CHAPTER 3

Modernizing Questionnaire Survey Research and Ethnography

Traditional research methods for studying culture in the humanities and cultural anthropology were qualitative, rather than quantitative, and early anthropologists tended to assume that each distinct culture was uniform within itself and thus ethnography did not require statistical sampling procedures. However, for 70 years, the Human Relations Area Files at Yale University have collected quantitative data “to promote understanding of cultural diversity and commonality in the past and present.”1 The journal World Cultures emphasizes a related methodological approach, the analysis of “aggregate cultural data in coded form on a large sample of societies.”2 For decades, cultural anthropologists have been aware that no culture is totally uniform and that the degree of uniformity varies greatly.3 Survey researchers may cogently argue that their methods offer both a broader and more rigorous perspective than that possessed by an ethnographic field researcher who is embedded among some of the people under study. However, in recent years, surveys have faced increasing difficulty studying systematic samples of the population, and their advantages measuring large-scale public opinion may be of limited value in studying subcultures or the realities of daily life. In addition to online questionnaires, there now are many kinds of rather systematic data, even quantitative, available for researchers, using Internet as a virtual hilltop from which to observe cultures.

Variations on the Theme of Musicology

Since 1972, the General Social Survey (GSS) has been a mainstay of sociological research, although its character changed somewhat as it matured, and only briefly did it collect data about music. Originally, it could be described as a tool of the Social Indicator movement that wanted to maximize the benefit of U.S. government programs and policies, and therefore needed a reliable way to assess the condition of the general population. In so doing it would supplement surveys by the Census Bureau and engage academics in the debates about where social life in America was going. To accomplish this, it would periodically send interviewers to the homes of a cluster sample of Americans to ask the same questions to see if the answers were changing, such as “Last week were you working full time, part time, going to school, keeping house, or what?” In 1972 this question was asked of 1,613 respondents, and by 2016 the total over the years had reached 62,447. “Keeping house” had dropped from 26.8 percent to 9.9 percent, reflecting the gradual entry of more women into the workforce.

All the GSS data are available online for anybody to analyze, at the Survey Documentation and Analysis (SDA) website:

SDA is a set of programs for the documentation and Web-based analysis of survey data. SDA was developed, distributed and supported by the Computer-assisted Survey Methods Program (CSM) at the University of California, Berkeley until the end of 2014. Beginning in 2015, CSM is managed and supported by the Institute for Scientific Analysis, a private, non-profit organization, under an exclusive continuing license agreement with the University of California.4

In 1995 I was honored to manage the initial funding through National Science Foundation grant 9422785:

This is a prototype Internet service for the General Social Survey employing NCSA Mosaic. The project will develop a system to provide enhanced access to survey data, using the General Social Survey for implementation of these integrated services, which will subsequently be extended to a variety of other survey data sets. These services will provide facilities for hypertext viewing and searching of complete survey documentation, customized and documented extracts from data sets, statistical analysis, and File Transfer Protocol delivery of full or extracted data sets. The General Social Survey is an ideal source of survey material to develop the system, because it is a highly diverse large dataset of complex structure, extensively documented in terms of publications based on each item, and has already been the basis of more than three thousand scientific publications and dissertations.5

NCSA Mosaic was the first widely available web browser, and this project was a pilot study on how the web could be used for educational purposes and as an active archive for scientists and scholars.

The Social Indicator movement was somewhat controversial, because significant fractions of the public and the political world were opposed to government social activism, but it also turned out that social scientists wanted the GSS to ask many new kinds of questions that they could analyze for academic publications. Therefore, the GSS began adding topical modules in addition to a somewhat limited set of constant social indicators. In 1993, the GSS included a culture module with a section measuring preference for styles of music that can illustrate some of the factors that shape cultural science. The GSS asked respondents to rate each of 18 genres of music on this scale from 1 to 5: dislike it very much, dislike it, mixed feelings, like it, or like it very much. Table 3.1 reports correlations among eight of these genres, selected because their structure offers generalizable insights. The correlation coefficients are tau-b, in which an item correlates 1.00 with itself and 0.00 with another genre with which it has absolutely no relationship, and has negative correlations if liking one genre implies hating the other. The rows and columns of the table are arranged in terms of the correlations with classical music.

Table 3.1 Preference correlations among music genres

	Classical	Opera	Broad-way	Jazz	New Age	Gospel	Rock	Metal
Classical	1.00	0.50	0.40	0.23	0.17	0.04	0.02	–0.01
Opera	0.50	1.00	0.38	0.20	0.17	0.09	–0.01	0.01
Broad-way	0.40	0.38	1.00	0.22	0.08	0.08	0.05	–0.10
Jazz	0.23	0.20	0.22	1.00	0.25	0.02	0.22	0.10
New Age	0.17	0.17	0.08	0.25	1.00	–0.12	0.26	0.34
Gospel	0.04	0.09	0.08	0.02	–0.12	1.00	–0.17	–0.18
Rock	0.02	–0.01	0.05	0.22	0.26	–0.17	1.00	0.32
Metal	–0.01	0.01	–0.10	0.10	0.34	–0.18	0.32	1.00
Like it	49.6%	21.9%	52.6%	51.1%	18.0%	59.5%	55.1%	11.2%
Answered	1,529	1,510	1,515	1,553	1,300	1,559	1,544	1,526

Not surprisingly, there is a strong 0.50 correlation between liking classical music and liking opera. However, they are not equally popular, 49.6 percent of the respondents saying they liked classical music, compared with only 21.9 percent who like opera, combining the “like” and “like very much” responses. Broadway musicals are American, rather than European as in the case of most classical music, but drew upon the classical tradition. Classical music has essentially no correlations with gospel, contemporary rock, or heavy metal music, which have their own distinctive correlation patterns. While 1,529 respondents answered the question about liking classical music, only 1,300 answered the one about New Age, probably because some respondents were not familiar with it. Those who did respond gave it rather low ratings, only 18.0 percent liking it, while it correlated moderately with both contemporary rock, which got a favorable 55.1 percent preference score, and heavy metal, which was least popular at 11.2 percent.

The most popular musical genre in this set of eight, liked by 59.5 percent of the respondents, was gospel music. It lacks strong positive correlations, but has negative correlations with new age, contemporary rock, and heavy metal. Here we encounter again one of the most difficult problems in the analysis of preference data. Respondents may vary in terms of response bias, some tending to give favorable ratings and others giving more unfavorable ratings, which would have the effect of inflating positive correlations and deflating negative ones. Among the many ways cultural science can deal with response bias is to consider other items as well. For example, a standard GSS item asks “How often do you attend religious services?” Of the 250 respondents who never attend religious services, 44.4 percent like gospel music, compared with 84.4 percent of the 135 who attend more than once a week. Of those who never attend religious services, 19.8 percent like heavy metal, compared with only 3.1 percent of the most frequent attenders. The example is a good one, because it is easy to comprehend in cultural terms. Obviously, gospel music is connected to traditional religion, while heavy metal is secular.

A second general observation is that the traditional plan of administering a questionnaire or interview to a statistically representative random sample of the general population may be too costly but also not necessary, if the goal is to explore cultural differences. The current NSF grant supporting the GSS cost fully $15,726,165.6 This has supported two main GSS surveys, four GSS components of the International Social Survey Program, plus “designing an extensive, rigorous experiment comparing an Internet GSS survey to the standard GSS.” Already in 2001, Tom W. Smith, the GSS leader, reported the results of comparing Internet-based and in-person surveys, finding somewhat limited differences in his particular dataset.7 By 2014 he was advocating a multilevel, multisource (ML-MS) approach to improving survey research, at one point commenting:

Human societies are complex, multi-faceted entities and social-science research designs need to measure that complexity. A key requirement is contextualization. Most people live in households which are nested in neighborhoods, communities, and countries. They are not isolated individuals, but interact with and are notably influenced by their families, neighborhoods, communities, nations, etc. Surveys need to collect information on each of these levels from the individual to the nation so the contexts in which people live are understood and that through multi-level analysis the impacts of these different levels can be measured and modeled.8

In 1998 I joined the team led by James Witte that deployed two vast online studies, Survey2000 and Survey2001, to explore issues of how well online respondents represented the general population.9 In 2002 I published an exploration of this difficult topic, using a subset of the data based on responses from children aged 13 to 15.10 The main finding was that often, if not always, subsamples with different biases may show roughly the same correlations between variables. Of course, members of a well-organized subculture may exhibit much finer distinctions between elements of their own culture than do outsiders, which implies that cultural scientists may often decide to use purposive samples rather than random samples. The early-teenage data offered a first step toward this methodology because the respondents had been recruited in different ways. Of 2,942 respondents aged 13 to 15, 51.3 percent completed the questionnaire at home, largely recruited voluntarily through the National Geographic websites, while 40.6 percent did so at school, mostly in an assignment that National Geographic had organized at two schools in each U.S. state and Canadian province. Table 3.2 reports results for music preferences, comparing these two very different samples.

Table 3.2 Music preferences from an online survey

	Like or like very much		Correlation with female
	Home	School	Home	School
Classic, symphony, and chamber music	51.7%	23.4%	0.08	0.02
Opera	14.4%	7.8%	0.12	0.00
Broadway musicals/show tunes	46.5%	29.1%	0.29	0.20
Jazz	45.8%	33.1%	0.00	–0.04
Big band/swing	50.0%	42.7%	0.09	0.03
Mood/easy listening	37.6%	26.5%	0.09	0.09
Country and western	17.6%	18.4%	0.11	0.06
Bluegrass	10.3%	9.5%	0.05	–0.11
Hymns/gospel	26.1%	20.1%	0.12	0.08
Rhythm and blues	41.3%	33.8%	–0.01	–0.02
Rap/hip-hop	41.3%	61.4%	0.07	–0.01
Dance music (e.g., electronica)	45.8%	54.2%	0.16	0.24
Caribbean (e.g., reggae, soca, calypso)	38.7%	31.1%	–0.02	–0.03
Latin (e.g., mariachi, salsa)	23.2%	17.0%	0.06	0.06
Music of ethnic/national tradition	38.2%	27.9%	0.07	0.07
Modern folk/singer-songwriter	28.2%	16.3%	0.17	0.07
Contemporary pop/rock	67.1%	53.7%	0.07	0.15
Alternative rock	65.2%	61.7%	0.06	0.01
Oldies/classic rock	54.5%	38.2%	0.12	0.09
Heavy metal	25.2%	37.2%	–0.08	–0.18

The first four music genres are listed in the same order as in
Table 3.1, starting with classical music that is vastly more popular among children who responded at home rather than in school. Two decades ago, when these data were collected, prosperous and well-educated households were much more likely to have good Internet connectivity, but even today we might expect big social class differences in the orientation of families toward elite or classical artistic culture. Going down the list of 20 genres, only 4 show significantly greater popularity among the children responding through the computers at their schools: rap/hip-hop, dance music (e.g., electronica), alternative rock, and heavy metal. Country and western music is about equally popular. However, as in the case of New Age mentioned earlier, some genres are not universally familiar. While there are variations between the two sets of respondents in the gender difference in a genre’s popularity, it is noteworthy that the two clearest cases of female preference, Broadway musicals/show tunes and dance music (e.g., electronica), were preferred differently by the home versus school respondents.

The higher ratings for several genres among children who responded online at home may reflect Richard A. Peterson’s omnivore theory. He proposed that people vary in the diversity of their cultural tastes, some being omnivores who like many different styles, and others being univores whose interests are very tightly focused.11 In connection with Witte’s research team, Peterson wrote an article collaborating with John Ryan about how music preferences might be evolving, given the ease of downloading it from Internet, perhaps rendering everybody omnivores. In their introduction, they noted that musical culture has already undergone a major transformation:

As recently as two centuries ago, all popular music was embodied. The medium of transmission was the human voice as well as the breath or hand on the instrument. Transmission and reception were face-to-face and full of potential for intimacy and interactivity. Music was “live” and disappeared as soon as it was performed. The only way in which to hear music was to attend when and where a performance was taking place, and so most people knew practically nothing about other people’s music.12

From that perspective, by allowing everybody to listen to all kinds of music, Internet may be erasing the cultural boundaries between musical genres and ethnic traditions. Conceivably, the only categories that would still be meaningful would be the personality variables of the individual members of the audience, for example, extraverted versus introverted music, and some recent research has begun to explore movie preferences in terms of the personality of the viewer.13 While I am not convinced by that argument, it can be tested through future online surveys, and the ethnographic research technique can be conceptualized as a form of intellectual individualism.

The term survey research is often used as a synonym for questionnaire research, and yet the term survey can refer to systematic observation in the natural world, not just asking questions. At many online services, people engage in public behavior following systematic procedures, analogous to but possibly more genuine than a verbal response to a researcher’s question. Among the clearest examples are several web-based services designed for musicians to share recordings, perhaps sell them directly online to listeners, and advertise themselves for local performance gigs. Here we will briefly look at data from Bandcamp and ReverbNation, but others include Sonicbids, Spotify, and Soundcloud.14

The motto at the top of Bandcamp’s website urges: “Discover amazing new music and directly support the artists who make it.”15 As Wikipedia reports,

Artists and labels upload music to Bandcamp and control how they sell it, setting their own prices, offering fans the option to pay more (which they do 40% of the time) and selling merchandise... Bandcamp’s website offers users access to an artist’s page featuring information on the artist, social media links, merchandising links and listing the artist’s available music.16

Bandcamp’s search engine permits searching for performers by 28 genres and 25 cities where many of them perform. Table 3.3 reports May 2019 data for 6 illustrative genres in the 11 cities outside the United States.

Table 3.3 Bandcamp music listings, showing regional cultural variations

	Classical	Devo-tional	Folk	Jazz	Rock	Electronic	Total N
Berlin	3.2%	3.3%	9.2%	10.3%	25.1%	49.0%	3,255
London	3.0%	3.0%	13.9%	9.0%	33.1%	38.0%	3,170
Vancouver	1.7%	2.6%	16.6%	4.8%	39.0%	35.3%	3,016
Sydney	1.7%	2.1%	9.8%	9.0%	33.7%	43.6%	2,478
Madrid	2.3%	2.1%	5.8%	4.1%	53.6%	32.1%	2,104
Manchester	1.1%	2.2%	9.9%	4.2%	24.2%	58.4%	1,662
Dublin	2.0%	1.2%	15.8%	4.6%	28.7%	47.7%	989
Buenos Aires	0.4%	2.0%	9.2%	2.9%	79.1%	6.4%	762
Glasgow	1.5%	0.9%	16.3%	3.4%	44.5%	33.4%	326
Amsterdam	1.2%	1.2%	8.1%	17.1%	26.7%	45.7%	322
Mexico City	1.9%	0.0%	5.6%	2.3%	54.1%	36.1%	266

For example, we see that about half of the 3,255 performers located in Berlin, Germany, perform electronic music, frankly a term that seems modern but is a century old and has many variants, as Wikipedia correctly reports:

Electronically produced music became prevalent in the popular domain by the 1990s, because of the advent of affordable music technology. Contemporary electronic music includes many varieties and ranges from experimental art music to popular forms such as electronic dance music. Today, pop electronic music is most recognizable in its 4/4 form and more connected with the mainstream culture as opposed to its preceding forms which were specialized to niche markets.17

The table also reports that a quarter of the Berlin performers in this subset of genres work in the “American” rock tradition, and only 3.2 percent in their national Bach–Beethoven classical genre. This low percentage partly just reflects the fact that Bandcamp is primarily for solo players and small bands, but it also reflects a cultural divide within a city, and the Berlin Philharmonic Orchestra has its own elaborate website.18

A serious study of these online music social media would want to chart the changing statistics over time, for example, monitoring the Buenos Aires performers over the next few years to see how many of them adopted the high-tech electronic systems. Looking more deeply at the performers’ individual Bandcamp pages could define what each city means by “folk music,” for example, looking for possible Irish influences in Glasgow folk music, given the wide popularity of Irish dance music at the present time and western Scotland’s proximity to Ireland. National data from the comparable ReverbNation website in Table 3.4 may contain clues, in the Celtic column, which indicates that this Irish–Scottish tradition has some popularity in France and Germany. Yes, in India the local folk-pop Bollywood genre is vastly more popular than anywhere else, but the Christian/gospel genre is more popular in India than in France or Germany. The orientation of these sites is primarily toward independent performers, and the ReverbNation website quotes this testimonial: “It’s fair to say that much of our recent success wouldn’t have happened without ReverbNation. It’s exactly what every starting band needs.”19

Table 3.4 National variations in musical culture, from ReverbNation listings

	Bolly-wood/Tolly- wood	Celtic	Christ-ian rock	Christ-ian/gospel	Classi-cal	Coun-try	Folk	Latin	Total N
United States	0.1%	1.6%	5.4%	25.1%	3.8%	40.6%	17.8%	5.6%	14,772
Canada	0.2%	2.2%	1.5%	10.7%	7.2%	44.1%	30.5%	3.3%	1,228
United Kingdom	0.3%	2.4%	1.4%	16.0%	11.5%	24.1%	42.5%	1.9%	1,064
Australia	0.2%	2.5%	1.6%	8.0%	4.6%	44.3%	36.3%	2.5%	438
India	49.3%	0.0%	4.1%	16.2%	14.9%	6.1%	9.1%	0.3%	296
Germany	0.0%	4.8%	1.6%	6.9%	18.0%	26.5%	32.8%	9.5%	189
France	0.0%	5.9%	1.3%	5.3%	23.7%	10.5%	42.1%	11.2%	152
Italy	0.0%	2.9%	0.7%	8.8%	27.2%	14.0%	28.7%	17.6%	136
Mexico	0.0%	0.0%	4.7%	15.7%	7.9%	9.4%	11.8%	50.4%	127

To be sure, each online source of data such as these will reflect a particular social organization as well as subculture, and in many fields with strong commercial aspects. But each offers links to sources of other data. For example, it would be easy to take random or purposive samples of the performers listed by Bandcamp and ReverbNation and then interview them or send them questionnaires. Perhaps the Glasgow folk music is performed on high-technology bagpipes! Oh, and intrepid researchers could also attend public performances! That raises the point that their audiences may also be worthy of study, and that in today’s online world communities of fans are easy to find.

Comparative Surveys Across Subcultures

No form of culture, other than perhaps family structures, has been the focus of social scientific research more than religion. Sociologists tend to name a Frenchman, Emile Durkheim, and a German, Max Weber, as the founders of their field, although British and American schools of sociology were already in existence when those two began publishing. The 1897 book Suicide, by Durkheim, contrasts Protestants with Catholics in their rates of self-destruction, while Weber contrasted their differences with respect to economics in his 1904 book The Protestant Ethic and the Spirit of Capitalism.20 The U.S. government conducted censuses of religious organizations in 1890, 1906, 1916, 1926, and 1936, which more recent researchers have found very valuable for statistical analysis. Data from the four 20th-century surveys can be freely downloaded from the digital library of the Association of Religion Data Archives, in a form I contributed after Rodney Stark and I had used them in our 1996 book, Religion, Deviance and Social Control.21 This dataset compiles information regarding the religious composition of 378 cities in the United States from 1906 to 1936 and contains 90 variables, including observations on church membership, growth, and suicide rates.22

Another dataset I contributed to ARDA was a 244-item questionnaire completed by 1,025 members of the Endtime Family, or Children of God (CoG), a religious group that is in high tension with its surrounding cultural environment. This dataset assesses the validity of applying survey data techniques to unusual religious groups. Additionally, most of the variables were taken from the GSS, enabling comparisons between the Endtime Family and the general population.23 This illustrates a general principle of cross-cultural research: A scientific instrument developed to chart the structure of one culture can subsequently be adapted for comparison with another culture. I had done extensive ethnographic and interview research in many communes of the CoG in the United States, plus one each in Canada and France, and the resultant book combined methodologies to understand its history, social structure, and beliefs.24

One of the questions asked, “How close do you feel to God most of the time?” Of GSS respondents who answered, 31.8 percent said “extremely close,” while 42.1 of CoG gave the same response. The percentages answering “somewhat close” were about the same, 52.5 percent and 51.6 percent, reflecting the fact that the CoG were struggling to get extremely close to God, but many felt they had not yet succeeded. The group’s Wikipedia page reports it

initially spread a message of salvation, apocalypticism, spiritual “revolution and happiness” and distrust of the outside world, which the members called The System. In 1976, it began a method of evangelism called Flirty Fishing, that used sex to “show God’s love and mercy” and win converts, resulting in controversy.25

Wikipedia even has a page for flirty fishing that calls it

a form of evangelistic religious prostitution ... The term is derived from Matthew 4:19 from the New Testament, in which Jesus tells two fishermen that he will make them “fishers of men”. Cult leader David Berg extrapolated from this that women in his movement should be “flirty fishers” (also called “bait” or “fisherwomen”). The targeted men were called “fish”. The cult published several documents with exact instructions. Flirty Fishing was defined as using sex appeal for proselytizing.26

Wikipedia is not entirely clear about the fact that the CoG considered sexual intercourse to be a holy act, even perhaps believing that Jesus or God inhabited the souls of any two people of opposite genders, regardless of age, who joined in this way. Thus flirty fishing brought the fisherwomen closer to God, some of them believed. Clearly, this was a different culture from conventional Christianity.

This immediately raises a question: How did the CoG conceptualize their sacred parent? The questionnaire included a dozen items, taken from the 1983 and 1984 GSS, a set that begins with this question: “When you think about God, how likely are each of these images to come to your mind?” Table 3.5 compares the percentages who responded “extremely likely” in eight subsets of respondents, divided by gender and within the more numerous GSS respondents by how close they feel to God, combining in “not close” three rarer responses: “not very close,” “not close at all,” and “does not believe” that God exists.

Table 3.5 Extremely likely an image of God would come to mind

	GSS males by closeness to God			GSS females by closeness to God			Children of God
	Extremely	Some what	Not close	Extremely	Some what	Not close	Males	Females
Creator	90.0%	81.5%	56.7%	94.0%	84.9%	46.5%	89.8%	86.6%
Judge	71.6%	46.5%	25.3%	59.8%	43.4%	21.4%	26.4%	17.2%
Redee-mer	76.4%	60.9%	27.7%	80.9%	62.4%	24.4%	71.0%	62.4%
Lover	61.1%	43.0%	15.7%	56.8%	40.0%	15.3%	58.4%	62.8%
Master	76.2%	52.3%	29.0%	75.3%	53.4%	23.2%	69.8%	62.6%
Mother	37.5%	20.4%	9.8%	36.2%	23.7%	12.8%	18.6%	15.3%
Father	82.2%	58.6%	25.0%	82.2%	63.1%	33.3%	85.3%	85.2%
Spouse	28.6%	16.6%	6.0%	24.6%	12.2%	8.9%	29.1%	44.7%
Friend	77.4%	56.0%	31.9%	80.2%	66.1%	27.2%	81.3%	86.0%
King	71.5%	48.2%	23.1%	71.4%	50.1%	28.8%	86.5%	86.9%
Liberator	63.7%	39.9%	16.2%	66.2%	40.8%	16.4%	64.2%	60.0%
Healer	84.2%	68.3%	35.1%	87.5%	72.7%	32.3%	76.1%	74.9%
Creator N	432	610	127	622	1,064	314	400	606

For several of the dozen God descriptors, the CoG respondents gave similar answers to those of the GSS respondents who felt extremely close to God, or between the GSS extremely and somewhat close columns of figures. For example, 90.0 percent of the 432 GSS male respondents who felt extremely close to God said it would be extremely likely that “creator” would come to mind, as did 89.8 percent of 400 CoG males, almost exactly identical fractions. But members of the CoG were very reluctant to conceptualize God as their judge, compared with religious ordinary citizens. This may reflect their history coming out of the counterculture of the 1960s and the unconventional lifestyles members experienced both before and after joining.

Pagan religions of Europe and the Middle East assigned important sacred roles to goddesses and priestesses, while the Judeo-Christian-Islamic tradition conceptualized God as a male—Father rather than Mother—and until recently limited the full role of clergy to men. In today’s world, one might expect equality of the supernatural genders again, and at least one Protestant denomination has begun to consider a gender-neutral conception of God.27 Even around 1997 when the CoG data were collected, and 1983 to 1984 when the GSS data were collected, this discussion had not yet begun. But the CoG members were very unlikely to think of God as their Mother, rather than Father. Among female members of this group, a relatively high percentage could imagine God being their spouse, probably a reflection of the theology behind flirty fishing.

What do the examples of the old U.S. religious censuses, the GSS, and the Association of Religion Data Archives tell us about the future of questionnaire survey research? First, we must face difficult questions about what organizations can legitimately and successfully administer research on cultural topics like religion about which the general public differs. After the 1936 religious census, the U.S. government came to the provisional view that it was unconstitutional for the U.S. government to conduct or support research on religion. The famous First Amendment says,

Congress shall make no law respecting an establishment of religion, or prohibiting the free exercise thereof; or abridging the freedom of speech, or of the press; or the right of the people peaceably to assemble, and to petition the Government for a redress of grievances.28

Freedom of speech presumably allows social scientists to administer questionnaires or interviews about religion to willing respondents, but concerns were expressed that government support of them could function as indirect establishment of religion. For that reason, when Charles Glock and Rodney Stark conducted a major questionnaire study of Anti-Semitism, they could not apply for government support, and had to seek funds from individual donors instead.29

Very few research projects supported by the NSF, other than the GSS which covers a diversity of topics, concern religion. Entering the word “God” into the online grant abstract system and reading the results turned up just 28 research projects, the first dating from 1993. Just 4 of the 28 were studies explicitly of modern religious phenomena, as suggested by their titles: “Clergy, Parishioners, and Politics: A Survey of ELCA and Episcopal Church Ministers and Parishioners,” “Religious Orthodoxy and New Media Technologies,” “Hearing Religious Language, Making Political Choices,” and “Religion and Financial Market Behavior.” The last two of these were very low budget grants supporting the necessary costs of a doctoral dissertation research project. Of the 24 other grants, 14 could be described as anthropology or history, and thus do not directly concern religion in American society today. One does, however, suggest implications that religion of the past had for the current structure of scientific concepts, reporting the assumption behind the still-significant but nearly three-century-old Linnaean system for classifying animals: “Linnaeus assumed that God created an initial pair of organisms for each species and after God’s original creation no new species could arise.”30

The Phenomenological Approach to Ethnography

Scientists tend to place a high value on objectivity, yet research on human culture requires a very personal investment. The posthumous publication of the personal diary of anthropologist Bronislaw Malinowski caused intense controversy, some critics branding him prejudiced and others praising him as a pioneer of “intensive personal fieldwork.”31 Malinowski studied the Trobriand “savages” who lived on islands near New Guinea, while here we shall explore the equally exotic virtual version of New England in The Secret World (TSW), a massively multiplayer online role-playing game. As Malinowski well illustrated, cultural anthropology was an intellectual extension of European colonialism in the sense that it viewed indigenous cultures from the perspective of the home culture of the anthropologist, and it was an intellectual corrective to the extent that it took the indigenous culture seriously as a valid perspective in its own right. In the extreme, the home culture of a scientist is not his or her nation, but his or her self. Thus, cultural anthropology is prepared to converge with phenomenology, the most self-centered school of thought. In his 1637 book, Discourse on the Method, Descartes presented his influential principle: “I think, therefore I am.”32 That is, one should doubt beliefs unless they are absolutely irrefutable, and one’s own existence is a logical precondition for having doubts or any other kind of thoughts.

Three centuries later such Cartesian meditations were the basis of phenomenology, which in turn influenced the birth of ethnomethodology, perhaps the most individualistic form of sociology, especially as presented by George Psathas, my main teacher when I was an undergraduate at Boston University around 1970. Yet ethnomethodology emphasizes role-playing, so in participant observation research one is able to adopt for research purposes the perspective of a specific other person, for example, a social theorist like Psathas.33 TSW seemed very appropriate for this kind of phenomenological exploration because it depicts our real world today, but under the assumption that most people are completely ignorant of the horrifying reality that exists just outside their consciousness. Produced by a Norwegian company, it offers rich literary culture in three geographical areas: New England, Egypt, and Transylvania. Especially prominent are American horror writers in the New England tradition, including Nathaniel Hawthorne, Edgar Allan Poe, and Stephen King. Howard Phillips Lovecraft’s intellectual influence was the greatest, as symbolized by the fact that a prominent virtual location was an archetype of a New England prep school named Innsmouth Academy, given that Innsmouth was a postapocalyptic town in Lovecraft’s mythos that had been attacked by monsters called Deep Ones, an obvious source also for the fictional town in TSW named Kingsmouth.34

I decided initially to take the role of my great-grandfather William Folwell Bainbridge, calling the first of three research avatars Folwell. He had been the clergyman for First Baptist Church in Providence, Rhode Island, in the 1870s, had explored Egypt extensively in 1867, and then in 1879 to 1880 did a globe-circling research study of American Protestant Missions and the cultures they sought to convert that resulted in two sociological books.35 His behavior became erratic, his career and marriage disintegrated, and when he died in 1915 his surgeon son dissected his father’s brain seeking exculpatory evidence to soothe this family shame. A powerful connection to the horror variant of phenomenology was symbolized by the fact that he was buried in Swan Point Cemetery, where the body of horror genius Lovecraft would later be interred. In “Surreal Impersonation,” an essay about the value of phenomenological fieldwork, I summarized Folwell’s findings thus:

The past is dead, but lives. To understand the horrors of our existence, we must dig up the past, but full awareness would drive us mad, so we must also bury the past. This theory very much fits the worldview of H. P. Lovecraft, especially his novella, “At the Mountains of Madness,” that explores the remains of an ancient civilization in Antarctica, incredibly old, incredibly advanced, and incredibly corrosive of our mundane reality. The work of archaeologists and other scientists is dangerous, Lovecraft asserted, because “some day the piecing together of dissociated knowledge will open up such terrifying vistas of reality, and of our frightful position therein, that we shall either go mad from the revelation or flee from the light into the peace and safety of a new dark age.” One of the responsibilities of secret groups like the Illuminati is to prevent ordinary people from learning the truth, and thus being driven insane. Lovecraft once said, “All my tales are based on the fundamental premise that common human laws and interests and emotions have no validity or significance in the vast cosmos at large.”36

This frankly pessimistic view of life relates to the twin sociological traditions of phenomenology and ethnomethodology in several ways. Lovecraft never was rewarded by publishers or the general public for his profound creativity, dying in abject poverty. Thus his sense that life was meaningless was an empirical finding from his own personal science, the active exploration of reality on the basis of one’s own private experience. Although not having Lovecraft in mind, Psathas advocated a sociological approach that was a form of personal science:

The distinction between natural science and social science ... is based on the fact that men are not only objects existing in the natural world to be observed by the scientist, but they are creators of a world, a cultural world, of their own. In creating this world, they interpret their own activities. Their overt behavior is only a fragment of their total behavior. Any social scientist who insists that he can understand all of man’s behavior by focusing only on that part which is overt and manifested in concrete, directly observable acts is naive, to say the least. The challenge to the social scientist who seeks to understand social reality, then, is to understand the meaning that the actor’s act has for him.37

In his Boston University classes as well as writings, Psathas referred to publications like Studies in Ethnomethodology by Harold Garfinkel, The Social Construction of Reality by Peter Berger and Thomas Luckman, and Cartesian Meditations: An Introduction to Phenomenology by Edmund Husserl.38 The classic work that held these radical philosophical statements together was The Phenomenology of the Social World by Alfred Schutz.39 Originally published in German in 1932, this English edition came a third of a century later when the intellectual rebellions of the 1960s were at their peak. Schutz was especially influential in promoting a very radical version of cultural relativism, the concept of multiple realities, not only suggesting that each human individual lived in a different universe, but that roles and episodes in each person’s life were distinct realities. In a book surveying the sociological literature, I summarized how Schutz considered the large subset of realities that were detached from the intense social interactions of daily life:

A person cannot move smoothly from one of these worlds to another. Rather, a person’s consciousness must leap across, and the transition from one reality to another is always a mental shock. Compared with the paramount reality of everyday life, these finite provinces of meaning lack coherent social interaction. The experience of time is different, perhaps utterly so, and we cannot share time coherently with another person. Objects are not stable, and the results of our actions typically vanish. Schutz considers the worlds of phantasms (fiction, fantasies, myths, jokes, and the like), dreams, and scientific theory, finding each distinctly different from and yet connected to the world of everyday life.40

TSW certainly overflows with phantasms, and with that in mind I created a new avatar named Psathas to explore its reflection of New England again, while I simultaneously surveyed the numerous external phantasms it connected to. There are many ways in which this virtual world is a collection of multiple realities, most notably that it currently exists in two forms, the original 2012 version titled The Secret World, and a version with a much simpler interface designed to attract new players, Secret World Legends (SWL), that was launched in 2017 and was explored by my Psathas avatar. In addition to Folwell, I already had a secondary avatar in the original TSW, named Dionysius Bainbridge after a probable distant family member who had been the political mentor of his stepson, Guy Fawkes, famous for attempting to destroy the British parliament building in 1605.41 Wikipedia gives his name as both Dionis Baynbrigge and Denis Bainbridge.42 In 2015, a musical group called the Gunpowder Plot Conspirators released a memorial piece titled “Farewell Dionis Baynbrigge,” illustrated by this band wearing the Guy Fawkes masks associated with hacker-radicals like the online activist Anonymous cabal.43

In order to compare the very different interface systems and the contrasting sets of variables that define the avatar’s characteristics, I revived Dionysius in New England, alternating him with Psathas, as they attempted to explore their different realities in the same geography. Since I was already familiar with the original interface, I added the experiment of seeing how difficult it was for Dionysius to progress without taking on any more missions for nonplayer characters (NPCs), even as Psathas did as many missions as he could to document his interactions with friendly NPCs. Table 3.6 reports that Folwell visited all 228 areas available in the three regions of this virtual world, while Dionysius and Psathas were limited to locations in New England, Dionysius visiting far fewer because he had no motivation to wander in search of missions to perform. Fragments of legends could be learned by finding a bright node somewhere in the geography and touching it. Dionysius killed 5,505 monsters in New England, rather more than the 4,042 by Folwell or 3,128 by Psathas, because that was the only way he could gain experience or equipment because he earned no further mission or lore rewards.

Table 3.6 Three parallel explorations of a virtual reality

Description	Folwell	Dionysius	Psathas
Location exploration sites:
Solomon Island (New England)	65	25	58
Valley of the Sun God (Egypt)	66	0	0
Transylvania (Romania)	97	0	0
Legend fragments learned (lore):
Solomon Island (New England)	55	9	51
Valley of the Sun God (Egypt)	22	0	0
Transylvania (Romania)	1	0	0
Missions completed:
Kingsmouth Town (Solomon Island)	55	20	41
The Savage Coast (Solomon Island)	46	0	39
The Blue Mountain (Solomon Island)	27	0	25
The Scorched Desert (Valley of the Sun God)	33	0	0
City of the Sun God (Valley of the Sun God)	16	0	0
Transylvania	0	0	0
Humanoid monsters killed in New England:
Zombies: animated corpses lacking self-awareness	2,657	4,356	2,063
Familiars: former companions of students of magic	555	724	455
Draug: the cause of a hundred ghost ship tales	282	328	239
Rakshasa: demons in the Hindu–Buddhist tradition	116	0	29
Spectres: terrible mass, terrifying physicality	109	89	90
Wendigo: cannibal vestiges of humans	104	4	56
Filthy Humans: driven mad by an epidemic disease	70	1	48
Golems: monsters created from inanimate matter	39	2	43
Deep Ones: from the cold depths of the Atlantic	33	0	41
Shades: evil ghosts who escaped the underworld	33	0	26
Scarecrows: manifestations of fear and hatred	30	0	30
Revanants: craving misery, pestilence, and death	14	1	8

Psathas paid close attention to the process of receiving missions and being rewarded when he completed them. Location exploration simply meant going to a particular small area within a region, and Folwell found all of them in all three lands. Thus, no social interaction was required. Also distributed across the landscape were nodes where pieces of text could be collected, which assembled into lore-based legends. Commenting on a blogsite article, Hikari Kenzaki wrote: “Secret World Legends has some wonderfully confusing lore. I say wonderfully because it’s intended to confuse you. Your handlers and even the Buzzing will outright lie to you through the course of the game.”44 Handlers are the high-status NPCs that give orders to the player’s avatar, usually lore-related, within one of three mutually tolerant factions: the Illuminati in New York City, the Templars in London, and the Dragons in Seoul, Korea. The legend fragments are like scraps of paper torn from a book, thus impersonal whoever the author may have been. Handlers may be interacted with in person or at a distance, but are personified. The Buzzing are multizone, cross-faction, and mysterious:

We are the Education Protocol. We climb the twisted ladder of your cells; we haunt your digital text; we hide in your hat. We are the jagged teeth that trip the tumblers of your mind. You will not know our triggers. For all the world’s a cypher. And everything is true.45

Many missions are given by local NPCs, more or less separately from the faction-centered handlers. Innsmouth Academy is clearly connected to the Illuminati, yet its staff have their own needs, chiefly protection against familiars, a subclass of zombies who were former companions of students of magic. Some NPCs are more economic in function, buying and selling equipment or potions that confer special powers. A player’s avatar may craft equipment, usually from raw materials obtained by processing other pieces of equipment looted from defeated enemies. Other NPCs may be called extras, by analogy with the crowds in movies who lack names or important roles, sometimes relevant to missions, but often merely standing in the background. Another category is departed characters, whether they died or simply ran away, who left artifacts such as cell phones bearing messages that would give the avatar a mission.

Missions usually require killing enemies, but it is also easy to gain experience points and resources by killing enemies without any complex story. Some of them function individually, while others are in teams of two, three, or four, who must be battled simultaneously. Both inside and outside missions there are boss characters, individual enemies who are dangerous and difficult to kill. Some are animals or mysterious lights in the air, but Table 3.4 lists the total of 12,675 humanoid monsters I killed in New England, not to tally all the mummies Folwell killed in Egypt. Now that I am no longer playing Folwell, Dionysius, and Psathas, perhaps they have joined the ranks of the familiars, and plague today’s students at Innsmouth Academy. I am left contemplating the culture of my native New England, in the so-called real world, counting the bosses and lights in the air I personally encountered. More generally, the New England horror literature tradition is really a major part of the region’s dominant culture, and the Choate School I attended was not very different from Innsmouth Academy.

Observing Cultural Structure Statistically Through Interlocks

Research on the social structure of industries has often traced interlocking directorates, the ties between corporations represented by people who hold leadership positions in two or more.46 In my earlier book in this series, Virtual Local Manufacturing Communities, I analyzed the membership connections between guilds in A Tale in the Desert, an online social simulation of the construction of Ancient Egypt.47 Unlike most online games, Tale does not allow violence, does not require membership in hostile factions, and therefore permits each person to belong to many groups. The research focus was on how group membership was a function of two factors, the functions of the groups and their geographic distribution across a huge simulated Egypt. This methodology can be applied as well to the study of online communities that are focused on genres of culture, and here we will illustrate this opportunity through a cluster of Facebook groups oriented toward the 1967 to 1968 British television series The Prisoner that dramatizes many of the dangers that humanists believe information technology presents to humanity.

Patrick McGoohan, co-creator of the series, played the role of a person named only Number 6, who is held captive in a beautiful but mysterious seaside location named The Village, as a sequence of dictators named Number 2 seek to control his mind. The main Wikipedia page for The Prisoner had 1,123,086 views, July 1, 2015, through March 1, 2019; the page listing the 17 episodes had 199,940, and a page about the mysterious village where Number 6 is imprisoned had 100,493.48 A common salutation spoken by many characters, “Be seeing you,” is reminiscent of “Big Brother is watching you” from George Orwell’s novel Nineteen Eighty-Four.49 Each episode begins with the following opening sequence:

Number Six: Where am I?

Number Two (not identified as yet): In the village.

Six: What do you want?

Two: Information.

Six: Whose side are you on?

Two: That would be telling. We want information ...information ... information!!!

Six: You won’t get it!

Two: By hook or by crook, we will.

Six: Who are you?

Two: The new Number Two.

Six: Who is Number One?

Two: You are Number Six.

Six (running on the Village’s beach): I am not a number; I am a free man!!!

Two: [Laughter]50

The largest Facebook group devoted to The Prisoner, as of April 18, 2019, was simply named The Prisoner, with 6,648 identifiable members, 1,864 of whom had been members for “about 10 years” or longer. Officially, Facebook was launched on February 4, 2004, in the United States and expanded to other nations in the following year.51 Thus it is remarkable that The Prisoner group was created on December 3, 2004, so early in the history of this social medium. The FB profile of the founder, David Bowers, suggests he had just graduated from high school and was a freshman at Pennsylvania State University, thus indicating how The Prisoner group harmonized initially with the original focus of Facebook on college students. He continues to be listed as one of three admins of the group, the other two apparently more active at the present time. Anthony Rooney joined the group on September 16, 2007, and by 2013 had taken the main responsibility for adding new members. The third admin, Shane Poole, joined much later, on October 5, 2014, but has been very active, notably serving as Podcast Host at In the Village.52 Its goal is to teach newcomers about the subculture, being an “introcast about the classic tv show The Prisoner that ran from 1967-1968. Join our newbies as they discover the mystery of the village for the very first time.”53 The Prisoner’s “about” page asks “What’s it all about?” then offers this introduction:

When a man resigns from a highly classified job, he is abducted to a mysterious location known only as The Village, where his captors try to find out the reasons for his resignation. Has he been imprisoned by his own side, trying to protect their secrets, or is it the other side trying to extract them? The Prisoner, now unwillingly designated “Number Six,” must fight to retain his own identity in a society geared to total sublimation of the individual, while making unceasing efforts to escape.54

Currently it is possible to purchase The Complete Prisoner Megaset of DVDs from Amazon.com for $250, and the site indicates their original publication date was September 25, 2001. If one has a Prime membership with Amazon, one can watch online for free. However, two sources uploaded the 17-episode series for free viewing at YouTube. As of March 1, 2019, each episode had its own Wikipedia page, with a total of 475,971 pageviews, and on the two YouTube channels they had received 219,779 and 135,704 views, respectively. After joining nine Facebook groups devoted to The Prisoner, I posted this question: “What do you think of the fact that all episodes are freely available on YouTube now, apparently uploaded by people ... who believe the copyright ran out after 50 years?” A veteran member replied, “The copyright hasn’t run out, it will run out 75 years after the death of the last director (so around 2084). The series is copyright ITV Global Entertainment.” Indeed, when I checked back three months later, both sets of videos had been removed from YouTube. It is ironic that one can borrow books and physical videos for free from lending libraries, but online sharing may violate copyright, or as the current rhetoric calls them, intellectual property rights. Arguably, prior to industrialization all elements of culture were shared, and intellectual property rights did not exist. Patents and copyrights were necessary to promote creativity and hard work during the era of mass production. Now that we have entered postindustrial society, some writers suggest it is time to again permit unlimited sharing.55

Dutiful residents of The Village often criticized Number 6 for being unmutual, meaning that he refused to freely share everything with everyone. If we want information, one good place to get it is The Unmutual, The Prisoner news website.56 As of May 20, 2019, its history page lists 813 major updates since it launched April 10, 2004, “as replacement for the old ‘tripod’ Unmutual site.” It links to other websites, fan fiction, and a vast diversity of essays about The Prisoner. A page listing local clubs begins:

For many years, various “Prisoner Local Groups” have been active across the UK. These are small informal meetings (usually in pubs) where fans can attend to socialize with other Prisoner enthusiasts who share their interests. None of the groups listed here require membership to any fan club or society, and are not controlled by any such committee. If you have an easy going outlook and don’t want the narrow intensity of fans who take The Prisoner far too seriously then these are the meetings for you.57

Here our emphasis is on the nine Facebook groups, virtual clubs that take The Prisoner at least somewhat seriously, listed here in descending order of size and giving their years of creation:

1. 1. The Prisoner (2004) “dedicated to the classic Sixties TV series created by and starring PATRICK McGOOHAN.”
2. 2. The Prisoner & Portmeirion (2015) “A group for fans of The Prisoner and the village of Portmeirion, setting of the cult ‘60s TV series.”
3. 3. The Prisoner & Patrick McGoohan Memorabilia Group (2016) “This group is for Memorabilia related to the TV series The Prisoner and the actor Patrick McGoohan who was The Prisoner and did many films and TV series.”
4. 4. Everything McGoohan (2016) “This group is for everything Patrick McGoohan has been involved with or about people he has worked with. Also for discussion about Portmeirion and anything that was made there!”
5. 5. The Prisoner (1967–68) TV Series, Including Danger Man (2014) “A Group to discuss aspects of the Classic 60’s Television Series the Prisoner, Starring Patrick McGoohan as Number 6 with various guest actors as Number 2. Now including all series of Danger Man from the earliest half hour Episodes up to the two colour episodes Koroshi and Shinda Shima.”
6. 6. Patrick McGoohan (2008) “For fans of Danger Man star and Prisoner creator, Patrick McGoohan.”
7. 7. Patrick McGoohan is ... Danger Man !! (2015) “Danger Man, featuring the talents of the great Patrick McGoohan as John Drake, was a British TV series produced (& transmitted) between 1959 & 1968. The series was also a success in the USA, being shown there under the name ‘Secret Agent’.”
8. 8. Portmeirion & The Prisoner Memories (2017) “Its our family album to share your snaps. This space is for lovers of Portmeirion and the cult 60’s TV series ‘The Prisoner.’ Portmeirion is located in North Wales and this was the location used for filming ‘The Prisoner.’”
9. 9. When is PMG on your TV? (2018) “When is PMG (Patrick McGoohan) on your TV is to let every PMG fan know when he is on TV and also to let us know when you are watching him on your device!”

Portmeirion is the name of the more-or-less real Village where the series was filmed, and Danger Man is the name of the earlier spy-oriented adventure series Patrick McGoohan had starred in, called Secret Agent in the United States, and that provided some background. To get a sense of how these groups connected into a subculture, I copied the names of all their members, with data about which groups each person belonged to. A total of 9,796 people are represented in this dataset, and 7,746 belong to only 1 of the 9 groups. Another 1,160 belong to 2 groups, and 433 belong to 3. The remainder are rather deeply embedded in this subculture: 202 to 4 groups, 123 to 5 groups, 68 to 6 groups, 35 to 7 groups, 19 to 8 groups, and 10 people belong to all 9 groups. Table 3.7 summarizes this structure.

Table 3.7 Nine Facebook groups that dwell in the village

	Group 1	Group 2	Group 3	Group 4	Group 5	Group 6	Group 7	Group 8	Group 9
Group 1	6577	24.9%	24.7%	25.1%	24.2%	25.1%	19.9%	28.8%	21.9%
Group 2	904	2508	28.9%	27.8%	23.7%	19.3%	21.0%	38.0%	28.7%
Group 3	490	460	1166	33.3%	21.8%	22.2%	20.6%	34.5%	31.7%
Group 4	444	405	363	1024	28.7%	28.1%	30.7%	30.9%	35.3%
Group 5	341	285	205	256	790	20.1%	26.8%	22.5%	23.1%
Group 6	351	230	208	249	158	781	27.1%	18.2%	21.8%
Group 7	171	163	136	195	156	157	460	15.5%	25.4%
Group 8	163	201	163	142	97	78	56	296	30.7%
Group 9	35	45	48	53	34	32	35	39	81

The diagonal from upper left to lower right gives the membership of each group, slightly lower than the totals one sees at the main Facebook page for each, 6,577 rather than 6,648 in the case of the biggest Group 1, removing a few with multiple identities and others who apparently wished greater privacy. The numbers below and to the left of that diagonal are the members of one group that also belonged to the other, 904 member interlocks in the case of the first two groups. The numbers above and to the right express the interlocks as the average percentage of the two groups: 904 is 13.745 percent of 6,577 and 36.045 percent of 2,508; the average of those two percentages, rounded off, is 24.9 percent.

The most important finding is that the interlock percentages are substantial, ranging from 15.5 percent to 38.0 percent, keeping in mind that randomly selected Facebook groups of this size range would tend to have 0.0 percent interlocks. The 38.0 percent interlock is between the two groups that have Portmeirion in their names, indicating that people with a strong interest in the real Village might join both. The second highest interlock is between two groups focused on Patrick McGoohan, and that also have the same administrators. So, the differences in the interlock percentages are meaningful. But the main observation is that all the interlocks are strong enough to indicate that this is a coherent subculture.

Conclusion

Each example of quantitative data summarized in this chapter has some value for mapping a subculture, yet the unit of analysis varied. The GSS was not a random sample of individuals, but interviewed clusters of respondents in a representative set of geographic locations; the CoG respondents were opinion leaders within a radical religious movement; Bandcamp and ReverbNation posters were ambitious musicians, and the meaning of membership in Facebook groups varies from person to person and group to group. These observations refute the hope of decades past that absolutely objective science can be done via surveys, yet the data remain valuable through convergence of humanities and the social information sciences. Instead of relying upon one or two professional critics to describe an artwork or genre to us, we can use methods like surveys to gain much richer multidimensional descriptions. Each research method has its own limitations, yet they offer researchers a diversity of choices when planning a study, and may be combined in many ways. Adding the phenomenological approach to the mix, we can realize that it can be valid and valuable to chart how a subculture relates to ourselves, placing the researchers on the map beside the cultists and musicians, even counting the interlocks between groups of sociologists, artists, and computer programmers.

---