CIVIL WAR, SOCIAL INTEGRATION AND :NTAL

HEALTH IN CROATIA

Robert M . Kunovich and Randy Hodso nOhio State University

The National Council for Eurasian and East European Researc h910 17 th Street, N.W.

Suite 300Washington, D .C. 20006

TITLE VIII PROGRAM

Project Information *

Contractor :

Ohio State University

Principal Investigator :

Randy Hodson

Council Contract Number :

814-21

Date :

October 27, 1999

Copyright Information

Individual researchers retain the copyright on their work products derived from research funde dthrough a contract or grant from the National Council for Eurasian and East European Researc h(NCEEER). However, the NCEEER and the United States Government have the right to duplicat eand disseminate, in written and electronic form, reports submitted to NCEEER to fulfill Contract o rGrant Agreements either (a) for NCEEER's own internal use, or (b) for use by the United StatesGovernment, and as follows : (1) for further dissemination to domestic, international, and foreigngovernments, entities and/or individuals to serve official United States Government purposes or (2)for dissemination in accordance with the Freedom of Information Act or other law or policy of th eUnited States Government granting the public access to documents held by the United State sGovernment . Neither NCEEER nor the United States Government nor any recipient of thi sReport may use it for commercial sale .

The work leading to this report was supported in part by contract or grant funds provided by th eNational Council for Eurasian and East European Research, funds which were made available b y

the U.S. Department of State under Title VIII (The Soviet-East European Research and Trainin gAct of 1983, as amended) . The analysis and interpretations contained herein are those of the author .

Executive Summary

Research has shown that social relationships are generally beneficial for mental health (Thoit s

1995) . However, few scholars have examined this association after the occurrence of a significant shoc k

to the social system as a whole . The purpose of this article is to examine the relationship between socia l

integration and war-related distress in Croatia immediately following the recent civil war . Does social

integration decrease war-related distress? Does social integration buffer the effect of traumatic events o n

war-related distress? We analyze these questions using nationally representative survey data collected i n

Croatia in 1996 . Results suggest that social integration has both positive and negative direct effects o n

distress . Being a member of informal organizations, such as sports clubs, and participating in socia l

activities are beneficial for mental health . On the other hand, being a member of some formal

organizations, such as church organizations and unions, is detrimental to mental health . There is little

support for the idea that social integration buffers the effect of traumatic events on distress . Only one o f

thirty-six possible interactions is significant and supports the buffer hypothesis . Frequent participation i n

social activities buffers the effect of experiencing violence on war-related distress . Also, some forms of

social integration appear to aggravate the effect of traumatic events on war-related distress. In sum, socia l

integration does affect war-related distress after a system shock, but in complex and sometime s

unexpected ways .

Does social integration affect psychological well-being? Scholars have been interested in thi s

question for decades . Research suggests that social integration has a direct and positive effect on mental

health (Cohen and Wills 1985 ; Thoits 1995) . More specifically, people with more social ties and bette r

quality social ties are healthier psychologically than those with fewer social ties and worse quality socia l

ties . There are few analyses . however, that focus on the relationship between social integration and

mental health where there has been a significant shock to the social system . The purpose of this article i s

to examine the relationship between social integration and mental health in Croatia following the recen t

war of national separation . Does social integration have a direct effect in reducing war-related distress i n

Croatia immediately following the war? Does social integration buffer the effect of traumatic events o n

war-related distress?

One of the goals of social research is to develop parsimonious theories that apply to a variety o f

cases . Part of this process involves examining phenomena in a variety of social settings in the hope tha t

differences will refute prior hypotheses and lead to the improvement of theory . This article is an attempt

to extend sociological knowledge on social relationships, mental health, and collective trauma . Croatia is

a useful case for this purpose because of the recent civil war . The war in Croatia, lasting from 1991 t o

1995, resulted in the death of over ten thousand people, created enormous refugee flows, and led to th e

massive destruction of property and state infrastructure . In such a setting, what role – if any – does socia l

integration play in psychological well-being? In other words, are social relationships capable o f

improving mental health when the entire fabric of a society has been significantly frayed ?

In the first section of this article, we define social integration and war-related distress . We then

briefly outline the war in Croatia and present statistics illustrating the frequency of traumatic events, th e

This article is forthcoming, in Journal of Health and Social Behavior . We wish to thank the International Researc hand Exchanges Board (IREX) for providing support to collect the survey data for this article and the Center of theInvestigation of Transition and Civil Society for assisting with data collection. We also wish to thank Mersho nCenter at The Ohio State University and the National Council for Eurasian and East European Research forproviding additional support for analyzing the data . Finally, we wish to thank Catherine Ross, Angelo A . Alonzo ,Sheri Kunovich, and Kazimierz M . Slomczynski for their helpful advice . Please direct all correspondence to RobertM. Kunovich, 300 Bricker Hall, 190 North Oval Mall, Columbus, OH 43210-1353 . Telephone : (614) 292-6681 .FAX: (614) 292-6687 . Internet : kunovich .2@osu .edu .

1

scope of negative reactions to them, and changes in social integration from 1989 to 1996 . Next, we

develop a theoretical model of war-related distress in Croatia focusing on the roles of social integration ,

traumatic events, stratification position, and background variables . We test the theoretical model with

survey data from Croatia and conclude with a discussion of the complex and sometimes unexpected way s

that social integration affects war-related distress.

Social Integration

Most scholars make a distinction between the structures and the functions of social suppor t

(Barrera 1986; Cohen and Wills 1985 ; Haines and Hurlbert 1992 ; House and Kahn 1985 ; House,

Umberson and Landis 1988 ; Thoits 1995) . Structures of social support refer to the existence, number .

and characteristics of social relationships . For example, scholars are typically interested in the presence

of social ties, such as being married, having siblings, and the number of close friends, as well a s

participation in organizations and the properties of one's social networks (Barrera 1986) . Functions of

social support refer to the resources and demands made available by social relationships . Three function s

are identified by House et al . (1988) : social relationships provide support in the form of instrumental aid ,

emotional caring, and information ; they involve demands and conflict with others ; and they regulate and

control behavior through social pressure .

Our focus is on social integration, which is a form of structural social support . In this article,

social integration, or one's level of embededness, refers to participation in organizations, the frequency o f

social activities, and the existence of close personal relationships .

Social Integration and Mental Health

How does social integration affect mental health? House et al . (1988) hypothesize that social

relationships may alter perceptions of stressful events, fulfill a basic human need for social interaction ,

and promote healthy behaviors such as exercise . Others suggest that social relationships improve well -

being by facilitating help-seeking or other coping strategies ; improving self-esteem, self-identity, an d

2

mastery ; or providing others with whom one can talk (see Belle 1987 ; Berkman 1985 ; Brown 1978 ;

Dunkel-Schetter et al . 1987; Holahan and Moos 1987 ; Pearlin 1985; Ross and Mirowsky 1989 ; Thoit s

1995) .

The Negative Side of Social Integratio n

To this point, we have confined our discussion to the positive aspects of social integration .

However, scholars are beginning to recognize that social integration, in some situations, may also b e

detrimental to psychological well-being (House et al . 1988; Rook 1998; Spitzberg and Cupach 1998) .

House et al . (1988), for example, suggest that one characteristic of social support is that it involve s

demands and conflicts with others . Empirical evidence supports these ideas . Rook (1984) demonstrate s

that negative social outcomes, such as having one's privacy invaded or having promises of help broken ,

decrease psychological well-being . Moreover, Rook (1984) demonstrates that negative social outcomes

have more powerful effects on mental health than positive social outcomes . In sum, "it is clear that

personal relationships can function as a source of conflict, strain, and disappointment, and trouble d

personal relationships threaten health and well-being" (Rook 1998 : 370) .

War-Related Distres s

Our focus is on war-related distress as a negative indicator of psychological well-being in Croati a

in the aftermath of civil war, which is a specific form of post-traumatic stress .1 Post-traumatic stress is an

anxiety disorder resulting from exposure to an extreme traumatic event . Social relationships are generall y

Post-traumatic stress involves reliving the traumatic event, avoiding stimuli associated with the event, emotiona lnumbness, and increased arousal (American Psychiatric Association 1994) . To support the diagnosis of post -traumatic stress disorder, responses to traumatic events must involve fear, hopelessness, or horror (America nPsychiatric Association 1994) . Symptoms must exist for longer than one month to be considered indicators of post -traumatic stress (American Psychiatric Association 1994) . Moreover, the disturbance must cause distress orimpairment in daily functioning (American Psychiatric Association 1994) . Post-traumatic stress may be classifie das acute if symptoms are reported for more than one month, but less than three months or as chronic if symptom spersist for greater than three months (American Psychiatric Association 1994) . Some individuals also experience adelayed onset of post-traumatic stress – that is, they only develop post-traumatic stress after at least six months hav epassed between the traumatic event(s) and the onset of post-traumatic stress (American Psychiatric Associatio n1994) .

3

beneficial – that is, they decrease post-traumatic stress – for individuals who are exposed to extrem e

traumatic events . This finding has been supported among such diverse groups as : burn patients (Perry et

al . 1992), car accident victims (Buckley, Blanchard, and Hickling 1996), children (LaGreca et al . 1996 ;

Vernberg et al . 1996), disaster survivors (Solomon et al . 1993), police officers (Stephens and Long 1997) ,

former prisoners of war (Engdahl et al . 1997 ; Neria, Solomon, and Dekel 1998), resident physicians

(Klamen, Grossman, and Kopacz 1995), veterans (Boscarino 1995 ; Fontana, Rosenheck, and Horvath

1997 ; Fontana and Rosenheck 1998; Sutker et al . 1995), women with breast cancer (Andrykowski and

Cordova 1998), battered women (Kemp et al . 1995 ; Perrin et al . 1996), and among the general public i n

community studies (Davidson et al . 1991) .

The benefits of social relationships in situations where entire communities have suffere d

traumatic events are of particular interest for the current analysis . Erikson introduces the concept of

collective trauma in his book, Everything in its Path (1976) . The substantive core of the book is a set of

excepts from interviews with respondents from Buffalo Creek, West Virginia, who survived a

catastrophic flood on February 26, J972 . The significance of the flood lay not only in the collectiv e

experience of death and destruction, but also later in the scarring of the community . Collective trauma,

such as in Buffalo Creek in 1972 and Croatia from 1991 to 1995, is :

a blow to the basic tissues of social life that damages the bonds attaching people together an d

impairs the prevailing sense of communality. . .it is a form of shock, a gradual realization that the

community no longer exists as an effective source of support (Erikson 1976 : 154) .

Can social integration affect psychological well-being after society has been damaged by broadl y

traumatizing events? Does social integration have a negative or positive effect on psychological well -

being following collective trauma? Does social integration buffer the effect of traumatic events on

psychological well-being following collective trauma or does it aggravate the impact traumatic events o n

distress?

4

The War in Croatia

Croatia provides a useful case in which to examine the relationship between social integration an d

mental health following collective trauma because of the widespread social disruption caused by the war .

The extent of the damage in human and material terms is staggering: out of a population of about 4 . 7

million, almost 14,000 people were killed or reported as missing, over 37,000 people were injured ,

hundreds of thousands of people were made homeless, and approximately $27 billion (U .S. Dollars) o f

public and private property were destroyed (Keesing's 1996 : 40913). The war also touched less tangibl e

aspects of people's lives . Conflict with ethnic "others" led to heightened nationalism, polarization, an d

the redefinition of ethnic and religious boundaries and gender roles . In sum, the war infiltrated public an d

private life and left its mark on society beyond the physical and emotional destruction that it left behind .

In the following section, we briefly outline the events leading from the dissolution of Yugoslavia to civi l

war in Croatia (for a more detailed description of the events, see Bugajski 1995 : 98-132; Radan 1995; and

Ramet 1996 : 197-221) .

A number of factors have been implicated in the dissolution of Yugoslavia and the civil war tha t

followed – for example, the death of Tito, increasing pressure from the Croatian, Slovenian, Bosnian, an d

Macedonian republics for more autonomy within the Yugoslav federal structure, and the politica l

dismantling of the Eastern Bloc and the Soviet Union (see Cohen 1993 : Dentich 1994 ; Ramet 1996;

Sekulic 1997; Woodward 1995) . Violence in Croatia between the Croats and Serbs began early in 1991 .

The fighting was small in scale and took place largely between Croat police forces and Serb militants wh o

had the backing of the Yugoslav Federal Army . Following the Republic of Croatia's declaration o f

independence on June 25, 1991, the Yugoslav Federal Army and Serb militants began an offensive to tak e

control of regions in Croatia dominated by the Serb minority with the ultimate goal of incorporating them

into a greater Serbia (Bugajski 1995) . In July and August, the Serb Autonomous Regions of Western

Slavonija and Slavonija, Baranja, and Western Srem were declared and immediately incorporated into the

Repubic of Serb Krajina . Figure 1 (at the end of this paper) shows the location of the republics of the

former Yugoslavia and the different sub-regions within Croatia .

5

The first phase of fighting came to an end by January 1992 . The truce would hold, despite minor

skirmishes, until 1995 . After a year fighting, between 6,000 and 10,000 people had been killed, 400,00 0

were homeless, over $18 .7 billion (U.S. Dollars) in material damage had been incurred, and roughly 30 %

of Croatia was occupied by Serb forces (Ramet 1996) .

The final phase of the war in Croatia began in May 1995 . On May 1, the Croatian army launche d

an offensive in Western Slavonija. Croatian Serb forces retaliated by shelling Zagreb, the capital o f

Croatia . On May 2, however, the Croatian army won a decisive victory by regaining control of western

Slavonija . Fighting re-intensified at the end of July as Croatian troops overran several key towns near

Knin, the capital of Serb Krajina . Finally, on August 4 . 1995, Croatian troops launched a final offensive .

Within two days, Krajina was occupied and the Croatian Serb army was largely destroyed . In this final

offensive, Croatian armed forces overran Serb Krajina sending about 150,000 Serbs fleein g toward

Belgrade, roughly 20,000 homes were destroyed, and thousands of lives were lost (Massey, Hodson, an d

Sekulic 1998) .

The scope of the devastation and violence can be seen in the survey data collected in 1996 –

seven months after the end of the war. Almost half of the respondents had a friend or relative who ha d

been in life-threatening danger; over a third of the respondents had been in life-threatening dange r

themselves ; and one in five had a friend killed in the war (see Table 1 at the end of this paper) . The data

also shed light on the psychological impact of the war . Thirty-seven percent of the sample had a

persistent sense of a foreshortened future, one third of the sample had recurrent and bothersome thought s

or memories about a traumatic war-related event, and one person in ten reported significant impairment i n

social, occupational, or other important areas of functioning (see Table 2) .

A comparison of the frequency of social activities, organizational memberships, and clos e

personal relationships in 1989 and 1996 sheds some light on how social relationships were affected by th e

war. Membership in sports clubs and local associations dropped by roughly a half and three quarters ,

respectively, from 1989 to 1996 (see Table 3) . In 1996, more people reported never going to the cinema ,

never going to restaurants, and never going on trips . On the other hand, more people reported going to a

6

café or on trips weekly . Also, there were dramatic shifts in the frequency of visiting friends and family

members . The percent visiting friends and family weekly rose from 65% and 49%, respectively, in 198 9

to 84% and 76% in 1996 . The utility of these visits as a form of social support, however, remains to b e

seen .

Theoretical Model of War-Related Distress in Croatia

Our theoretical model focuses on the effects of three forms of social integration on war-related

distress in Croatia, controlling for traumatic events, position in the stratification system . and background

variables . We also examine the possibility of a buffering effect on war-related distress – that is, a

negative, non-additive effect from the interaction of social integration and traumatic events .

Social Integration

We hypothesize that social integration – that is, being a member of organizations, participating i n

social activities, and having close personal relationships – decreases war-related distress . Moreover, w e

expect that the effect of traumatic events on war-related distress is weaker for those who are socially

integrated . Why might social integration decrease war-related distress and/or buffer the effect o f

traumatic events on distress? First, social integration may decrease alienation . With more and better

quality social relationships, people may develop a sense of mastery and meaning . Second, social

integration may lead to coping behaviors that allow respondents to deal more effectively with traumati c

events . Third, social integration may be associated with emotional, material, and informational resource s

that help people to overcome traumatic events .

Traumati c Events

Experiencing violence, experiencing damage to personal property, and fighting in the war ar e

expected to increase war-related distress . All of these traumatic events are similar to events outlined i n

the Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association 1994) ,

7

which are known causes of post-traumatic stress . Specifically, these events fall outside the range o f

normal human experience and they involve a direct threat to the respondent or to the respondent's famil y

and friends .

Stratification Positio n

People with more education, white-collar occupations, and more family income are expected t o

have lower levels of war-related distress . Studies routinely find that high socioeconomic statu s

individuals are better off than low socioeconomic status individuals in terms of mental health (Mirowsk y

and Ross 1989 ; Slomczynski et . al 1999) . Scholars argue that high socioeconomic status individuals ar e

healthier psychologically because they have more favorable work and economic conditions, more social -

psychological resources, and healthier lifestyles (Ross and Wu 1995) and because they are less likely t o

experience economic hardship (Ross and Huber 1985) . Individuals with higher levels of education hav e

also been shown to have lower levels of post-traumatic stress (Breslau et al . 1991) .

Background Controls

We expect ethnic minorities, women, the aged, and residents of war regions to have higher level s

of war-related distress than Croats, men, the young, and residents of regions outside the immediate wa r

zones, because of greater vulnerability to the events and lesser availability of options and resources . We

control for urbanism as a potentially important demographic indicator. Urbanism serves as an indirect

indicator of war experiences as most of the fighting revolved around the defense and capture of regiona l

centers .

Gender and War-Related Distress

In addition to examining the effect of gender on war-related distress, we also examine non -

additive effects of gender and social integration, gender and traumatic events, and three-way interaction s

among gender, social integration, and traumatic events on war-related distress . Examination of thes e

8

interactions is based on several considerations . First, scholars have consistently found that women repor t

higher levels of psychological distress than men (Mirowsky and Ross 1989) . Women are also more likely

to develop post-traumatic stress even though there are no significant differences between men and women

in exposure to traumatic events (Breslau et al . 1991 ; Breslau et al . 1997). In addition, scholars have

demonstrated that the direct and buffering effects of social networks and social support on mental health

differ by sex (Dressler J985; Haines and Hurlbert 1992) .

Second, Croatian women and men experienced the dissolution of Yugoslavia and the demise o f

socialism in significantly different ways . During socialism, women's share of life in the political and

economic spheres was significantly less than men's share (Ramet 1996). With the demise of socialism in

Croatia and the dissolution of Yugoslavia, women began to fall even further behind men as growin g

conservativism channeled women into more traditional roles . Ramet (1996) attributes the decline in th e

status of women to the increased influence of Christian churches, arguments by political leaders that there

were more pressing matters to worry about than gender equality, the rejection of all things espoused b y

socialism (e .g., gender equality), and nationalism . Ramet (1996) and Benderly (1997) argue that

nationalism heightens the role of women as instruments for the biological reproduction of the nation an d

for the maintenance of the family as a cradle of cultural values .

We also expect differences between men and women in the effects of social integration, traumati c

events, and in the buffering effect of social integration on war-related distress because of the use of rap e

as a psychological weapon during the war. While the frequency of rape during the war in Croatia i s

unknown, it has been estimated that over 20,000 women were raped in Bosnia and Herzegovina by 199 3

(Ramet 1996) . The organized and systematic rape and forced impregnation of women was accepted if no t

explicitly ordered during the war in Bosnia and Herzegovina and, to a lesser degree, in Croatia (Rame t

1996; Stiglmayer 1994) . According to Ramet (J996), "rape . . . is given a specifically national content . . . [i t

is] . . . used to act out, in symbolic terms, the subjection of one nation by another . At the same time, rap e

affirms the subordination of gender issues . . . to nationalist concerns" (284) .

9

In sum, women in Croatia experienced the dissolution of Yugoslavia and the civil war in Croati a

in a significantly different way than men . Growing conservatism and nationalism forced women bac k

into traditional roles of mother, wife, and mother of the Croatian nation . The use of rape as a weapon in

the war reinforced these trends . Do women have higher levels of war-related distress than men i n

Croatia? Are the effects of social integration and war-related events similar for women and men? Doe s

social integration buffer war-related distress differently for women ?

Major Hypotheses

The major hypotheses guiding the analysis can be summarized in the following propositions .

Proposition 1 : Individuals who are socially integrated – for example, those who belong to organizations ,

those who frequently participate in social activities, and those who have close persona l

relationships – have lower levels of war-related distress .

Proposition 2: Social integration buffers the effect of traumatic events on war-related distress .

These propositions are displayed in Figure 2 . Note that socially isolated individuals are expected to have

higher levels of distress and to experience traumatic events more negatively .

Data and Measurement

We test these hypotheses using nationally representative survey data collected by the Center o f

the Investigation of Transition and Civil Society2 in March and April, 1996 – seven months after the en d

of the war in Croatia . Ninety-five out of a possible 350 opcine (singular opcina), or counties, wer e

selected to form the sampling base from which households and, then, individuals were randomly sampled .

Opcine were selected in a purposeful manner to maximize variation relating to people's experience of th e

war. The face-to-face survey was administered to a random sample of 2,305 respondents within thes e

selected opcine. Of the 2,202 completed interviews, 409 respondents fought in the war and 17 2

1 0

respondents are ethnic minorities (non-Croatians) . The age, education, and sex distributions closel y

approximate those reported in the 1990 Yugoslav census .

War-Related Distress

War-related distress is measured by a 15-item index (Cronbach's Alpha= .87) . The 15 items,

which were derived from the post-traumatic stress disorder diagnostic criteria listed in the Diagnostic and

Statistical Manual of Mental Disorders (American Psychiatric Association J994), were summed, yielding

an unweighted, continuous index ranging from 0 to 15 . The war-related distress index (see Table 2) has a

mean of 2 .567 and a standard deviation of 3 .306 .

Social Integration

We focus on three types of social integration : organizational memberships, the frequency of

social activities, and the existence of close personal relationships . Seven dummy variables are include d

as measures of membership in political, economic, cultural, and religious organizations . Specifically, we

measure membership in political parties, unions, church organizations, non-governmental organizations ,

art and culture societies, sports clubs, and local associations (e .g ., local community improvemen t

associations or neighborhood chess clubs) .

Frequency of social activity is measured by a six-item scale (see Appendix A for factor analysis) .

This standardized and weighted scale (Cronbach's alpha= .76) indicates the frequency of going to the

cinema, restaurants, and cafés, going on trips, participating in sports, and attending concerts .

Four variables measure the existence of close personal relationships : the frequency of visiting

friends and relatives, being married, having children (that live in the home), and living with othe r

2 The Center of the Investigation of Transition and Civil Society was established by a group of researchers from th eInstitute for Social Research in Zagreb, Croatia in the early 1990s . The center is a private, non-profit organizationfinanced by private American and European foundations .

1 1

relatives . The frequency of visiting friends and relatives is a two-item, summed index . 3 Being married,

having children at home, and living with other relatives are dummy variables (1=yes) .

Traumatic Events

Three variables are included to measure the degree to which individuals were exposed t o

traumatic events during the civil war. The violence index consists of the sum of 18 war-related

experiences, such as being wounded, attacked, or forcefully emigrated (see Table 1) . The property

damage index consists of the sum of six items measuring the degree of destruction (no destruction, partia l

destruction, total destruction) to one's home, vacation home, economic/farming property, househol d

property, livestock, and crops. Direct participation in the war as a soldier or a member of a militia i s

measured by a dummy variable (1 des) .

Stratification Position

Three variables are included as indicators of stratification position : education, occupation, and

family income. Education is measured by the level of education completed . Response choices. arranged

from less education to more education, are : "no schooling, elementary school, basic high school, technica l

school, middle school, high school, university, and post-university ." Occupation is a dummy variable i n

which white-collar occupations are coded as 1 . Family income is measured by household monthly

income. Household monthly income is logged to decrease skew and standardized to facilitate

interpretation .

Background Controls

In addition to the analysis variables, we control for ethnicity, sex, age, urbanism, and residence i n

a war region . Ethnicity and sex are measured by dummy variables in which ethnic minorities (non -

Visiting friends and relatives were included in exploratory factor analyses of the frequency of social activities .However, these two items consistently load together as a second factor and the reliability of the social activitie s

1 2

Croatians) and women are coded as 1 . Age is measured in years . Urbanism is measured by self-reported

estimates of the size of the city of current residence . Response choices, in ascending order, are : "village,

local town, large local town, regional city, large regional city, or republic center ." War region is a

dummy variable in which the regions of Slavonija, Dalmacia, and Krajina – where most of the fightin g

occurred – are coded as 1 (see Figure 1) .

Results

The Main Effects Model

In support of the first proposition, some forms of social integration decrease war-related distres s

(see Table 4, column 3) . People who frequently participate in social activities and members of sport s

clubs, on average, have significantly lower levels of war-related distress . For example, a one standard

deviation unit increase in the frequency of social activities decreases distress by .106 standardized units .

However, the effects of social integration are not always in the anticipated direction . Specifically ,

members of economic and religious organizations (unions and church organizations), on average, hav e

significantly higher levels of war-related distress than non-members . People with close personal

relationships (e .g ., those who are married or have other relatives in the household) are no better off than

those without such close personal relationships .

Experiencing violence and fighting in the war are both significant predictors of war-relate d

distress . A one standard deviation unit increase in the level of violence experienced increases distress b y

.285 . In addition, those who fought in the war, on average, have significantly higher levels of distres s

than those who did not fight . The strength of these effects actually increases once we control for social

integration .

Our results also suggest that stratification position is a key predictor of war-related distress . Afte r

controlling for the main effects of all other variables in the model, individuals with more education an d

1 3

scale is significantly improved by removing them .

more family income have lower levels of war-related distress . A one standard deviation unit increase in

education and logged family income decrease distress by .087 and .057, respectively .

Finally, our results suggest that background factors explain a significant amount of the variatio n

in war-related distress . All else being equal, ethnic minorities and women have higher levels of distres s

than Croatians and men . The effect of age on war-related distress is curvilinear . As people age, they

have higher levels of distress, however, this effect weakens with increasing age . Also, people living i n

larger cities, on average, have higher levels of war-related distress. A one standard deviation uni t

increase in urbanism increases distress by .192 .

The Buffer Hypothesis Mode l

In addition to examining the additive effects of social integration on war-related distress, we als o

explore the possibility of non-additive, buffering effects on distress – that is, an interaction between socia l

integration and traumatic events. While additive effects tell us how social integration affects war-related

distress for everyone, the interaction between social integration and traumatic events gives us mor e

insight into how the effect of social integration on distress varies by exposure to traumatic events .

All possible interactions between social integration and traumatic events were examined . These

interactions were examined in groups segregated by type of social integration and type of traumatic event .

and, subsequently, in a unified model . Of the 36 possible interactions, three are statistically significant at

the .05 level : experiencing violence and party membership, experiencing violence and participating i n

social activities, and experiencing property damage and visiting friends and relatives (see Table 5, colum n

l) .

These results provide little support for the second proposition that social integration buffers th e

effect of traumatic events on war-related distress . In support of the buffer hypothesis, the negative effec t

of experiencing violence is significantly weaker for members of political parties (see Figure 3) . Also, the

effect of experiencing violence is weaker for those who frequently participate in social activities .

However, upon closer inspection, only the violence/social activities interaction resembles the buffe r

1 4

hypothesis . Individuals who frequently participate in social activities have lower levels of distress and the

effect of violence on distress is weaker compared to individuals who do not frequently participate i n

social activities . While party membership buffers the effect of violence on distress, members of political

parties begin with higher levels of war-related distress . Thus, only individuals who have experienced

extreme levels of violence benefit from membership in a political party .

As in the main effects model, it is again apparent that social integration is not always beneficial t o

psychological well-being . Rather than buffering, some forms of social integration may actually intensify

the effect of traumatic events on distress. While individuals who frequently visit friends and relative s

have lower levels of distress, the effect of experiencing property damage on distress is greater for them

(see Figure 3) . On the other hand, individuals who less frequently visit friends and relatives have highe r

levels of distress, but the effect of property damage on distress is negative . Thus, visiting friends an d

relatives is partially detrimental to psychological well-being if an individual experienced property damag e

during the war. Friends and family do not help one to forget property loss, rather they aggravate and

intensify the loss .

The Sex Interaction Mode l

In addition to examining the additive effect of sex on war-related distress, we explore th e

possibility of differential effects of social integration and traumatic events on war-related distress by sex .

This is based on several considerations . First, results from other analyses suggest that men and women

report different levels of psychological distress and that the effects of social support on well-being diffe r

for men and women (Dressler 1985 ; Haines and Hurlbert 1992 ; Mirowsky and Ross 1989) . Second,

women experienced a decrease in status and were forced into more traditional roles with the dissolution of

Yugoslavia and the resulting civil war (Ramet 1996) . Third, women were linked symbolically to th e

nation and targeted for rape during the war (Ramet 1996 ; Stiglmayer 1994) .

1 5

All possible interactions between sex and traumatic events and sex and social integration were

examined. These interactions were examined in groups segregated by traumatic events and types o f

social integration, and, subsequently, in a unified model . Only one interaction is consistently significan t

at the .05 level (see Table 5, column 2) . Specifically, the effect of fighting in the war on distress i s

significantly weaker for women than men (see Figure 4) .4 Among individuals who did not fight in th e

war, women have higher levels of war-related distress than men . However, women who fought in the war

score, on average, 1 .111 lower on distress than women who did not fight. Also, men who fought in the

war score, on average, .672 higher than men who did not fight. Thus, among individuals who did fight i n

the war, men have higher levels of distress than women . 5

Discussion

Social integration affects war-related distress after a powerful societal shock . However, there are

many caveats. Not all forms of social integration are beneficial for mental health . Members of economic

and religious organizations have significantly higher levels of distress than non-members . In addition ,

visiting friends and relatives appears to be detrimental to psychological well-being for those wh o

experienced significant property damage . Despite this, the news is not all bad . Individuals who are

members of some organizations, sports clubs for example, have lower levels of war-related distress .

Also, frequent participation in social activities is beneficial for reducing distress . Finally, frequently

participating in social activities buffers the effect of experiencing violence on distress . Our results, then,

suggest there is little support for the idea that social integration effectively buffers the effect of traumati c

events on distress following collective trauma. Clearly, the relationship between social integration ,

traumatic events, and war-related distress is complex in a setting of societal disruption .

The effect of social relationships on war-related distress, and psychological well-being mor e

generally, has usually been examined after the occurrence of traumatic personal events . Example s

include car accidents, heart attacks, or violent crime . Many studies have also examined social integratio n

4 In our sample over 6% of those who fought in the war are women .5 We examined additional three-way interactions to test whether social integration buffers the effects of traumati cevents on war-related distress similarly for men and women . None of the possible three-way interactions ar e

1 6

and mental health among groups of individuals traumatized outside of their usual social networks, fo r

example soldiers fighting in a foreign war . The larger network of social relationships, in all of thes e

analyses, remains relatively in tact . In other words, homecoming and recovery involve returning t o

supportive social networks .

Croatia is different because of the collective nature of the trauma . This may help to explain why

being a member of some organizations decreases war-related distress while membership in other s

increases distress . We suggest that members of economic and religious organizations have higher level s

of war-related distress than non-members because of the functions of these organizations and the

collective nature of the destruction . Economic and religious organizations, as parts of the institutional

structure of society, reflect the state of the larger society . After the war in Croatia, it can be expected that

discussions of the hardship and loss infiltrated and perhaps even dominated the activities of these formal

organizations. In other words, like the situation for survivors of the Buffalo Creek flood (Erikson 1976) ,

the scale of the disaster in Croatia is such that those remaining are trapped in its memory . Formal

organizations may come to serve as forums where traumatic events are revived and relived well beyond

the point where such sharing is of psychological value .

On the other hand, informal organizations, such as sports clubs, generally exist for purpose s

removed from national politics . Provision of opportunities for physical recreation, for example, is th e

central function of such local organizations . Such activities are removed from national politics and the

national trauma. Informal organizations may, therefore, be more useful in the societal healing proces s

than more traditional civic organizations .

The significant, positive interaction between property damage and visiting friends and relatives

also highlights the potentially negative aspect of social integration . Following collective trauma, friend s

and family members may, themselves, be unable to provide the emotional sympathy and strength neede d

to lessen or buffer psychological distress . While supportive functions of family and friends can break

1 7

significant at the .05 level . Thus, there are no significant sex differences in the buffering of traumatic events o ndistress by social integration .

down with collective trauma, negative functions associated with social integration do not dissipate an d

may even intensify . In sum, following collective trauma, talking to others who have also been affected by

the war may create a spiral of despair for those who have experienced traumatic events .

Our results also highlight some significant differences between men and women, both in th e

degree to which they report war-related distress and in the effects of traumatic events on war-relate d

distress . Net of social integration, exposure to traumatic events, stratification position, and other

background variables, women have significantly higher levels of distress than men . However, women ar e

less negatively affected by active participation in the war. Specifically, active participation in the war i s

more beneficial for women's mental health than for men's . The benefit for women from activ e

participation may be due to the fact that such acts go against the traditional roles increasingly assigned t o

women since the dissolution of Yugoslavia and the resulting civil war . In other words, being an active

participant, as opposed to a passive victim, may offset the marginalized status assigned to women an d

reduce war-related distress .

Conclusion

This article helps to answer two important questions about the relationships between traumati c

events, social integration, and mental health following collective trauma — does social integration have a

direct effect in reducing war-related distress and does social integration buffer the effect of war-relate d

events on distress? The answer is that social integration has both positive and negative direct effects on

war-related distress following collective trauma and that there is little support for the buffer hypothesis .

These results also raise new questions. Why are some forms of social integration detrimental fo r

mental health while others have beneficial effects? We have suggested that these differences are due to a

combination of the different functions of the organizations involved, the form of social integration, an d

the degree of societal destruction . In instances of societal-level trauma, informal local clubs appear to b e

more effective in the healing process than more formal societal-level organizations . Similarly, going to

the café or the movies seem to provide relief for those affected by the war, while frequently visiting wit h

1 8

friends and relatives does not. Future analyses should examine in greater detail the effects of membership

in various types of organizations and participation in various types of activities, paying attention to th e

role of the specific activities involved .

Further, how exactly does social integration affect mental health? To date, there have been only a

handful of analyses that have examined the possible mechanisms, such as efficacy and help-seeking, tha t

connect social integration and mental health . However, these studies have produced conflicting results .

Future studies should search for mechanisms that both positively and negatively affect psychologica l

well-being paying particular attention to the role of formal and informal organizations and different type s

of social activities . It is our hope that this and future analyses will reveal useful ways to facilitate

individual and social healing following societal destruction .

1 9

Reference s

American Psychiatric Association . 1994 . Diagnostic and Statistical Manual ofMental Disorders (4thedition) . Washington D.C. : American Psychiatric Association .

Andrykowski, M.A. and M.J . Cordova. 1998 . "Factors Associated with PTSD Following Treatment fo rBreast Cancer: A Test of the Andersen Model ." Journal of Traumatic Stress J1 : 189-203 .

Barrera, Manuel . 1986. "Distinctions Between Social Support Concepts, Measures, and Models . "American Journal of Community Psychology 14 : 413-445 .

Basoglu, M., M. Parker, E. Ozmen, O . Tasdemir, and D. Sahin . 1994 . "Factors Related to Long-TermTraumatic Stress Response in Survivors of Torture in Turkey ." Journal of the American Medica l

Association 272 : 357-363 .Belle, Deborah . J987. "Gender Differences in the Social Moderators of Stress ." Pp. 257-277 in Gender

and Stress, edited by R.C . Barnett, L . Biener, and G.K. Baruch . New York : Free Press .Benderly, Jill . 1997 . "Rape, Feminism, and Nationalism in the War in Yugoslav Successor States ." Pp .

59-72 in Feminist Nationalism, edited by Lois A . West. New York : Routledge .Berkman, Lisa F . 1985. "The Relationship of Social Networks and Social Support to Morbidity an d

Mortality." Pp . 241-262 in Social Support and Health, edited by S . Cohen and S .L. Syme .Orlando, FL : Academic Press .

Boscarino, Joseph A . 1995 . "Post-Traumatic Stress and Associated Disorders Among Vietnam Veterans :The Significance of Combat Exposure and Social Support ." Journal of Traumatic Stress 8 : 317-336 .

Breslau, Naomi, Glenn C . Davis, Patricia Andreski, Edward Peterson . 1991 . "Traumatic Events an dPosttraumatic Stress Disorder in an Urban Population of Young Adults ." Archive ofGeneralPsychiatry 48: 216-222 .

Breslau, Naomi, Glenn C . Davis, Patricia Andreski, Edward Peterson, and Lonni R . Schultz. 1997. "SexDifference in Posttraumatic Stress Disorder ." Archive of General Psychiatry 54 : 1044-1048 .

Brown, B . Braford. 1978. "Social and Psychological Correlates of Help-Seeking Behavior Amon gUrban Adults ." American Journal of Community Psychology 6 : 425-439 .

Buckley, T.C ., E .B. Blanchard, and E .J . Hickling . 1996. "A Prospective Examination of Delayed Onse tPTSD Secondary to Motor Vehicle Accidents ." Journal ofAbnormal Psychology 105 : 617-625 .

Bugajski, Janusz. 1995 . Nations in Turmoil : Conflict and Cooperation in Eastern Europe . Boulder:Westview Press .

Cohen, Lenard J . 1993 . Broken Bonds : Yugoslavia's Disintegration and Balkan Politics in Transition .

Boulder: Westview Press .Cohen, Sheldon and Thomas Ashby Wills . 1985 . "Stress, Social Support, and the Buffering Hypothesis . "

Psychological Bulletin 98: 310-357 .Davidson, J .R.T., D. Hughes, D .G. Blazer, and L .K. George . 1991 . "Posttraumatic-Stress-Disorder in the

Community : An Epidemiologic-Study ." Psychological Medicine 21 : 713-721 .Denitch, Bogdan . 1994 . Ethnic Nationalism: The Tragic Death of Yugoslavia . Minneapolis : University

of Minnesota Press .Dressler, William W. 1985 . "Extended Family Relationships, Social Support, and Mental Health in a

Southern Black Community ." Journal of Health and Social Behavior 26 : 39-48 .Dunkel-Schetter, Christine, Susan Folkman, and Richard S . Lazarus . 1987. "Correlates of Social Support

Receipt ." Journal of Personality and Social Psychology 53 : 71-80 .Engdahl, B ., T.N. Dikel, R . Eberly, and A . Blank. 1997. "Posttraumatic Stress Disorder in a Communit y

Group of Former Prisoners of War: A Normative Response to Severe Trauma." AmericanJournal of Psychiatry 154 : 1576-1581 .

Erikson, Kai T. 1976 . Everything in its Path . New York: Simon and Schuster .Fontana, A . and R. Rosenheck. 1998. "Duty-Related and Sexual Stress in the Etiology of PTSD Amon g

Women Veterans Who Seek Treatment ." Psychiatric Services 49 : 658-662 .

2 0

Fontana, A ., R. Rosenheck, and T . Horvath. 1997. "Social Support and Psychopathology in the Wa rZone." Journal of Nervous and Mental Disease 185 : 675-681 .

Haines, Valerie A . and Jeanne S . Hurlbert . 1992. "Network Range and Health ." Journal ofHealth and

Social Behavior 33 : 254-266 .Holahan, Charles J . and Rudolf H . Moos. 1987. "Personal and Contextual Determinants of Coping

Strategies." Journal ofPersonality and Social Psychology 52 : 946-955 .House, James S . and Robert L . Kahn. 1985 . "Measures and Concepts of Social Support ." Pp. 83-108 in

Social Support and Health, edited by S . Cohen and S .L. Syme. Orlando, FL: Academic Press .House, J .S ., D. Umberson, and K.R. Landis . 1988. "Structures and Processes of Social Support ." Annua l

Review ofSociology 14: 293-318 .Keesing's Record of World Events . 1996. "Tudjman's State of Nation Address -- War Casualty Figures . "

42 : 40913 . Avenel, N.J . : Cartermill Publishing .Kemp, A., B .L. Green, C . Hovanitz, and E .I. Rawlings . 1995. "Incidence and Correlates of

Posttraumatic-Stress-Disorder in Battered Women: Shelter and Community Samples ." Journal of

Interpersonal Violence 10 : 43-55 .Klamen, D.L., L .S. Grossman, and D . Kopacz . 1995 . "Posttraumatic-Stress-Disorder Symptoms in

Resident Physicians Related to their Internship ." Academic Psychiatry 19: 142-149 .LaGreca, A .M., W.K. Silverman, E .M . Vernberg, and M .J . Prinstein . 1996. "Symptoms of Posttraumatic

Stress in Children After Hurricane Andrew : A Prospective Study." Journal of Consulting and

Clinical Psychology 64 : 712-723 .Massey, Garth, Randy Hodson, and Dusko Sekuli . 1998. "Ethnic Nationalism and Liberalism in Post -

War Croatia ." Presented at the Inter-University Consortium in Dubrovnik, Croatia, May, 1998 .Mirowsky, John and Catherine E . Ross . 1989 . Social Causes of Psychological Distress . New York :

Aldine de Gruyter.Neria, Y., Z. Solomon, and R. Dekel . 1998 . "Eighteen-Year Follow-Up Study of Israeli Prisoners of War

and Combat Veterans ." Journal of Nervous and Mental Disease 186 : 174-182 .Pearlin, Leonard I . 1985 . "Social Structure and Processes of Social Support ." Pp. 43-60 in Social

Support and Health, edited by S . Cohen and S .L . Syme . Orlando, FL: Academic Press.Peron, S ., V.B. Van Hasselt, I . Basilio, and M . Hersen . 1996 . "Assessing the Effects of Violence on

Women in Battering Relationships with the Keane MMPI-PTSD Scale ." Journal of Traumati c

Stress 9 : 805-816 .Perry, S ., J . Difede, G . Musngi, A .J . Francis, and L . Jacobsberg . 1992 . "Predictors of Posttraumatic -

Stress-Disorder After Burn Injury . " American Journal of Psychiatry 149 : 931-935 .Radan, Peter . 1995. "Secessionist Self-Determination : The Cases of Slovenia and Croatia ." Pp . 141-15 7

in Nationalism and Postcommunism : A Collection of Essays, edited by A. Pavkovi, H .Koscharsky, and A . Czarnota . Brookkkfield, VT : Dartmouth Publishing Company .

Ramet, Sabrina P . 1996 . Balkan Babel: The Disintegration of Yugoslavia from the Death ofTito to

Ethnic War. Boulder : Westview Press.Rook, Karen S . 1984. "The Negative Side of Social Integration : Impact on Psychological Well-Being . "

Journal ofPersonality and Social Psychology 46 : 1097-1108 .Rook, Karen S . 1998. "Investigating the Positive and Negative Sides of Personal Relationships : Through

a Lens Darkly?" Pp . 369-393 in The Dark Side ofPersonal Relationships, edited by Brian H .Spitzberg and William R. Cupach . Mahwah, NJ : Lawrence Erlbaum Associates .

Ross, Catherine E. and Joan Huber. 1985 . "Hardship and Depression." Journal of Health and Socia l

Behavior 26 : 312-327.Ross, Catherine E. and John Mirowsky. 1989. "Explaining the Social Patterns of Depression : Contro l

and Problem Solving -- or Support and Talking?" Journal of Health and Social Behavior 30 :206-219 .

Ross, Catherine E. and Chia-ling Wu. 1995 . "The Links between Education and Health ." American

Sociological Review 60: 719-745.

2 1

Sekulic, Dusko. 1997 . "The Creation and Dissolution of the Multinational State : The Case ofYugoslavia ." Nations and Nationalism 3 : 165-179 .

Slomczynski, Kazimierz M ., Krystyna Janicka. Bogdan W. Mach, and Wojciech Zaborowski . 1999 .

Mental Adjustment to the Post-Communist System in Poland. Warsaw, Poland: Institute o fPhilosophy and Sociology Publishers .

Solomon, S .D., M. Bravo, M. Rubiostipec, and G . Canino . 1993 . "The Effect of Family Role onResponse to Disaster ." Journal of Traumatic Stress 6 : 255-269 .

Spitzer, Brian H. and William R . Cupach (editors) . 1998 . The Dark Side of Personal Relationships .

Mahwah, NJ : Lawrence Erlbaum Associates .Stephens, C . and N. Long. 1997. "The Impact of Trauma and Social Support on Posttraumatic Stres s

Disorder: A Study of New Zealand Police Officers ." Journal of Criminal Justice 25 : 303-314 .

Stiglmayer, Alexandra (editor) . 1994. Mass Rape: The War Against Women in Bosnia-Herzegovina .

Lincoln, NE: University of Nebraska Press .Sutker, P .B., J .M. Davis, M . Uddo, and S.R. Ditta . 1995. "War Zone Stress, Personal Resources, an d

PTSD in Persian-Gulf-War Returnees ." Journal ofAbnormal Psychology 104 : 444-452 .

Thoits, Peggy A . 1995. "Stress, Coping, and Social Support Processes: Where Are We? What Next? "Journal of Health and Social Behavior Extra Issue : 53-79 .

Verberg, EM, A.M. LaGreca, W.K. Silverman, and M.J . Prinstein . 1996 . "Prediction of Posttraumati cStress Symptoms in Children After Hurricane Anderw ." Journal of Abnormal Psychology 105 :

237-248 .Woodward, Susan L . 1995 . Balkan Tragedy : Chaos and Dissolution After the Cold War. Washington

D.C . : The Brookings Institute .

22

Table 1 . Violence in Croatia (1996, N=2,202).Did you have any of these war- Standardrelated experiences? % Yes Mean DeviationViolence Index (Range=0-18) 3 .721 4.076

Forcefully emigrated 6.5%Life endangered 36.0%Relative in life-threatening danger 44.3%Friend in life-threatening danger 46.8%Relatives disappeared 8.6%Friends disappeared 14.3%Relative wounded 17.7%Friends wounded 27.7%Family member attacked 11 .2%Relatives attacked 20.O%Friends attacked 23 .8%Relatives captured 6.4%Friends captured 10.0%Family forcefully emigrated 13 .5%Relatives forcefully emigrated 25 .4%Friends forcefully emigrated 25 .5%Relatives killed 13 .7%Friends killed 20.8%

23

Table 2. War-Related Distress in Croatia (1996, N=2,202) .Which, if any, of the following symptoms have you experienced for over one month induration during the past year? % Yes Mean

StandardDeviation

War-Related Distress Index (Range=0-15) 2 .567 3 .306

Recurrent and bothersome thoughts or memories about a traumatic war-related event 33 .2%Recurrent distressing dreams about a war-related event 18.7%A recurrent sense of reliving past war trauma in the present such as flashback kinds of 13.4%

experiencesPersistent intense emotional or physical distress at exposure to cues that symbolize or resemble 18 .1 %

an aspect of war traumaPersistent avoidance behaviors, thoughts, or feelings related to war trauma such as avoiding 15 .4%

certain conversations, ideas, or activities that arouse painful memorie sPersistent loss of memory for important aspects of war trauma 4.5 %Markedly diminished interest or participation in usual activities 12 .5 %Persistent feelings of being detached or estranged from others such as family members with 8.5%

whom you have felt closePersistent reductions in the ability to feel your emotions 8 .9%A persistent sense of a foreshortened future 36 .9%Persistent difficulty falling or staying asleep 16 .9%

Persistent irritability or outbursts of anger 16 .9%Diminished ability to concentrate 22 .3%Being easily startled or panicked frequently 19 .1%Significant impairment in social, occupational, or other important areas of functioning 11 .4%

2 4

Table 3 . Social Integration in Croatia (1989 a , N=2,510 ; 1996, N=2,202) .1989 1996

Organizational Memberships:Political Party b 7 .6%Union 74 .8 %Church Organization 4 .0 %Non-Governmental Org . --- 2 .5 %Art and Culture Society 3 .1% 3 .2 %Sports Club 12 .5% 6 .5%Local Association 13 .3% 3 .6%Frequency of Social Activities :Going to Cinem a

Never 62 .7% 71 .9 %Monthly 30 .1% 17 .9 %Weekly 7 .1% 10 .2%

Going to ConcertNever 76 .2 %Monthly 17 .3 %Weekly 6 .5%

Going to RestaurantNever 63 .0% 69 .2 %Monthly 27 .7% 19 .0%Weekly 9 .3% 11 .8%

Going t o CaféNever 56 .5% 54 .0%Monthly 19 .7% 11 .1 %Weekly 23 .7% 34 .9 %

Going on a TripNever 50 .4% 60 .5%Monthly 41 .4% 24 .7%Weekly 8 .2% 14 .8%

Playing Sport sNever 71 .7% 72 .8%Monthly 12 .2% 10 .1 %Weekly 16 .1% 17 .1 %

Close Personal Relationships :Visiting Friend s

Never 5 .6% 7.8%Monthly 29 .1% 8.4%Weekly 65 .3% 83 .8%

Visiting Relative sNever 7 .1% 8.8%Monthly 43 .9% 15 .6%Weekly 49 .0% 75 .7%

a. The 1989 data are from a Yugoslav-wide survey conducted just prior to the dissolution of the country . Similarquestions on social integration are contained in both the 1989 and 1996 surveys . We present results in Table 3only for those respondents living in Croatia .

b. Data pertaining to membership in political parties, unions, church organizations, and non-governmenta lorganizations as well as the frequency of concert attendance are not available in the 1989 data .

25

Table 4 . Ordinary Least Squares Regressions of War-Related Distress on Social Integration . Traumatic Events, Stratificatio nPosition, and Background Controls (1996, N=2.202) .

Distress Distress Distres sVariables b(s .e .) p b(s .e.) p b(s .e.) pConstant 2 .207( .939) 1 .161 (.905) 1 .520( .975 )

Background Control sEthnic Minority (1=yes) .843(.253)*** .068 .725(.243)** .059 .780( .242)** .06 3Female (1=yes) .758(.138)*** .114 .880( .148)*** .133 .737( .153)*** .11 1Age .103( .025 )* * * .507 .108(.024)*** .530 .057( .027)* .28 2

Age2 - .001( .000)* * * - .515 - .001( .000)*** - .476 -.001( .000)* - .29 8

Urbanism .388(.040)*** .217 .317(.039)*** .177 .344( .040)*** .19 2War Region (l yes) .603( .138)*** .091 .086( .138) .013 .049( .139) .00 7

Stratification PositionEducation - .205( .054)* * * - .097 - .226(.052)*** - .107 - .185( .054)*** - .08 7

White-Collar (1=yes) - .298( .172) - .041 - .264( .164) - .037 - .139( .172) - .01 9

Family Income (logged) - .378( .099)*** - .091 - .356( .095)*** - .085 - .236( .105)* - .05 7

Traumatic EventsViolence Index .226( .0 18)*** .279 .23 1( .018)*** .28 5

Property Damage Index .040( .046) .018 .038( .045) .01 8

Fought in War (1=yes) .425( .200)* .050 .500( .201)* .05 9

Social Integratio nOrganizational Memberships

Political Party (1=yes) .432( .259) .03 5

Union (1=yes) .370( .178)* .04 9Art and Culture Society (I =yes) .559(.378) .03 0

Sports Club (1=yes) - .631( .279)* - .04 7

Local Association (1=yes) - .265(.362) - .01 5Non-Governmental Org. (1=yes) - .240( .423) - .01 1Church Organization (1--yes) 709( .333)* .04 2

Frequency of Social Activitie s

Social Activities Scale - .391( .105)*** - .106

Close Personal Relationship s

Visiting Friends and Relatives - .048( .035) - .02 7

Married (1=yes) .124( .168) .01 8Children in House (1=yes) - .183( .194) - .02 8

Other Relatives in House (1=yes) - .174( .204) - .023

26

R 2

.08 5*** p<.001 ; ** p< .01 ; * p< .05 (two-tailed t-test) .

220 2.164

2202.18 0

N

2202

Table 5 . Ordinary Least Squares Regressions of War-Related Distress on Social Integration, Traumatic Events . StratificationPosition, and Background Controls with Interactions (1996, N=1202) .

Buffer Hypothesis Model

Sex Interaction ModelVariables

b(s .e .)

R

b(s .e .)

R

Constant 1 .663(.973) 1 .434 ( .974 )

Background ControlsEthnic Minority (1 =y-es) .778(.241)** .063 .790(.242)** .06 4Female (1-=yes) .740(.152)*** .112 .839( .157)*** .12 7Age .058(.027)* .285 .055( .027)* .27 0

Age2 - .001(.000)* - .291 -.001(.000)* - .28 3Urbanism .3 36(040)* * * .188 .344(040)*** .19 3War Region (1=yes) .044(.139) .007 .050(138) .00 7

Stratification Positio nEducation - .187(054)*** - .089 - .183(054)*** - .08 6White-Collar (1=yes) - .122( .172) - .017 - .120( .172) - .01 7Family Income (logged) - .230(.104)* - .055 -.231( .104)* - .05 5

War-Related EventsViolence Index .246(019)*** .303 .232(018)*** .28 6Property Damage Index - .258( .124)* - .120 .037( .045) .01 7Fought in War (lames) .592(.201)** 070 .672(210)** .07 9

Social IntegrationOrganizational Membership s

Political Party (1=yes) 1 .018(.377)** .082 .465(259) .03 7Union (1-=yes) .398( .177)* .052 .373( .177)* .04 9Art and Culture Society (1=yes) .543( .376) .029 .629( .378) .03 3

Sports Club (l yes) - .630( .278)* - .047 - .641( .278)* - .04 8Local Association (1-yes) - .300( .362) - .017 - .264( .362) - .01 5Non-Governmental Org . (1=yes) - .212( .421) - .010 - .270(422) - .01 3Church Organization (1=yes) .721( .332)* .043 .715( .333)* .04 3

Frequency of Social Activities

Social Activities Scale - .155( .132) - .042 - .390(105)*** - .105

Close Personal Relationship s

Visiting Friends and Relatives - .084( .037)* - .048 - .051(035) - .02 9Married (1=yes) .113( .167) .016 .153(168) .02 2Children in House (1=yes) - .210( .193) - .032 - .198(194) - .03 0

Other Relatives in House (1=yes) - .208( .203) - .028 - .166(203) - .02 2

Interactions

Violence * Political Party - .136( .057)* - .07 2Violence * Social Activities Scale - .050( .017)** - .08 2Damage * Visiting .045( .017)** 15 0

Female * Fought in War -1 .783( .632)** - .058

N

2202R 2

.18 8*** p< .001 ; ** p<.01 ; * p<.05 (two-tailed t-test) .

220 2.183

27

a The bolded regions within South Central Europe are former Yugoslav Republics . Thebolded regions in Croatia are the areas where the principal fighting occurred .

Figure 2 . The Buffer Hypothesis : Traumatic Events, Social Integration, and War-RelatedDistress .

Figure 3 . Interactions : Traumatic Events and Social Integration.

Violence Index(Mean=3 .721, S .D.=4 .076)

Figure 3. (Continued)

Violence Index(Mean=3 .721, S .D.=4 .076)

Figure 3 . (Continued)

Property Damage Index

(Mean=.506, S .D.=1 .541)

Figure 4 . Interactions : Sex and Fought in the War .

Fought in the War(Mean=.186, S .D.=.389)

No Yes

APPENDIX A. SOCIAL ACTIVITIES SCALEFACTOR ANALYSIS (CROATIA, 1996, N=2,202).Variable

Factor ScoreFrequency :Going to Cinema .720Going to Concert .603Going to Restaurant .660Going to Café .61 0Going on a Trip .58 8Participate in Sports .544

N 2202Eigenvalue 2.93 2% Variance 48.863Cronbach's Alpha .764

APPENDIX B . ZERO ORDER CORRELATIONS CROATIA, 1996, N=2,202 .(l) ( 2 ) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12)

(1) War-Related Distress 1 .000(2) Ethnic Minority (1=yes) .077* 1 .000(3) Female (1=yes) .135* - .042 1 .000(4) Age .082* .055* .036 1 .00 0(5) City Size .152* .039 .032 - .051* I .00 0(6) War Region (1 = yes) .059* - .021 - .005 - .005 - .163* 1 .00 0(7) Education - .079* .003 - .I24* - .372* .359* - .012 1 .00 0(8) White-Collar (1=yes) - .072* - .030 .031 - .367* .162* - .029 .429* I .00 0(9) Family Income (logged) - .099* - .059* - .057* - .388* .222* - .027 .399* 349* 1 .000(10) Violence Index .290* .036 .003 - .180* .117* .226* .138* .075* .08I* 1 .00 0(I I) Damage Index .I16* - .022 - .004 .017 - .058* .239* - .033 - .032 - .0I7 .339* I .00 0(12) Fought (1 =yes) .016 - .043* - .441* - .28I* .009 .086* .144* .082* .I3I* .236* .099* I .000(13) Political Party (1=yes) .009 - .058* - .186* - .020 - .094* .078* .059* - .011 .023 .080* .060* .193 *(14) Union (1=yes) .106* .009 .103* .416* - .094* - .034 - .299* - .445* - .316* - .079* .007 - .172 *(15) Art/Culture Society (1 =yes) .008 - .024 - .009 - .070* .006 - .0I8 .096* .046* .034 .028 - .038 .040(16) Sports Club (1=yes) - .087* - .00I - .207* - .148* .022 .007 .128* .042* .082* .055* - .017 .I96 *(I7) Local Association (1=yes) - .025 - .01I - .095* - .019 - .048* .003 .079* .019 .055* .043* - .006 .I03 *(18) Non-Governmental (1-yes) .006 .008 - .0I7 .008 .020 .031 .124* .023 .015 .080* .006 .051 *(19) Church Org . (1 =yes) .040 - .034 .011 .015 - .006 .070* .071* .013 .007 .029 - .012 .00 9(20) Social Activities Scale - .128* - .029 - .204* - .543* .235* - .059* .485* .304* .396* .183* - .012 .274 *(21) Frequency Visiting - .043* .011 - .036 - .242* .038 .038 .111* .107* .156* .094* - .011 .078 *(22) Married (1=yes) .043* .0I2 - .080* .143* - .026 .014 .010 - .020 .084* - .009 .008 - .03 5(23) Children in House (1=yes) - .028 .018 - .078* - .286* .061* .048* .154* .168* .338* .064* - .016 .103 *(24) Relatives in House (1=yes) - .029 - .057* .085* - .093* - .083* - .033 - .034 - .013 .060* - .005 .040 .00 7

Mean 2 .567 .078 .525 46 .155 3 .169 .449 3 .842 .298 7 .684 3 .721 .506 .186Standard Deviation 3 .306 .268 .500 16 .262 1 .849 .498 1 .562 .457 .793 4 .076 I .541 .389

* p< .05 (two-tailed) .

APPENDIX B. (CONTINUED)(13) (14) (1 ) (16) (17) (18) (19) (20) (21) (22) (23) (24)

(1) War-Related Distres s(2) Ethnic Minority (1=yes)(3) Female (1 =yes )(4) Age(5) City Size(6) War Region (1=yes)(7) Educatio n(8) White-Collar (1=yes )(9) Family Income (logged )(10) Violence Index(II) Damage Inde x(12) Fought (1 =yes)(13) Political Party (1=yes) l .00 0(14) Union (1=yes) - .050* l .00 0(15) Art/Culture Society (1 =yes) .075* - .038 I .00 0(16) Sports Club (1=yes) .153* - .076* .057* I .000(17) Local Association (1=yes) .230* - .029 .104* .118* 1 .00 0(18) Non-Governmental (1 =yes) .107* - .008 .120* .052* .110* 1 .000(19) Church Org. (1=yes) .045* .002 .134* .049* .109* .056* 1 .00 0(20) Social Activities Scale .067* - .351* .152* .285* .079* .063* .031 1 .00 0

(21) Frequency Visiting - .001 - .099* .008 .059* .031 - .013 .002 .222* 1 .00 0

(22) Married (1=yes) .043* .097* - .056* - .072* .035 .022 - .019 - .235* - .017 1 .00 0

(23) Children in House (1 =yes) .025 - .121* .005 .046* .042 - .022 .022 .152* .095* .339* 1 .000

(24) Relatives in House (1-yes) - .022 - .018 .011 .012 - .014 - .002 - .038 .009 - .019 - .316* - .578* l .00 0

Mean .076 .748 .032 .065 .036 .025 .040 .000 6 .754 .661 .487 .260

Standard Deviation .266 .434 .176 .247 .186 .156 .197 .894 l .898 .474 .500 .43 9

* p< .05 (two-tailed) .