Social Cohesion in Australia: Comparing Church and Community

Abstract

In a context of increasing ethnic and religious diversity, Australia’s future prosperity may depend, in part, on the ability to maintain social cohesion. Drawing on the framework developed by the Scanlon Foundation Social Cohesion Research Program, this study examines data from the 2016 National Church Life Survey and the 2016 Australian Community Survey to compare levels of social cohesion among Australian churchgoers and among the general population. Social cohesion metrics were stronger among churchgoers than the wider population across the domains of belonging, social justice, civic participation, acceptance of others and worth. Differences were also observed between Christian denominations on most domains, but with few exceptions, social cohesion among churchgoers from each denomination was still higher than that observed for all Australians. The findings suggest that Christian groups play a positive role in the promotion of social cohesion by building both bridging and bonding social capital among those who participate, but that these groups are unlikely to be a significant source of agitation to prevent some of the greatest contemporary threats to social cohesion.

1. Introduction

With the global movement of people and ideas, societies have become increasingly diverse, including in terms of religious diversity (Bouma and Halafoff 2017). Australia is one of the most multicultural nations in the world, and migration patterns continue to increase ethnic and religious diversity. In this changing context, Australia’s future prosperity may depend, in part, on the ability to maintain social cohesion.

Classical post-Westphalian social theories saw diversity as a problem to be overcome, a challenge to social cohesion (Jupp et al. 2007, pp. 9–20). Preoccupations with social cohesion presume that internal conflict undermines the capacity of a society, group or organisation to cooperate to survive and prosper. This is evidenced in violence among elements of a society, falling standards of health and well-being, and flagging economic productivity. But what is social cohesion?

Defining social cohesion presents a very interesting problem. A review of sociological texts reveals that definitions of social cohesion are rare, and far from consistent (Jupp 2018). Books on the social policy of social cohesion provide lists of factors that are necessary for a social democracy to hold together but not definitions of social cohesion. One summation of what is meant by social cohesion, and the definition that we use in the present paper, is the following: “social cohesion refers quite simply to the capacity of a society or a group to so organize its resources and people to produce what it needs to sustain and reproduce itself” (Bouma and Ling 2007, p. 80).

However, this simply describes a society that is working well. Then we must clarify what we mean by “well”. What kind of society? Totalitarian societies appear cohesive. China has raised millions from poverty in the past 30 years. Liberal democracies appear unsteady. Whose interests feature in the definition? There is an inherent conservative core to concerns for social cohesion (Jupp 2018). How is change to be managed? Whose power is being conserved? Are we referring to a particular set of relational practices that are democratic, participatory, and egalitarian—whether these be in everyday life, or in legal or political decision making? Are we referring to the processes by which a society organizes itself and its resources? It is clear that the underlying concept of the ideal society that shapes the meaning of social cohesion varies from context to context.

Concerns about social cohesion appear to be driven, in part, by fears of what would happen, or is happening as a society changes and develops. Will I/we be part of what is coming? What will I/we lose? Will competition for scarce goods and services become violent? Each fear or concern motivates a domain of the issues that are bundled together in the concept of social cohesion. Alternatively, social cohesion can be seen to refer to one of several ideal forms of society—assimilation and similarity, mutual understanding and respect, harmonious intergroup relations, and productive cooperation. Fears and ideals such as these form the backdrop to questions about the belonging, worth, participation, social justice and acceptance dimensions of social cohesion adopted for this paper.

In a context of increasing ethnic and religious diversity, the role of religion in relation to social cohesion is worthy of attention. Participation in religious communities has been shown both to promote and undermine social cohesion (Bouma 1994, 1997; Bouma et al. 2001; Akbarzadeh 2001). Among Christian churchgoers, participation has been seen to develop both bonding social capital, which refers to networks of reciprocity and trust between people in the same social group (in this case the church), and various forms of bridging social capital, which concerns networks between social groups (Dixon 2010; Dixon and Arunachalam 2018; Leonard and Bellamy 2010). However, in some contexts, the relationships, norms and values within a group may serve to build walls between that group and the wider society (Appleby 2000).

In order to test whether there is a relationship between attending Christian congregations and social cohesion in Australia, this paper compares levels of reported social cohesion among churchgoers with that among the general population. Do those who attend different Christian denominations vary in their degree of expressed social cohesion both in comparison with each other and with the wider society? Our approach is based on the ongoing Scanlon Foundation Social Cohesion Research Program (Markus and Arunachalam 2008; Markus 2018) and employs data from two surveys conducted in 2016 by NCLS Research (Powell et al. 2016; Powell and Pepper 2016). First, the Australian demographic context of this study is described, with a particular focus on religious diversity. Then, the Scanlon Foundation framework and measures of social cohesion are outlined and our hypotheses presented, followed by a description of the methodology for the two surveys. Finally, the data are presented and analyzed in relation to the hypotheses, the wider literature, and the framing of social cohesion used for this study.

1.1. Australia’s Increasing Diversity

Australia’s population growth and composition has largely been driven by immigration (Bouma 1995; Jupp 2009, 2018). The total population in 2016 was over 24 million and the annual population growth was 1.4%. Figure 1 indicates that the contribution of net overseas migration (NOM) to growth has been higher than the contribution from natural increase for over a decade (rising to 61.4% due to NOM and 38.6% due to natural increase in 2018, ABS 2019). This pattern of migration has resulted in substantial ethnic diversity. The 2016 Census of Population and Housing showed that more than a quarter (26%) of the population was overseas-born, and that 45% had at least one overseas-born parent (ABS 2017a). Leading countries of birth for migrants were the United Kingdom, New Zealand, China, India, the Philippines, Vietnam and Italy. Over the last thirty years, an increasing proportion of immigrants have been drawn from South and Southeast Asia. By way of comparison in 2014 OECD figures for those “foreign-born” were: 13% for OECD on average, 44% in Luxemburg; 29% in Switzerland; 20% in Canada; 13% in Germany, 13% in the United States, 13% in the United Kingdom, and 12% in France (OECD 2016).

Australia is one of the most religiously diverse nations in the world (Bouma 2016) which is not surprising given the sources of post-war migrants.

Table 1. Top 20 religions in Australia in 2016 and 2011.

2016	2011
No religion—30.1% 1	Catholic—25.3%
Catholic—22.6%	No religion—22.3%
Anglican—13.3%	Anglican—17.1%
Uniting Church—3.7%	Uniting Church—5.0%
Christian (not further defined)—2.6%	Presbyterian and Reformed—2.8%
Islam—2.6%	Eastern Orthodox—2.6%
Buddhism—2.4%	Buddhism—2.5%
Presbyterian and Reformed—2.3%	Islam—2.2%
Eastern Orthodox—2.1%	Christian (Not further defined)—2.2%
Hinduism—1.9%	Baptist—1.6%
Baptist—1.5%	Hinduism—1.3%
Pentecostal—1.1%	Lutheran—1.2%
Lutheran—0.7%	Pentecostal—1.1%
Sikhism—0.5%	Judaism—0.5%
Other Protestant—0.5%	Jehovah’s Witnesses—0.4%
Judaism—0.4%	Sikhism—0.3%
Jehovah’s Witnesses—0.4%	Seventh–day Adventist—0.3%
Seventh-day Adventist—0.3%	Other Protestant—0.3%
Latter-day Saints—0.3%	Salvation Army—0.3%
Oriental Orthodox—0.2%	Latter-day Saints—0.3%
Total Christian—52.1%	Total Christian—61.1%
Total Other Religions—8.2%	Total Other Religions—7.2%

Source: ABS (2017c). ¹ In 2016 for the first time, “No religion” was placed at the top of the response options on the Census form.

presents the proportions of the population identifying with major religious groups in Australia, comparing census results for 2016 and 2011. In international comparison, Australia stands out for having three substantial minority religious communities at or above 2% and two at about 0.5%. There is of course a great deal of ethnic diversity within each of these groups. Muslims have come from over 60 countries, Catholics have been strengthened by Italian, Dutch, Vietnamese, Philippine and other sources. Hinduism, Sikhism, Islam, and Buddhism are all increasingly substantial and vibrant religious communities largely due to recent migration from South and Southeast Asia and, for Muslims, earlier migration from the Middle East. Detail on the history and characteristics of Australia’s religions may be found in Jupp (2009).

According to the 2016 Census, older age groups (65+) were more likely than younger groups to identify with Christianity, whereas young adults aged 18–34 were more likely to identify with religions other than Christianity (12%) and to report not having a religion (39%) (ABS 2017b). In a study in which teenagers themselves were asked about their religion, over half (52%) said they had “no religion” (Singleton et al. 2018).

Because ethnic and religious diversity is increasing, it is important to look more closely at the role of diverse religious groups in producing social cohesion. In this paper we focus on the largest religious group in Australia—Christianity. The Scanlon Foundation Social Cohesion reports have published differences in some areas of social cohesion by religious identification. But what about religious participation? Are those who attend religious services more or less likely to feel part of Australia? While the Scanlon survey instruments from 2009 through to 2013 included religious participation, few results on its influence have been released.

Sharpening our focus from adherents to attendees is important for several reasons. First, religious groups vary widely in the percentage of adherents who attend. For example, attendance rates are very low among Anglicans, a little higher among Catholics and high among Pentecostals (comparing Table 1 with the estimates of Powell et al. (2017) of the number of people attending churches in a given week). Second, a higher proportion of churchgoers are overseas-born than is the case for the wider Australian population. In contrast, a lower proportion of Christian adherents were born overseas (McAleese et al. 2018). Third, attendance is a measure of exposure to the teachings, practices and values of the religious group; an indicator of the effect of congregations on the perspectives and actions of individuals (Gill 1999).

Around 20 Christian denominations account for some 95% of the weekly churchgoers in Australia, with the Catholic Church being the largest (Powell et al. 2019). These denominations can be grouped into four denominational types: Catholic, Mainline Protestant (Anglican, Lutheran, Presbyterian, and Uniting Church—an amalgamation of Methodist, Congregational and some Presbyterians), Pentecostal and Other Protestant (largely conservative and evangelical). Other Christian denominations are very small and include many Orthodox churches linked with migrant communities as well as independent churches. With the exception of the Pentecostal churches, all other denominational groups have an older age profile when compared with the Australian population. Women are over-represented in all four denominational types as are churchgoers with a university-level education. The country of birth of churchgoers varies significantly by denomination. Catholics have the highest proportion born in non-English-speaking countries, Pentecostals and Other Protestants are similar to the wider population and Mainline Protestants are lower than average. For example, Lutherans are heavily Australian-born (McAleese et al. 2018).

1.2. Defining and Measuring Social Cohesion

In order to measure variation in reported social cohesion, the NCLS Research surveys of social cohesion in church and community draw on work undertaken over more than a decade by the Scanlon Foundation (Markus and Arunachalam 2008; Markus and Arnup 2010; Markus 2018). While social cohesion has a long tradition in academic enquiry and increasing interest in recent decades, there is no agreed definition of social cohesion (Markus and Arunachalam 2008). However, Markus and colleagues identify three common elements across the literature:

• “Shared vision: Social cohesion requires universal values, mutual respect and common aspirations or identity shared by their members”;
• “A property of a group or community: Social cohesion describes a well-functioning core group or community in which there are shared goals and responsibilities and a readiness to co-operate with the other members”; and
• “A process: Social cohesion is generally viewed not simply as an outcome, but as a continuous and seemingly never-ending process of achieving social harmony” (Markus and Arunachalam 2008, p. 25).

Measures of social cohesion in liberal democracies, or postindustrial societies, tend to focus on factors taken to produce social cohesion rather than measuring cohesion directly or focussing on outcomes of social cohesion. This is true of the Scanlon Foundation surveys which measure inputs to social cohesion, framed around five domains:

• “Belonging: Shared values, identification with Australia, trust”;
• “Social justice and equity: Evaluation of national policies”;
• “Participation: Voluntary work, political and co-operative involvement”;
• “Acceptance and rejection, legitimacy: Experience of discrimination, attitudes towards minorities and newcomers”; and
• “Worth: Life satisfaction and happiness, future expectations” (Markus and Arunachalam 2008, p. 25).

Several measures are grouped into these domains, validated by factor analysis (Markus and Arnup 2010, pp. 40–41).

1.3. Hypotheses

We anticipate that churchgoers will express higher levels of social cohesion than the wider Australian population due to the bonding and bridging capital provided by participation in a substantial social organisation (Putnam 2000; Dixon 2010; Leonard and Bellamy 2010). We expect this to be particularly expressed through civic participation and trust in others. Churchgoers are more highly educated than the wider population (McAleese et al. 2018), which may also contribute to higher social cohesion in some domains (Markus and Arunachalam (2008) report this effect for the worth and acceptance domains). Allport’s (1954) contact hypothesis leads us to anticipate that the social contact with culturally different others that is likely to result from the higher proportion of immigrants in churches than in Australia at large would lead to higher levels of social cohesion among churchgoers specifically with regards to attitudes toward newcomers to Australia. Finally, religious service participation has been shown to relate positively to subjective wellbeing across a large number of studies (Koenig et al. 2012). There are a range of explanations for this finding: religions provide cognitive resources for coping with stress, they have rules and regulations for behavior that reduce the likelihood of some types of stressful life events, and they encourage prosocial action and virtues that enhance social relationships and positive emotions (Koenig 2012). An enhancement on the worth domain should therefore be evident among churchgoers.

2. Materials and Methods

2.1. Data Collection and Samples

Data from two surveys conducted by NCLS Research are used to answer the research questions.1 Both surveys received ethics clearance from Australian Catholic University (Ethics Register Number 2016-186E).

2.1.1. 2016 National Church Life Survey

The Australian National Church Life Survey (NCLS) is a five-yearly quantitative survey of thousands of Christian churches and hundreds of thousands of churchgoers in approximately 20 Australian denominations (Catholic, Anglican and other Protestant). The survey is based in local churches (congregations and parishes), with approaches to recruitment and sampling varying across the denominations (attempted census, random sampling, or opt-in; paid for by the local church or by the denomination) (Pepper et al. 2018). Participating local churches ask their attendees aged 15 years and over to complete a confidential hard copy form, directly after or during a worship service. In 2016, an online survey option was also available. The 2016 NCLS Attender Survey consisted of a four-page main survey of demographics, Christian faith and practice and church health, which was completed by most individual participants and a series of smaller four-page surveys, each of which was a random sample of the total participants. The 2016 NCLS Small Sample Attender Survey D (“2016 Attender D”) covered the majority of the questions from the Main Attender Survey, together with a suite of questions on multiculturalism, social cohesion, intercultural communication and cultural values.

Catholic NCLS data are random samples, however in Protestant denominations there are self-selection biases in church participation related to church size, locality and theological tradition, with larger urban churches of an evangelical flavor over-represented in the datasets (Pepper et al. 2018). Nevertheless, the datasets have national coverage and denominational diversity, and churches from a wide diversity of traditions participate. The relatively low level of religiosity of the Australian population means that it is rarely possible to use national population studies to address research questions that concern Christian faith and practice across the diverse landscape of Australian churches—subsample sizes are too small. The NCLS fills a particular niche in this regard (Pepper et al. 2018). In the present case, we are in a good position to examine differences in views about social cohesion among the institutional churches. For a more detailed explanation of the NCLS methodology, participation rates, and the strengths and limitations of the NCLS datasets, see Pepper et al. (2018).

The 2016 Attender D dataset (Powell et al. 2016) comprised churchgoers aged 15 years and over who were a random subsample of respondents in the total 2016 NCLS dataset (total N of approximately 260,000 people from 3000 congregations and 20 denominations). Results were weighted to adjust for variations in survey participation levels between denominations and between churches of different sizes within the Protestant denominations. Churchgoers from 14 denominations were sampled well enough to be included in the weighted analysis. These denominations account for some 95% of the weekly churchgoers in Australia (not including Orthodox, independent or house churches). Other denominations which were inadequately sampled or absent are not included in the analysis. The final sample size was N = 1442.

2.1.2. 2016 Australian Community Survey

The 2016 Australian Community Survey (ACS) was an online anonymous survey of 1258 respondents aged 18 years and over. The survey was distributed by Online Research Unit (ORU), to a sample drawn from their Australian Consumer Panel. ORU meets ISO 20252 and ISO 26362 standards for market research and panel work. The survey instrument was around 60 questions, including demographics, measures of religion and religiousness, social cohesion, civic participation, attitudes to religion and Christian churches, and contact with churches. Quotas were set for age, gender and location, derived from the 2011 Census of Population and Housing, and a quota was also set for education (given the tendency for people with higher levels of education to respond to survey invitations). Panelists received a blind email invitation and followed a link to the survey introduction page, which included information about the survey, a participant information statement and the terms of consent. Participants received entries into ORU cash and gift card prize draws, as is standard for participation in surveys offered through the Australian Consumer Panel. The dataset is weighted to reflect the demographic profile of the Australian population aged 18+ on age, gender and education, according to the 2016 Census, by applying a methodology similar to that used for weighting the Australian Survey of Society Attitudes datasets (Evans 2017).

Table 2. Demographics of National Church Life Survey (NCLS) and Australian Community Survey (ACS) samples.

Variable	Percentage
	NCLS	ACS
Age
15–29	13	20
30–49	23	37
50–69	37	28
70+	27	14
Gender
Female	60	51
Male	40	49
Country of birth
Australia	67	77
Other English-speaking	7	12
Non-English-speaking	26	11
Educational attainment
School	39	45
Trade certificate/diploma	24	32
University degree	37	23
Denomination 1
Catholic	48	18
Anglican	11	16
Uniting Church	8	4
Baptist	8	2
Pentecostal	15	3
Other Protestant	10	N/A
Total Christian	N/A	49

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ In the case of the NCLS, denomination is an attribute of the church attended by the respondent.

lists the demographics of the NCLS and ACS samples. Comparisons with unweighted samples are given in

Table A1. Demographics of NCLS and ACS samples (weighted and unweighted).

Variable	Percentage 2
	NCLS Weighted	NCLS Unweighted	ACS Weighted	ACS Unweighted
Age
15–29	12.9	14.8	20.5	20.0
30–49	23.2	23.1	37.5	39.4
50–69	36.7	36.3	28.4	29.3
70+	27.2	25.8	13.7	11.4
Gender
Female	60.0	58.7	51.2	50.7
Male	40.0	41.3	48.8	49.3
Country of birth
Australia	67.3	70.5	76.6	74.1
Other English-speaking	7.2	8.2	12.3	12.8
Non-English-speaking	25.5	21.3	11.1	13.1
Educational attainment
School	38.6	36.7	45.4	34.4
Trade certificate/diploma	24.2	23.2	31.6	37.7
University degree	37.2	40.1	23.0	27.9
Denomination 1
Catholic	48.5	29.9	17.8	18.4
Anglican	10.9	19.4	16.1	15.7
Uniting Church	7.7	8.8	3.8	3.8
Baptist	7.7	17.1	2.4	2.5
Pentecostal	15.0	5.3	2.8	2.7
Other Protestant	10.2	19.4	N/A	N/A
Total Christian	N/A	N/A	49.0	49.4

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ In the case of the NCLS, denomination is an attribute of the church attended by the respondent. ² Percentage of valid responses. In the NCLS, the proportion missing on each question ranged from 1.4% to 3.6%. In the ACS, a response was required on each question, except for religious identification, which included “prefer not to say” as a response option. Some 4.8% of respondents selected this option.

.

2.2. Measures

This study uses the questions developed by Markus and colleagues (Markus and Arunachalam 2008; Markus and Arnup 2010) to measure social cohesion, with some modifications—namely the addition of neutral response options on symmetrical Likert scales.

Belonging: Three items were used to evaluate the belonging domain. Respondents were asked to what extent they take pride in the Australian way of life, to what extent they have a sense of belonging in Australia, and whether they agreed or disagreed that maintaining the Australian way of life is important.

Social justice and equity: Three items tested perceptions about social equity in terms of the evaluation of the gap between high and low incomes, the perception of Australia as a land of economic opportunity and whether those on low incomes receive enough government support. Respondents were also asked about their degree of trust in government.

Acceptance/rejection, legitimacy: Respondents were asked to agree or disagree about whether accepting immigrants from many different countries makes Australia stronger and whether the government should assist in maintaining the customs and traditions of ethnic minorities. A general question about experience of discrimination in the previous year was included, and finally, respondents were asked about their future prospects.

Participation: Respondents were asked to identify which, if any, of five forms of political action they had done over the last three years.

Worth: Two items operationalized the worth domain. The first asked about level of financial satisfaction and the second was an indicator of happiness over the last year.

The full wordings for the social cohesion measures are given in Appendix B.

2.3. Analysis

Two basic comparisons are conducted for all questions across the five domains of social cohesion. First the results from the two groups—Australian churchgoers and all Australians—are compared. Second, the responses from the five largest denominations of churchgoers are compared to see if type of religious community makes a difference—Catholic, Anglican, Baptist, Uniting Church and Pentecostal churches. Results for Pentecostal respondents should be treated with caution due to the low number of unweighted cases upon which they are based.

In these comparisons Chi-square tests are used to evaluate if any differences are statistically significant (p < 0.05). To quantify the magnitude of the effect, the effect size was calculated using Phi and Cramer’s V.

Comparisons for NCLS and ACS on indicators for each measure are given in

Table 3. Social cohesion indicators, NCLS versus ACS.

Domain, Measure and Indicator	Percentage		Effect Size 1
	NCLS	ACS
Belonging
Pride in Australian way of life: great/moderate extent	92	81	0.19 ***
Sense of belonging in Australia: great/moderate extent	96	84	0.31 ***
Maintaining Australian way of life is important: strongly agree/agree	81	72	0.13 ***
Social Justice and Equity
Income gap is too large: strongly agree/agree	70	72	0.07 **
Australia is a land of economic opportunity: strongly agree/agree	77	55	0.26 ***
People on low incomes receive enough support: strongly agree/agree	44	30	0.18 ***
Trust government to do right thing: almost always/most of the time	44	29	0.23 ***
Participation
Voted in an election	86	76	0.12 ***
Signed a petition	53	42	0.11 ***
Contacted an MP	20	14	0.09 ***
Joined a boycott	9	9	ns
Attended a protest	7	5	0.05 **
Acceptance/Rejection, Legitimacy
Accepting immigrants makes Australia stronger: strongly agree/agree	66	41	0.28 ***
Ethnic minorities should be given assistance: strongly agree/agree	27	17	0.20 ***
Experienced discrimination due to skin colour, ethnicity or religion	12	13	ns
Life prospects in 3–4 years: much improved/a little improved	49	33	0.26 ***
Worth
Satisfaction with financial situation: very satisfied/satisfied	69	35	0.37 ***
Happiness over last year: very happy/happy	82	52	0.33 ***

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ Phi for the full crosstabulation. * p < 0.05, ** p < 0.01, *** p < 0.001, ns indicates not significant, determined for the full crosstabulation.

, with full results showing individual response options and the effects of weighting given in

Table A2. Item wordings and results for social cohesion domains, NCLS and ACS (weighted and unweighted).

Domain and Measure 1	Percentage 3
	NCLS Weighted	NCLS Unweighted	ACS Weighted	ACS Unweighted
BELONGING
To what extent do you take pride in the Australian way of life and culture?
To a great extent	52.2	48.7	37.8	38.2
To a moderate extent	39.9	41.3	43.4	43.6
Only slightly	6.7	8.7	13.4	12.9
Not at all	1.1	1.4	5.3	5.3
To what extent do you have a sense of belonging in Australia?
To a great extent	77.0	74.4	49.2	49.8
To a moderate extent	19.5	21.6	35.1	35.1
Only slightly	3.2	3.5	10.6	10.4
Not at all	0.3	0.5	5.1	4.7
In the modern world, maintaining the Australian way of life and culture is important
Strongly agree	38.7	35.0	34.0	33.0
Agree	42.8	42.7	38.4	40.2
Neutral/unsure 2	14.5	17.4	21.6	21.1
Disagree	3.8	4.5	4.4	4.4
Strongly disagree	0.2	0.4	1.7	1.4
SOCIAL JUSTICE AND EQUITY
In Australia today, the gap between those with high incomes and those with low incomes is too large
Strongly agree	29.8	28.6	34.1	34.4
Agree	40.1	42.3	38.0	37.4
Neutral/unsure 2	23.8	22.9	20.9	21.1
Disagree	4.5	4.6	6.1	6.0
Strongly disagree	1.8	1.7	0.9	1.1
Australia is a land of economic opportunity where in the long run, hard work brings a better life
Strongly agree	26.4	21.8	12.5	12.9
Agree	50.2	51.9	42.6	42.7
Neutral/unsure 2	16.6	18.7	27.7	27.3
Disagree	5.5	6.1	12.9	12.8
Strongly disagree	1.3	1.5	4.3	4.4
People living on low incomes in Australia receive enough financial support from the government
Strongly agree	12.3	11.7	6.5	6.4
Agree	31.4	31.3	23.5	23.6
Neutral/unsure 2	27.6	30.1	30.5	29.7
Disagree	22.0	20.0	24.6	25.8
Strongly disagree	6.7	6.9	14.8	14.5
How often do you think the government in Canberra can be trusted to do the right thing for the Australian people?
Almost always	4.9	3.8	3.1	2.9
Most of the time	38.6	38.5	25.5	26.6
Only some of the time	47.7	49.6	47.3	46.8
Almost never	8.8	8.1	24.2	23.8
PARTICIPATION
The following are some different forms of political action people can take. Which, if any, have you done over the last three years or so? (Mark ALL that apply)
Voted in an election	86.1	86.2	76.4	78.0
Signed a petition	52.8	53.3	41.8	43.0
Written or spoken to a Federal or State MP	20.5	21.3	13.9	15.0
Joined a boycott of a product or company	8.5	9.7	9.2	9.3
Attended a protest, march or demonstration	7.1	6.5	4.6	4.7
None of the above 3	10.6	10.0	16.0	15.1
ACCEPTANCE/REJECTION, LEGITIMACY
Accepting immigrants from many different countries makes Australia stronger
Strongly agree	22.2	21.4	11.4	12.9
Agree	43.5	43.6	30.0	31.6
Neutral/unsure 2	25.9	26.1	33.8	31.5
Disagree	6.8	6.8	15.4	15.6
Strongly disagree	1.7	2.0	9.4	8.4
Ethnic minorities in Australia should be given Australian government assistance to maintain their customs and traditions
Strongly agree	6.1	5.8	3.0	3.1
Agree	21.4	20.8	13.7	14.7
Neutral/unsure 2	36.1	37.5	32.9	32.5
Disagree	26.3	25.4	27.2	27.5
Strongly disagree	10.2	10.5	23.1	22.2
Have you experienced discrimination because of your skin colour, ethnic origin or religion over the last 12 months?
Yes	11.6	11.8	12.8	14.1
No	88.4	88.2	87.2	85.9
In three or four years, do you think that your life in Australia will be:
Much improved	22.9	19.3	9.0	10.1
A little improved	25.9	28.0	24.3	26.1
The same as now	40.3	42.8	39.8	38.8
A little worse	9.3	8.4	20.0	18.7
Much worse	1.6	1.5	6.8	6.4
WORTH
How satisfied or dissatisfied are you with your present financial situation?
Very satisfied	18.5	20.2	5.8	6.0
Satisfied	50.6	50.1	29.3	30.4
Neither satisfied nor dissatisfied 2	17.8	17.1	26.8	27.2
Dissatisfied	10.3	9.9	26.2	24.9
Very dissatisfied	2.8	2.7	11.9	11.4
Taking all things into consideration, would you say that over the last year you have been…
Very happy	23.8	22.0	10.9	11.6
Happy	57.9	58.7	41.3	43.5
Neither happy nor unhappy 2	12.5	12.5	29.0	25.7
Unhappy	4.6	5.2	13.9	14.0
Very unhappy	1.1	1.5	4.9	5.2

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ Measures obtained from Markus and Arnup (2010). ² The inclusion of a neutral option differed from the Scanlon Foundation surveys. ³ Percentage of valid responses. In the NCLS, the proportion missing on each question ranged from 4.6% to 7.7%. In the ACS, a response was required on each question, except for political participation and experience of discrimination, which included “prefer not to say” as a response option. The 1.9% and 5.1% of respondents respectively who selected this option are treated as missing.

.

Table 4. Descriptive statistics for social cohesion indicators, NCLS denominational breakdown.

Domain, Measure and Indicator	Percentage					Effect Size 1
	Catholic	Anglican	Baptist	Uniting	Pentecostal
Belonging
Pride in Australian way of life: great/moderate extent	94	85	89	93	94	0.08 *
Sense of belonging in Australia: great/moderate extent	97	95	94	96	97	0.09 **
Maintaining Australian way of life is important: strongly agree/agree	86	68	74	82	81	0.11 ***
Social Justice and Equity
Income gap is too large: strongly agree/agree	70	68	69	80	68	ns
Australia is a land of economic opportunity: strongly agree/agree	80	65	72	72	81	0.11 ***
People on low incomes receive enough support: strongly agree/agree	44	42	49	35	43	0.08 *
Trust government to do right thing: almost always/most of the time	42	50	39	40	48	0.08 *
Participation
Voted in an election	85	88	83	93	84	ns
Signed a petition	46	59	54	60	61	0.14 ***
Contacted an MP	19	24	21	30	14	0.11 **
Joined a boycott	8	14	10	14	3	0.13 ***
Attended a protest	7	9	3	14	7	0.09 *
Acceptance/Rejection, Legitimacy
Accepting immigrants makes Australia stronger: strongly agree/agree	69	63	64	65	59	ns
Ethnic minorities should be given assistance: strongly agree/agree	34	25	24	24	14	0.11 ***
Experienced discrimination due to skin colour, ethnicity or religion	12	10	12	5	17	0.09 *
Life prospects in 3–4 years: much improved/a little improved	45	42	56	33	76	0.17 ***
Worth
Satisfaction with financial situation: very satisfied/satisfied	69	70	71	76	64	ns
Happiness over last year: very happy/happy	80	78	81	88	87	ns

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ Cramer’s V for the full crosstabulation. * p < 0.05, ** p < 0.01, *** p < 0.001, ns indicates not significant, determined for the full crosstabulation.

shows comparisons between NCLS denominations, with full results in

Table A3. Item wordings and results for social cohesion domains, NCLS denominational breakdown (weighted).

Domain and Measure 1	Percentage
	Catholic	Anglican	Baptist	Uniting	Pentecostal
BELONGING
To what extent do you take pride in the Australian way of life and culture?
To a great extent	57.0	45.5	45.5	50.1	52.1
To a moderate extent	36.8	39.9	43.3	43.3	41.5
Only slightly	5.0	12.3	9.2	5.2	6.4
Not at all	1.2	2.3	2.0	1.4	0.0
To what extent do you have a sense of belonging in Australia?
To a great extent	74.4	71.4	71.6	82.2	87.3
To a moderate extent	22.6	23.5	22.6	14.0	9.7
Only slightly	2.7	5.0	4.1	3.0	3.1
Not at all	0.2	0.0	1.7	0.8	0.0
In the modern world, maintaining the Australian way of life and culture is important
Strongly agree	40.5	35.8	32.6	34.2	45.1
Agree	45.8	32.3	41.7	47.5	36.1
Neutral/unsure 2	12.2	25.0	18.8	13.9	12.1
Disagree	1.5	6.9	5.7	3.6	6.7
Strongly disagree	0.0	0.0	1.2	0.7	0.0
SOCIAL JUSTICE AND EQUITY
In Australia today, the gap between those with high incomes and those with low incomes is too large
Strongly agree	30.2	29.2	27.1	37.4	30.9
Agree	40.1	39.1	41.5	42.2	36.9
Neutral/unsure 2	23.7	25.9	21.6	13.5	25.1
Disagree	3.8	4.9	8.6	3.9	6.3
Strongly disagree	2.3	1.0	1.2	3.0	0.8
Australia is a land of economic opportunity where in the long run, hard work brings a better life
Strongly agree	28.5	19.6	19.8	13.2	39.2
Agree	51.1	45.7	52.1	58.5	42.0
Neutral/unsure 2	15.5	25.3	21.1	15.2	11.5
Disagree	3.8	6.3	5.7	11.6	6.3
Strongly disagree	1.0	3.2	1.4	1.5	0.9
People living on low incomes in Australia receive enough financial support from the government
Strongly agree	11.1	10.2	15.1	9.5	15.1
Agree	33.4	31.9	33.6	26.0	27.7
Neutral/unsure 2	27.1	29.1	29.2	31.2	25.5
Disagree	21.6	18.2	18.8	20.9	28.7
Strongly disagree	6.8	10.6	3.3	12.4	3.0
How often do you think the government in Canberra can be trusted to do the right thing for the Australian people?
Almost always	5.3	4.1	2.0	2.4	7.1
Most of the time	37.1	46.2	37.2	37.7	41.1
Only some of the time	49.9	41.0	53.9	43.0	40.0
Almost never	7.8	8.7	6.9	16.9	11.8
PARTICIPATION
The following are some different forms of political action people can take. Which, if any, have you done over the last three years or so? (Mark ALL that apply)
Voted in an election	85.4	88.5	82.7	93.1	84.4
Signed a petition	46.0	59.3	53.8	60.3	61.3
Written or spoken to a Federal or State MP	19.1	24.4	20.6	30.3	13.9
Joined a boycott of a product or company	7.9	14.2	9.8	14.1	3.0
Attended a protest, march or demonstration	6.7	8.7	3.1	13.6	6.9
None of the above 2	10.9	7.4	12.0	4.6	14.9
ACCEPTANCE/REJECTION, LEGITIMACY
Accepting immigrants from many different countries makes Australia stronger
Strongly agree	23.5	20.5	22.3	23.0	23.2
Agree	45.6	42.9	41.5	42.1	36.2
Neutral/unsure 2	22.8	26.4	29.0	28.1	30.2
Disagree	6.3	7.5	4.2	5.3	9.7
Strongly disagree	1.8	2.7	2.9	1.5	0.7
Ethnic minorities in Australia should be given Australian government assistance to maintain their customs and traditions
Strongly agree	7.6	6.3	6.8	3.8	3.2
Agree	26.3	18.9	17.3	20.6	10.8
Neutral/unsure 2	35.9	36.7	38.0	42.0	30.8
Disagree	22.3	24.5	25.6	22.6	43.4
Strongly disagree	7.8	13.5	12.3	11.0	11.8
Have you experienced discrimination because of your skin colour, ethnic origin or religion over the last 12 months?
Yes	11.7	9.6	12.3	5.5	16.9
No	88.3	90.4	87.7	94.5	83.1
In three or four years, do you think that your life in Australia will be:
Much improved	20.6	13.2	21.4	11.1	51.2
A little improved	24.6	29.2	34.3	21.9	24.8
The same as now	43.4	49.4	35.9	53.9	14.4
A little worse	10.2	7.2	7.1	11.7	6.4
Much worse	1.3	1.0	1.3	1.5	3.2
WORTH
How satisfied or dissatisfied are you with your present financial situation?
Very satisfied	16.2	22.5	23.0	24.6	16.4
Satisfied	52.5	47.5	48.3	51.1	47.9
Neither satisfied nor dissatisfied 2	18.2	15.4	18.2	14.7	19.3
Dissatisfied	10.9	11.4	7.6	9.6	10.9
Very dissatisfied	2.3	3.2	3.0	0.0	5.5
Taking all things into consideration, would you say that over the last year you have been…
Very happy	24.3	19.4	21.3	22.0	26.0
Happy	55.7	59.0	59.4	66.4	61.2
Neither happy nor unhappy 2	14.9	12.3	13.1	7.0	6.6
Unhappy	4.3	7.4	5.2	3.9	4.9
Very unhappy	0.8	1.9	1.0	0.7	1.4

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ Measures obtained from Markus and Arnup (2010). ² The inclusion of a neutral option differed from the Scanlon Foundation surveys.

.

3. Results

3.1. Comparisons between Churchgoers and All Australians

Indicators of belonging were high for both samples. Churchgoers had higher levels of social cohesion in this domain when compared to all Australians (Table 3). Differences were statistically significant on all three measures, and were strongest for sense of belonging, where 96% of churchgoers reported a sense of belonging to a great or moderate extent, compared with 84% of all Australians. Some 92% of churchgoers reported pride in the Australian way of life to a great or moderate extent (81% of all Australians), and 81% of churchgoers agreed that maintaining the Australian way of life and culture is important (72% of all Australians).

Churchgoers were more likely than all Australians to agree that people on low incomes receive enough financial support from government (44% versus 30%), to see Australia as a land of opportunity (77% versus 55%), and to trust the government to do the right thing by the Australian people (44% versus 29%). Churchgoers were slightly less likely than all Australians to agree that the difference between those with high and low incomes is too large (70% versus 72%).

Social participation in terms of political action varied depending on the type of action. With the exception of boycotts, for which there was no significant difference, churchgoers reported higher levels of political participation than all Australians, although the magnitude of the effects were small compared with the other social cohesion domains.

Churchgoers were more likely than all Australians to agree that accepting migrants from many different countries makes Australia stronger (66% versus 41%), and to support government assistance for ethnic minorities to maintain their customs and traditions (27% versus 17%). Churchgoers were also more positive about life prospects, with 49% expecting their life to be improved in the next several years, compared with 33% of all Australians. There was no difference between the two samples in terms of incidence of discrimination in the previous 12 months.

Worth indicators were much higher among Australian churchgoers when compared to all Australians, displaying the highest effect sizes across all domains. Some 69% of churchgoers were satisfied with their present economic situation, compared with 35% of all Australians, and 82% indicated that they had been happy over the previous year, compared with 52% of all Australians.

3.2. Comparisons between Denominations

An examination of denominational differences shows significant differences across 13 of the 18 social cohesion measures (Table 4).

In the belonging domain, particularly high levels of social cohesion were observed among Catholics for the importance of maintaining the Australian way of life and culture (86% strongly agree or agree) and pride in the Australian way of life and culture (57% to a great extent, 37% moderate extent). At 68%, the result for Anglicans who agreed with the importance of maintaining the Australian way of life was relatively low and slightly below that for all Australians (72%). Anglicans also reported the lowest levels of pride among the denominations (85% to a great or moderate extent). The result for sense of belonging in Australia was very strong among Pentecostals: 87% experienced a sense of belonging to a great extent, 10% to a moderate extent.

With regard to the social justice and equity domain, relatively low levels of agreement were observed among Uniting Church attendees on three of the four measures. The difference across denominations on the question of whether the gap between those on high and low incomes is too large was not statistically significant. Pentecostals and Catholics were the most likely groups to agree that Australia is a land of economic opportunity where hard work brings a better life (81% and 80% respectively).

The highest levels of political participation were reported by Uniting Church attendees. Pentecostals were relatively low on contacting a member of Parliament and participating in boycotts (14% and 3% respectively).

Attitudes regarding cultural difference were particularly positive among Catholics and least positive among Pentecostals. At 34%, the proportion of Catholics who agreed that ethnic minorities should be given government assistance was very strong, compared with 14% of Pentecostals and 17% of all Australians. Some 69% of Catholics agreed that accepting immigrants from many different countries makes Australia stronger (59% of Pentecostals, although the differences between the denominations were not statistically significant for this item). At 17%, Pentecostals were the most likely denominational group to report discrimination. Expectations for an improved life in the next several years ranged from a low of 33% of Uniting Church participants to a high of 76% of Pentecostals.

There were no significant differences between denominations in the worth domain.

4. Discussion

What role does Christian congregational involvement have in relation to social cohesion? As was anticipated, this study comparing churchgoers with all Australians shows that church attendance makes a contribution to higher levels of reported social cohesion across all five domains: belonging, worth, participation, social justice/equity and acceptance. The difference was particularly strong for the worth domain (satisfaction with financial situation and happiness) and for the measure of sense of belonging in Australia. Religious participation generates social capital, leading to higher social cohesion.

Our findings for worth mirror the consistent reports in the literature that religious service participation enhances subjective wellbeing (Koenig et al. 2012). The strength of the result for the specific measure of sense of belonging in Australia (stronger than pride in the Australian way of life and the importance of maintaining the Australian way of life) warrants further investigation.

The results concerning the acceptance domain, which we expected from Allport’s (1954) contact hypothesis, are consistent with Markus and Arnup’s (2010, p. 77) finding for the effect of attendance on views toward migration, in which the most positive views were expressed by those who attended religious services at least monthly. The Scanlon Foundation surveys have also explored more specific views about cultural diversity, such as attitudes towards people from particular religions (e.g., Markus 2012) and whether there should be discrimination in the intake of migrants on the basis of religion (e.g., Markus 2015). Whether the positivity towards cultural diversity remains among churchgoers when the issues are sharpened is a question for future research.

Higher social cohesion among churchgoers on the economic measures in the social justice and equity domain indicates a relatively high degree of economic conservatism among churchgoers, who were more likely than all Australians to consider that hard work can bring a better life and that people on low incomes receive enough financial support from the government. This finding is consistent with the decades-long trend of a strongly conservative vote among churchgoers (Pepper et al. 2019). Allegiance with conservative parties may not be due to those parties’ neoliberal economic policies but rather to matters such as same-sex marriage and abortion; when it comes to views about public policy, it is on matters of the family that churchgoers think their churches should be most active (Powell and Pepper 2014). However, voters often take their cues on policy matters from the parties with which they identify (Brader and Tucker 2012; Margolis 2018). It is interesting that the evaluation of the gap between people on high and low incomes was similar for churchgoers and all Australians, given the strength of the result for churchgoers on the two other economic questions.

Do members of different denominations vary in their views? NCLS data was used to compare five denominations: Catholic, Anglican, Baptist, Uniting Church and Pentecostal churches. Statistically significant differences were observed across most measures of social cohesion, except for those in the worth domain. However, with a few exceptions, cohesion among participants from each denomination was still higher than that observed for all Australians. Catholic and Pentecostal attendees expressed the strongest feelings of belonging, and Anglicans the lowest. Uniting Church participants were relatively low on social justice and equity, but still higher than Australians at large. Pentecostals and Catholics were the groups most likely to think that Australia is a land of economic opportunity. The Catholic Church has the highest proportion of migrants among the large denominations, which may explain aspirational economic views. Among Pentecostals, in addition to the impact of their higher levels of migrant churchgoers, emphasis on this-worldly rewards coming to the faithful (Bowler 2013; Hunt 2000) may explain the finding. The high political participation in the Uniting Church is consistent with a relatively strong emphasis on advocacy in the Uniting Church, whose councils repeatedly take public stances on matters of social justice and environmental issues and invite church members to do likewise (Pepper and John 2014; UCA NSW & ACT Synod n.d.). A high proportion of Pentecostals also signed petitions, which may reflect the young age profile of this denomination (McAleese et al. 2018). Catholics’ strong support for migration and multiculturalism could again reflect the high proportion of migrants among Catholics, but also the decades of solid teaching in support of welcoming and accepting others of difference faiths following the promulgation of Nostra Aetate (Paul VI 1965). The Uniting Church has the oldest age profile among the large denominations (McAleese et al. 2018), which explains the low expectation of improved life circumstances. The very high expectations reported by Pentecostals is likely to be due in part to the young age profile and larger immigrant population in Pentecostal churches. The extent to which demographic differences explain the variations between the denominations on social cohesion is a subject for future research.

To return to the questions raised in the introduction to this paper, social cohesion is a conservative construct. Whose interests feature in its definition? Whose power is being conserved? How is change to be managed? On the issue of interests and power, churchgoers are relatively well educated and thus relatively prosperous. They are conservative in their political outlook, and in a country which has seen conservative governments in power for a large majority of the years since Federation, their interests appear to be well-served. Churchgoers are “good citizens” whose trust of government is relatively high and who are more active than the wider population in using the democratic channels that are available to them to enact social change. They embrace (ethnic) diversity to a relatively high degree. These results suggest that they may be relied upon to help strengthen or maintain social cohesion when it is under threat. But, with their adherence to the status quo, they are unlikely to be a significant source of agitation to prevent some of the greatest contemporary threats to social cohesion before their worst excesses are felt. For example, the transgression of ecological limits and global climate change, which if unchecked are likely to undermine the stability and sustainability of our society (IPCC 2014; Steffen et al. 2017, 2018).

There are limitations of this study that relate to sampling in both surveys. First, the ACS was not a probability sample and cannot be claimed to be representative. Second, there are sampling limitations associated with the NCLS, including the non-probabilistic nature of Protestant samples, the under-representation of certain groups and the lower likelihood that people with lower levels of formal education complete the surveys. We consider each of these issues in turn.

To address the limitations with the ACS sample, we compared the NCLS results with the Scanlon Foundation Social Cohesion survey results from 2016 (Markus 2016), although this is hampered by methodological differences. Generally, the comparison shows a similar direction of the effect of church attendance, with the exception of political participation, which may have been due to the different mode of delivery of the participation question, which is the only multiple response question in the social cohesion set. It is striking that, even though neutral responses were not prompted in the Scanlon survey, on the questions with symmetrical response options (agree/disagree, satisfied/dissatisfied etc.), the results for churchgoers on the combined positive response options (e.g., strongly agree plus agree) were similar in most cases to the Scanlon results. This is further evidence of the strength of social cohesion among churchgoers.

Protestant NCLS data are nonprobability samples. There is a likely participation bias related to church health, which may have been partially corrected through weighting (Pepper et al. 2018). It is unclear whether or how this bias relates to social cohesion. While weighting addresses the low participation of Pentecostal churches in the overall results, small unweighted numbers of Pentecostal respondents means that the results for this group should be treated with caution. It is possible that people with lower levels of formal education were undersampled in the NCLS, due to literacy difficulties. No substantial investigation has been conducted to date on education biases in NCLS datasets, although some research does suggest that those with lower levels of formal education may be less likely to fill out the survey forms to completion (Pepper and Leonard 2016, p. 6). However, a test of ACS versus NCLS results for social cohesion at the same level of education (school, trade qualification, university degree) shows that higher social cohesion persists among churchgoers in almost all cases (

Table A4. Social cohesion indicators, NCLS versus ACS by education level.

Domain, Measure and Indicator Education Level	Percentage		Effect Size 1
	NCLS	ACS
BELONGING
Pride in Australian way of life: great/moderate extent
School	95	81	0.23 ***
Trade certificate/diploma	93	80	0.23 ***
University degree	90	83	0.15 ***
Sense of belonging in Australia: great/moderate extent
School	97	84	0.33 ***
Trade certificate/diploma	97	83	0.37 ***
University degree	96	87	0.26 ***
Maintaining Australian way of life is important: strongly agree/agree
School	85	73	0.16 ***
Trade certificate/diploma	86	73	0.17 ***
University degree	75	69	ns
SOCIAL JUSTICE AND EQUITY
Income gap is too large: strongly agree/agree
School	71	75	0.17 ***
Trade certificate/diploma	71	73	ns
University degree	69	65	ns
Australia is a land of economic opportunity: strongly agree/agree
School	78	54	0.30 ***
Trade certificate/diploma	81	56	0.30 ***
University degree	72	57	0.17 ***
People on low incomes receive enough support: strongly agree/agree
School	43	28	0.22 ***
Trade certificate/diploma	45	31	0.20 ***
University degree	43	33	ns
Trust government to do right thing: almost always/most of the time
School	41	24	0.25 ***
Trade certificate/diploma	46	28	0.23 ***
University degree	45	39	0.16 ***
PARTICIPATION
Voted in an election
School	84	77	0.08 **
Trade certificate/diploma	92	75	0.23 ***
University degree	86	77	0.10 **
Signed a petition
School	43	40	ns
Trade certificate/diploma	58	42	0.16 ***
University degree	60	45	0.14 ***
Contacted an MP
School	16	9	0.10 ***
Trade certificate/diploma	23	18	ns
University degree	24	18	0.07 *
Joined a boycott
School	5	7	ns
Trade certificate/diploma	11	10	ns
University degree	11	12	ns
Attended a protest
School	3	3	ns
Trade certificate/diploma	6	5	ns
University degree	11	8	ns
ACCEPTANCE/REJECTION, LEGITIMACY
Accepting immigrants makes Australia stronger: strongly agree/agree
School	56	35	0.28 ***
Trade certificate/diploma	66	38	0.30 ***
University degree	75	59	0.18 ***
Ethnic minorities should be given assistance: strongly agree/agree
School	24	12	0.24 ***
Trade certificate/diploma	23	17	0.17 ***
University degree	33	26	0.11 *
Experienced discrimination due to skin colour, ethnicity or religion
School	6	10	0.08 **
Trade certificate/diploma	15	14	ns
University degree	15	17	ns
Life prospects in 3–4 years: much improved/a little improved
School	43	29	0.27 ***
Trade certificate/diploma	47	31	0.25 ***
University degree	57	45	0.19 ***
WORTH
Satisfaction with financial situation: very satisfied/satisfied
School	67	34	0.37 ***
Trade certificate/diploma	64	33	0.34 ***
University degree	75	41	0.35 ***
Happiness over last year: very happy/happy
School	79	46	0.36 ***
Trade certificate/diploma	82	54	0.33 ***
University degree	84	62	0.26 ***

Source: Powell et al. (2016), Powell and Pepper (2016). ¹ Phi for the full crosstabulation. * p < 0.05, ** p < 0.01, *** p < 0.001, ns indicates not significant, determined for the full crosstabulation.

). While education factors may reduce the size of differences between church and the wider community, the differences remain.

Our findings could be further validated through additional analysis of the Scanlon data. Pooling the national Scanlon samples from 2009 to 2013 (the five surveys in which religious service attendance was asked, Dorman 2009; Blackmore and Steel 2010; Blackmore and Balasubramanian 2011; Blackmore et al. 2012; Blackmore and Balasubramanian 2013) should yield a large enough sample of regular Christian churchgoers to examine denominational differences for most of the large denominations, provided there is sufficient stability in the social cohesion measures across survey waves. However, Pentecostalism was not a religious group that was prompted in the surveys.

In addition to investigating attitudes to religious diversity more specifically, other research questions that could also be explored in future include the relative contribution of different types of factors (such as demographics) to social cohesion in both church and community datasets, including the extent to which demographics explain denominational differences. A more detailed analysis could also be conducted on the NCLS data to examine the ways in which age, religiousness and strength of bonding of the individual to the congregation influence social cohesion. Social cohesion items could also be included in future waves of the NCLS in order to track changes over time and to explore relationships with other constructs such as worldviews. Replicating the items would also allow these current findings to be further validated.

In summary, this research has found significant differences between the expressions of churchgoers and those of the general population on issues related to social cohesion. This suggests that, on balance, Christian groups play a positive role in the promotion of social cohesion, building both bridging and bonding capital among those who participate.