Table 2 Empirical evidence by types of studies and countries: Single level studies

Kawachi, Kennedy, Locher & Prothrow-Stith, 1997 [17]

A cross-sectional ecological study based on data from 39 US states. Socioeconomic data were obtained from 1990 US Census Population and Housing Summary Tape File 3A.

Using General Social Survey (GSS) responses on social trust, perceived lack of fairness, perceived helpfulness of others and memberships in groups- taking each one separately.

'Aggregated' social capital (individual level responses aggregated to the US state level).

Income inequality and mortality in US states.

Each of the four measures was found to be positively associated with mortality.

Wilkinson, Kawachi & Kennedy, 1998 [74]

A cross-sectional study data used from the US General Social Surveys (1986–90) and National Center for Health Statistics (1981–1991).

Social mistrust was used as indicator of social capital.

'Aggregated' social capital (individual level responses aggregated to US state level)

Mortality and violent crime rate in 39 US states.

Social mistrust closely related with mortality and violent crime rates.

Lochner, Kawachi, Brennan & Buka, 2003 [75]

A cross-sectional study design for persons 45–64 years, indicators of neighborhood social capital were obtained from the 1995 Community Survey of the Project on Human Development in Chicago Neighborhoods.

Measured by reciprocity, trust, and civic participation.

'Aggregated' social capital (individual level responses aggregated to neighborhood level)

Mortality rates in 342 Chicago neighborhoods, USA.

Higher levels of neighborhood social capital were associated with lower neighbourhood death rates for total mortality as well as death from heart disease and "other" causes for death.

Holtgrave & Crosby, 2003 [76]

An ecological state level study where data used from Putnum, 2001 [US General Social Surveys (1974–1994); DDB Needham archive (1975–1998); Roper Social and Political Trends archive (1974–1997)] Youth Risk Behavior Surveillance Survey (1999).

Includes 14 variables that span the domains of community organizational life, involvement in public volunteerism, informal sociability and social trust and called this "comprehensive social capital index".

Combination of 'aggregated' social capital (individual level responses aggregated to the US state level) and contextual social capital

Case rates of gonorrhoea, syphilis, Chlamydia and AIDS in 48 US States.

Social capital index inversely associated with gonorrhoea, syphilis, Chlamydia and AIDS case rates.

Milyo & Mellor, 2003 [77]

An ecological analysis based on state-level data came from several publicly available sources. Crude and age-adjusted mortality rates for 1990 are obtained from the US Centers for Disease Control. The Gini-coefficient for family income and percent of persons below the federal poverty line are come from the US Bureau of Labor Statistics. Other than social capital all covariates were obtained from the US Census Bureau.

Social capital was measured by Robert Putnam's index of state social capital and by an index of social mistrust derived from responses to the General Social Survey, following the method described in [17]

Combination of 'aggregated' social capital (individual level responses aggregated to US state level) and contextual social capital

Age-adjusted mortality rates, defined as deaths per 100,000 in 1990.

The study did not find significant association between mortality and either minority racial concentration, or social capital. Authors conclude that different age-adjustment methods can cause a change in the sign or statistical significance of the association between mortality and other socioeconomic factors.

Holtgrave & Crosby, 2004 [78]

An ecological state level study data came from Putnum, 2001 [US General Social Surveys (1974–1994); DDB Needham archive (1975–1998); Roper Social and Political Trends archive (1974–1997)] Youth Risk Behavior Surveillance Survey (1999).

Includes 14 variables that span the domains of community organizational life, involvement in public volunteerism, informal sociability and social trust and called this "comprehensive social capital index".

Combination of 'aggregated' (individual level responses aggregated to the US state level) and contextual social capital

Tuberculosis case rates in 48 US States.

Social capital index inversely associated with tuberculosis case rates.

Veenstra, 2000 [79]

A cross-sectional study, data used from a survey administered to randomly selected citizens within randomly selected households from eight health districts of Saskatchewan in 1997.

Constructing different indices for overall civic participation, trust in government, trust in neighbours, trust in people from respondents' communities, trust in people from respondents' part of Saskatchewan, and trust people in general.

Individual level social capital

Self-rated health within Saskatchewan, Canada.

Social capital was not significantly related to self-rated health among the general population but positive impact among the elderly.

Veenstra, 2002 [80]

A cross-sectional study, data came from the Canadian National Population Health Survey, considered 30 health districts in Saskatchewan, Canada.

The social capital index incorporated associational density and civic participation.

Individual level social capital

Mortality rates; low birth weight rate; proportion of residents receiving mental health services etc.

Except low birth weight rate social capital was inversely and weakly related to age-adjusted mortality rates, and other outcomes.

Veenstra et al . 2005 [81]

A cross-sectional study based on an individual level data from a telephone survey of a random sample of adults (N = 1504) neighborhood of residence in the city of Hamilton, Canada The survey contained a range of questions designed to capture participation in social and community networks, health status and behaviors, use and access of health services and socio-demographic factors was administered to respondents between November 2001 and April 2002.

Social capital measured by constructing an index focusing specifically on breadth and depth of involvement in voluntary associations where respondents were asked if they belonged to an association and assessed its type, i.e., religious, cultural/historic, community, social services/health, sports/athletics, pastimes/social/artistic, professional or political assessing degree of involvement in the association. The minimum score on this index was zero (no groups were mentioned) and the maximum was 6.0.

Individual level social capital

Various measures of individual health, such as- self-rated health, chronic conditions, emotional distress and body-mass index.

Overall involvement in voluntary associations was significantly related to emotional distress and almost significantly related to self-rated health before and after controlling for age, gender and neighbourhood of residence. More participation in voluntary associations apparently had a positive association with these measures of health.

Kennedy, Kawachi & Brained, 1998 [8]

Cross-sectional, ecological analysis across 40 regions of Russia. The study based on a stratified random sample of the population, survey conducted by the All-Russian Center for public opinion in April through June, 1994

Indicators used for social capital such as civic engagement, and trust in government and social cohesion (divorce rate, per capita crime rate, conflicts in workplace) ;

Both 'aggregated' social capital (individual level responses aggregated to the region of Russia) and contextual social capital

Mortality in Russia.

Social capital and cohesion indicators were strongly linked with age-adjusted mortality for both sexes.

Rose, 2000 [82]

Based on a nation-wide cross-sectional survey (New Russia Barometer Survey) conducted in March-April, 1998.

Sense of self-efficacy, trust of others, inclusion or exclusion from formal and informal networks, social support and social integration,

Individual level social capital

Self-reported physical and emotional health in Russia.

Both human and social capital were associated with improved self-reported health.

Skrabski, Kopp & Kawachi, 2003 [83]

A cross-sectional ecological study based on The Hungarostudy II, a national survey representing the Hungarian population over the age of 16 conducted for the 20 counties in 1995.

Measured by three indicators: lack of social trust, reciprocity between citizens and help received from civil organization

'Aggregated' social capital (Individual level responses aggregated to the counties of Hungary)

Gender specific mortality rates for the middle aged population in the 20 counties of Hungary Life satisfaction, suicide rates for 50 countries.

All of the social capital variables were significantly linked with middle age mortality.

Skrabski, Kopp & Kawachi, 2004 [84]

A cross-sectional ecological study based on The 'Hungarostudy' survey 2002 and Hungarian Central Statistical Office (1996–2000).

Social trust, organization membership and reciprocity were used as indicators of social capital.

'Aggregated' social capital (Individual level responses aggregated to the sub region of Hungary)

Gender specific all cause specific mortality rates in the 150 sub-regions of Hungry.

Social distrust positively and other two indicators inversely and significantly associated with all cause mortality rates.

McCulloch, 2001 [85]

A representative cross-sectional study based on the British Household Panel Study for the years 1998 and 1999.

Summed individual responses to eight questions about different community problems and classified them into low, medium, high and very high levels of social disorganization.

Individual level social capital

Examined psychiatric morbidity using the 12 item general health questionnaire for a representative cross section of British households.

People in the lowest categories of social capital had increased risk of psychiatric morbidity.

Kelleher, Timoney, S Friel & McKeown, 2002 [86]

A cross-sectional ecological study employed three data sources namely,1996 census data from the Central Statistics Office, 1997 general election first preference voting data in all 41 constituencies were aggregated to county level and the National survey on lifestyles, attitudes and nutrition (SLAN) data. The study comprised adults over 18 years sampled by post using the electoral register from 273 representative district electoral divisions.

Party political affiliation and general election voting pattern was used to measure vertical social capital.

Contextual social capital

Standardised mortality ratios (SMR) and selected reported measures of health status, lifestyle in Ireland.

There was no significant relation between SMR and voting pattern for the two main political parties but a significant relation with left wing voting. There was a positive significant relation between left wing voting and dissatisfaction with health and rate of smoking.

Hyyppä & Mäki, 2001b [88]

An individual level cross sectional study based on a randomly selected samples of Finnish-speakers (N = 1,000), and Swedish-speakers (N = 1,000), representing all adults living in bilingual Ostrobothnian municipalities. 75,000 Finnish speakers and 78,000 Swedish-speakers from national registers of the Social Insurance Institution of Finland (KELA) and stratified by gender, age, and municipality.

Social capital operationalized by social ties and integrity operationalized by asking four questions on friendship, voluntary neighbourhood assistance, reciprocal civic trust, and civic engagement.

Individual level social capital

Self-rated good health.

Swedish speaking community appeared to hold a better stock of social capital than Finnish speaking counterparts which were significantly and positively associated with good self-rated health.

Hyyppä & Mäki, 2003 [89]

An individual level cross- sectional study used randomly selected samples of Finnish-speakers (N = 1,000), and Swedish-speakers (N = 1,000), representing all adults living in bilingual Ostrobothnian municipalities. 75,000 Finnish speakers and 78,000 Swedish-speakers from national registers of the Social Insurance Institution of Finland (KELA) and stratified by gender, age, and municipality.

Active participation in voluntary associations, friendship ties, religious involvement and hobby club activity and trust were used as measures of social capital.

Individual level social capital

Self-rated good health.

Active participation in voluntary associations, friendship ties and trust associated with self-rated good health.

Liukkonen, Virtanen, Kivimäki, Pentti & Vahtera, 2004 [90]

The study based on a prospective cohort of 6028 public sector employees in Finland. In the 10-Town Study sent out a questionnaire to all full-time permanent employees who were at work at the time of survey in the eight towns participating in the study in 1997.

Employment security and social support two indicators reflecting employment type and co-worker support, and combined them into a variable indicating the amount of 'social job capital'.

Individual level social capital

A 5-point scale of self-rated health and Psychological distress was measured by the 12-item version of the General Health Questionnaire.

The results indicated that a low level of 'social job capital' is associated with poor health only in the age-adjusted model in women. However, after accounting for baseline health differences and other background variables, the significant associations were disappeared both in women and in men.

Bolin, Lindgren, Lindstrom & Nystedt, 2003 [18]

The study employed a set of individual panel data based on the Swedish survey of living conditions (ULF). The panel consisting of about 3800 individuals, for the years 1980/81, 1988/89 and 1996/97.

As an indicator of social capital the authors considered whether individual had a close friend outside his or her household.

Individual level social capital

Self-rated health in Sweden.

Social capital had a positive effect on self-assessed health.

Sundquist, Lindström, Malmström, Johansson & Sundquist, 2004 [91]

A cross-sectional follow-up study based on data from the Swedish Annual Level-of-Living Survey (SALLS). During 1990 and 1991, 6861 women and men aged 35–74 were interviewed.

Neighbors talk often in the area, whether attended mutual activity in the neighbourhood and whether socialize with neighbors at least once every three months were used as indicators of social capital.

Individual level social capital

Coronary heart disease (CHD) morbidity and mortality in Sweden.

Persons with low social participation in the social participation index exhibited an increased risk of CHD.

Lindström, 2004 [92]

A cross-section study, data came from the population investigated by a postal questionnaire in Scania in southern Sweden during November 1999-February 2000, the public health survey in Scania 2000. The postal questionnaire was sent to 24,922 randomly chosen persons aged 18–80 (born in 1919–1981) that were registered as living in Scania.

Social participation measured as how actively the person takes part in the activities of formal and informal groups in society during last year and Generalised trust to other people is a self-reported indicator that reflects the person's perception of generalised trust to other people With the combination of social participation and trust, four alternatives social capital levels are identified such as, high-social participation/high trust (high social capital), high-social participation/low trust ("the miniaturisation of community"), low-social participation/high trust (traditionalism), and low-social participation/low trust (low-social capital).

Individual level social capital

Siahpush & Singh, 1999 [93]

A state level ecological study and data complied from several Australian Bureau of Statistics documents for the years 1990–1996.

Five indicators of social integration namely percentage of people living alone, divorce rate, unemployment rate, proportion of people who are discouraged job-seekers and unionization rate were used as proxy of social capital.

Contextual social capital

State level six Cause-specific mortality, all cause mortality and sex specific life expectancy in Australia.

Higher levels of social capital for most of the indicators were significantly associated with mortality rates and life expectancy.

Chavez, Kemp & Harris, 2004 [94]

A household level cross-sectional survey based study originally developed as evaluation tool for neighborhood based interventions and used for two disadvantaged neighborhood in south-western Sydney, Australia.

Six common social capital components such as, neighbourhood attachment, support networks, feelings of trust and reciprocity, local engagement, personal attachment to the area, feelings about safety and pro-activity in the social context were used.

Individual level social capital

Self-reported health.

It is revealed that with the exception of feelings of trust and reciprocity, no other social capital component made significant contributions to explaining health variance and those macro-level factors such as housing conditions and employment opportunities emerged as key explanatory factors.

Ziersch, Baum, MacDougall & Putland, 2005 [95]

Data came from a broader study, the Health Development and Social Capital Project (HDSCP), undertaken in the Western suburbs of Adelaide in 1997. Two sources of data come from HDSCP through a questionnaire (n = 2400) and in-depth interviews (n = 40).

Social capital operationalized by neighbourhood connections, neighbourhood trust, reciprocity, neighbourhood safety, local civic action.

Individual level social capital

Physical and mental health as measured by SF-12 (a summary scores used as response variable) for the resident live in Adelaide, Australia

Only perceived neighborhood safety was related to physical health and neighbourhood connections and neighbourhood safety were positively associated with mental health, with those with stronger neighbourhood connections and higher levels of perceived neighbourhood safety, having better mental health.

Lynch, Davey Smith, Hillemeier, Shaw, Raghunathan & Kaplan, 2001 [96]

A cross-sectional study data came from the World Values Survey (1990–1991); UN Human Development Report and WHO mortality data base (1991–1993).

Used social distrust, organization membership and volunteering as indicators of social capital.

'Aggregated' social capital (Individual level responses aggregated to different country level)

Life expectancy, mortality, low birth-weight and self-rated health for 16 OECD countries.

Difference between countries in levels of social capital showed weak and inconsistent associations with age-specific and cause-specific mortality rates.

Kennelly, O'Shea & Gavey, 2003 [97]

The study used a panel data set covering three time periods and the trust data comes from the three waves of the World Values Survey: 1981–84, 1990–93 and 1995–97.

Measured by the proportion of people who say that they generally trust other people and by membership in voluntary organization.

'Aggregated' social capital (Individual level responses aggregated to different country level)

Population health in 19 countries in the OECD.

Found very little statistically significance evidence of social capital had a positive effect on population health.

Smith & Polanyi 2003 [98]

A cross-sectional study data came from the 1995–97 World Values Survey conducted in a variety of countries including Australia, Sweden and Norway (n = 5,096).

Social capital operationalized through socially oriented norms and behaviors.

Individual level social capital

The gradient between income and self-rated health across three different welfare countries.

The study found variation in the level of social capital measures across the three different countries. Socially oriented norms were not strongly correlated with each other, or with socially oriented behaviors. And existence of socially oriented norms or behaviors did not reduce the likelihood of lower income groups reporting poor self-rated health, relative to the highest income groups.

Carlson, 2004 [99]

An ecological cross-sectional study based on data from the World Value Survey conducted in 1995–1997 and based on data from 18 European countries and from respondents aged 18 years and over.

Measured from an individual perspective, where the individual's trust in people, confidence in the legal system or membership of organizations are investigated.

Individual level social capital

Self-rated health constructed as 'Very good' and 'good' were defined as good health and 'satisfactory', 'poor' and 'very poor' were defined as poor health.

Both economic factors and some aspects of social capital played a role in the area differences in self-rated health. Economic factors appeared to be more important. People in the countries in central and eastern Europe tended to be worse off than in western Europe, both in terms of economy and in terms of social capital.

Helliwell & Putnam, 2004 [100]

A cross-sectional study, data came from three different sources of survey data. The first source was the World Value Survey (WVS) of the years 1980, 1991–1992 and 1995–1997, covered 49 countries, and used a three-wave panel of roughly 84,000 observations. The second data source was the Social Capital Benchmark Survey in the US includes about 29, 000 observations drawn from a national random sample supplemented by samples from many participating communities The third source was the Canadian data were drawn from two national waves and two special over-samples of a survey sponsored by the Social Sciences and Humanities Research Council of Canada and the sample used in the analysis was about 7500.

Social capital operationalized by the strength of family, neighbourhood, religious and community ties.

Individual level social capital

Life satisfaction, happiness and self assessed health status measured on the same five point scale used in all three surveys.

Social capital was strongly associated with subjective well-being through many independent channels and in several different forms. All indicators of social capital appeared independently and robustly related to happiness and life satisfaction, both directly and through their impact on health.

Pollack & Knesebeck, 2004 [101]

A cross-national study based on Germany and the United States in the years 2000 and 2001. Data obtained by computer assistance telephone interviews (CATI) conducted in Germany (N = 682) and the United States (N = 608) with probability samples of non-institutionalized persons aged 60 and older was used.

Social capital operationalized by both norms (reciprocity and civic trust) and behaviors (participation). Participation was assessed by whether people attended a church, charity group, sports club, self-help group, or other local activity at least once a month.

Individual level social capital

Three self-reported health indicators overall health, depression (CES-D) and functional limitations.

Lack of reciprocity was associated with poorer self-rated health and depression in both countries and civic mistrust was associated with poorer self-rated health in both countries. Lack of participation was, associated with poorer self-rated health and depression in Germany, The effect of norms is stronger in the US than in Germany. Participation in community groups, however, is more strongly associated in Germany.