fairlyrare event, occurring roughly once per 10,000people each year across the entire UnitedStates. With such wide geographical dispersion,any data set that uses a cluster sampling ap-proach, no matter how many cases are collectednationally, is likely to miss a large enough shareof completed suicides to make a multilevelinvestigation unlikely. Furthermore, even astudy as large as Add Health could nevertally a sufficient number of suicides, and so


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


adolescents who commit suicide are droppedfrom the sample, making the data missing ordependent on reports by others. Finally, theissue of what is the appropriate comparisongroup remains. Researchers who use mortalityfiles compare individuals who committedsuicide with those who died from other causes.Using national-level data that covers allgeographical regions (e.g., national mortalityfiles) fundamentally changes the comparison todying by suicide versus dying by other causes(Wray et al. 2008). Although the multileveldata problem is removed, the question ofmeaning remains. What does it mean to com-pare the correlates, whether at the individual orcontextual level, of individuals who died fromsuicide (an intentional act of self-destruction)with those who died from all other causes (noneof which is intentional)? Clinical researchershave used matched controls. Again, however,how would such controls who did not commitsuicide be selected? In essence, the appropriatecomparison group needed for macro-microresearch is all those who have not committedsuicide, even if the data are right truncated.

Sociologists, however, have demonstrated ahigh level of creativity in surmounting data dif-ficulties and in finding innovative ways to ag-gregate data at different levels to test uniquesocial hypotheses, offering insights into press-ing societal questions without abandoning thegoals of building the stock of basic sociologicalknowledge (e.g., see King & Bearman 2009, Liuet al. 2010a,b on the social factors that explainrising autism rates).3. Clarifying the nature and mechanismsof social and cultural influence. Here weare advocating a reconsideration of sociolog-ical theories of suicide outside the dominantDurkheimian perspective and the careful ap-plication of newer directions in sociologicalthought. For example, from a life course per-spective, there is no necessary theoretical an-tagonism between imitation theories and thosethat pose other mechanisms. The former mayprovide an understanding of the timing of sui-cides while the latter targets those underly-ing forces or events that set up a predisposi-

tion to suicide. Recent studies provide evidencefor suicide contagion, particularly among youth(Gould et al. 1990, 1989; Romer et al. 2006)and suggest that social networks are implicatedin a surprising number of different kinds of“contagion,” including suicidality (Christakis &Fowler 2009, Bearman & Moody 2004).

These findings, coupled with a consider-ation of contemporary theoretical insights,suggest that one useful approach would beto take Durkheimian theory back to firstprinciples and translate it through a socialnetwork perspective. In the 1990s, Pescosolidoand colleagues took steps in that direction,showing that findings on the religious influenceon suicide rates are consistent with the natureof coreligionists’ networks and the sets oforganizational supports in regional hubs indifferent U.S. denominations (Pescosolido &Georgianna 1989; Pescosolido 1990, 1994).In line with the work by other sociologistswho suggested the tight correspondence, if notidentity, between Durkheim’s groups and con-temporary network theory’s ties (e.g., Bearman1991), she suggested that a social networkelaboration of Durkheim’s theory allows forgreater specificity and cross-fertilization withcontemporary social theory. With Durkheim’s“societies” translated into the operation ofdifferent networks, solidarity comes from thepresence (or absence) of strong, interlockingsocial relationships.

As depicted in Figure 2, reimagingDurkheim’s theoretical scheme in networkterms produces a safety net that has distinctivetopographical features.

One dimension, running from left toright, represents integration. Another dimen-sion, running from back to front, representsregulation. Both dimensions run from highto low, and their interaction creates the fourtypes of suicide. When individuals live insocial structures characterized by too littleintegration or regulation, the threads of thesocial safety net are too far apart to catchthem when crises destabilize their equilibrium.Egoistic and anomic suicides are “diseasesof the infinite” because of the extreme gaps


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


AnomieAnomieAnomie

Level of regulationIn social networks

AltruismEgoism

Fatalism

Level of integration in social networks

Figure 2Networks and the Durkheimian Theory of Suicide.

in the societal safety net that normally sup-ports individuals during times of individualor community crisis. Conversely, the socialsafety net closes up when social structuresare overregulated or overintegrated. With noflexibility or give in the safety net, individualswho experience crises hit a wall that shattersrather than supports. It is in the center of thenet, where ties are moderately integrated andregulated that individuals can be safely caughtand restrained from their suicidal impulses.6

This approach that produces a theoreticalprediction plane for two aspects of networkstructures on which Durkheim focused (i.e.,integration and regulation) has been recentlyconsidered in light of concerns with multilevelcontexts (Pescosolido 2010). However, thisis just one example of how contemporarysociological theory can be brought to bear onrethinking and testing fundamental sociologicalideas on suicide. Although some have recentlyargued that the Durkheimian approach has

6With examples ranging from Jonestown (Coser & Coser1979) to Fischer’s (1982) communities in NorthernCalifornia, this approach undermines many of the early theo-retical criticisms of Durkheim’s theory (e.g., Johnson 1965).Yet this translation into network theory remains in the earlystages, making concerns about useful data and methodolog-ical creativity even more pressing.

been “falsified” (Nolan et al. 2010), our read isdifferent. The stockpile of sociological researchsuggests that, despite sometimes conflictingand weaker-than-expected findings, there issufficient theoretical power and empiricalsupport for Durkheim’s theory. This does notmean that all of the hypotheses that Durkheimconsidered are applicable in contemporary soci-ety; however, the framework of connectednessand well-being remains central to sociology’sbasic contributions and finds resonance incontemporary population health research (Hall& Lamont 2009). It is no longer realistic to takeDurkheim’s hypotheses from the nineteenthcentury and expect them to be reasonable basesfor empirical investigation in the twenty-first.However, it is also premature and unwise toclaim that the nature of social structure doesnot influence suicide rates.4. Bringing the full scope of sociolog-ical tools to bear. Our review clearlyindicates that past and present research onsuicide has been overwhelmingly dominated byquantitatively oriented approaches, with all butone (Timmermans 2005) of the 41 articles wereviewed fitting this description. Alternativemethodologies—qualitative or mixed methodsapproaches—are practically nonexistent. Al-though the Durkheimian legacy, the nature ofeasily accessible data, and funding priorities


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


may partially account for this pattern, theextent to which qualitative methods are un-derrepresented is surprising, especially giventhe long history of valuable, qualitatively basedcontributions in related areas of study, such asthe sociology of death and dying (see Fontana& Keene 2009 for an overview).

Contemporary ethnographic accounts ofthose who deal with suicide on a daily basis—the trauma surgeons, suicide hotline volunteers,physicians and workers in emergency psychi-atric wards, and first responders—would offerinsights to both theoretical and methodolog-ical innovation. Even as suicide attempts tendto have different correlates than completedsuicide, interviews with those who have madesuicide attempts (especially medically seriousattempts) or with friends and family memberswho experienced a loss to suicide could offermuch needed data. Content analyses of suicidenotes or of suicide-related social networkingopen up further possibilities for interpretiveaccounts. Recent efforts by Scourfield andcolleagues (2010) to develop a “sociologicalautopsy” that complements the psychologicalautopsy draw from both quantitative andqualitative methods within a single analyticalframe.

In the more traditional vein of research onthe stages in the classification of suspiciousdeaths, information on how various medico-legal systems differentially code suicides and onwhether such differences generate random orsystematic sources of bias remains elusive anda good candidate for additional qualitative andmixed method investigation. Similarly, natural(or quasi-natural) experiments hold promise forevaluating and potentially validating official sui-cide statistics [e.g., see demography’s exampleof examining racial disparities in birth outcomes(Lauderdale 2006)].

We have a unique opportunity to take theWeberian and interactionist approach advo-cated by Douglas (1967) and more thoroughlyinvestigate the social meaning(s) of suicide, thecontexts from which these meanings are derivedas well as the patterns they follow, and howthese patterns vary according to existing fault

lines of stratification. In fact, Durkheim’s fourtypes of suicide suggest research into individualmeaning. For example, what are the differentmeanings of suicide for those who contemplateand eventually attempt it? Historians offer ex-cellent examples of this kind of work, situatingdetailed, individual-level case studies withinsociological frameworks (Anderson 1987,MacDonald & Murphy 1991, Weaver 2009).

CONCLUSION

In the United States, suicide accounts for morethan 10% of total deaths among individualsbetween the ages of 15 and 34. Moreover,for White men within the same age range,suicide ranks as the second leading cause ofdeath (Heron 2010). Sociologists are longaccustomed to believing that suicide ratesreflect underlying levels of social integrationand cohesion. We are probably right aboutthat. The durability of the Durkheimiantheory is truly impressive. Carefully crafted,problem-oriented sociological research canprovide strategies for change that can con-tribute to the sense of collective efficacy thatcommunities need to solve their most pressingsocial problems, of which suicide is often one.

Our brief history of sociological researchsuggests that two different and often separategoals underlie research on suicide: namely,a manifest concern with the developmentand testing of sociological theory and a morelatent concern for the amelioration of socialproblems. Although Burawoy’s (2004) notionof public sociology has argued that theseare not fundamentally at odds, in practicethey have often led to different communitiesof researchers in sociology, communitiesin which interaction is quite limited, if notabsent. As has been argued elsewhere, boththe integration of insights from the disci-pline’s mainstream and from outside thediscipline can serve to advance progress insociology’s contributions (Pescosolido &Kronenfeld 1995, Pescosolido 2006). Further-more, translating scientific innovations thathave occurred in other disciplines into the


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


language and practice of sociology targets thefocus on complexity that is increasingly part ofthe larger agenda of the sciences ( Jasny et al.2009). Adopting a multidisciplinary agenda aswell as an open-source model of informationand idea sharing will make sociological contri-butions more visible and amenable to use andreuse by other suicide researchers, regardless ofdiscipline. Taking a public sociology approachto suicide, treating it as a significant socialproblem in and of itself (not merely a symptomof larger social pathologies), aims strategicallyat intervening in public discourse and policydebates around suicide. No article we reviewedin the top-tier journals takes this aim asprimary; few even mention it as secondary. Itis time to take that more latent concern andmake it more manifest.

Basic insights from sociologists have beenused to reduce catastrophic mistakes and errorsin complex organizations and professions andto understand disaster-related deaths (Bosk1979, Klinenberg 2002, Vaughan 1996), butwe have yet to offer many direct insights intohow our theories and methods might be usedto power suicide prevention research. This is amost opportune time to move forward in thisdirection given the Centers for Disease Con-trol’s (2008) major policy statement on the roleof “connectedness” in suicide prevention andtheir call for intervention proposals along thisline. Furthermore, the massive groundswell of

activism advocating community-based preven-tion programs is reflected in the acceleratedexpansion of the dozens of community- andcampus-based chapters of national suicideprevention organizations such as the AmericanFederation of Suicide Prevention, the JasonFoundation, and the Jed Foundation; thefederal funding priorities of the Garrett LeeSmith Memorial Act; and high-profile effortsby the Veterans’ Administration to reducesuicide among soldiers returning from warsin Iraq and Afghanistan (U.S. Army 2010).The challenges are significant, and the answersare far from obvious. However, sociologicalunderstandings of suicide are not only relevantbut are essential to these prevention efforts.

The new directions for research we haveoutlined require that we simultaneously standback and reassess Durkheim and the compet-ing theories of his time and move forward alongnew avenues of inquiry and collaboration to em-brace the complexity of suicide. Choosing notto pursue these new directions means contin-uing on the well-worn path that sociology hasfollowed, with few deviations, since the close ofthe nineteenth century. In that event, we fearnot only that future sociologists will becomeless interested in pursuing answers to the rid-dle of suicide, but also that the role for soci-ology in the multidisciplinary study of suicidewill eventually die, as it were, a self-inflicteddeath.

DISCLOSURE STATEMENT

The authors are not aware of any affiliations, memberships, funding, or financial holdings thatmight be perceived as affecting the objectivity of this review.

LITERATURE CITED

Agerbo E, Stack S, Peterson L. 2009. Social integration and suicide trends in Denmark, 1906–2006. Presented atAnnu. Meet. Am. Assoc. Suicidol., 42nd, San Francisco

Almgren G, Guest A, Immerwahr G, Spittel M. 1998. Joblessness, family disruption, and violent death inChicago, 1970–90. Soc. Forces 76:1465–93

Anderson O. 1987. Suicide in Victorian and Edwardian England. New York: Oxford Univ. PressArango V, Mann JJ. 1992. Relevance of serotonergic postmortem studies to suicidal behavior. Int. Rev. Psy-

chiatr. 4:131Atkinson JM. 1978. Discovering Suicide: Studies in the Social Organization of Sudden Death. London: MacMillan


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Baller RD, Richardson KK. 2002. Social integration, imitation, and the geographic patterning of suicide. Am.Sociol. Rev. 67:873–88

Baron JN, Reiss PC. 1985. Same time, next year: aggregate analyses of the mass media and violent behavior.Am. Sociol. Rev. 50:347–63

Baudelot C, Establet R. 2008 [2006]. Suicide: The Hidden Side of Modernity. Boston: PolityBearman PS. 1991. The social structure of suicide. Sociol. Forum 6:501–24Bearman PS, Moody J. 2004. Suicide and friendships among American adolescents. Am. J. Public Health

94(N1):89–96Beck A, Steer R, Kovacs M, Garrison B. 1985. Hopelessness and eventual suicide: a 10-year prospective study

of patients hospitalized with suicidal ideation. Am. J. Psychiatr. 142:559–63Berkman LF, Glass T, Brissette I, Seeman TE. 2000. From social integration to health: Durkheim in the new

millennium. Soc. Sci. Med. 51:843–57Bielby DDV, Bielby WT. 1984. Work commitment, sex-role attitudes, and women’s employment. Am. Sociol.

Rev. 49:234–47Bosk CL. 1979. Forgive and Remember: Managing Medical Failure. Chicago: Univ. Chicago PressBranas CC, Nance ML, Elliott MR, Richmond TS, Schwab CW. 2004. Urban-rural shifts in intentional

firearm death: different causes, same results. Am. J. Public Health 94:1750–55Breault KD. 1986. Suicide in America: a test of Durkheim’s theory of religious and family integration, 1933–

1980. Am. J. Sociol. 92:628–56Brunner J, Stalla GK, Stalla J, Uhr M, Grabner A, et al. 2001. Decreased corticotropin-releasing hormone

(CRH) concentrations in the cerebrospinal fluid of eucortisolemic suicide attempters. J. Psychiatr. Res.35:1–9

Burawoy M. 2004. The critical turn to public sociology. In Enriching the Sociological Imagination: How RadicalSociology Changed the Discipline, ed. R Levine, pp. 309–21. Boston: Brill Acad.

Burr JA, Hartman JT, Matteson DW. 1999. Black suicide in U.S. metropolitan areas: an examination of theracial inequality and social integration-regulation hypotheses. Soc. Forces 77:1049–80

Campbell ME, Troyer L. 2007. The implications of racial misclassification by observers. Am. Sociol. Rev.72:750–65

Cavan R. 1928. Suicide. Chicago: Univ. Chicago PressCavanagh J, Carson A, Sharpe M, Lawrie S. 2003. Psychological autopsy studies of suicide: a systematic review.

Psychol. Med. 33:395–405Centers for Disease Control and Prevention. 2008. Promoting Individual, Family, and Community Connectedness

to Prevent Suicidal Behavior. Atlanta, GA: Cent. Dis. Control Prev.Christakis NA, Fowler JH. 2009. Connected: The Surprising Power of Our Social Networks and How They Shape

Our Lives. New York: Little, BrownClaassen CA, Yip PS, Corcoran P, Bossarte RM, Lawrence BA, Currier GW. 2010. National suicide rates a

century after Durkheim: Do we know enough to estimate error? Suicide Life-Threat. Behav. 40:193–223Coleman JS. 1986. Social theory, social research, and a theory of action. Am. J. Sociol. 91:1309–35Conwell Y, Duberstein P, Cox C, Herrmann J, Forbes N, Caine E. 1996. Relationships of age and Axis I

diagnoses in victims of completed suicide: a psychological autopsy study. Am. J. Psychiatr. 153:1001–8Coser RL, Coser LA. 1979. Jonestown as perverse utopia. Dissent 26:158–263Douglas JD. 1967. The Social Meanings of Suicide. Princeton, NJ: Princeton Univ. PressDuberstein PR, Conwell Y, Conner KR, Eberly S, Evinger JS, Caine ED. 2004. Poor social integration and

suicide: fact or artifact? A case-control study. Psychol. Med. 34:1331–37Durkheim E. 2006 [1897]. On Suicide. London: Penguin ClassicsEasterlin RA. 1987. Birth and Fortune: The Impact of Numbers on Personal Welfare. Chicago: Univ. Chicago

PressEccles JS. 1987. Gender-roles and women’s achievement. Psychol. Women Q. 11:135–72Elder GH Jr. 1994. Time, human agency, and social change: perspectives on the life course. Soc. Psychol. Q.

57:4–15Ellison CG, Burr JA, McCall PL. 1997. Religious homogeneity and metropolitan suicide rates. Soc. Forces

76:273–99


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Fedden H. 1938. Suicide: A Social and Historical Study. London: P DaviesFischer CS. 1982. To Dwell Among Friends. Berkeley: Univ. Calif. PressFontana A, Keene JR. 2009. Death and Dying in America. Cambridge, UK: PolityGarfinkel H. 1967. Studies in Ethnomethodology. Englewood Cliffs, NJ: Prentice-HallGerson K. 1986. Hard Choices: How Women Decide About Work, Career, and Motherhood. Berkeley: Univ. Calif.

PressGibbs JP. 2000. Status integration and suicide: occupational, marital, or both? Soc. Forces 78:949–70Gibbs JP, Martin WT. 1964. Status Integration and Suicide: A Sociological Study. Eugene: Univ. Or. PressGibbs JP, Martin WT. 1974. A problem in testing the theory of status integration. Soc. Forces 53:332–39Giddens A. 1965. The suicide problem in French sociology. Br. J. Sociol. 16:3–18Girard C. 1993. Age, gender, and suicide: a cross-national analysis. Am. Sociol. Rev. 58:553–74Goffman E. 1969. The insanity of place. Psychiatry 32:357–88Goldsmith S, Pellmar T, Kleinman A, Bunney W. 2002. Reducing Suicide: A National Imperative. Washington,

DC: Natl. Acad. PressGold M. 1958. Suicide, homicide, and the socialization of aggression. Am. J. Sociol. 63:651–61Gould MS, Wallenstein S, Davidson L. 1989. Suicide clusters: a critical review. Suicide Life-Threat. Behav.

19:17–29Gould MS, Wallenstein S, Kleinman M. 1990. Time-space clustering of teenage suicide. Am. J. Epidemiol.

131:71–78Halbwachs M. 1978 [1930]. Les Causes du Suicide. London: RoutledgeHall PA, Lamont M. 2009. Successful Societies: How Institutions and Culture Affect Health. New York: Cambridge

Univ. PressHedstrom P, Liu K, Nordvik MK. 2008. Interaction domains and suicide: a population-based panel study of

suicides in Stockholm, 1991–1999. Soc. Forces 87:713–40Henry AF, Short JF. 1954. Suicide and Homicide: Some Economic, Sociological and Psychological Aspects of Aggression.

New York: Free PressHeron M. 2010. Deaths: leading causes for 2006. Natl. Vital Stat. Rep. Vol. 58(14), U.S. Dep. Health Hum.

Serv., Bethesda, MD. http://www.cdc.gov/nchs/data/nvsr/nvsr58/nvsr58_14.pdfHuff-Corzine L, Corzine J, Moore DC. 1991. Deadly connections: culture, poverty, and the direction of lethal

violence. Soc. Forces 69:715–32House JS, Landis KR, Umberson D. 1988. Social relationships and health. Science 241(4865):540–45Jamison KR. 1999. Night Falls Fast: Understanding Suicide. New York: KnopfJasny BR, Zahn LM, Marshall E. 2009. Introduction: connections. Science 325:405Johnson BD. 1965. Durkheim’s one cause of suicide. Am. Sociol. Rev. 30:875–86Kawachi I, Kennedy BP. 1997. Health and social cohesion: why care about income inequality? BMJ 314:1037–

40King M, Bearman P. 2009. Diagnostic change and the increased prevalence of autism. Int. J. Epidemiol.

38:1224–34Klinenberg E. 2002. Heat Wave: A Social Autopsy of Disaster in Chicago. Chicago: Univ. Chicago PressKnox KL, Conwell Y, Caine ED. 2004. If suicide is a public health problem, what are we doing to prevent it?

Am. J. Public Health 94:37–45Kposowa AJ, Breault KD, Singh GK. 1995. White male suicide in the United States: a multivariate individual-

level analysis. Soc. Forces 74:315–25Krull C, Trovato F. 1994. The Quiet Revolution and the sex differential in Quebec’s suicide rates: 1931–1986.

Soc. Forces 72:1121–47Kubrin CE, Wadsworth TP, DiPietro S. 2006. Deindustrialization, disadvantage and suicide among young

Black males. Soc. Forces 84:1559–79Lauderdale DS. 2006. Birth outcomes for Arabic-named women in California before and after September 11.

Demography 43:185–201Lehmann JM. 1995. Durkheim’s theories of deviance and suicide: a feminist reconsideration. Am. J. Sociol.

100:904–30Liu K, King M, Bearman PS. 2010a. Social influence and the autism epidemic. Am. J. Sociol. 115:1387–434


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.

http://www.cdc.gov/nchs/data/nvsr/nvsr58/nvsr58_14.pdf


Liu K, Zerubavel N, Bearman P. 2010b. Social demographic change and autism. Demography 47:327–43MacDonald M, Murphy TR. 1991. Sleepless Souls: Suicide in Early Modern England. New York: Oxford Univ.

PressMaimon D, Kuhl DC. 2008. Social control and youth suicidality: situating Durkheim’s ideas in a multilevel

framework. Am. Sociol. Rev. 73:921–43Maltsberger JT, Goldblatt MJ. 1996. Essential Papers on Suicide. New York: NY Univ. PressMann JJ, Apter A, Bertolote J, Beautrais A, Currier D, et al. 2005. Suicide prevention strategies: a systematic

review. JAMA 294:2064–74Maris RW. 1981. Pathways to Suicide: A Survey of Self-Destructive Behaviors. Baltimore, MD: Johns Hopkins

Univ. PressMaris RW, Berman AL, Maltsberger JT, Yufit RI. 1992. Assessment and Prediction of Suicide. New York:

GuilfordMasaryk TG. 1970 [1881]. Suicide and the Meaning of Civilization. Chicago: Univ. Chicago PressMaume DJ. 2006. Gender differences in restricting work efforts because of family responsibilities. J. Marriage

Fam. 68:859–69Menninger KA. 1938. Man Against Himself. New York: Harcourt, BraceMiller M, Hemenway D. 2008. Guns and suicide in the United States. New Engl. J. Med. 359:989–91Minois G. 1999. History of Suicide: Voluntary Death in Western Culture. Baltimore, MD: Johns Hopkins Univ.

PressMorselli EA. 1882. Suicide: An Essay on Comparative Moral Statistics. New York: AppletonNolan P, Triplett J, McDonaugh S. 2010. Sociology’s Suicide: a forensic autopsy? Am. Sociol. 41:292–305Norstrom T. 1995. The impact of alcohol, divorce, and unemployment on suicide: a multilevel analysis. Soc.

Forces 74:293–314Pampel FC. 1998. National context, social change, and sex differences in suicide rates. Am. Sociol. Rev. 63:744–

58Pampel FC, Williamson JB. 2001. Age patterns of suicide and homicide mortality rates in high-income nations.

Soc. Forces 80:251–82Pescosolido BA. 1990. The social context of religious integration and suicide: pursuing the network explana-

tion. Sociol. Q. 31:337–57Pescosolido BA. 1992. Beyond rational choice: the social dynamics of how people seek help. Am. J. Sociol.

97:1096–138Pescosolido BA. 1994. Bringing Durkheim into the 21st century: a network approach to unresolved issues in

the sociology of suicide. In Le Suicide 100 Years Later, ed. D Lester, pp. 264–95. Philadelphia: CharlesPescosolido BA. 2006. Of pride and prejudice: the role of sociology and social networks in integrating the

health sciences. J. Health Soc. Behav. 47:189–208Pescosolido BA. 2011. Organizing the sociological landscape for the next decades of health and health care

research: the Network Episode Model III-R as cartographic subfield guide. In Handbook of the Sociology ofHealth, Illness, and Healing: A Blueprint for the 21st Century, ed. BA Pescosolido, JK Martin, JD McLeod,A Rogers, pp. 39–66. New York: Springer

Pescosolido BA, Georgianna S. 1989. Durkheim, suicide, and religion: toward a network theory of suicide.Am. Sociol. Rev. 54:33–48

Pescosolido BA, Kronenfeld JJ. 1995. Health, illness, and healing in an uncertain era: challenges from and formedical sociology. J. Health Soc. Behav. 35:5–33

Pescosolido BA, Mendelsohn R. 1986. Social causation or social construction of suicide? An investigation intothe social organization of official rates. Am. Sociol. Rev. 51:80–100

Phillips DP. 1974. The influence of suggestion on suicide: substantive and theoretical implications of theWerther effect. Am. Sociol. Rev. 39:340–54

Phillips DP. 1979. Suicide, motor vehicle fatalities, and the mass media: evidence toward a theory of suggestion.Am. J. Sociol. 84:1150–74

Phillips DP, Carstensen LL. 1986. Clustering of teenage suicides after television news stories about suicide.New Engl. J. Med. 315:685–89

Phillips DP, Paight DJ. 1987. The impact of televised movies about suicide: a replicative study. New Engl. J.Med. 317:809–11


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Pickering WSF, Walford G. 2000. Durkheim’s Suicide: A Century of Research And Debate. New York: Taylor &Francis

Pope W. 1976. Durkheim’s Suicide: A Classic Analyzed. Chicago: Univ. Chicago PressPorterfield AL. 1949. Indices of suicide and homicide by states and cities: some Southern-non-Southern

contrasts with implications for research. Am. Sociol. Rev. 14:481–90Pridemore WA. 2006. Heavy drinking and suicide in Russia. Soc. Forces 85:413–30Quetelet AJ. 1842. A Treatise on Man and the Development of His Faculties. Edinburgh, UK: ChambersRehkopf DH, Buka SL. 2006. The association between suicide and the socio-economic characteristics of

geographical areas: a systematic review. Psychol. Med. 36:145–57Robinson WS. 1950. Ecological correlations and the behavior of individuals. Am. Sociol. Rev. 15:351–57Romer D, Jamieson PE, Jamieson KH. 2006. Are news reports of suicide contagious? A stringent test in six

U.S. cities. J. Commun. 56:253–70Roy A, Segal NL, Centerwall BS, Robinette CD. 1991. Suicide in twins. Arch. Gen. Psychiatr. 48:29–32Sacks H. 1995. Lectures on Conversation. New York: Wiley-BlackwellSafer DJ. 1997. Adolescent/adult differences in suicidal behavior and outcome. Ann. Clin. Psychiatr. 9:61–66Sainsbury P. 1955. Suicide in London: An Ecological Study. Maudsley Monogr. No. 1. London: Chapman &

HallSainsbury P, Jenkins JS. 1982. The accuracy of officially reported suicide statistics for purposes of epidemio-

logical research. J. Epidemiol. Community Health 36:43–48Schmid C. 1928. Suicide in Seattle: 1914–1925. Seattle: Univ. Wash. PressScourfield J, Fincham B, Langer S, Shiner M. 2010. Sociological autopsy: an integrated approach to the study

of suicide in men. Soc. Sci. Med. doi: 10.1016/j.socscimed.2010.01.054. In pressSennett R. 2006. Introduction. In On Suicide, by E Durkheim, pp. xi–xxiv. London: PenguinSimpson M. 1998. Suicide and religion: Did Durkheim commit the ecological fallacy or did van Poppel and

Day combine apples and oranges? Am. Sociol. Rev. 63:895–96Smelser NJ, Warner RS. 1976. Sociological Theory: Historical and Formal. Morristown, NJ: General LearningStack S. 1983. The effect of religious commitment on suicide: a cross-national analysis. J. Health Soc. Behav.

24:362–74Stack S. 1994. 100 Years Later: how Durkheim would reformulate his theory. In Centennial of Durkheim’s Le

Suicide, ed. D Lester, pp. 237–49. Philadelphia: The Charles PressStack S, Gundlach J. 1992. The effect of country music on suicide. Soc. Forces 71:211–18Stafford MC, Gibbs JP. 1985. A major problem with the theory of status integration and suicide. Soc. Forces

63:643–60Statham DJ, Heath AC, Madden PA, Bucholz KK, Bierut L, et al. 1998. Suicidal behaviour: an epidemiological

and genetic study. Psychol. Med. 28:839–55Stockard J, O’Brien RM. 2002a. Cohort effects on suicide rates: international variations. Am. Sociol. Rev.

67:854–72Stockard J, O’Brien RM. 2002b. Cohort variations and changes in age-specific suicide rates over time: ex-

plaining variations in youth suicide. Soc. Forces 81:605–42Tarde GD. 1903 [1895]. The Laws of Imitation. New York: H. HoltThorlindsson T, Bjarnason T. 1998. Modeling Durkheim on the micro level: a study of youth suicidality. Am.

Sociol. Rev. 63:94–110Timmermans S. 2005. Suicide determination and the professional authority of medical examiners. Am. Sociol.

Rev. 70:311–33Timmermans S. 2006. Postmortem: How Medical Examiners Explain Suspicious Deaths. Chicago: Univ. Chicago

PressTrout DL. 1980. The role of social isolation in suicide. Suicide Life-Threat. Behav. 10:10–23Trovato F. 1998. The Stanley Cup of hockey and suicide in Quebec, 1951–1992. Soc. Forces 77:105–26U.S. Army. 2010. Health Promotion, Risk Reduction and Suicide Prevention ReportU.S. Public Health Service. 1999. The Surgeon General’s Call to Action to Prevent Suicide. Washington, DC:

DHHSvan Poppel F, Day LH. 1996. A test of Durkheim’s theory of suicide—without committing the “ecological

fallacy.” Am. Sociol. Rev. 61:500–7


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


van Poppel F, Day LH. 1998. Reply to Simpson. Am. Sociol. Rev. 63:896–99van Tubergen F, te Grotenhuis M, Ultee W. 2005. Denomination, religious context, and suicide: neo-

Durkheimian multilevel explanations tested with individual and contextual data. Am. J. Sociol. 111:797–823Vaughan D. 1996. The Challenger Launch Decision: Risky Technology, Culture, and Deviance at NASA. Chicago:

Univ. Chicago PressWadsworth T, Kubrin CE. 2007. Hispanic suicide in U.S. metropolitan areas: examining the effects of immi-

gration, assimilation, affluence, and disadvantage. Am. J. Sociol. 112:1848–85Weaver JC. 2009. A Sadly Troubled History: The Meanings of Suicide in the Modern Age. Montreal, Quebec:

McGill-Queen’s Univ. PressWender PH, Kety SS, Rosenthal D, Schulsinger F, Ortmann J, Lunde I. 1986. Psychiatric disorders in the

biological and adoptive families of adopted individuals with affective disorders. Arch. Gen. Psychiatr.43:923–29

Whitt HP. 2006. Where did the bodies go? The social construction of suicide data, New York City, 1976–1992.Sociol. Inq. 76:166–87

Wray M. 2007. Interviewing the dead: suicide and the limits of the psychological autopsy. Presented at Annu. Meet.Soc. Study Soc. Probl., New York

Wray M, Miller M, Gurvey JE, Carroll J, Kawachi I. 2008. Leaving Las Vegas: exposure to Las Vegas andrisk of suicide. Soc. Sci. Med. 67:1882–88

Yang Y. 2008. Social inequalities in happiness in the United States, 1972 to 2004: an age-period-cohortanalysis. Am. Sociol. Rev. 73:204–26

Yang Y, Schulhofer-Wohl S, Fu WJ, Land KC. 2008. The intrinsic estimator for age-period-cohort analysis:What it is and how to use it. Am. J. Sociol. 113:1697–736

Zilboorg G. 1936. Suicide among civilized and primitive races. Am. J. Psychiatr. 92:1347–69


Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.

SO37-Frontmatter ARI 11 June 2011 11:38

Annual Reviewof Sociology

Volume 37, 2011

Contents

Prefatory Chapters

Reflections on a Sociological Career that Integrates Social Sciencewith Social PolicyWilliam Julius Wilson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 1

Emotional Life on the Market FrontierArlie Hochschild � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �21

Theory and Methods

Foucault and SociologyMichael Power � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �35

How to Conduct a Mixed Methods Study: Recent Trends in a RapidlyGrowing LiteratureMario Luis Small � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �57

Social Theory and Public OpinionAndrew J. Perrin and Katherine McFarland � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � �87

The Sociology of StorytellingFrancesca Polletta, Pang Ching Bobby Chen, Beth Gharrity Gardner,

and Alice Motes � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 109

Statistical Models for Social NetworksTom A.B. Snijders � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 131

The Neo-Marxist Legacy in American SociologyJeff Manza and Michael A. McCarthy � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 155

Social Processes

Societal Reactions to DevianceRyken Grattet � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 185

v

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Formal Organizations

U.S. Health-Care Organizations: Complexity, Turbulence,and Multilevel ChangeMary L. Fennell and Crystal M. Adams � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 205

Political and Economic Sociology

Political Economy of the EnvironmentThomas K. Rudel, J. Timmons Roberts, and JoAnn Carmin � � � � � � � � � � � � � � � � � � � � � � � � � � � � 221

The Sociology of FinanceBruce G. Carruthers and Jeong-Chul Kim � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 239

Political Repression: Iron Fists, Velvet Gloves, and Diffuse ControlJennifer Earl � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 261

Emotions and Social Movements: Twenty Years of Theoryand ResearchJames M. Jasper � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 285

Employment Stability in the U.S. Labor Market:Rhetoric versus RealityMatissa Hollister � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 305

The Contemporary American Conservative MovementNeil Gross, Thomas Medvetz, and Rupert Russell � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 325

Differentiation and Stratification

A World of Difference: International Trends in Women’sEconomic StatusMaria Charles � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 355

The Evolution of the New Black Middle ClassBart Landry and Kris Marsh � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 373

The Integration Imperative: The Children of Low-Status Immigrantsin the Schools of Wealthy SocietiesRichard Alba, Jennifer Sloan, and Jessica Sperling � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 395

Gender in the Middle East: Islam, State, AgencyMounira M. Charrad � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 417

Individual and Society

Research on Adolescence in the Twenty-First CenturyRobert Crosnoe and Monica Kirkpatrick Johnson � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 439

vi Contents

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.


Diversity, Social Capital, and CohesionAlejandro Portes and Erik Vickstrom � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 461

Transition to Adulthood in EuropeMarlis C. Buchmann and Irene Kriesi � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 481

The Sociology of SuicideMatt Wray, Cynthia Colen, and Bernice Pescosolido � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 505

Demography

What We Know About Unauthorized MigrationKatharine M. Donato and Amada Armenta � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 529

Relations Between the Generations in Immigrant FamiliesNancy Foner and Joanna Dreby � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 545

Urban and Rural Community Sociology

Rural America in an Urban Society: Changing Spatialand Social BoundariesDaniel T. Lichter and David L. Brown � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 565

Policy

Family Changes and Public Policies in Latin America [Translation]Brıgida Garcıa and Orlandina de Oliveira � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 593

Cambios Familiares y Polıticas Publicas en America Latina [Original,available online at http://arjournals.annualreviews.org/doi/abs/10.1146/annurev-soc-033111-130034]Brıgida Garcıa and Orlandina de Oliveira � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 613

Indexes

Cumulative Index of Contributing Authors, Volumes 28–37 � � � � � � � � � � � � � � � � � � � � � � � � � � � 635

Cumulative Index of Chapter Titles, Volumes 28–37 � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � � 639

Errata

An online log of corrections to Annual Review of Sociology articles may be found athttp://soc.annualreviews.org/errata.shtml

Contents vii

Ann

u. R

ev. S

ocio

l. 20

11.3

7:50

5-52

8. D

ownl

oade

d fr

om w

ww

.ann

ualr

evie

ws.

org

by U

nive

rsity

of

Tex

as -

Aus

tin o

n 10

/06/

11. F

or p

erso

nal u

se o

nly.

the sociology of suicide · 2018-07-31 · (fatalism) toa now infamous footnote wherehesuggeststhat...

Documents