Global Studies of Religiosity and Spirituality: A Systematic Review for Geographic and Topic Scopes

Abstract

This paper advances the global study of religiosity by conducting a systematic review of the geographic scope, religious traditions, levels of analysis, and topics investigated within contemporary scientific studies of religion, paying particular attention to intersections with generosity. The analysis builds upon a meta-analysis of 30 years of scientific studies of religion that was published ten years ago and engages a similar framework to analyze the most recent ten years of research on religiosity and spirituality. Specifically, this analysis codes for the potential for Western-centrism, Christian-centrism, and congregational-centrism, all while attending to ways to study the potential intersection between religiosity and generosity, especially during the formative youth development life stage. Two data sources inform this analysis: the international data catalog of the Association for Religious Research Archives (ARDA) and the Journal for the Scientific Study of Religion (JSSR). The results indicate that centrism remains, though perhaps to a lesser extent than in the previous decades, with the notable exception of a remaining inequality in the geographic scope. Implications for research are discussed, including practical implications to implementing a better geo-tagging process to more overtly identify the scope of data and make U.S. scope less implicit.

1. Introduction

This paper aims to advance global studies of religiosity and spirituality, and their intersection. Within the United States, at least, many studies have established a link between religiosity and generosity during the formative years of child and youth development. For example, when adult Americans explain how they learned to give, they often refer retroactively to key experiences during their childhood or adolescence when they observed people giving within a religious setting, or describe having seen their parents be religious and participate in giving in generous ways, to religious or other causes (e.g., Herzog and Price 2016). Notably, the causal directions are complex in this relationship: people could learn to give through religious experiences, or people who are more inclined to give could select into religious participation as part of a broader prosocial orientation.

Being precise in identifying the presence, or lack of presence, of a causal relationship between religiosity and generosity often requires scholars to distinguish religious forces from other social forces. In quantitative approaches, there are distinctions between studies that conceive of religiosity as the primary independent variable that predicts social outcomes, versus a non-religious social force as the primary independent variable that predicts religious outcomes (for more on this: Smilde and May 2010 conducted a meta-analysis of 30 years of publications in five social science journals that categorized analytical designs into religiosity as the independent versus dependent variable).1

However, the reality is that religious and social forces often co-occur in complex ways and can mutually develop throughout the life course. The potential for causal bidirectionality does not negate the ability to study religious and social forces quantitatively, nor empirically. Rather, the existence of such complexities underscores the need to rely on information amassed from multiple data sources and engaging all methodologies. This does not mean that each individual study needs to engage both quantitative and qualitative designs, in a mixed methods approach (e.g., Headley and Plano Clark 2020; Hesse-Biber and Johnson 2015), but rather that the field generally needs to harness data from diverse inputs and analytical approaches, in order to study multiple angles of complex issues.

Indeed, contemporary approaches to social science more often move beyond unproductive paradigmatic or methodological debates, typically grounded in ‘either/or’ logics, to account for more robust, ‘both/and’ approaches. To do so, meta-analyses of the field are crucial for summarizing potential for biases in theoretical approaches, topics, geographic scopes, and methodological tendencies more generally. Grounded on years of existing research, meta-analyses can also identify fruitful avenues for future studies to address identified issues. Notably, Cadge et al. (2011) provided a meta-analysis of the scientific study of religion nearly a decade ago. Their analysis identified major biases with existing approaches in the field, namely wide-spread tendencies toward: (1) Western-centrism: focusing geographically on data collected in the Northern America (especially the United States) and Western Europe, and then assuming broader generalizability; (2) Christo-centrism: focusing topically on Christianity, and to the extent that other religions are studied, they are included in a comparative fashion, and often within populations where non-Christian religions are the minority (e.g., the United States and Western Europe); (3) congregational-centrism: focusing primarily on the organizational sites of religious congregations, such as how individuals relate to congregations, leaders form religious communities, and/or how congregations are shaped by larger national or denominational cultures; and (4) positive-centrism: focusing primarily and uncritically on religiosity as a positive social force rather than assuming that, as with any social force, there are potentials for both negative and positive associations (e.g., religious wars), and even co-occurrence of both (e.g., strict churches providing members a sense of community by clearly excluding others from participation: McKendry-Smith 2016; Ellis et al. 2016; Olson and Perl 2001).

Within this context, the current paper seeks to de-center and re-center the study of religion, as identified by Cadge et al. (2011), specifically by conducting systematic searches that intentionally seek to: (1) provincialize the United States and Western Europe by focusing on geographies; (2) limit tendencies toward Christian-centrism by parsing studies by topic and approach, naming the identifiable religious traditions within each study; (3) amass studies that investigate religiosity from a range of levels and approaches, especially outside of congregations alone; and (4) critically engage the multiple ways in which religiosity can relate to prosocial actions, including promoting anti-social actions. The findings of this review indicate that several centrist trends and inequities persist and continue to plague the scientific study of religion, highlighting the need for more critical attention.

2. Research Questions

This paper is interested in assessing what existing contemporary approaches exist in globally studying the potential intersection between religiosity and spirituality during youth development. Before addressing that intersectional question, it is first necessary to account for each topic in turn. This paper focuses in particular on existing global studies of religiosity and spirituality. The analysis aligns with the critiques of the field offered a decade ago by Smilde and May (2010) and Cadge et al. (2011) through iterating a series of hypotheses developed from their de-centering and re-centering framework. The first research question motivating the paper is: in the decade since 2010, has the science of religion de-centered and re-centered? This question is investigated empirically by operationalizing the three needed re-centers: geographic scope, religious tradition scope, and congregation scope. The hypotheses are built on the assumption that the ten years of research since these calls for change should have resulted in the desired changes. Thus, the three hypotheses are:

Hypotheses 1 (H1).

Since 2010, the geographic scope of scientific study of religion is not U.S.-dominant.

Hypotheses 2 (H2).

Since 2010, the tradition scope of scientific study of religion is not Christian-dominant.

Hypotheses 3 (H3).

Since 2010, the topic scope of scientific study of religion is not congregation-dominant.

The null hypothesis for each of these is that the status quo in 2010 remains relatively unchanged in the decade of research since. Only after assessing the state of scientific studies of religion and spirituality, the focus of this paper, can a broader research question that attends to the intersection with generosity and philanthropy be addressed (the focus of another paper in this Special Issue).

3. Methods

The method engaged in this study is a systematic review of existing data sources. The overall approach, data sources, and measures are described further below.

3.1. Approach

First, the approach of this paper is to present the results of a systematic collection and meta-analysis of existing studies: datasets and publications (specified below). Rather than relying solely on the most prevalent and readily accessible studies to shape the topics and approaches of the reviewed studies, the database was constructed by employing recursive geographic keyword searching that de-prioritized studies focused solely on the United States or Western Europe. The heavy emphasis on Western geographies in the broader field presented some challenges to achieving this goal, which are explained further in the methods section. Additionally, inconsistencies in the ways countries and world regions are tagged in studies, especially publications, burdened the search requirements. To contribute to limiting this issue in future studies, the process developed in this paper is detailed below to facilitate broader replication, and the discussion section interprets and applies this process.

Second, broader tendencies in the field toward Christian-centrism (a high degree of attention to Christian religions alone) are limited in this paper through systematically tagging religiosities alongside geographies. Of particular importance, the iterations between discovery and systematic phases of reviewing revealed a major issue with available searches. The issue is that the United States is particularly unlikely to be named in readily accessible meta-data of studies, especially in publications, namely: the title, abstract, and keywords. Indeed, even after considerable review of the full text, even the details of the methodology section, it is often still not obvious in most U.S.-based publications that the geographic scope is the U.S. Complicating matters, the keywords can range from United States to U.S., USA, and one of the most common notations is American, which requires extra scouting to identify as meaning the U.S. and not other American or Northern America countries, e.g., Canada. Reminiscent of media crime coverage listing non-white racial identities more often, despite more white perpetrators of crime in the U.S., the first indication that a publication is U.S.-specific is the absence of a geo-tag. The same is often true of Christianity. To limit implicit biases in tagging absences, the included publications were explicitly tagged for geography and tradition.

Third, to advance beyond heavy congregational emphasis, the numerous studies in the U.S. that focus on congregations are intentionally excluded, whenever possible. Fourth, the discussion section critically considers how these centrism issues may be inadvertently perpetuated and identifies opportunities for fruitful future approaches, including advancing the role of open access, international journals in promoting a global community of scholars.

3.2. Data Sources

For the purposes of this analysis, the scientific study of religion is operationalized through two primary data sources: the Association for Religion Data Archives (ARDA) and the Journal for the Scientific Study of Religion (JSSR). Both of these are well-regarded, interdisciplinary outlets for scientific study of religion research, and both have made concerted efforts to expand beyond the United States to include international scholarship. The ARDA website catalogs and provides data downloads for a wide variety of studies, largely collected through surveys. The data archive presents users the option to browse datasets by category. The JSSR is the journal for the Society for the Scientific Study of Religion (SSSR), which is an interdisciplinary scholarly association supporting sociologists, political scientists, economists, psychologists, and others studying religion.

3.3. Measures

The primary measures for this analysis are indicators of: geographic scope, religious tradition scope, congregational scope, and topic scope. Each of these are discussed further below.

3.3.1. Geographic Scope

Several existing schemas for country groupings exist, and the third step involved an attempt to synthesize these approaches into a geo-tagging framework that could be applied consistently. Appendix A reviews five geographic schemas to exemplify inconsistencies across studies. The conclusion of the geographic scope analysis presented in the Appendix A is that, in order to achieve consistency in world region groupings, the ideal is to tag studies at the country level. Country-level geo-tags facilitate counts and comparisons by country, and also undergirds the ability to group multiple studies under a shared world region schema. If achieved, it would be possible to categorize countries according to multiple world regions schemas to facilitate cross-study learning. Thus, geographic scoping is important for establishing a field of global studies of religion and spirituality that assembles knowledge across a range of approaches. In the interim,

Table 1. Country to World Region Clusters in the United Nations World Statistical Area Codes (n = 222).

Northern America (n = 5 Countries)	Latin America (n = 52 Countries)	Africa (n = 60 Countries)	Asia (n = 50)	Europe (n = 50)
United States	Central America (n = 8)	Northern Africa(n = 7 countries)	Central Asia(n = 5)	Eastern Europe(n = 10)
Canada	Mexico	Algeria	Kazakhstan	Belarus
Greenland	Belize	Egypt	Kyrgyzstan (Kyrgyz Republic)	Bulgaria
Bermuda	Costa Rica	Libya	Tajikistan	Czechia (Czech Republic)
Saint Pierre and Miquelon	El Salvador	Morocco	Turkmenistan	Hungary
	Guatemala	Sudan	Uzbekistan	Poland
Oceania (n = 5)	Honduras	Tunisia	Eastern Asia (n = 7)	Moldova, Republic of
Australia	Nicaragua	Western Sahara	China (also Tibet, Taiwan)	Romania
New Zealand	Panama	Sub-Saharan Africa(n = 53 countries)	Hong Kong	Russian Federation (Russia)
Melanesia	South America(n = 16)	Eastern Africa(n = 22)	Macao	Slovakia
Fiji	Argentina	British Indian Ocean Territory	North Korea (Democratic	Ukraine
New Caledonia	Bolivia	Burundi	People’s Republic of Korea)	Western Europe(n = 9)
Papua New Guinea	Bouvet Island	Comoros	South Korea (Republic of Korea)	Austria
Solomon Islands	Brazil	Djibouti	Japan	Belgium
Vanuatu	Chile	Eritrea	Mongolia	France
Micronesia	Colombia	Ethiopia	South-eastern Asia(n = 11)	Germany
Guam	Ecuador	French Southern Territories	Brunei Darussalam	Liechtenstein
Kirbati	Falkland Islands (Malvinas)	Kenya	Cambodia	Luxembourg
Marshall Islands	French Guiana	Madagascar	Indonesia	Monaco
Micronesia, Republic of	Guyana	Malawi	Lao People’s Democratic	Netherlands
Northern Mariana Islands	Paraguay	Mauritius	Republic (Laos)	Switzerland
Palau	Peru	Mayotte	Malaysia	Southern Europe(n = 16)
U.S. outlying islands	South Georgia and	Mozambique	Myanmar (Burma)	Albania
Polynesia	the South Sandwich Islands	Reunion	Philippines	Andorra
American Samoa	Suriname	Rwanda	Singapore	Bosnia and Herzegovina
Cook Islands	Uruguay	Seychelles	Thailand	Croatia
French Polynesia	Venezuela	Somalia	Timor-Leste	Gibraltar
Niue	Caribbean(n = 28)	South Sudan	Vietnam (Viet Nam)	Greece
Pitcairn	Anguilla	Tanzania, United Republic of	Southern Asia(n = 9)	Holy See (Vatican)
Samoa	Antigua and Barbuda	Uganda	Afghanistan	Italy
Tokelau	Aruba	Zambia	Bangladesh	Malta
Tonga	Bahamas	Zimbabwe	Bhutan	Montenegro
Tuvalu	Barbados	Middle Africa(n = 9)	India	Macedonia, North
Wallis and Futuna Islands	Bonaire, Sint Eustatius and Saba	Angola	Iran	Portugal
	British Virgin Islands	Cameroon	Maldives	San Marino
	Cayman Islands	Central African Republic	Nepal	Serbia
	Cuba	Chad	Pakistan	Slovenia
	Curacao	Congo	Sri Lanka	Spain
	Dominica	Democratic Republic of the Congo	Western Asia(n = 18)	Northern Europe(n = 15)
	Dominican Republic	Equatorial Guinea	Armenia	Aland Islands
	Grenada	Gabon	Azerbaijan	Channel Islands
	Guadeloupe	Sao Tome and Principe	Bahrain	(Guernsey, Jersey, Sark)
	Haiti	Southern Africa(n = 5)	Cyprus	Denmark
	Jamaica	Botswana	Georgia	Estonia
	Martinique	Eswatini (Swaziland)	Iraq	Faroe Islands
	Montserrat	Lesotho	Israel	Finland
	Puerto Rico	Namibia	Jordan	Iceland
	Saint Barthelemy	South Africa	Kuwait	Ireland
	Saint Kitts and Nevis	Western Africa(n = 17)	Lebanon	Isle of Man
	Saint Lucia	Benin	Oman	Latvia
	Saint Martin	Burkina Faso	Qatar	Lithuania
	Saint Vincent and the Grenadines	Cabo Verde	Saudi Arabia	Norway
	Sint Maarten	Côte d’Ivoire (Ivory Coast)	Palestine, State of	Svalbard and Jan Mayen Islands
	Trinidad and Tobago	Gambia	Syrian Arab Republic	Sweden
	Turks and Caicos Islands	Ghana	Turkey	United Kingdom of Great Britain
	United States Virgin Islands	Guinea	United Arab Emirates	and Northern Ireland
		Guinea-Bissau	Yemen
		Liberia
		Mali
		Mauritania
		Niger
		Nigeria
		Saint Helena
		Senegal
		Sierra Leone
		Togo

Source: Author compilation from UN (2020).

lists the world regions.

3.3.2. Religious Tradition Scope

To assess the degree of Christian-centrism in the contemporary scientific study of religion, studies were tagged for any particular religiosities explicitly investigated. The coding for this is based entirely on the wording that scholars engage in describing the religious tradition(s) of their study. Not all studies explicitly name a religious tradition, and not all studies are focused on traditions. Moreover, relying on the coding employed by the existing publication is not an endorsement of this approach. In particular, more careful attention and coding is needed for indigenous religious traditions that do not fit neatly into major categories. Indeed, what is considered a religious ‘tradition’ versus indigenous religions is highly influenced by Western and Christian-centric notions influenced by colonialism. Nevertheless, the original religious tradition of the extant publication is employed.

3.3.3. Congregational Scope

The process for assessing congregational scope was first to tag the level for the unit of analysis. Levels of analysis are relatively commonly known and can be a helpful way to categorize and amass interdisciplinary scholarship into a larger body of knowledge (e.g., Barman 2017). In simple terms, macro-level units of analysis in social sciences typically focus on large-scale governmental, legal, and economic institutions, whereas micro-level units of analysis in social sciences typically collect data from individuals for the purposes of studying individual similarities and differences within the same population. Between those approaches are meso-level studies that collect data about organizations, social networks, or groups. Congregational studies are one type of meso-level approach, and thus it is necessary to assess the level of study before identifying the subset of congregational studies.

3.3.4. Topic Scope

The goal of this meta-analysis is to advance further study of the potential intersections between religiosity and generosity during youth development. Assessing the state of religiosity and spirituality research will facilitate a better linking with studies of generosity and philanthropy. One of the risks of interdisciplinary and intersectional research is that the depth of domain knowledge of a particular topic is traded for breadth. This can lead to scholars who are studying a topical intersection to recreate existing, even resolved, problems that are ladened within a field of knowledge to which the scholars are not exposed. This may already be occurring with the intersectional study of prosocial behaviors, in which religiosity and spirituality are sometimes studied but not typically by scholars who participate in the larger field of the scientific study of religiosity. This paper seeks to limit this issue by providing a thorough scan of existing research on religiosity.

4. Analysis and Results

This section presents the analysis and results of both data sources: ARDA datasets and JSSR publications since 2010. Each data source is analyzed for the three re-centering scopes and for topic.

4.1. ARDA Datasets

The top-level categories in the ARDA archives are: International Surveys and Data, U.S. Church Membership Data, and U.S. Surveys. The total number of objects cataloged in the international category is 116, whereas the two U.S.-focused lists catalog a combined 987 objects. This already indicates one answer to the first research question: there does not appear to be a balance in geographic focus among the datasets. To answer the sub-tier of this research question, identifying which countries are under-studied, this analysis focused on further parsing the first of these categories: international data. The sub-categories under international data are: (1) cross-national data, (2) multiple nation surveys, and (3) single nation surveys. Each of these are analyzed for geographic scope, religious tradition scope, congregation scope, and topic scope.

4.1.1. Cross-National Data

The ARDA cross-national data catalog returns 33 objects. However, a closer inspection reveals that the majority of these are multiple years of the same dataset. Several of the datasets compile data that were collected entirely before 2010. After parsing for both these parameters, there are six unique datasets, with data collected since 2010:

• ARDA National Profiles (Harris et al. 2010): includes data from the International Religious Freedom Reports (IRFR)
• Cross-National Socio-Economic and Religion Data (Harris et al. 2011): includes data from the Central Intelligence Agency (CIA) World Factbook, U.S. Department of State International Religious Freedom Report (IRF), and the United Nations Human Development Report (HDR)
• The Religion and State Project (Williams and Fox 2016)
• World Religion Project (Williams and Fox 2016; Williams et al. 2016): includes data from the Global Religion Dataset (GRD)
• Government Religious Preference 2.0 (Williams and Brown 2017): includes data from the Global Religion Dataset (GRD) and the Global Restrictions on Religion Data (GRRD)
• Religious Characteristics of States (Williams et al. 2017)

In reporting data across these sources, ARDA provides users with a feature for international datasets called National Profiles. This returns a hyperlinked list of all countries that populate a data dashboard that presents a brief national profile, pie chart of the largest religious groups in the nation, and reports for four major indexes: Religious Regulation Index, Religious Minority Discrimination Index, State Funding of Religions, and Societal Discrimination of Minority Religions. A brief, one-paragraph history and a map are also provided. The four data sources regularly drawn upon in these profiles are: The Religious Characteristics of States Dataset Project, and the Religion and State Project.

This indicates the importance that ARDA places on this cross-national category in synthesizing country data. It also indicates an emphasis on macro-level approaches that focus on religion and the state, with governmental, political, legal, and economic institutions as the primary unit of analysis. In fact, all six of these datasets are macro-level approaches that focus on religious demography by studying religious affiliation and categorizing national and world region populations by number of adherents within each major world religion. Notably, all of these datasets treat world religions and adherents versus non-adherents as mutually exclusive categories into which all individuals within each country can be aggregated.

Geographic Scope

In terms of geographic scope, the datasets include 175–252 countries within all major world regions. For example, the Cross-National Socio-Economic and Religion Data project (Harris et al. 2011) reports that the data are from 252 countries, which is more countries than many sources list as the total number of countries in the world. This is due to the fact that many territories of nations are coded as unique geographic entities. For example, the Christmas, Keeling, and Norfolk islands are all listed as individual entities, whereas other sources (such as the UN) group these within Australia. Additionally, the UN groups within Melanesia the following: Fiji, New Caledonia, Papua New Guinea, Solomon Islands, and Vanuatu, whereas this dataset lists each individually. After parsing the 32 territories within nations, there are 220 non-territory countries, which aligns with the 222 countries reported by the UN (see Table 1). Initial support is found for Hypothesis 1: these datasets are not U.S.-dominant.

Religious Tradition Scope

As for religious tradition, these datasets readily return analyses of non-Christian traditions, alongside Christian traditions. For example, the World Religion Project (Williams and Fox 2016; Williams et al. 2016) undergirds and provides analysis of Animism, Atheism, Baha’i, Buddhism, Catholicism, Christianity, Confucianism, Hinduism, Islam, Jainism, Judaism, Shintoism, Sikhism, Syncretism, Taoism, and Zoroastrianism. Initial support is found for Hypothesis 2: these datasets are not Christian-dominant.

Congregational Scope

None of these datasets focus on congregations as a site of data collection. They all engage the same macro-level approach: collect micro-level data from individuals and aggregate individual data within national geographies to compare national level, between-nation trends. Initial support is found for Hypothesis 3: these datasets are not congregation-dominant.

Topic Scope

For the potential intersection between religiosity and generosity during youth development, no dataset provides the ability to study this. All six datasets are focused on adult populations, and none of the datasets include either: (a) measures of charitable giving, volunteering, other generous activities, or (b) measures of the philanthropic or nonprofit sector. Thus, despite the wide geographic scope within these religion datasets, the topics of the broader intersectional study are not supported in these sources.

In summary, the ARDA cross-national data catalog provides a wealth of data on macro-level data best suited for investigating religion and the state, based upon mutually exclusive categories for affiliation with major religious traditions, as well as being a religious adherent or not. These data do not facilitate the study of the potential intersection with generosity and philanthropy, nor a study of youth or the development of social orientations or actions.

4.1.2. Multiple Nation Data

The ARDA multiple nation data catalog returns 37 objects. However, the majority of these are multiple years of the same dataset or data that were collected entirely before 2010. After parsing for both these parameters, there are six unique datasets with data collected since 2010:

• Carnegie Middle East Governance and Islam (CMEGI: Tessler 2010): includes data from the Arab Barometer, several country-specific datasets funded by the National Science Foundation (Algeria, Jordan, Kuwait, Morocco, Palestine, and Yemen), and the World Values Survey (WVS)
• Party Variation in Religiosity and Women’s Leadership (PVRWL: Kassem 2010)
• Perception and Acceptance of Religious Diversity among the European Population (PARDEP: Pollack 2010)
• Caucasus Barometer (CB: Zurabishvilli 2012)
• World’s Muslims Data Set (WMD: Bell 2012): includes data collected by the Pew Research Center
• A Cross-National Survey of Muslim Attitudes (ACNSMA: Woodward 2013)

Geographic Scope

In terms of geographic scope, the datasets above include 7–26 countries within most major world regions: Africa, Asia, Europe, Asia, and the Middle East (but not Northern America, Latin America, and only one on Eastern Europe). In terms of the particular countries, the CMEGI includes data on: Algeria, Bahrain, Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Morocco, Palestine, Qatar, Saudi Arabia, Turkey, and Yemen. The PVRWL includes data on: Afghanistan, Albania, Algeria, Austria, Bahrain, Bangladesh, Belgium, Bosnia and Herzegovina, Comoros, Djibouti, Egypt, Germany, Indonesia, Israel, Italy, Jordan, Kuwait, Lebanon, Mauritania, Morocco, the Netherlands, Palestine, Senegal, Tunisia, Turkey, and Yemen. The PARDEP includes data on: Denmark, France, Germany, the Netherlands, and Portugal. The Caucasus Barometer (CB) includes data on three countries: Armenia, Azerbaijan, and Georgia. The WMD includes data on Afghanistan, Albania, Algeria, Azerbaijan, Bangladesh, Bosnia and Herzegovina, Egypt, Indonesia, Iran, Iraq, Jordan, Kazakhstan, Kosovo, Kyrgyzstan, Lebanon, Malaysia, Morocco, Niger, Pakistan, Palestine, Russia, Tajikistan, Thailand, Tunisia, Turkey, and Uzbekistan. Finally, the ACNSMA includes data on seven countries: France, Germany, Indonesia, Malaysia, Niger, Nigeria, and Senegal. This provides further support for Hypothesis 1: not U.S.-centric.

Religious Tradition Scope

As for religious tradition, these datasets readily return analyses of non-Christian traditions; in fact, their data are considerably more weighted toward Islam than any other religious tradition. Specifically, four of the six datasets—CMEGI, PARDEP, WMD, and ADNSMA—list Islam readily in the meta-data, whereas no other religious traditions are specified. Additionally, the CB includes Islam within the four most prevalent religious affiliations for the three countries studied: Armenia, Azerbaijan, and Georgia. The four most prevalent religious traditions are: the Armenian Apostolic Church (n = 2256 and 91% of the Armenian sample); Georgia, Russian, or Greek Orthodox Church (n = 2136 and 85% of the Georgian sample); Islam, Sunni Islam, and Shia Islam (n = 2065 and 98% of the Azerbaijan sample). This provides further support for Hypothesis 2: these datasets are not Christian-dominant.

Congregational Scope

None of these datasets focus on congregations as a site of data collection. They all engage either a macro-level approach (typically individual-level data aggregated within national geographies to compare national level, between-nation trends, mostly based on affiliation with a religious tradition) or a micro-level approach (typically within-nation studies of individual differences/similarities).

Notably, the Caucasus Barometer (CB) includes more measures of religiosity than only affiliation with religious traditions. For example, the survey includes a measure of: (a) religious salience that asks: how important is religion in your daily life?; (b) religious distrust that asks: please tell me, how much do you trust or distrust (country’s) religious institutions to which you belong?; (c) religious service attendance frequency that asks: apart from special occasions, such as weddings and funerals, about how often do you attend religious services nowadays?; (d) religious behavior that asks: how often do you fast when it is required by your religious traditions?; (e) religious intensity that asks: overall, how religious would you say you are?; (f) religious out-group contact that iterates a series of questions about religious traditions to which the respondent does not belong that ask: about how often do you personally have contact with members of the following religious groups [insert each out-group religion independently]?; (g) religious out-group objections that iterates a series of questions about reactions if a family member had major interactions with a member of a religious tradition to which the respondent does not belong that ask about both (i) business dealings and (ii) marriage that asks: how would you react if a family member [insert ‘did business with’ or ‘married’] a member of the following religious group [insert each out-group religion independently]?

These additional religiosity measures provide rich information beyond affiliation alone. For example, despite 85–98 percent of respondents indicating adherence with a religious tradition, a considerably smaller proportion indicated high religious salience: when asked how important their religion is in daily life, the following proportions said very important: 56% of Armenians, 45% of Georgians, and 32% of Azerbaijanis. This provides further support for Hypothesis 3: these datasets are not congregation-dominant in their approach to studying religiosity.

Topic Scope

In terms of the focus of the larger study: studying the potential intersection between religiosity and generosity during youth development, five of the six datasets do not include either: (a) measures of charitable giving, volunteering, other generous activities, or (b) measures of the philanthropic or nonprofit sector. As a notable exception, the Caucasus Barometer (CB) is the only of the six datasets that provides the ability to study charitable giving, with a question that states: ‘I will read a list of activities. Could you please tell me which of these activities have you been involved in during the past six months: made a contribution to a charity, including donations by SMS and giving money to beggars?’ Additionally, an average of 57% of all respondents said that they would strongly object to a family member marrying a member of an out-group religion. The highest objections were for marriage to an atheist person (61% strongly object), and marriage to a Jewish person (73% strongly object).

These additional religiosity measures reveal more rich contours than measuring religious tradition affiliation alone. Moreover, partnering these religiosity measures with sub-region, nationally representative data indicate considerable within-region variation, even among these three former Soviet Union countries that are generally considered to have a high degree of cohesion. While this may be true of certain measures, the religious salience (importance of faith) measure alone reveals important distinctions within the region by country. In terms of the ability to study youth development, all six datasets are focused on the adult population, and thus none provide the ability to study the potential intersection of religiosity and generosity during youth development.

In summary, despite the wide geographic scope within these religion datasets, the intersectional topical focus of the current study is not supported. The ARDA multiple nation data catalog provides a wealth of macro-level data best suited for investigating religion and the state and based upon mutually exclusive categories for affiliation with major religious traditions, as well as being a religious adherent or not (four out of six datasets). Despite the wealth of information, these data do not facilitate the study of the potential intersection of religiosity with generosity and philanthropy, nor a study of youth or the development of social orientations or actions. The two exceptions to this are the micro-level data in the PARDEP and CB datasets, yet the former does not include data on generosity or philanthropy, leaving CB as the single exception that provides data on the potential intersection, but only for adults.

4.1.3. Single Nation Data

The ARDA single nation international data catalog returns 46 objects. However, again the majority of these are multiple years of the same dataset (e.g., National Survey on Discrimination in Mexico, Caucasus Barometer on each of Armenia, Azerbaijan, and Georgia independently) and data that were collected entirely before 2010 (e.g., Brazil Religion Survey 2007, English Church Census 2005, National Congregations Study Switzerland 2008, Northern Ireland Life and Times Survey 2008, Project Canada 1995, Religion in Italy 1994, Religious Coping in Iran 2006, Scottish Church Census 2002, Spiritual Life Study of Chinese Residents 2007, Taiwan Social Change Survey 2009, the Comparative Study of Religious Experience in Taiwan 2009, and Typology of Dutch Catholic Parishes 2003). After parsing for these parameters, there are four unique datasets with data collected since 2010:

• Study of Mysticism in Chinese Buddhist Monks and Nuns (SMCBMN: Chen 2011)
• Noordin Top Terrorist Network Data (NTTND: Williams et al. 2011)
• Culturally Adapted Spiritually Oriented Trauma-Focused Cognitive Behavioral Therapy for Child Survivors of Restavek (CASOTFCBTCSR: Wang 2015)
• Israel’s Religiously Divided Society (IRDS-Pew: Williams 2015): data from the Pew Research Center

Geographic Scope

In terms of geographic scope, the datasets include a total of four countries within most major world regions, specifically: countries within Asia, Middle East, and the Caribbean (but not Northern America, Latin America, Europe, or Eastern Europe). In terms of the particular countries included in each study, the SMCBMN studied China; the NTTND studied Indonesia; the CASOTFCBTCSR studied Haiti; and the IRDS-Pew studied Israel. This provides some support for Hypothesis 1: the data are not U.S.-centric.

Religious Tradition Scope

As for religious tradition, these datasets readily return analyses of non-Christian traditions. Specifically, one studied Buddhism (Chen 2011), one studied Islam, Judaism, and Christianity (Williams 2015), and only one was predominantly Christianity (Wang 2015). The fourth dataset (Williams et al. 2011) does not appear to be coded by religious tradition. Instead, the focus on terrorist networks facilitates investigation of whether respondents received formal education or religious instruction within a series of universities, high schools, and boarding schools, such as: Adelaide, Airlangga, Al-Husein-pesantren-Indramayu, West Java, al-Irsyad, al-Islam-pesantren, al-Mutaqien-Indramayu, al-Muttagien-Jepara, Bogor Agricultural, Brawijaya, Darul Fitroh, Darusysyahada-pesantren, Gontor, Luqmanul Hakeim-pesantren. Only one of these is identifiably Christian: The Christian University of Malang. This provides additional support for Hypothesis 2: the datasets are not Christian-centric.

Congregational Scope

None of these datasets focused on congregations as a site of data collection. All but two engaged either a macro-level approach or a micro-level approach, as described above. The first two meso-level studies were introduced in this data catalog. One of these focused on religious leaders, specifically Buddhist monks and nuns (Chen 2011). The other focused on religious networks, as described in the religious tradition scope above. This provides additional support for Hypothesis 3: these datasets are not congregation-centric in their approach to studying religiosity.

Topic Scope

For the potential intersection between religiosity and generosity during youth development, no dataset appears to provide the ability to study this. All four datasets are focused on adult populations, and none of the datasets include either: (a) measures of charitable giving, volunteering, other generous activities, or (b) measures of the philanthropic or nonprofit sector.

In summary, despite the beyond U.S., Christian, and congregation-centric approaches of these datasets, the intersectional topics of the broader study are not supported. The ARDA single nation data catalog provides a range of data levels—spanning from macro to micro, and also with some meso-level attention to studying religious networks and groups. Notably, many of the datasets facilitate studying potential negative ramifications of religiosity. However, these data do not facilitate the study of the potential intersection of religiosity and generosity, nor a study of youth or life course development.

4.1.4. Summary: ARDA International Datasets

Cumulatively, the ARDA international catalog provides 17 datasets on non-U.S. geographies with data collected since 2010. These 17 datasets include seven datasets with data on every country with populations of 250,000 or more (and some with data on countries with smaller populations: see

Table A4. Countries in ARDA international datasets.

Country	Study Count	Country	Study Count
Algeria	3	Armenia	1
Egypt	3	Austria	1
Germany	3	Belgium	1
Indonesia	3	Comoros	1
Jordan	3	Denmark	1
Lebanon	3	Djibouti	1
Morocco	3	Georgia	1
Palestine	3	Israel	1
Turkey	3	Italy	1
Afghanistan	2	Kazakhstan	1
Albania	2	Kosovo	1
Azerbaijan	2	Kyrgyzstan	1
Bahrain	2	Mauritania	1
Bangladesh	2	Nigeria	1
Bosnia and Herzegovina	2	Pakistan	1
France	2	Portugal	1
Iran	2	Qatar	1
Iraq	2	Russia	1
Kuwait	2	Saudi Arabia	1
Malaysia	2	Tajikistan	1
Netherlands	2	Thailand	1
Niger	2	Uzbekistan	1
Senegal	2	Country Total	81
Tunisia	2
Yemen	2

Source: Author compilation from ARDA (2010–2020).

in the Appendix A for study counts by country). However, all of those datasets include religious tradition affiliation as the single measure of religiosity. In the 11 remaining datasets, a total of 81 countries are studied. Figure 1 maps the frequency distribution for each country included in these 11 international datasets (all non-listed countries = 0 for study count). This analysis provided an initial set of responses to the research questions. First, the attention to the United States and non-U.S. geographies is not equal, and countries within non-U.S. geographies receive unequal attention. Second, ARDA international datasets do not appear to be Christian-centric.

Specifically, it appears that only three of the 17 datasets focus on Christianity, all of these alongside other religious traditions, whereas seven appear to focus on Islam, three on Judaism, two on Confucianism and Hinduism, and two on a wide range of religious traditions. None of these 17 datasets have a congregational emphasis, largely due to the fact that 12 of the datasets are focused on macro-level religion and states institutional processes. An additional two datasets investigate meso-level aspects of religiosity, one with a focus on terrorist networks and the other with a focus on Buddhist monks and nuns. Three datasets investigate religion and the micro-level, one to study religion and wellbeing, another to study religious diversity, and a third (Zurabishvilli 2012) providing the only dataset that facilitates cross-national analysis of religiosity and generosity.

Cumulatively, the ARDA international datasets provide initial support for Hypotheses 1–3. However, when returning to the full data catalog, it is notable that international datasets comprise 116 of the catalog objects, whereas U.S. surveys comprise 960 objects, plus U.S. church membership data comprises another 27 objects, indicating that centrism remains within the full data catalog.

4.2. JSSR Publications

To build a database of relevant JSSR publications, the first step was to assess the total population size of all articles published since 2010. JSSR is published by Wiley, and the Wiley within-journal parser returned 881 objects since 2010, whereas Google Scholar returned 832 with the same parameters. The inconsistency appears to be due to two factors. One, Wiley included several objects that were published prior to 2010 but appear to have been uploaded online in 2011. Two, the search engines appear to have engaged in different treatment of non-article objects also published in the journal, such as book reviews, editorials, presidential addresses, erratum notes, acknowledgments of reviewers, and notes from the editor. To reconcile these, all objects were downloaded into a reference management software to parse articles from other objects. After removing non-article objects, the final sample of articles published since 2010 was 506.

4.2.1. Geographic Scope

The next step was to parse by geography. In both search engines, cross-checks were completed to ensure the accuracy of geographic keywords. To do so, Boolean logic was engaged to assess whether the presence and absence of a geo-term would tally to the same total sample. For example, “United+States” was sampled alongside “NOT+United+States.” In theory, if the articles were tagged for Boolean logic to be operative, a sample of 500 publications with 250 tagged as United States should return 250 as not United States. However, the logic did not uphold in practice. The geo-tag search indicated that 475 articles were tagged as United States, whereas 128 were returned for not United States, which does not match the total number of publications. Moreover, Africa returned 82 articles, yet engaging “Africa-American” in an attempt to remove articles with keyword African-American that are not about Africa, returned 23. Similarly, Asia returned 67 articles, whereas “Asian-American” returned 20 articles. In summary, the Boolean-logic identified samples did not align.

Complicating geo-tags further, the common abbreviation of U.S. returned 611 articles, and America 257. Some of these overlapped the United States sample above, and others did not. This indicates that geographies are not coded in a consistent way, by the journal or by individual scholars. Additionally, cross-checking for world regions alongside specific country names revealed a Venn diagram of both overlapping and mutually exclusive sets of articles. For example, in searching Africa and Egypt separately, some articles were returned only for Africa, others only for Egypt, and still others returned for both keywords. In summary, searching by geo-tag did not return reliable samples.

Yet another layer of geo-tagging obscurity was found in the ways that scholars use world region labels in inconsistent, and sometimes indeterminable, ways. Some scholars refer to the Middle East to describe countries such as Turkey, whereas others geo-tag the same country as Asia, Western Asia, or only Turkey. Similarly, Mexico is sometimes included in a general label for Latin America, and other times treated as part of the continent of North America. Additionally, some used terms such as the Global South, developing, developed, or less developed countries. Aside from the larger debates surrounding such terminology (e.g., Lewis et al. 1997), the problem for the purposes of this analysis is that the inconsistencies in these terms obscures clear identification of the geographic scope. To one scholar, Argentina could be included in less developed countries, whereas for another it is not. This issue is even more extreme for within-continent regions, and particular countries are especially vulnerable to being categorized in myriad ways across studies, for example: India, Mexico, and Israel.

Together, these issues indicate the need to search all world regions in terms of typical names for the world region, alongside names of each particular country within the world region. Conducting a systematic search in this way returned both true and false results, in terms of true being articles that focus on the intended geographies within the world region (e.g., Africa) and false being articles that engage a keyword for the world region (Africa, or in the case of Northern Africa: Middle East) or name a country within the article (e.g., Kenya) but do not focus on that world region or country as the scope of the data or analysis. For example, a scholar may describe Africa as an example of a larger trend that is discussed in the literature review but then proceed in the methods and analysis to focus on data with a United States scope. Additionally, a scholar may describe African origins of African Americans studied within a United States scope. Both of these constitute false positives for this geographic search.

It thus appears that the only way to parse true from false returns is to review all major sections of the full paper to engage cognitively intense, domain knowledge to identify the actual geographic scope of the data. The methods sections of publications are particularly helpful for this. Indeed, most scholars do not specifically address the geographic scope of the data until the methods section. Still more do not address the geographic scope anywhere in particular, but the table or analyses aid an evaluation of the geographic scope, if one offers the paper a deep, careful, thorough, and cumbersome review.

Yet another variety of a geographic scoping issue is that many of the publications reporting on large-scale, cross-national studies that accumulate data on many, even hundreds, of countries do not list the specific countries in the article. For example, a scholar explained that the World Values Survey (WVs) data were analyzed, and that the particular wave of the WVS employed contains 86 countries, but that the current analysis only includes the 80 countries that had a particular measure of interest, and that the final model includes results for 77 nations because three countries had to be included for a variety of reasons. This results in a sub-sample of WVS-included countries that, unless specifically listed within the article, are known only to the scholar. Since this is a pervasive problem, the examples below do not indict the absence but rather highlight exemplary publications that overcome this issue. These publications chart a fruitful way forward. For example, Olson and Li (2015) engaged WVS data and included a table within their article’s appendix that includes all national-level descriptive statistics for each country, which lists each of the 69 countries in the analysis sub-sample by name. Additionally, Shulgin et al. (2019) also analyzed WVS data and created a sub-sample of countries based on key constructs for their analysis. These construct restrictions reduced the total number of countries available to 16, and the scholars specifically name these countries in the article body and list them in a table.

To broaden from particular datasets, this issue is also pervasive in meta-analyses, such as the current paper, which focuses on existing studies as the data source. In synthesizing information across numerous publications, scholars often neglect the geographic scope of the reviewed articles. This is understandable, considering that most of the existing review articles are not focused on geographic scope within the reviewed articles. Nevertheless, this practice further contributes to geographic obscurities, as even review articles are not typically amassing knowledge from the field into larger meaningful geographies, nor enabling the ability to readily detect U.S.-centrism. Yet again, there are exemplary publications that evidence a fruitful approach. For example, Shor and Roelfs (2013) include in their meta-analysis of 312 publications a table (Shor and Roelf 2013’s Table 1) listing the country for each publication. Learning from these examples, all the countries in this paper are listed in Table 1.

Adopting these practices in a more wide-scale way in future studies, or in keyword tagging of existing studies, would aid in limiting the field from inadvertently continuing a Western-centric geographic focus. As it currently stands, this meta-analysis had to rely on a cumbersome procedure to seek to identify geographic scope. Due to the pervasiveness of these geo-tag complications, the current study focused on systematically parsing a smaller sub-set of publications. The goal was to achieve full geo-identification within two specific world regions, which, if successful, could be scaled to more world regions. The two world regions of focus for this systematic analysis are Africa and Asia. When discovered in the process of searching Africa and Asia, other world regions and countries were also tagged to facilitate later expansion to all world regions and countries. Thus, there are some results that include other geographies, yet those need to be thoroughly caveated as from publication samples that have not yet been systematically sampled for that particular geography.

4.2.1.1. Systematic Search for Africa

To systematically search for Africa, and account for all the complexities with geo-tagging discussed above, this analysis engaged a cumbersome but necessary approach of searching for the overall world region, alongside searching for each particular country within Africa. An initial search for Africa returned 19 publications, but some of those were focused on United States or Western Europe and merely mentioned Africa somewhere in the article (e.g., Hempel et al. 2012; Ribberink et al. 2017). Another was focused on Turkey and mentioned South Africa in the literature review (Gurses and Ozturk 2020) and thus was moved into the Asia and Middle East collections but out of the Africa collection. Systematically searching for each country in Africa identified several previously unidentified publications, while also returning several additional false positives. After parsing each of those articles through a thorough, substantive review of the content of the full-length paper, 35 were identified as geographically scoped to Africa.

In summary, at least half of the relevant sample of publications geographically scoped to Africa would have been missed in a search for Africa without also searching for each country name within Africa. Conversely, searching for Africa and each country name in Africa returned considerably more false hits than true hits. If not carefully parsed, this result could contribute to a general perception in the field that there are many more publications on Africa than there in fact are. This issue is discussed further in the cumulative systematic sample section below (Section 4.2.1.3).

4.2.1.2. Systematic Search for Asia

Mirroring the Africa search, the same procedures were engaged in systematically searching for publications that are geographically scoped to Asia. An initial search for Asia returned 64 publications, and the search for specific countries within Asia returned considerably more. For example, 157 were initially identified as potentially about countries in Eastern Asia: 83 for China, 11 for Hong Kong, 31 for Korea (North or South), 31 for Japan, 1 for Mongolia, and 0 for Macao. Similarly, searches for countries in south-eastern Asia returned several more for each country: 24 for Indonesia, 21 for Philippines, 17 for Vietnam, 16 for Singapore, 13 for Malaysia, 7 for Thailand, 3 for Cambodia, 2 for Timor-Leste, and 1 each for Burma, Myanmar, Lao, Cambodia, and Brunei.

However, this potential Asia sample was also plagued by the same false-returns issue described above, as many of these publications merely mentioned Asia, or a country in Asia, somewhere in the text but did not analyze data from an Asian geography. Moreover, many also stated Asian-American within data collected within the United States, again often not overtly naming the United States as the scope of the geography, sometimes anywhere and frequently not within the title, abstract, or keywords (even when including U.S. and American keyword varieties).

After parsing false returns, and thoroughly combing for true returns by searching for each country, the Asia sample contains 70 publications. Notably, some of the studies included in this sample were found only through the systematic search for Africa, revealing an Asian geography upon further inspection. This indicates that the relevant publications for Asia or Africa may still be under-represented, and that additional relevant publications may be revealed in subsequent systematic searches for the other world regions. Nevertheless, this sampling and parsing strategy returns a considerably more reliable collection of the number of publications in each area, and thus undergirds the comparison of publication frequency by world geography described below.

4.2.1.3. Cumulative Systematic Search

In the process of systematically searching for Africa and Asia, and not Northern America or Western Europe (NA-WE), several false returns were collected. To facilitate the replicability of this approach in future studies, the parsing process and results are described in detail in Appendix A. Only a high-level summary of the results is presented here. After parsing the false positives, there are a total of 91 publications identified as focusing on geographies outside of NA-WE alone. Of these, 83 attend to Africa and/or Asia, and only 65 of those are exclusively in non-NA-WE geographies. After disambiguating, 13 of these address Africa without Asia, while 48 address Asia without Africa. This analysis revealed that Asia was studied considerably more than Africa in JSSR since 2010, and Africa was more often studied in conjunction with Asia than without. JSSR articles remain considerably more focused on Northern Europe and Western Europe than other geographies. Of the 219 articles systematically sampled using Africa and Asia keywords, 128 were revealed to still focus on NA-WE geographies alone, leaving only 91 that studied geographies outside of, or in addition to NA-WE geographies: 35 in Africa and 70 in Asia. In particular, the United States remains considerably more studied in JSSR since 2010 than other geographies, and indeed more studied than all the countries in other world regions combined (Figure 2). Thus, Hypothesis 1 is rejected: U.S.-centrism persists.

4.2.2. Religious Tradition Scope

Within the total of 219 publications that were systematically sampled for Africa and Asia geo-tags, at least 40 were overtly focused on Christianity, 43 of these were overtly focused on Islam (16 of these studied Christianity and Islam together), 19 focused overtly on Protestantism, 23 focused overtly on Catholicism, 14 focused overtly on Judaism (7 studied Christianity and Judaism together), 12 focused overtly on Buddhism (six of these studied Christianity and Buddhism together), six studied atheism (two of these studied Christianity and atheism together), five focused overtly on Mormonism (three of these studied Christianity and Mormonism together), five focused overtly on Hinduism (four of these studied Christianity and Hinduism together), five focused overtly on ‘folk religions’ (two of these studied Christianity and folk religions together), three focused overtly on Taoism (two of these studied Christianity and Taoism together), three focused overtly on New Religious Movements (none of these also studied Christianity), two focused overtly on Confucianism (one of these studied Christianity and Buddhism together), two focused overtly on African spirituality (one of these studied Christianity and African spirituality together), two focused overtly on Unitarian Universalism (one of these studied Christianity and Unitarian Universalism together), two focused overtly on Mysticism (one of these studied Christianity and Mysticism), one focused overtly on Baha’i (also with Christianity), one focused on Jainism (also with Christianity), one focused on Shintoism (also with Christianity), one focused on Sikhism (also with Christianity), one focused on Neo-Paganism (also with Christianity). Taken together,

Table 2. Religious tradition in Africa-Asia JSSR articles.

Religious Tradition	#	# +Xn	# −Xn
Christianity	40	40	0
Protestantism	19	19	0
Catholicism	23	23	0
Islam	43	16	27
Judaism	14	7	7
Buddhism	12	6	6
Atheism	6	2	4
Mormonism	5	3	2
Hinduism	5	4	1
Folk Religions	5	2	3
Taoism	3	2	1
New Religious Movements	3	0	3
Confucianism	2	1	1
African Spirituality	2	1	1
Unitarian Universalism	2	1	1
Mysticism	2	1	1
Baha’i	1	1	0
Jainism	1	1	0
Shintoism	1	1	0
Sikhism	1	1	0
Neo-Paganism	1	1	0
21 total religions	191	133	58

Source: Author compilation from JSSR (2010–2020).

presents the degree of Christian-centrism within these Africa and Asia sampled articles.

The results indicate that 28 articles were not overtly about a particular religion or set of religions. Of the 191 that were, Christianity and Protestantism were studied more than any other religion (n = 59), yet Islam was also overtly studied in 43 articles. Notably, the majority of the articles studying non-Christian religions did so in a comparative fashion by also including Christianity alongside other religious traditions (n = 133). Only 58 of the articles studied a non-Christian religion alone, without also studying Christianity. Thus, Hypothesis 2 is rejected: religious tradition imbalance persists, with a remaining centeredness around Christianity, even within the reduced sub-sample of publications systematically searched for Africa and Asia.

4.2.3. Congregational Scope

Within the total of 219 systematically Africa and Asia sampled publications, 214 were coded for level (five were undetermined). The majority of these focused on micro-level units of analysis, typically individuals not aggregated into organizations, groups, or nations: 119 articles. Another 43 focused on macro-level units of analysis, typically nations to study governmental or cultural institutions and compare between nations. The remaining 52 articles focused on meso-level units of analysis, such as organizations, networks, and groups. As discussed above, this is the sub-set of particular interest to assess the degree of congregationalism. Of these meso-level articles (n = 52), 22 are explicitly focused on congregations or denominations as the unit of analysis.

The remaining 30 meso-level approaches are as follows: seven studied religious networks, five studied religious organizations as sites of civic engagement, four studied religion and education, three studied religion and occupations (e.g., religion and science, or religion in healthcare chaplains), two studied atheism as a form of religious community, two studied experimentally assigned groups and religious dynamics, one studied religion and gender, one studied religious mission trips, one studied religious engagement in international NGO development efforts, one studied religion and socio-athletic organizations, one focused on funding for religious research, one studied online web forums and religiosity, and one studied cross-border religious groups.

In summary, despite congregation sites remaining prevalent generally, congregationally focused studies are considerably less prevalent within publications systematically sampled for their attention to Africa or Asia. More than half attended to the micro-level, and another fifth attended to the macro-level. While religious affiliation is a prevalent measure in both macro and micro-level studies, people in these approaches are not explicitly grouped within organizational units, such as congregations. A quarter of the articles focused on organizations, and less than half of those attended to congregations. However, the wide array of other organizational approaches leaves congregational approaches as the single largest grouping within meso-level studies. Thus, the evidence for Hypothesis 3 is not conclusive, yet there is some indication that global studies are less congregation-centric than Northern American or Western European studies.

4.2.4. Topic Scope

The subset of publications that explicitly attend to the intersection between religiosity and generosity is relatively small. Within the set of Africa and Asia searched articles, only 27 obviously attended to this topic. There could be more with a more thorough review of the topic of every publication, but since the focus here is on the visibility of approaches, this initial sample is based upon most evident keywording, such as generosity, philanthropy, charitable giving, volunteering, civic engagement, helping, and prosocial behavior.

Within this subset of 27 articles studying religiosity and generosity, seven studied Africa, 10 studied Asia, 16 studied Northern America, 9 studied Europe, three studied Eastern Europe, and one studied Latin America without conjunction with the other world regions. Only three of these also attended to youth or emerging adults. Thus, based on this evidence, there is a dearth of existing studies investigating the intersection of religiosity and generosity during youth development. However, caution in over-generalizing this finding is warranted. This caution is discussed further below, as well as in other papers in this Special Issue.

4.2.5. Summary: JSSR International Publications

Cumulatively, the international scope of JSSR appears to still be limited. Figure 3 displays a map of the geographic scopes within the sample of systematically searched for Africa and Asia publications. Similar to Figure 2 (above), the densities in Figure 3 represent the number of articles with data for each country, and high density in the United States, Canada, and Western Europe reveal that these Western geographies remain prevalent, even within this subset of publications sampled due to their Africa and Asia keywords. The high density of United States and Canada publications was discussed above; also of note is the prevalence of Western Europe scoped publications.

Table A5. Western Europe in JSSR publication sample.

France	17
Germany	14
Netherlands	11
Belgium	10
Austria	9
Switzerland	8
Luxembourg	4
Liechtenstein	0
Monaco	0

Source: Author compilation from JSSR publications (2010–2020).

in the Appendix A indicates that the top four most prevalent countries in Western Europe are: France with 17 publications, Germany with 14, Netherlands with 11, and Belgium with 10. The map and table visualize the overall inequality in geographic scope by world region, as well as the within-region inequalities by country. To further visualize this key conclusion, Figure 3 maps only the Africa and Asia geographies systematically sampled within all JSSR publications since 2010. The high density of attention paid to China is evident (n = 21). Turkey (n = 9) and India (n = 7) are also more prevalent. Within Africa, the most-studied countries are: Tanzania (n = 7), Nigeria (n = 6), Uganda (n = 6), South Africa (n = 5), Egypt (n = 5), Kenya (n = 5), and Burkina Faso (n = 5).

In summary, within contemporary JSSR publications sampled based on Africa and Asia keywords and systematically parsed, the United States and Western Europe remain over-studied, while Asia and Africa are comparably under-studied. Moreover, within-region inequalities exist such that: within Northern America, the United States is studied considerably more than Canada; within Western Europe, the western-most countries are studied more than the eastern and southern countries; within the Africa and Asia combined sample, Asia is studied more than Africa; within Asia, China is studied considerably more than any other country, then Turkey and India; and within Africa, Tanzania, Nigeria, Uganda, South Africa, Egypt, Kenya, and Burkina Faso are studied more than the rest of the countries in Africa. This evidence indicates the need to reject Hypothesis 1: geographic inequality remains. Additionally, Hypothesis 2 is also rejected: Christian-centrism remains, though it has seemingly lessened in contemporary scholarship and is less of an issue than geographic scope. There is not enough evidence to support Hypothesis 3, though the subset of meso-level, organizational approaches to studying religiosity appear to be less centric on congregations than in the past, and this centrism is less of an issue than geographic and religious tradition scopes.

5. Discussion

This paper presented results of a systematic review of scientific studies of religion data sources. The larger goal of the broader project was to assess approaches to studying the potential intersection between religiosity and generosity during youth development. However, that larger goal was hindered by three issues. One, most studies investigating the potential for an intersection between religiosity and generosity assumed that a relationship would be positive, such that religiosity promotes generosity. Relatedly, many scholars that do not specialize in the study of religion presume that studying a potential intersection among these topics implies both that the curious scholar thinks that: (a) a relationship exists, and (b) that the relationship is positive. Indeed, in inviting scholars and practitioners to join a global youth development network engaged in attending to the potential intersection of religiosity and generosity, considerable explanation had to be repeatedly given to assure invitees that attending to the potential for an intersection did not necessitate that one exists, nor that it would be statistically significant, nor substantively meaningful, nor necessarily positive.

For lack of a better way to describe this, the team came to refer to this phenomenon as the religion “hibbie jibbies,” as no other social scientific topic seems to engender such an extreme and assumption-laden response. For example, it never is presumed when a scholar states that they are interested in studying the potential intersection between education and political affiliation that the scholar must be of a particular political persuasion and must necessarily be interested in promoting the political party for that affiliation by using education to foster more political party adherents. This seems absurd, and yet a long history of scholarship identifies a potential link between advanced levels of education and liberal political affiliation (e.g., Ladd and Lipset 1975; Gross and Fosse 2012). Certainly, scholars can bring their own personal backgrounds, interests, affiliations, and biases into scientific inquiry on any topic, within the natural or social sciences. However, why should the study of religiosity raise considerably more concern than other lines of inquiry? Indeed, some of the most interesting research in the social scientific study of religion identifies blatant and inadvertent issues with religiosity, and Emile Durkheim, widely considered to be one of the founding fathers of sociology, refers to religion as simultaneously the source of both the greatest social solidarity and social conflicts (e.g., Durkheim [1912] 2008), and he was an atheist himself. Similarly, Hervieu-Léger (2000) described religion as a chain of memory and tradition that bonds people in ‘emotional communities’ that affirm collective identities of ethno-religious groups. Acknowledging the solidarity felt by participants in such groups does not imply that the theological tenets of those affinities are valid, nor beneficial for society generally, but it does validate participant experiences.

More generally, this problematic assumption not only indicates logical flaws, but it also limits an influx of critical inquiry in the study of religiosity and generosity, independently and intersected. Additionally, if scholars only study religiosity when there is considerable reason for theorizing that a strong relationship exists between religiosity and another topic of interest, then the cumulative body of studies of religiosity will falsely inflate the documented intersection of religiosity and other social forces. Indeed, a healthy field, in any social science discipline, would be characterized by equal attention to religiosity and its intersection with other topics as any other social phenomenon. Only then can the lack of an intersection, or weak intersections, be identified. Plus, strong and negative relationships may exist. For example, it is logically feasible that religiosity contributes to anti-social behaviors at a considerably higher rate than pro-social behaviors, but it would be impossible to know this if empirical studies only investigate the intersection in situations when religiosity has a high likelihood of promoting pro-social actions. In summary, scholars who are not experts in the study of religiosity risk ‘recreating the wheel’ by engaging religiosity in ways that are only assumed as being positive, or by ignoring the role of religiosity completely. Both approaches are problematic, and general scholarship needs to more fully attend to scholarship on religiosity, which better accounts for the ways in which religiosity is also a neutral and negative social force.

Setting external issues aside, problems also exist within the field of scientific studies of religion. As Cadge et al. (2011) and Smilde and May (2010) identified about a decade ago, the field has had historical issues with several centric tendencies, including unequal attention to Western geographies, Christian religious traditions, and congregational sites of data collection. This study investigated a set of hypotheses that these issues should have lessened or equalized in the ten years since identification. The evidence across both data sources presents some mixed results. The ARDA international data catalogs returned some initial evidence in support of changes to these historical trends. However, when shifting from data to publications, the JSSR article sample revealed continuation of several of the trends. In particular, a high degree of Western centrism remains prevalent, even in articles geo-tagged for Asia and Africa. The most concerning finding related to geographic scope is that the inconsistencies in the ways countries and world regions are tagged obscures the considerable inequalities, in terms of under-studied world regions and in terms of within-region inequities in between-countries attention. Indeed, there appears to be a level of tokenism, in particular countries being regularly studied as representative of their larger world region, while other countries within that region receive considerably less or no attention.

The other centric tendencies also remain prevalent to some degree, but it appears that they may be somewhat less intense than in studies conducted prior to 2010. Clarity on this issue hinges on gathering studies across all levels of analysis. For example, focusing only on the U.S.-specific datasets in ARDA, or the general population of JSSR publications, would support an impression of congregational-centrism and Christian-centrism. However, focusing only on the international data catalog in ARDA, or the country and world region tagged publications from geographies other than the United States and Western Europe, reveals a greater degree of studies that investigate religiosity across multiple levels, with all studies at the macro and micro-levels by necessity not focusing on congregations as the unit of analysis. When attending specifically to meso-level studies that are systematically sampled to parse out Western scopes, there is also more of a range of approaches to other types of organizational sites that do not necessarily focus on congregations. This issue deserves more attention, and another paper in this Special Issue reviews this topic in greater depth. Most importantly, in terms of the emphasis in the current study, there are few extant studies that facilitate an adequate global inquiry into the potential intersection between religiosity and generosity during youth development. Nevertheless, several approaches are relevant and can inform such an inquiry. Likewise, there is a considerable dictionary of available measures, indices, scales, and nationally normed survey questions to draw upon in constructing future studies on this topic.

5.1. Limitations and Future Studies

Despite the important results gleaned from this analysis, several limitations exist. Most notably, the analysis was conducted in English and sampled publications were all published in English, at least. Though several of the publications were also published in a second language, the requirement to minimally be published in English could limit the results in important ways. However, English publications are the most likely to inform scholars located in English-speaking countries about the data and results generated within other locales and thus the analysis represents the biases present within publications written in English. It is entirely possible that scholars outside the U.S. and Western Europe are publishing research in journals that are relevant to those geographies, and not in the sources sampled for this analysis. For example, a sociologist conducting research on India could publish the results in Contributions to Indian Sociology rather than JSSR. For this reason, a second paper in this Special Issue expands the scoped sources beyond the initial two presented in this paper, and thus continues the de-centering and re-centering process begun in this analysis.

In future studies, it will be important to build upon the initial steps of this study, especially to scale the scope of the approach begun here to include all world regions and countries in a systematic search. Similarly important will be to expand to additional data sources. ARDA and JSSR provide an initial view into the trends of the broader social scientific study of religion, but they are also limited insofar as no one particular source can represent all the available studies on religiosity. For example, other important sources include the Journal of the American Academy of Religion (JAAR), Review of Religious Research (RRR), Journal of Religion (JOR), and Religions. Additionally, journals that could facilitate a focus on the intersectional topic in particular include the Nonprofit and Voluntary Sector Quarterly (NVSQ), Voluntas, and Nonprofit Management and Leadership. For these searches, a full list of religiosity-related keywords will be especially important, which is the matter that a subsequent paper in this Special Issue attends to as a next step for scaling the parameters of the systematic search begun here. Moreover, there are a range of disciplinary specific journals that could also be included, such as the sociological approaches in the Sociology of Religion journal (SOR), psychological approaches in the Psychology of Religion and Spirituality (PRS), or business and management approaches in the Journal of Management, Spirituality, and Religion (JMSR). Attention to these sources would further advance the systematic approach begun here and eventually enable a comparison in scopes across data sources. Moreover, expanding beyond publications in English, and synthesizing non-English publications within a translated meta-analysis published in English, would considerably advance the field.

5.2. Conclusions

In conclusion, this analysis provides a systematic review of the geographic scope, religious traditions, levels of analysis, and topics investigated within ten years of contemporary scientific studies of religion. Important databases were examined and reviewed in order to extract and retrieve contemporary knowledge about religious traditions all over the world and the relationship of these traditions to generosity. Western-centrism, Christian-centrism, and congregational-centrism were categories that were especially examined in this study. In particular, the investigation included the examination of: (1) available international data derived from ARDA catalogue and (2) the Journal for the Scientific Study of Religion (JSSR) archives. The findings indicate that several centrist trends identified about a decade ago persist, despite concerted critical attention being paid to the inequities that plague the scientific study of religion. The findings indicate that embedded practices contribute to the persistence of these inequities, most notably several limitations with accurately geo-tagging publications based on country scope, and larger disagreements and inconsistencies regarding world region clustering naming and inclusions. It seems that a few relatively simple practical steps could be taken to lessen these issues, such as scholars being intentional about identifying the geographic scope of their paper, especially when it is focused on the United States, and preferably in the readily available meta-data: title, keyword, abstracts. Additionally, journal editors could request geo-tagging during the publication process, and editorial boards may be of service in aiding journals and their associations in enacting other procedures that would more explicitly limit tendencies toward the geographic, tradition, or congregational centrism described here. Additionally, journal databases, such as Wiley, Sage, JSTOR, and Google Scholar, could geo-tag publications for wider consumption.

While the focus of this study was centrally on religiosity and spirituality, this paper has implications for the study of the intersection between religiosity and generosity. Namely, the analysis reveals a problematic assumption, among scholars that are not experts on religiosity, that religiosity and generosity are naturally related. This not only indicates logical flaws, but also limits an influx of critical inquiry in the study of religiosity and generosity as conducted in many cases by researchers who have a priori ideas that religiosity and generosity are naturally and strongly inter-related. In addition, the authors contend that if scholars only study religiosity, then there is a natural bias in favor of the theory that a strong relationship exists between religiosity and generosity. This may result in a cumulative body of studies of religiosity that will falsely inflate the documented intersection of religiosity and generosity. Thus, scholars need to examine the relationship between religiosity and generosity and not only analyze it on the basis of preconceived assumptions about when a relationship is likely to exist and be positive. This review aids that endeavor.