Lizardo, 2016

Poetics 58 (2016) 117
Contents lists available at ScienceDirect
Poetics
journal homepage: www.elsevier.com/locate/poetic
Why cultural matters matter: Culture talk as the

mobilization of cultural capital in interaction
Omar Lizardo
Department of Sociology, University of Notre Dame, 810 Flanner Hall, Notre Dame, IN, 46556, United States
A R T I C L E I N F O A B S T R A C T
Article history:
Received 5 July 2016 In this paper, I provide evidence for the substantive relevance of a phenomenon that has
Accepted 9 September 2016 been argued to be crucial in connecting cultural inequalities to stratication outcomes:
Available online 20 September 2016 culture talk, or the use of popular culture and the arts as conversation subjects with
friends and acquaintances. In spite of its theoretical importance, this outcome has not
received much empirical attention. To address this gap, I document the socio-demographic
correlates of culture talk using survey data from a probability sample of Americans. I report
three main ndings. First, culture talk frequency across ve culture consumption domains
is stratied by socioeconomic status, primarily be education and secondarily by income,
net of self-reported culture-consumption frequency. Second, the socioeconomic status
effect on culture talk is strongest for those who are most culturally active. Finally,
engagement in traditionally high status cultural pursuits (e.g. the ne arts) is a better
predictor of culture talk across all domains than consumption in less class stratied
domains (e.g. movies). I close by elaborating the theoretical and substantive implication of
these results.
2016 Elsevier B.V. All rights reserved.
1. Introduction
Research in culture and stratication has been primarily concerned with exploring inequalities linked to social position in
access to and engagement with symbolic goods (Katz-Gerro, 2004). The main question animating this literature is whether
taste and culture consumption patternsas constitutive of the lifestyles that mark Weberian status groupscontinue to be
associated with social position This organizing concern can be appreciated whether the construct of social position is
conceptualized in terms of educational attainment (Warde, Wright, & Gayo-Cal, 2008), membership in broadly dened
occupational groups (Tomlinson, 2003), aggregate measures of social class (Katz-Gerro, 2006), or constructed scalar indices
of social status based on patterns of interaction (e.g. intermarriage) across occupational groups (Chan and Goldthorpe, 2007).
The main conclusion of this still vibrant line of research is that lifestyle practices continue to be strongly shapedalbeit in
increasingly complex waysby social position however it is dened (Bennett et al., 2009). One basic nding is now well-
established: members of the educated, salaried professional elite are more likely to actively appropriate a wider variety of
aesthetic offerings, within and across industry-dened genre categories and institutional contexts of dissemination and
reception (Lizardo, 2008; Peterson, 2005; Prieur and Savage, 2011). In essence, what it means to be high-status in many
contemporary societies is to be an equal opportunity connoisseur of diverse cultural offerings and aesthetic experiences
across domains (e.g. literature versus lm, and within domains across substyles), forms of delivery (e.g. non-prot versus
E-mail address: olizardo@nd.edu (O. Lizardo).
http://dx.doi.org/10.1016/j.poetic.2016.09.002
0304-422X/ 2016 Elsevier B.V. All rights reserved.
2 O. Lizardo / Poetics 58 (2016) 117
commercial systems) and external legitimation (e.g. prescribed or authorized culture versus pop culture, folk culture
or kitsch) (Lizardo & Skiles, 2008, 2009; Peterson & Kern, 1996).
The persistent focus of the cultural stratication literature on the question of whether lifestyle practices continue to be
unevenly segmented according to class and status markers has resulted in a relative lack of attention to an equally important
question: what is the role of aesthetic consumption in everyday life? More precisely, given that high status persons are more
likely to engage with and be more generally knowledgeable about popular culture and the arts, how is this knowledge
implicated in their everyday behavior? In essence, now that we know that eclecticism is the form that good taste takes in
contemporary societies (Ollivier, 2008), it is time to askwith Erickson (1991)what is [this kind of] good taste good for?
This gap in our knowledge is surprising in light of the fact that the status of popular culture and the arts as a form of class
culture was, and continues to be, central to the theoretical tradition from which most analysts of the culture-stratication
link build on (Collins, 2009; DiMaggio, 1987; Douglas and Isherwood, 1996). In this respect, Gary Alan Fines basic complaint
about research of popular culture in the early 1970sthat this research generally ignored the behavioral implications of
everyday engagement with culture and the arts especially as it pertains to face-to-face interaction (Fine, 1977, p. 453)can be
repeated vis a vis research on cultural stratication today.
One of the key analytic lessons of some of these foundational work is that general familiarity with a given class culture is a
useful symbolic resource when it can be used as a signal providing an entry-point into delimited arenas of association (
writings Mohr & DiMaggio, 1995). This is the reason why, in their inuential paper, Lamont and Lareau (1988) linked the
notion of cultural capital to an explicit conceptual framework highlighting the active deployment of elite cultural signals for
the purposes of drawing symbolic boundaries. DiMaggio (1987) was also adamant building on the early work of Collins
(2009) in proposing that the major differences across status groups in the role that engagement with symbolic goods play
in everyday social settings is connected to the routine exploitation of cultural knowledge in social interaction within
everyday settings (Erickson, 1996; Lizardo, 2006). From this perspective, one important way that cultural aptitudes and
interests become actualized as cultural capital that is convertible into other resources (Bourdieu, 1986) is via their link
to everyday processes of informal socialization, patterns of endogamy, social closure, and status group reproduction in the
Weberian sense (DiMaggio & Mohr, 1985).
2. Culture talk as sociability
We can begin to characterize the empirical phenomenon that will be the focus of analysis via the rather common
observation that conversational rituals usually require some sort of content for their successful initiation and maintenance
(Collins, 2004; DiMaggio, 1987). Whenever the primary conversational subject matter animating a given interaction episode
moves towards the arts and the symbolic products of the culture industry, we may consider this an instance of culture talk.
Accordingly, I dene culture talk as the (more or less) routine deployment of cultural knowledge associated with aesthetic goods
as a resource to generate conversation. In a nutshell: the discussion of cultural matters with friends and acquaintances.
From a Simmelian perspective, we may consider culture talk one of the primary examples of sociability in late modern
societies. Sociability is social interaction that is not driven by explicit instrumental purposes, but that is keyed to the
continuing repetition of episodes of social interaction for their own sake (Simmel, 1949). Sociability thus represents the
limiting case and because seemingly devoid of explicit purpose, the play form of the more general process of association
(coming together from some purpose). In this respect, sociability is a form of social intercourse explicitly dissociated from
any sort of goal-directed pursuit. As a social form, sociability is thus partially independent of the particular content that
drives it in a given context, making culture talk a fairly general (and generic) phenomenon. However, this partial
independence of form from content does not mean that just any content will do.
2.1. The importance of aesthetic discourse as conversational content
Recent theorizing on the culture-network link has moved beyond the early formalism of traditional network analysis to
propose that specifying the form of association is not enough; instead, particular forms of association are constituted by
specic cultural contents that provide meaning to those forms (Fuhse, 2009). Here I propose that in the contemporary
context talk related to (1) aesthetic consumption experiences (Deighton, 1992; Holbrook & Hirschman, 1982), and (2) general
knowledge of the accepted discursive conventions governing expressions of aesthetic categorization, evaluation, and
enjoyment (Goldberg, Hannan, & Kovcs, 2016; Lizardo, 2008) become one of the most important sources of meaningful
content making possible sustained episodes of sociability.
As Simmel noted, sociability cannot be driven by purposive (e.g. goal-directed) or practical contents (e.g. work,
politics, etc.), but requires instead ludic content that serves to simply sustain the focus of attention in the interaction itself
(Collins, 2004). The reason for this is that sociability is a kind of social intercourse which does not have a strictly economic or
business purpose but access to which (and restrictions on) is a key marker of the status group boundaries (Weber, 1994p.
114) although in other contexts similar kinds of seemingly purposeless content are useful in relaxing those same status
boundaries (Ikegami, 2005). Building on Simmel, I argue that culture talk is the primary manifestation of sociability in
contemporary class-differentiated societies, precisely because conversation related to symbolic goods is the ideal ludic
content that ts this form of interaction. In comparison to aesthetic talk, other types of conversational content such as for
O. Lizardo / Poetics 58 (2016) 117 3
instance religion or politics are miserable failures, because they tend to suppress conversation and tend to highlight latent
social divisions (Eliasoph, 1998).
2.2. Culture industries as the source of aesthetic content
Where does aesthetic content come from? Cultural theorists have persuasively argued that media-culture and arts
dissemination institutions play a key role in producing an overabundant supply of (sometimes ephemeral) content useful for
generating discussion that serves as a resource for conversation in the context of small groups (Fine, 1977). The recurrently
reproduced micro-orders based on face-to-face (and increasingly electronically mediated) interaction require a constant
stream of novel content in order to be recurrently recreated and thus reproduce themselves over time (DiMaggio, 1987). In
this respect, the formal and informal system of production of aesthetic goods provide as input the very content that serves
as the scaffolding around which everyday episodes of talk and sociability are built.
In the same vein, popular culture scholars have proposed that [p]art of the pleasure of popular culture is talking about it;
part of its meaning is . . . talk (Frith, 1998). Accordingly, civic membership and even belonging to neighborhoods and
communities is partially constituted via talk, and [m]uch of this talk is about the mass media and its cultural commodities
(Fiske, 1987, p. 78). For instance, Elizabeth Long (2003) in her study of womens reading groups in Houston, Texas, nds that
cultural knowledge and taste become the primary content of social interaction. Consistent with the claim that connects the
consumption of aesthetic products and sociability, she nds that reading group members tend to press books into service
for the meanings that they transmit and the conversations they generate (2003, p. 73). Similarly, Benzecry and Collins (2014,
pp. 316317) point to the crucial role of culture talk in constituting the aesthetic experience for consumers of Opera in
Buenos Aires. Among Opera fans, talk is almost constant, and this talk revolves around the music, past performances, and
tales of their own heroic commitment to the art form. Through this type of culture talk, Opera fanatics ritually construct
and maintain their relationship to other fans, while at the same time marking the boundary separating true connoisseurs
from less serious followers.fanatics
2.3. Cultural content as a strategic interactional resource
Why is aesthetic talk effective at relationship building and maintenance? As noted in both classic (DiMaggio, 1987; Fine,
1977) and more recent (Cardon & Granjon, 2005; Lewis, Kaufman, Gonzalez, Wimmer, & Christakis, 2008) work, talk about
culture and the arts infuses local conversation-based interaction rituals with a common focus of attention (Collins, 2004). If
the establishment of this common focus is successful, then culture talk may serve as a conduit for either deep, engrossing
conversations (Frith, 1998) when others form part of more intimate circles or for least-common denominator talk
useful for sustaining brief interactions (DiMaggio, 1987, p. 443), depending on the pragmatic context (Schultz and Breiger,
2010).
In contrast to material goods which are physically present and visible and which may serve as barriers for interaction
via the creation of invidious distinction, the consumption of aesthetic goods is invisible once it has occurred. This
evanescent quality makes artistic experience, described and exploited in conversation, a portable and thus potent medium of
interactional exchange. (DiMaggio, 1987, pp. 443444). Because commercial (popular) culture and the arts (along with
fashion, cuisine and sport) are the closest analog to a common cultural currency in late modern societies, they have the
capacity to be a smooth facilitator of interaction across situational settings, social occasions and social positions (Erickson,
1996).
Sociological considerations of the role of culture talk in the creation and maintenance of social relations via consumption
and interaction rituals, converges with work in marketing and consumer culture studies (Deighton, 1992). According to
Gainer (1995), art consumers use conversation about the arts to make what could potentially be a purely private experience
part of their public interactions with others. Via culture talk, art lovers jointly build both their own personal identities as art
consumers and their interpersonal relations with other art acionados. In this way, social relationships are often created
and maintained through rituals of reproduction based on consumption which individuals join in shared performances with
symbolic meaning (Gainer, 1995, p. 254). Consumption of the arts, as mediated via culture talk, allows individuals to pursue
interpersonal goals connected to companionship and friendship. This also helps to build bridges to socially distant
acquaintances, via the above-referenced generation of interaction rituals for work conversation (Lizardo, 2013). Via the same
mechanism, culture talk can help to establish criteria for co-membership in exclusive social circles (Kadushin, 1966), while at
the same time managing and reinforcing individual and community identities.
3. The stratication of culture talk
3.1. Culture talk and socioeconomic status
The Simmelian link of culture talk and sociability, while revealing, also obscures a fact that is relatively uncontroversial
for contemporary theorists of the link between culture and stratication processes. Rather than simply being a formal or
neutral form of association, opportunities to engage in the type of non-instrumental sociability that Simmel points to are
unlikely to be equitably distributed as access to and practical familiarity with the content in question is itself inequitably
distributed (Bourdieu, 1984). This type of entertainment talk (Collins, 2009, p. 81), precisely because it is dissociated from
explicit instrumental pursuits and because it is not driven by extrinsic rewards but by friendship and personal liking,
comes to serve as one of the primary ways in which vertical and horizontal status-group boundaries are dynamically built
and rebuilt. These boundaries are recurrently recreated mainly via processes of inclusion and exclusion at the interactional
level.
Accordingly, when the very knowledge and associated experiences that are required for engaging in this form of
interaction are unequally distributed coming to form the basis for the formation of partially segregated status cultures
(Collins, 2009) then we should nd that members of high-status groups should be more likely to routinely engage in
culture talk. As noted at the outset, this proposition receives support from a long line of research in the sociology of taste
showing the continuing existence of strong socio-demographic differences in rates of familiarity and engagement with
culture and the arts, especially across groups dened by such status markers as education, and to a lesser extent, income.
If aesthetic experiences are translated into culture talk in mundane conversational settings, then we should expect that
individuals who have a higher-baseline probability of engagement with culture and the arts should also be more likely to
engage in culture talk. A strong version of this argument would predict that status-linked differences in the likelihood of
engaging in culture talk are a byproduct of differential opportunities to engage in cultural experiences. Under this
formulation, status-linked differences in cultural engagement are the primary source of individual differences in cultural
knowledge. Therefore, there should be no residual differences in culture talk propensities across status groups within levels
of consumption. In statistical terms, the direct effect of socio-economic status on culture talk is fully mediated by
consumption frequency.
Several considerations lead us to expect that this last version of the argument is not on the right track. First, persons
endowed with large stocks of cultural capital should be able to participate in conversation featuring culture as a topic even
when they are not active consumers in the domain in question. That is, they may demonstrate what Peterson (1992, p. 252)
referred to as passing knowledge of a given cultural domain on the spot. This is in fact a form of social skill that is almost a
requirement among members of the young, highly-educated, new middle class (DiMaggio, 1996).
Recent work on the culture-stratication link suggests that the ability to deploy cultural knowledge in interaction in
partial dissociation from direct consumption can be generated via multiple, interlinked mechanisms: (1) past experience in
the educational system scholastic capital in Bourdieus (1984) sense where traditional high-status culture (and
increasingly popular culture) have become established parts of the curriculum (Lizardo, 2008); (2) past experience in the
upper-middle and middle-class household domestic capital where extra-curricular forms of aesthetic engagement and
arts-training have become an institutionalized element of the socialization process (Weininger, Lareau, & Conley, 2015); (3)
routine interaction with members of the personal network who are themselves avid consumers of the cultural forms in
question and who thus make culture talk a recurrent conversational topic allowing ego to gather cultural knowledge that
may be used in other conversational interactions (Cardon & Granjon, 2005); and nally, (4) person-to-organization
connections that allow high-status individuals to participate in elite urban social circles where culture talk is a prescribed
conversational resource (Ostrower, 1998).
This alternative conceptualization of the link between culture talk and social position suggests the existence of
(relatively) strong structuration of culture talk by socio-economic status and other markers of difference, even after
adjusting for the relative propensity to participate in the cultural realms in question. These residual socio-demographic
differences in culture talk propensities should be strongest precisely for those social markers most clearly associated with
status and privilege (Khan, 2011), such as education and income stratum.1 However, insofar as mechanisms associated with
network connectivity and social capital may also be involved in generating the capacity to participate in interaction rituals
featuring culture talk (independent of consumption), other axes of difference partially orthogonal to socio-economic status
but correlated with social connectivity should retain their predictive power even after adjusting for levels of cultural
consumption.
Hypothesis 1. A persons socioeconomic status predicts culture-talk frequency even after adjusting for levels of cultural
participation.
Nevertheless, we should not expect all dimensions of socio-economic status to be equally good predictors of culture talk.
As argued above, culture talk is deeply tied to a persons capacity to appropriate cultural goods and then use this experiential
content as discursive resource in interaction. In this respect, culture talk is more closely connected to a given persons stock of
(embodied) cultural capital (Bourdieu, 1986) than it is to economic resources. Accordingly, we should nd that those
dimensions of socio-economic status that more closely track a history of accumulation of cultural capital (such as
educational qualications) should be better predictors of culture talk than those dimensions of socio-economic status that
index material advantage (such as income).
1
As noted by (Bennett et al., 2009), this does not imply that those persons who are disengaged from cultural consumption do not have active social lives.
My argument does imply however, that the social lives of the aesthetically disengaged will be structured in distinct ways, by for instance being relatively
more tied to neighborhood, kinship, and locality (Prieur and Savage, 2011).
Hypothesis 2. Indicators of cultural capital (e.g. education) are better predictors of culture-talk frequency than indicators
of economic capital (e.g. family income).
3.2. Cumulative advantage processes
Insofar as being an active participant in the arts is a clear index of cultural capital, a variation of this argument would
propose that the differences in cultural talk propensities that are traceable to differences in socio-economic status
themselves not independent from a persons status as an active arts participant. Instead we should expect that the predictive
power of socio-economic status should change depending on whether we are talking about active or inactive culture-
consumers. More specically we should expect that status-linked differences in the likelihood of engaging in culture talk
should be larger among active culture-consumers than among inactive ones. This would be consistent with a world in which
even when exposed to a similar set of experiences, persons would not only differ in the capacity to translate those
experiences into a conversational resource based on social location, but this differential capacity should be most evident
among persons who are more likely to engage in culture consumption in the rst place. This is consistent with a cumulative
advantage (DiPrete & Eirich, 2006) mechanism linking consumption, status position and culture talk. If this is correct,
members of the high status groups should not only be more likely to engage in culture talk than their lower-status
counterparts but this difference should be largest among the most active consumers.
Hypothesis 3. A persons socioeconomic status is a better predictor of culture talk frequency for culturally active persons.
3.3. Cross-domain differences in the stratication of culture talk
Does the stratication of culture talk vary across domains? The fact that the traditional ne arts dissemination
institutions, such as galleries and museums, symphony orchestras and theaters, continue to be patronized mainly by a highly
selected, educated elite is a well-established empirical nding (DiMaggio and Mukhtar, 2004). We also know that attempts
to democratize access to the ne arts on the part of these institutions have met with equivocal results, and in most cases
have failed outright (Alexander, 1996). The dependence of nonprot arts organizations on a primarily economically and
culturally advantaged consumer base provides high-status elites with recurrent occasions to meet others in the same circles
at sponsored social gatherings (Ostrower, 1998). These are interaction opportunities that are ideally suited for the
deployment of culture talk.
In addition, it is clear that the eld of ne arts production possesses a long-standing legacy of discursive tools developed
by professional critics and lay enthusiasts for the critical evaluation and the intellectualized expression of taste judgments
(Baumann, 2007; Weber, 1977). The availability of established linguistic conventions with which to categorize and evaluate
cultural products may facilitate the mobilization of this type of content as a conversational resource among elites (DiMaggio,
1987). In all, if we presume that routine access to and knowledge about the visual and performing arts is certainly more
strongly stratied by education and socio-economic status than is the case for more commercialized domains, then we
should nd that culture talk involving this type of content should also be more highly stratied by status.
Hypothesis 4. A persons socioeconomic status predicts culture-talk frequency in ne arts domains but not in commercial
arts domains.
Recent studies of the consumption of the popular arts, inspired by Bourdieus (1993) work on elds of cultural production,
has begun to conceive of the popular and commercial arts as equally capable of being the subject of attempts to impose
hierarchies of value and taste, the construction of exclusionary canons, and the establishment systems of cultural currency
(Baumann, 2007; Frith, 1998; Regev, 2007; Thornton, 1996). Rather than operating according to a logic that is antithetical to
that which structures ne arts consumption elds, as presumed by an earlier generation of cultural studies, the realm of
commercial culture production and consumption comes to be structured according to hierarchies of perception and
evaluation that are premised upon categories of perception and vocabularies of appreciation developed in autonomous elds
of cultural production to the products of the culture industry. In this respect, artistic hierarchies, which rank producers
according to their aesthetic or expressive value . . . are becoming a central structuring force in a growing number of elds of
production including that of rock and roll, a quintessential instance of the commercial arts (Regev, 1994).
More recently, other scholars have begun to extend this analysis of the dynamics whereby other popular culture elds-
such as lm come to acquire a hierarchical form modeled on the artistic eld (e.g. Allen & Lincoln, 2004). Thus, as Baumann
(2007) has argued, certain commercial forms (such as lm) have developed an intellectualized vocabulary of appreciation
that matches that of their ne arts counterparts. The adoption of this ideology of autonomous art by producers and gate-
keepers of the popular arts mirrors the increasing propensity of (culturally advantaged) consumers of popular culture to
adopt a similar stance (Thornton, 1996).
Just like in the ne arts eld, where struggles to dene artistic value based on what is dened as art and not art are in
play, a similar use of accumulated knowledge and discriminatory skill is apparent in low cultural forms, and has the same
hierarchical effect. Low culture, that is to say, generates its own capitalmost obviously perhaps, in those forms . . . which
are organized around exclusiveness, but equally signicant for the fans: . . . of even the most inclusive forms (Frith, 1998, p.
9). Insofar as the popular arts are subject to this intellectualization process, and insofar as critical gatekeepers develop a
language of appreciation and historical canonization, then they will be likely to gure as fodders for culture talk among high-
status classes. Thus, in the very same way in which both the commercial and the traditional arts can be appropriated
aesthetically by members of high status classes (Holt, 1998; Lizardo & Skiles, 2012), we should nd that high status groups
are equal opportunity culture talkers across domains:
Hypothesis 5. A persons socioeconomic status predicts culture-talk frequency across all cultural domains.
If culture talk is spurred by the extent to which a person is familiar with intellectualized discourses of cultural
appreciation, then we should also expect that there should exist predictable asymmetries in the extent to which the
consumption of ne arts spurs cultural talk in commercial domains as opposed to the reverse. More specically, we should
nd that the consumption of the ne arts (as an index of familiarity with the discursive conventions of the art eld) should be
a good predictor of culture talk in so-called commercial domains as the consumption of commercial art, but not vice versa.
That is, ne arts consumers should be as (or more) likely to engage in culture talk in commercial domains, but commercial
arts consumers should be less likely than ne arts consumers to engage in culture talk in ne arts domains.
Hypothesis 6. A persons ne arts consumption is a better predictor of culture talk across both commercial and ne arts
domains than commercial arts consumption.
4. Data and variables
4.1. Data source
In the analysis that follows, I consider data obtained from the Arts and Culture Supplement to the National Social Survey
elded in 2002 (N = 1002) and 2004 (N = 801) (Miringoff, Miringoff, Opdycke, & Hoynes, 2016) The NSS is a weighted (by age,
education, marital status, gender, race, ethnicity) probability sample of respondents residing the United States collected
through computer-aided telephone interviews. The NSS is particularly appropriate for my purposes because both waves
contain information not only about culture consumption experiences, but also about the likelihood that individuals engage
in informal social interaction featuring cultural content taken from consumption in the same domains. Because the items of
interest were elded in both waves, I pool individuals for which I have complete data on all socio-demographic variables
from the 2002 (N = 1002) and 2004 (N = 801) surveys.
4.2. Outcome variables
As part of the survey, respondents were asked to report (during the past twelve months), approximately how often have
you had a discussion with someone you know about (1) a movie, (2) a musical performance, live or recorded, (3) a book, a
poem or a story, (4) an art show or a work of art, and (5) a play or other dramatic performance. Respondents were asked to
choose from the following response categories: 1) very often, 2) fairly often, 3) not very often, 4) or not at all. These are the
main empirical indicators of culture talk and thus will serve as the response variables in the analyses that follow.
4.3. Main predictors
4.3.1. Indicators of culture consumption frequency

Respondents were also asked to report how often (in the last year) they had been to the movies, listened to recorded
music at home, read a book, had gone to an art show or a museum, or had gone to any live performances. Because I am
interested in accounting for residual socio-demographic differences in culture talk net of consumption, I adjust for
consumption frequency in the analysis. Respondents were asked to choose from the following response categories: 1) very
often, 2) fairly often, 3) not very often, 4) or not at all.2
4.3.2. Indicators of socioeconomic status

I was limited by the availability of socio-demographic indicators in the survey: for instance, the NSS contains no
information on respondents current occupation or family background, both of which are theoretically relevant dimensions
of social differentiation. The main indicator of socio-economic status, as indicated above, is educational attainment and,
2
How does the NSS compare to other Arts participation Surveys? When it comes to book reading, the 2002 NSS estimates that about 55.7% of respondents
report reading books Fairly Often or Very Often. The corresponding estimate for the 2002 SPPA (Survey for Public Participation in the Arts) is about 58%.
When it comes to going to an art museum, about 12% of NSS respondents report going Fairly Often and about 3.7% of respondents report going Very
Often; these correspond well with SPPA 2002 data based on more specic questions on attendance frequencies: 13.4% of SPPA respondents report having
gone to the museum between 2 and 5 times in the last year, and 3.2% report having gone more than ve times. In the NSS, about 9% of respondents report
going to a live performance Very Often. In the 2002 SPPA about 7.3% of respondents report having gone out to a Jazz, Classical, Opera, Musical, Ballet or
Dance performance between 2 and 5 times in the last year. In all, the NSS appears to do a good job of capturing rates of arts participation that are comparable
to that obtained from other National Surveys.
secondarily, family income. The education variable is coded as ve ordered categories (from high-school dropout to graduate
or professional degree) and the income variable is coded in six ordered categories (from less than 20 K a year to 100 K or
more).3
4.4. Additional predictors
I also use the following vector of socio-demographic covariates: (1) respondents age (in seven ordered categories), (2)
respondents gender, and (3) urban Location (urban versus other location).4 Descriptive statistics corresponding to each of
the predictor variables included in the models are presented in Table A3 (Appendix A). For the sake of efciency, simplicity,
and to achieve comparability in the magnitude of the effects across the different types of predictors in the presentation of the
results, I enter all of the ordered categorical predictors as linear inputs (summarizing the effect of education, for instance, in a
single coefcient rather than (K1 indicator variables, where K is the number of ordered categories). This linearity constraint
results in some loss of information, but the substantive story remains the same even when the linearity constraint is relaxed
and predictors enter as unordered categorical inputs (results available on request). To facilitate comparison of the magnitude
of the effects across inputs, I standardized each of the predictors by subtracting the mean and dividing by two times the
standard deviation as recommended by Gelman (2008). This makes the interpretation of the (ordered categorical) numeric
predictors such as education, income, and age straightforward: the coefcient estimate indicates the effect of moving from
the lowest to the highest value, while the binary predictors (gender, urban and marital status) retain their usual
interpretation.
4.5. Analytic strategy
4.5.1. The additive model

As noted above, one obvious answer to the question of the socio-demographic structuration of culture talk suggests that
this outcome should be most strongly predicted by culture-consumption. Insofar as it is direct experiences with cultural
goods and performances that come to serve as the source of the available content for culture talk we should expect that
individuals who are active culture consumers should also evince a higher propensity to engage in culture talk. However, my
main interest is status-linked differences in culture talk propensity that are not directly traceable to diversity in
consumption behavior. Thus, I specify an ordered probit model (Long & Freese, 2006) holding constant self-reported
consumption frequency, in order to ascertain whether there is residual variation across status groups in culture talk
propensity within-levels of self-reported consumption. This model takes the form:
EPY it j aj g t Ei dt C it Sbk X ik 1
Where Yit is the frequency of culture talk for individual i in cultural domain t, Ei is the educational attainment (or family
income) of that person, and Cit is the self-reported consumption frequency for each individual in that domain. The effect of
these last predictor is expected to be large and positive (?? > 0). Finally, Bk is an i k vector of coefcient estimates
corresponding to the effect of the other predictors included in the model (Xik). I refer to this specication as the additive
model. If there is residual variation linked to socioeconomic status in culture talk then we should expect that?? > 0, even after
adjusting for self-reported consumption frequency.
4.5.2. The multiplicative model

Alternatively, we may expect that the status effect should itself vary according to the level of self-reported consumption.
This would be the case, for instance, if persons of high socioeconomic status engage in culture talk at higher rates than
persons of low-education even when they consume culture at the same rates. I examine the validity of this hypothesis by
specifying a model of the form:
4 4
EPY it j aj g t Ei S dmt C imt S dmt C imt Ei Sbk X ik 2
m2 m2
This is the multiplicative model. In this model I enter the frequency of consumption as (4 1 = 3) binary indicators (Cimt)
and interact each with the socioeconomic status variable (Ei). As Ai and Norton (2003), have noted, the empirical validity of
Hypothesis 2 which says that the effect of education varies across self-reported consumption status cannot be veried
simply by looking at the t-statistics corresponding to each of the multiplicative effects. This is due to the fact that in non-
linear (e.g. logit or probit) models actual interaction effects may be present even if the t-statistic for the multiplicative terms
3
Family income information was not available for about one fth of the cases (20.8%) across the two survey waves. For these cases (N = 375) I imputed
their income using the regression-based imputation procedure in Stata 11. Essentially this involves using an ordered probit model to predict the missing
income values using all remaining socio-demographic information for that respondent (e.g. education, age, marital status and selected interactions.
4
I experimented with models that included a vector of dummy variables coding for race/ethnic identication and marital status. These variables had for
the most part non-statistically signicant effects in all of the equations. To avoid over-parameterizing the models and for the sake of parsimony, I excluded
these variables from the presentation of the results. Tables showing the coefcient estimates and t-statistics for these variables are available upon request.
Table 1
Coefcient Estimates for Predictors of Expected Frequency of Culture-Talk, Ordered Probit Estimates, 2002, 2004 NSS.
Movies Music Books Arts Plays

Education (centered) 0.26* 0.33* 0.28* 0.22* 0.25*
(4.80) (5.97) (4.99) (3.63) (4.35)
Family Income (centered) 0.08 0.17* 0.10* 0.07 0.18*

(1.48) (3.14) (1.91) (1.18) (3.19)
Age (centered) 0.45* 0.27* 0.25* 0.16* 0.16*

(-8.11) (-5.18) (-4.72) (-2.86) (-3.03)
Gender (Woman = 1) 0.23* 0.06 0.33* 0.19* 0.27*

(4.32) (1.07) (6.18) (3.42) (5.12)
Urban Location (Yes = 1) 0.05 0.19* 0.10* 0.08 0.23*

(0.95) (3.62) (1.92) (1.46) (4.22)
Consumption Freq. 1.04* 0.62* 1.08* 1.28* 1.05*

(17.39) (11.29) (19.09) (21.60) (18.39)
Tau1 1.15* 0.70* 0.91* 0.13* 0.36*

(-28.94) (-21.11) (-24.97) (-4.17) (-11.26)
Tau2 0.31* 0.13* 0.04 1.05* 0.76*

(-9.55) (4.16) (-1.22) (27.10) (21.69)
Tau3 0.71* 1.00* 0.79* 1.85* 1.61*

(20.83) (27.65) (22.80) (34.03) (34.03)
N 1791 1795 1792 1790 1793
BIC 4367.4 4724.7 4498.9 3634.9 4052.3
Model Chi2 568.7 266.4 529.5 609.6 517.8
is smaller than the conventional level (e.g. 1.96) corresponding to a statistically signicant effect at p = 0.05. To deal with
this issue, after estimating the model, I compute the marginal effect of moving from the lowest to the highest education level
on the probability of engaging in culture talk very often. I then plot the corresponding point estimate and 90% condence
interval across groups of respondents dened by self-reported frequency levels. The presence of a statistically discernible
upward trend in the marginal effects estimates would constitute positive evidence of the expected interaction between
status and frequency of consumption for that domain. I estimated ve ordered probit models (one for each cultural domain
available) of the additive and multiplicative form with the culture talk indicator as the response variable, the frequency of
consumption for that cultural domain and the two indicators of socioeconomic status (family income and education) as the
key predictors, and a series of socio-demographic variables as the additional predictors.
5. Results
5.1. Status-linked differences in the frequency of culture-talk
Hypothesis 1 proposes that a persons socio-economic status should predict culture talk even after adjusting for
participation frequency. To test this hypothesis, I specied an additive model (Eq. (1)) predicting culture talk while adjusting
for self-reported frequency of consumption. Table 1 shows the point estimates and associated t-statistics corresponding for
each of the predictors from these models. The results are consistent with this hypothesis. Looking at the effect of education, I
nd that there exist strong levels of status-linked heterogeneity in culture-talk even after holding constant self-reported
consumption behavior (p < 0.01).5 Persons who have attained advanced educational qualications are more likely to engage
in culture talk featuring popular culture and the arts as content than persons who have not, within levels of self-reported
participation frequency. Family income also predicts culture talk net of education and consumption frequency, although it is
only statistically signicant in three of the ve cultural domains. In all, these results provide strong evidence for the
proposition linking culture talk to social position within levels of same-domain culture consumption.
Hypothesis 2 states that not all indicators of socioeconomic status are created equal when it comes to predicting culture
talk. Instead, we should nd that indicators of socioeconomic status closer to the cultural capital pole, such as education, are
5
For those interested in whether culture consumption is predicted by the same factors as culture talk, Table A1 in the Appendix A shows the results of
models with consumption frequency as the main outcome across all ve domains, with the same set of sociodemographic predictors.
better predictors of culture talk that dimensions of socioeconomic status associated with economic capital (Bourdieu, 1986).
To test this hypothesis, I proceed in two steps. First, I specify a model that constrains the coefcient estimates for education
and family income to be the same and compare their t (using Walds likelihood ratio criterion) to a model in which they are
allowed to vary freely for each of the ve cultural domains. The results provide partial support for this hypothesis. We can
reject the hypothesis that the magnitude of the coefcient estimates for education and income is the same for both Movies
and Books (p < 0.05), reject it at a borderline level for Music (p = 0.07) and fail to reject for the ne art domains of Plays and
Art, where the substantive magnitude of the income effect seems to rival the effect of education as a predictor of culture talk.
The likelihood ratio test is based only on the statistical signicance of the coefcient estimates. A complementary test of
Hypothesis 2 directly compares the effect sizes of the two competing predictors. To that end, I compute the marginal effects
of income and education from the models specied earlier. As noted above, the marginal effect is the shift in probability in
the dependent variable for a unit change in the independent variable (which, given the standardization procedure I followed,
is a shift from the lowest to the highest value). These results are plotted in Fig. 1. The results provide even stronger support for
Hypothesis 2. Education matters more in a substantive sense than income for four out of ve outcomes, with Plays being the
only exception.
5.2. Heterogeneity of socioeconomic status effects by levels of engagement
Hypothesis 3 proposes that indicators of socioeconomic status are better predictors of culture talk among consumers who
are most actively engaged in culture consumption. To test this hypothesis I specify a multiplicative version of the
specication shown in Table 1 for all ve outcomes (Eq. (2)). In this model, the effect of socioeconomic status is allowed to
vary freely by levels of culture consumption. I select education as the main indicator of socioeconomic status for this test,
since as we saw earlier, this is a more consistent predictor of the outcome. How does the additive model compare to the
multiplicative specication? Fig. 2 shows the marginal effect of education on the probability of engaging in frequent culture-
talk (dened slightly different for commercial and ne arts activities) for respondents partitioned according to the levels of
self-reported engagement in that domain. Because education is standardized by dividing by two times the standard
deviation (Gelman 2008), the marginal effect here essentially refers to the shift in the probability of frequent culture talk as
Fig. 1. Marginal effects of (ordinal) standardized income and education predictors of the probability of frequent culture-talk.
Fig. 2. Marginal effects of education on the probability of engaging in culture talk very often (for movies, music and books) and either very or fairly often
(for arts and plays) by levels of cultural engagement in each domain, 2002 and 2004 NSS.
Table 2
Loadings for each of the ve culture consumption items on rst two factors, principal factor analysis with promax rotation.
Factor 1 Factor 2 Uniqueness

Read Books (for Pleasure) 0.637 0.239 0.610
Go to Museums/Galleries 0.783 0.105 0.415
Go to Live Performances 0.650 0.264 0.426
Listen to Music 0.091 0.882 0.252
Go to Movies 0.389 0.458 0.553
we move from the lowest to the highest level of educational credentials. As discussed earlier, the strength of the
multiplicative interaction between education and consumption can be gauged by the extent to which we observe any trends
in the marginal effect estimate as we move up from less frequent to more frequent engagement patterns.
The results provide fairly strong levels of support for Hypothesis 3. Upward trends in the marginal effect of education on
frequent levels of culture talk are strong and clear in lm book-reading, and to a lesser extent music. They are much weaker in
the case of dramatic performances.6 In the case of the visual arts, I simply nd a categorical distinction between the most
active participants and everybody else, with education-based differentiation in culture talk being strong for the former
group and non-existent for the latter. In all, these results provide fairly broad support for the hypothesis that status linked
differences in culture talk are much stronger for the most active culture consumers.
5.3. Heterogeneity of socioeconomic status effects across cultural domains
Hypothesis 4 predicts that status-based differentiation should be stronger for the traditional high-status cultural
domains, while Hypothesis 5 predicts that we should nd status-based differentiation in all domains. Estimates of the
income and education effects plotted in Fig. 1 can help us adjudicate between these competing predictions. If we take
education as the relevant dimension of socioeconomic status, it is clear that Hypothesis 5 is on the right track: Highly
educated respondents are more likely to engage in culture talk across all cultural domains. This pattern of results supports
Hypothesis 5 and contradicts the status differentiation argument expressed in Hypothesis 4. However, if we take income as
the socioeconomic status dimension of interest, then it seems that Hypothesis 4 receives some (equivocal) support, as
income predicts culture talk with regards to Plays but fails to predict talk in other cultural domains (with the exception of
music). However, given the fact that education is in fact the strongest and most consistent predictor, it seems like Hypothesis
5, suggesting that all elds are subject to a logic of intellectualization conducive to culture talk is the most likely candidate.
6
In this test, I dene frequent as choosing very often for Movies, Music, and Books and choosing either fairly often or very often for Arts and Plays, to
allow for more or less comparable probabilities across domains.
Table 3
Coefcient Estimates for Selected Predictors of Expected Frequency of Culture-Talk, Ordered Probit Estimates, 2002, 2004 NSS.

Fine Arts Factor 0.24* 0.47* 0.30* 0.36* 0.52*
(7.44) (15.47) (8.15) (7.80) (12.67)
Commercial Arts Factor 0.18* 0.26* 0.17* 0.04 0.11*
(5.60) (4.72) (5.48) (1.42) (3.41)
N 1780 1779 1778 1780 1781
BIC 4263.1 4385.8 4341.5 3553.7 3875.0
Model Chi2 653.4 577.8 663.0 684.7 681.8
5.4. Culture talk differences across commercial and ne arts audiences
Hypothesis 6 predicts that ne arts consumption is a better predictor of culture talk across domains than is engagement
with the commercial arts. In order to test this hypothesis, I proceed in two steps. First, I create two variables designed to tap
the relative likelihood of each respondent to engage in the commercial and ne arts. These variables come from a principal
factor analysis of the ve consumption frequency items, in which I retain only the two main factors (those with an eigenvalue
larger than 0.9).7 I use a promax rotation to allow the factors to be correlated with one another. The factor loadings for the
rst two principal factors are shown in Table 2. Factor loadings conform to our a priori classication of cultural domains:
reading, visiting museums and galleries and attending the theater load on a common factor, while going to the movies and
listening to the music load on a second factor. I use the regression-based scores from these two factors are taken as indicators
of the relative likelihood to engage the ne and commercial arts respectively for each individual.
In a second step, I introduce these two variables as predictors of culture talk in an ordered probit model that adjusts for
the same variables as those shown in Table 1. The coefcient estimates for these models are shown in Table 3 (I omit the
presentation of the other predictors for the sake of economy). To test the hypothesis that ne arts consumption is a better
predictor of culture talk than commercial arts consumption across all domains, I estimate a constrained model that restricts
these two coefcients to be the same. I then use a likelihood ratio test to check whether there is a signicant loss of model t
in comparison to the unconstrained model. The results provide strong support for hypothesis 6. Across all domains except for
lm, being a ne arts consumer is a better predictor of culture talk (p < 0.05).
6. Discussion
6.1. Summary of the argument and ndings
This paper furthers current understandings of the mechanisms connecting culture and status group stratication by
bringing attention to the socio-demographic correlates of an important phenomenon. The literature on culture stratication
has been primarily focused on the question of the determinants of lifestyle consumption patterns, with a view to establishing
whether there still exists a correlation between cultural choices and social position. This one-sided concern has led to the
neglect of an equally important question: that of the behavioral implications of familiarity and engagement with the arts. In
this respect current research has generally failed to consider the existence of stratication-linked differences in the extent to
which cultural knowledge is differentially put to use by members of different social strata as a resource for interaction.
To shed light on these issues, I synthesized insights from classical theoretical sources in the culture and stratication
literature, the connection between culture and integration into Simmelian social circles, Simmels account of culture and
sociability, and recent empirical and conceptual advances in our thinking of the relationship between culture and social
networks. I argued that one of the most important behavioral implications of familiarity with culture and the arts is its role as
the type of content that is useful for the generation of conversational interaction; in short, culture talk. This is a phenomenon
that deserves as much empirical and theoretical attention the patterns of cultural choice that have traditionally
monopolized the attention of culture and stratication researchers.
This paper contributes to the study of culture and stratication in several ways: analytically, I developed a framework that
partially meets my self-imposed challenge by highlighting the theoretical and empirical importance of the culture talk
phenomenon. I began with the Simmelian legacy on the importance of seemingly purposeless (non-instrumental and non-
goal-oriented) conversational content for the emergence of sociability (or the play form of association) and connected it
7
At the suggestion of an anonymous reader, I also estimated models predicting culture talk across all ve domains using an overall (the rst unrotated
factor from a principal factor analysis) score tapping general consumption propensities across domains. The results are shown in Table A2 of the Appendix A.
The key results in this specication are that, not surprisingly, the overall consumption score is a better predictor of culture talk than the domain-specic one
and that a good portion (but not all, with the exception of music) of the effect of socioeconomic status on culture talk is mediated via the generalized
propensity to engage culture across domains.
to the more contemporary stress by cultural stratication theorists on entertainment talk as the cement that holds status
groups together as micro-relational communities of acquaintance and informal membership (Collins, 2009; DiMaggio, 1987)
Second, I advance research in the now well-established literature linking taste, lifestyle consumption and stratication. I
do this by examining the socio-demographic structuring of culture talk as a constitutive component of lifestyle-based
stratication in contemporary societies. The results show that, consistent with prior theory in cultural stratication, culture
talk is stratied by the same markers of social position and status that previous research has shown also characterize
audience segmentation in the corresponding cultural domains. More specically, I show that culture talk in all domains
whether commercial or the ne arts is stratied by education, income, age and gender, and to some extent urban location
net of self-reported engagement in those domains. Those individuals endowed with larger stocks of cultural and economic
capital are more likely to routinely engage in culture talk, the young are more likely to engage in culture talk than the old,
women are more likely to engage in culture talk than men, and urbanites (in some domains) are more likely to engage in
culture talk than their suburbanite/rural-dwellers counterparts.
Most importantly, socioeconomic status structures culture talk across all cultural domains that I examined, but it does so
in a way that is structured by the expertise that comes with cultural engagement. This happens in two ways. First,
socioeconomic status (especially education) has a much stronger effect on culture talk among those respondents reporting
the highest level of engagement in that domain. Second, high levels of engagement in traditional ne art domains (such as
art museums, galleries, and plays) predicts culture talk across all domains whether traditionally ne arts or commercial arts,
suggesting that culture talk is driven by the intellectualization of aesthetic discourse that seems characteristic of an
increasing number of culture production and dissemination elds (Allen & Lincoln, 2004; Baumann, 2007; Regev, 2007).
6.2. Limitations and suggestions for future research
This empirical study is not without its limitations. As noted earlier, a key limitation of the NSS data source is the lack of
information on respondents occupation and family background. These are key dimensions of social differentiation that
should be, according to the theoretical argument laid out above, clearly linked to variations in both the accumulation of
dispositions towards certain forms of culture consumption, and associated tendencies to use culture talk as a conversational
resource. For instance, as above Lareau (2002, p. 763) notes there continue to exist systematic differences in the use of
language between middle and working class homes, with middle class children being exposed to a constant stream of speech
since early in life. These children are also encouraged to engage in prolonged conversations about abstract topics pertaining
to events outside the home, in which they are encouraged, in contrast to working class children to develop their own
opinions, judgments, and observations. It is thus very likely that the seeds of the capacity to use culture talk in interaction
are rooted in the domestic cultural capital fostered in the middle class household (Bourdieu, 1984). Subsequent work on
the accumulation dynamics of cultural capital should investigate differences in the propensity of culture-talk use by social
background. This research may focus in culture talk propensity differences (both in terms of frequency and content), for
instance, between movers into high-education strata and stayers.culture-talk
Occupational eld should also be linked to systematic differences, both across and within levels of education, in the
tendency to use culture talk in interaction. Certain elds, especially those concerned with the abstract manipulation of
symbols as well as knowledge and culture production occupations, lend themselves more readily to the recurrent activation
of latent capacities for the objectication of cultural works required to engage in culture talk. Thus it is likely that even when
considering individuals with the same objective educational qualications (e.g. a Bachelors or a Postgraduate degree) it is
likely that depending on the eld of specialization (e.g. interior design versus engineering) we should nd systematic
differences in the relative likelihood that individuals deploy culture talk for purposes of construction, maintenance and
reactivation of social relationships. It is also very likely that a good portion of the residual gender effect on cultural talk
documented above, may have a lot to do with the concentration of women in certain culture production and other front-
stage occupational elds in which facility with high-status cultural resources is a precondition for both entry and success. As
Erickson (1996) has shown, male dominated industries may actually discourage certain forms of cultural engagement
culture consumption (and by implication culture talk), especially concerning certain aesthetic matters that are considered
feminine pursuits, further contributing to both gender and within-education differences keyed to occupation in the
likelihood of observing the deployment of culture-talk in interaction. Future research on culture talk should concentrate on
empirical investigating the existence of heterogeneity in culture talk propensity across ne-grained Durkheimian micro-
classes comprised of distinct occupational groups (culture consumption Collins, 2009).
Another important limitation of the present study concerns the measurement of the culture-talk outcome itself. The NSS
data limits the present study to broad outcomes associated with culture consumption domains (e.g. art, movies, music) but
there is no measurement of within-domain differences in culture-talk linked to ne-grained (e.g. genre-level) differentiation
between different culture consumption activities (e.g. kinds of books, musical genres, types of movies, and so on). In addition,
there is the issue of the extent to which culture-talk is centered around cultural goods traditional conceived (expressive
cultural objects and performances) to include other domains such as sports, fashion, and cuisine (cultural domains initially
included in DiMaggio, 1987). This links up to both empirical issues regarding the existence of similar differentiation
dynamics in culture-talk propensity across other horizontal axes of difference beyond socioeconomic status and social
class as well theoretical issues regarding the conceptual bounds of the notion of culture-talk. For instance, it is clear that
generational differentiation in both the rate and content of culture-talk is linked to corresponding differentiation in ne-
grained categories of engagement with music, literature, and the arts (e.g. Hip Hop and Modern Art versus Classical Music
and Impressionist Art) as these become linked to dynamically changing new and old forms of cultural engagement
(Savage & Gayo, 2011).
In terms of the rst issue, it is possible to extend the broad argument regarding the pervasiveness and functional utility
of culture talk for the construction of social relationships to realms of differentiation that go beyond (and cut across)
socioeconomic status narrowly dened. For instance, as Carter (2005) has noted Black youth differentiate themselves in
terms of their levels of investment in what she refers to as dominant and nondominant cultural capital. While the former
is composed of class differentiated set of cultural signals, linguistic habits, cultural tastes, and practices characteristic of
middle class Whites, the latter consist of identity conrming tastes and lifestyle practices, speech styles, and sartorial
choices (e.g. such as listening to Rap and Hip and Hop) that are seen as strong signals of cultural authenticity and as
reafrming ingroup boundaries. The most successful minority youth are those who become cultural straddlers being able
to code-switch between dominant and nondominant styles of self-presentation depending on the situation and institutional
setting. It is clear that in this case, culture-talk is deeply involved in the construction and activation of social relationships,
but this culture talk will be differentiated by content depending on whether the relationships pertain to relationships with
co-ethnic others or with members of the dominant group. In this sense, extending the culture-talk argument to deal with
differentiation by race and ethnicity would require information that goes beyond rates of culture talk in broad cultural
domains to include more ne-grained forms of cultural familiarity.differentiated
Theorizing how culture talk plays into the generation of gender-based inequalities would also require going beyond
culture as traditionally conceived to include realms of cultural engagement that are more explicitly gendered in a way
that benets men. It has been traditionally found that there is a female advantage when it comes to the accumulation of
explicitly aesthetic cultural resources such as the plastic and performing arts (Christin, 2012) as well as in terms of
membership in the reading class (E. Long, 2003). However, other forms of cultural practice such as familiarity with and
regular engagement with organized sports (especially in the U.S.) continue to be highly gendered in a way that advantages
men. Thus, while culture-talk can be deployed in a way that allows persons to function as bridges (Schultz & Breiger, 2010),
it can also be used to build fences and reify gender boundaries (Erickson, 1996). Consistent with this idea, Turco (2010) nds
that the experiences of two types of tokens in a high status segment of the nancial industry (Black Men and Women) is
qualitatively different and this difference is keyed to the extent that women are systematically excluded from sites of
sociability, homosociality and occupational advancement built around interaction rituals centered on sports talk. This type
of culture-talk is pervasive in these types of workplaces where sports language, sport metaphors, and even self-
conceptions become routinized in insidious ways. This is a dynamic that has also been uncovered in professional settings less
explicitly gendered, such as the retail fashion industry (sports languageand this Purcell, 2013).
In all, this research shows, in line with Erickson (1996) that in certain male-dominated contexts culture-talk with move
away from that centered on the culture industry and their products (where womens expertise is equal or exceeds men) to
realms of cultural competence in which men are able to showcase higher levels of expertise. The theoretical lesson seems to
be that in extending the argument above to other axes of stratication beyond social class and education to include gender,
race, age, ethnicity, nationality (and even religion) may require extending the bases of culture talk beyond expertise with the
formal and content-based properties of aesthetic goods. Yet, this does not mean that the culture-talk concept has no natural
bounds. As noted above, due to the particular exigencies required of the forms of interaction typical of modern sociability
(that it be removed from explicit instrumental pursuits) not all domains of experience can serve as its basis. In particular I
argued that ludic domains, where persons consume objects and experiences in play form (up to an included
entertainment, the arts, and sports) are the privileged experiential sites of culture talk (prototypicalculture Fiske, 1987).
An empirical implication of this is that most conversation for enjoyment (DiMaggio, 1987) should feature content
related to these domains as its basis, while the introduction of topics from more serious domains, such as politics, the
economy, or religion should actually serve as a deterrent for culture mediated interaction. We should also expect other forms
of particularistic talk (e.g. that related to experiences circumscribed to a given domain such as family and work) to not be
good candidates for culture-talk (and thus to be unlikely to be featured as conversational content across different
relationships) for the simple reason that they lack transposability and thus can only be used to construct relationships with
specic others at a given social site. These are of course all empirical and not purely theoretical questions, and should be the
subject of future research, one that builds on the now non-negligible amount of work demonstrating the cultural, social and
institutional potency of the phenomenon (e.g. Long 2003; Purcell 2013; Rivera 2012; Turco 2010).
6.3. Implications for culture and inequality studies
The argument and the results reported in this paper have broader implications for contemporary theorizing on the
connection between culture, connectivity and structured inequality (yErickson, 2008). In this respect, the type of routine
deployment of cultural knowledge in concrete interaction that I have labeled as culture talk can be thought of as an
important relational mechanism (Gross, 2009), linking inequalities in the propensity to use aesthetic content as a
conversational resource to inequalities in relational outcomes associated to social connectivity and status-based boundaries.
These long-standing inequalities in connectivity (both in terms of range and diversity) have been duly noted in the network
literature, but have yet to receive a satisfactory treatment. Building on the results reported in this paper, it can be argued that
different rates of engagement in culture talk provides a micro-interactional mechanism accounting for the dynamic
persistence of cross-group inequalities on the major dimensions of social connectivity.
First, under the proposed account, individuals who have a larger propensity to engage in culture talk should also have an
easier time during the relationship-formation process (Lazarsfeld & Merton, 1954). This is a plausible micro-mechanism that
can help us to account for the existence stubborn status-linked inequalities in social capital. For instance, it is well-known
that education has been and continues to be one of the best predictors of network size in the United States (DiPrete, Gelman,
McCormick, Teitler, & Zheng, 2011; Fischer, 1982; Marsden, 1987) and other national settings (Mollenhorst, Vlker, & Flap,
2011) today. As we have seen, the higher-educated are also more likely to engage in culture talk than the less-educated
(within-levels of consumption), and thus they should also be more likely to routinely use this as an interactional resource to
form new relationships. The same can be said for those socio-demographic groups known to have smaller networks than
average such as the less-educated elderly (Cornwell, Laumann, & Schumm, 2008). Culture talk propensities decline with age,
consistent with the claim that rates of relationship formation may underlie these differences.
Second, persons who belong to socio-demographic groups that exhibit a larger propensity to engage in culture talk should
also tend to be embedded in status segregated networks. Social ties that are partially constituted via culture talk retain this
property because they are dynamically sustained via conversational meanings and contents access to which is also status-
segregated (Collins, 2009), as well as the relationship level meaningful histories built around those culture-mediated
interactions (Fine & Kleinman, 1983). In this respect, differential rates of culture talk across socio-demographic groups
emerge as a plausible mechanism that accounts for the existence of status-segregated social networks even in the absence of
any strictly instrumental attempt on the part of individuals to select status-homogamous contacts.conscious
Third, it is well-established that the social-networks of Americans are characterized by high-rates of status-endogamy
specially premised on race and education (DiPrete et al., 2011), and that marital homogamy premised on education and
income has increased at the expense of marital endogamy premised on other axes of difference (Kalmijn, 1994). Persons who
belong to high-education or high-income groups are also more likely to engage in culture talk than the less-educated
(within-levels of consumption). Following the same line of reasoning, they should also be expected to be more likely to
routinely use this as an interactional resource to recurrently recreate status-segregated conversational rituals (Collins,
2004).
Finally, those individuals who have a more pronounced propensity to engage in culture talk should also tend to be more
likely to bridge structural and cultural holes (Lizardo, 2014; Pachucki & Breiger, 2010). This is a phenomenon that has been
referred by Schultz and Breiger (2010) as the strength of weak culture. Status-linked inequalities in culture talk
propensities can thus be proposed as a mechanism that explains the often noted fact that high-status individuals are more
likely to be connected to a more diverse pool of others than low status individuals (DiPrete et al., 2011; Erickson, 1996;
Marsden, 1987). With persons of high education and/or high-income more likely to engage in culture talk than the less-
educated, it stands to reason that they should also be able toin some settings (Purcell, 2013) use this cultural expertise to
make sporadic connections across group and status boundaries.
While the validity and scope of these hypotheses remains to be more fully explored, the results reported in this paper
along with convergent evidence from other sources (Rivera, 2012; Schultz & Breiger, 2010; Turco, 2010) strongly point in
this direction, suggesting exciting prospects for future research in this area. In this manner, a focus on the culture talk
phenomenon can serve to cross-fertilize work in the culture and stratication literature and the sociology of taste with
research on inequalities in network access and social capital.research
Acknolwedgements
None.
Appendix A
See Tables A1A3
Table A1
Coefcient Estimates for Predictors of Frequency of Cultural Participation, Ordered Probit Estimates, 2002, 2004 NSS.

Education (centered) 0.25* 0.09 0.36* 0.59* 0.43*
(4.51) (1.52) (6.58) (9.83) (7.76)
Family Income (centered) 0.28* 0.01 0.05 0.21* 0.24*

(5.18) (-0.14) (0.99) (3.54) (4.46)
Age (centered) 0.77* 0.60* 0.10* 0.10* 0.22*

(-14.34) (-9.76) (1.87) (-1.73) (-4.23)
Gender (Woman = 1) 0.12* 0.17* 0.36* 0.10* 0.04

Table A1 (Continued)
(2.37) (3.02) (6.93) (1.79) (0.70)
Urban Location (Yes = 1) 0.23* 0.04 0.10* 0.18* 0.23*

(4.44) (0.64) (1.87) (3.14) (4.30)
Tau1 0.51* 1.70* 0.90* 0.14* 0.34*

(-15.87) (-33.37) (-25.89) (4.47) (-10.91)
Tau2 0.59* 0.99* 0.18* 1.08* 0.60*

(18.26) (-27.46) (-6.08) (28.92) (18.68)
Tau3 1.33* 0.36* 0.43* 1.88* 1.40*

(32.33) (-11.77) (13.95) (33.17) (32.90)
N 1792 1798 1795 1792 1794

BIC 4401.5 3603.4 4847.0 3715.8 4472.5
Model Chi2 280.5 106.5 108.9 163.8 150.2
Table A2
Predicting Culture Talk Frequency Using Overall Consumption Factor Score.

Consumption Factor Score 0.67* 0.86* 0.69* 0.84* 0.94*
(16.29) (20.33) (16.88) (19.46) (21.59)
Education (centered) 0.12* 0.08 0.19* 0.22* 0.14*

(2.15) (1.38) (3.33) (3.62) (2.38)
Family Income (centered) 0.06 0.03 0.01 0.02 0.13*

(1.06) (0.62) (0.22) (0.40) (2.26)
Age (centered) 0.60* 0.21* 0.00 0.04 0.00

(-10.89) (-3.83) (-0.03) (0.69) (-0.08)
Gender (Woman = 1) 0.20* 0.00 0.41* 0.13* 0.20*

(3.76) (-0.04) (7.82) (2.41) (3.71)
Urban Location (Yes = 1) 0.03 0.06 0.03 0.03 0.19*

(0.61) (1.22) (0.61) (0.59) (3.46)
Tau1 1.14* 0.79* 0.89* 0.10* 0.39*

(-28.79) (-22.24) (-24.56) (-3.25) (-11.64)
Tau2 0.31* 0.14* 0.04 1.03* 0.80*

(-9.51) (4.40) (-1.18) (27.11) (22.12)
Tau3 0.71* 1.10* 0.77* 1.79* 1.69*

(20.68) (28.57) (22.35) (34.17) (34.36)
N 1780 1779 1778 1780 1781

BIC 4368.1 4378.9 4537.2 3700.7 3875.0
Model Chi2 533.4 569.7 452.3 522.8 666.8
Table A3
Descriptive statistics for predictor, control, and outcome variables included in the analyses.
Mean SD Min Max

Consumption Freq. (Books) 2.72 1.13 1 4
Consumption Freq. (Art) 1.65 0.83 1 4
Consumption Freq. (Plays) 2.01 0.97 1 4
Cult. Consumption (Music) 3.41 0.89 1 4
Consumption Freq. (Movies) 2.09 0.97 1 4
Cult. Talk (Movies) 2.72 1.05 1 4
Cult. Talk (Books) 2.55 1.10 1 4
Table A3 (Continued)
Mean SD Min Max
Cult. Talk (Plays) 1.98 0.95 1 4
Cult. Talk (Art) 1.80 0.91 1 4
Cult. Talk (Music) 2.38 1.05 1 4
Respondent's Educ. 4.04 1.21 1 6
Family Income 3.35 1.78 1 6
Respondent's Age 4.58 2.02 1 7
Gender (Woman = 1) 0.50 0.50 0 1
Urban Location 0.60 0.49 0 1
Observations 1774
References
Ai, C., & Norton, E. C. (2003). Interaction terms in logit and probit models. Economics Letters, 80(1), 123129.
Alexander, V. D. (1996). Pictures at an exhibition: Conicting pressures in museums and the display of art. The American Journal of Sociology, 101(4), 797839.
Allen, M. P., & Lincoln, A. E. (2004). Critical discourse and the cultural consecration of American lms. Social Forces, 82(3), 871894.
Baumann, S. (2007). Hollywood highbrow: From entertainment to art. Princeton University Press.
Bennett, T., Savage, M., Silva, E., Warde, A., Gayo-Cal, M., & Wright, D. (2009). Culture, class, distinction. London: Routledge.
Benzecry, C., & Collins, R. (2014). The high of cultural experience: Toward a microsociology of cultural consumption. Sociological Theory, 32(4), 307326.
Bourdieu, P. (1984). Distinc tion: A social critique of the judgement of taste. Harvard University Press.
Bourdieu, P. (1986). The forms of capital. In J. Richardson (Ed.), Handbook of theory and research for the sociology of education (pp. 241258).New York:
Greenwood Press.
Bourdieu, P. (1993). The field of cultural production: Essays on art and literature. New York: Columbia University Press.
Cardon, D., & Granjon, F. (2005). Social networks and cultural practices: A case study of young avid screen users in France. Social Networks, 27(4), 301315.
Carter, P. L. (2005). Keepin it real: School success beyond black and white. Oxford University Press.
Chan, T. W., & Goldthorpe, J. H. (2007). Social stratication and cultural consumption: Music in England. European Sociological Review, 23(1), 119.
Christin, A. (2012). Gender and highbrow cultural participation in the United States. Poetics, 40(5), 423443.
Collins, R. (2004). Interaction ritual chains. Princeton, NJ: Princeton University Press.
Collins, R. (2009). Conflict sociology: A sociological classic updated. Paradigm Publishers.
Cornwell, B., Laumann, E. O., & Schumm, L. P. (2008). The social connectedness of older adults: A national prole*. American Sociological Review, 73(2), 185
203.
Deighton, J. (1992). The consumption of performance. The Journal of Consumer Research, 19(3), 362372.
DiMaggio, P., & Mohr, J. (1985). Cultural capital, educational attainment, and marital selection. The American Journal of Sociology, 90(6), 12311261.
DiMaggio, P., & Mukhtar, T. (2004). Arts participation as cultural capital in the United States, 19822002: Signs of decline? Poetics, 32(2), 169194.
DiMaggio, P. (1987). Classication in art. American Sociological Review, 52(4), 440455.
DiMaggio, P. (1996). Are art-museum visitors different from other people? The relationship between attendance and social and political attitudes in the
United States. Poetics, 24(24), 161180.
DiPrete, T. A., & Eirich, G. M. (2006). Cumulative advantage as a mechanism for inequality: A review of theoretical and empirical developments. Annual
Review of Sociology, 32(1), 271297.
DiPrete, T. A., Gelman, A., McCormick, T., Teitler, J., & Zheng, T. (2011). Segregation in social networks based on acquaintanceship and trust1. The American
Journal of Sociology, 116(4), 12341283.
Douglas, M., & Isherwood, B. (1996). The world of goods: Towards an anthropology of consumerism. London: Routledge.
Eliasoph, N. (1998). Avoiding politics: How Americans produce apathy in everyday life. Cambridge University Press.
Erickson, B. H. (1991). What is good taste good for? Canadian Review of Sociology/Revue Canadienne De Sociologie, 28(2), 255278.
Erickson, B. H. (1996). Culture, class, and connections. American Journal of Sociology, 102(1), 217251.
Erickson, B. H. (2008). The crisis in culture and inequality. In S. J. Tepper, & B. Ivey (Eds.), Engaging art: The next great transformation of Americas cultural life
(pp. 343362).London: Routledge.
Fine, G. A., & Kleinman, S. (1983). Network and meaning: An interactionist approach to structure. Symbolic Interaction, 6(1), 97110.
Fine, G. A. (1977). Popular culture and social interaction: Production, consumption, and usage. Journal of Popular Culture, 11(2), 453466.
Fischer, C. S. (1982). To dwell among friends: Personal networks in town and city. University of Chicago Press.
Fiske, J. (1987). Television culture. Methuen.
Frith, S. (1998). Performing rites: On the value of popular music. Harvard University Press.
Fuhse, J. A. (2009). The meaning structure of social networks. Sociological Theory, 27(1), 5173.
Gainer, B. (1995). Ritual and relationships: Interpersonal inuences on shared consumption. Journal of Business Research, 32(3), 253260.
Gelman, A. (2008). Scaling regression inputs by dividing by two standard deviations. Statistics in Medicine, 27(15), 28652873.
Goldberg, A., Hannan, M. T., & Kovcs, B. (2016). What does it mean to span cultural boundaries? Variety and atypicality in cultural consumption. American
Sociological Review, 81(2), 215241.
Gross, N. (2009). A pragmatist theory of social mechanisms. American Sociological Review, 74(3), 358379.
Holbrook, M. B., & Hirschman, E. C. (1982). The experiential aspects of consumption: Consumer fantasies, feelings, and fun. The Journal of Consumer Research,
9(2), 132140.
Holt, D. B. (1998). Does cultural capital structure american consumption? The Journal of Consumer Research, 25(1), 125.
Ikegami, E. (2005). Bonds of civility: Aesthetic networks and the political origins of Japanese culture. Cambridge University Press.
Kadushin, C. (1966). The Friends and Supporters of Psychotherapy: On social circles in urban life. American Sociological Review, 31(6), 786802.
Kalmijn, M. (1994). Assortative mating by cultural and economic occupational status. The American Journal of Sociology, 100(2), 422452.
Katz-Gerro, T. (2004). Cultural consumption research: Review of methodology, theory, and consequence. International Review of Sociology, 14(1), 1129.
Katz-Gerro, T. (2006). Comparative evidence of inequality In cultural preferences gender, class, and family status. Sociological Spectrum: The Official Journal of
the Mid-South Sociological Association, 26(1), 6383.
Khan, S. R. (2011). Privilege. Princeton: Princeton University Press.
Lamont, M., & Lareau, A. (1988). Cultural capital: Allusions, gaps and glissandos in recent theoretical developments. Sociological Theory, 6(2), 153168.
Lareau, A. (2002). Invisible inequality: Social class and childrearing in black families and white families. American Sociological Review, 67(5), 747776.
Lazarsfeld, P. F., & Merton, R. K. (1954). Friendship as a social process: A substantive and methodological analysis. In M. Berger, R. M. MacIver, T. Abel, & C. H.
Page (Eds.), Freedom and control in modern society: (Vol. 18. pp. 1866).New York: Van Nostrand.
Lewis, K., Kaufman, J., Gonzalez, M., Wimmer, A., & Christakis, N. (2008). Tastes, ties, and time: A new social network dataset using Facebook.com. Social
Networks, 30(4), 330342.
Lizardo, O., & Skiles, S. (2008). Cultural consumption in the ne and popular arts realms. Sociology Compass, 2(2), 485502.
Lizardo, O., & Skiles, S. (2009). Highbrow omnivorousness on the small screen? Cultural industry systems and patterns of cultural choice in Europe. Poetics,
37(1), 123.
Lizardo, O., & Skiles, S. (2012). Reconceptualizing and theorizing omnivorousness: Genetic and relational mechanisms. Sociological Theory, 30(4), 263282.
Lizardo, O. (2006). How cultural tastes shape personal networks. American Sociological Review, 71(5), 778807.
Lizardo, O. (2008). The question of culture consumption and stratication revisited. Sociologica, Vol. 2(2008) Retrieved from http://www.rivisteweb.it/doi/
10.2383/27709.
Lizardo, O. (2013). Variety in cultural choice and the activation of social ties. Social Science Research, 42(2), 321330.
Lizardo, O. (2014). Omnivorousness as the bridging of cultural holes: A measurement strategy. Theory and Society, 43(34), 395419.
Long, J. S., & Freese, J. (2006). Regression models for categorical dependent variables using stata, 2nd ed. Taylor & Francis.
Long, E. (2003). Book clubs: Women and the uses of reading in everyday life. University of Chicago Press.
Marsden, P. V. (1987). Core discussion networks of Americans. American Sociological Review, 52(1), 122131.
Miringoff, M.-L., Miringoff, M., Opdycke, S., & Hoynes, W. (2016). National social survey [United States] (Version ICPSR35588-v2). Ann Arbor, MI: Inter-
university Consortium for Political and Social Research [distributor]. Retrieved from http://doi.org/10.3886/ICPSR35588.v2.
Mohr, J., & DiMaggio, P. (1995). The intergenerational transmission of cultural capital. Research in Social Stratification and Mobility, 14, 167200.
Mollenhorst, G., Vlker, B., & Flap, H. (2011). Context overlap in personal networks: Are publicand PrivateTwo separate worlds? European Sociological
Reviewjcr034.
Ollivier, M. (2008). Modes of openness to cultural diversity: Humanist, populist, practical, and indifferent. Poetics, 36(23), 120147.
Ostrower, F. (1998). The arts as cultural capital among elites: Bourdieus theory reconsidered. Poetics, 26(1), 4353.
Pachucki, M. A., & Breiger, R. L. (2010). Cultural holes: Beyond relationality in social networks and culture. Annual Review of Sociology, 36(1), 205224.
Peterson, R. A., & Kern, R. M. (1996). Changing highbrow taste: From snob to omnivore. American Sociological Review, 61(5), 900907.
Peterson, R. A. (1992). Understanding audience segmentation: From elite and mass to omnivore and univore. Poetics, 21(4), 243258.
Peterson, R. A. (2005). Problems in comparative research: The example of omnivorousness. Poetics, 33(56), 257282.
Prieur, A., & Savage, M. (2011). Updating cultural capital theory: A discussion based on studies in Denmark and in Britain. Poetics, 39(6), 566580.
Purcell, D. (2013). Baseball, beer, and bulgari: Examining cultural capital and gender inequality in a retail fashion corporation. Journal of Contemporary
Ethnography, 42(3), 291319.
Regev, M. (1994). Producing artistic value. The Sociological Quarterly, 35(1), 85102.
Regev, M. (2007). Cultural uniqueness and aesthetic cosmopolitanism. European Journal of Social Theory, 10(1), 123138.
Rivera, L. A. (2012). Hiring as cultural matching: The case of elite professional service rms. American Sociological Review, 77(6), 9991022.
Savage, M., & Gayo, M. (2011). Unravelling the omnivore: A eld analysis of contemporary musical taste in the United Kingdom. Poetics, 39(5), 337357.
Schultz, J., & Breiger, R. L. (2010). The strength of weak culture. Poetics, 38(6), 610624.
Simmel, G. (1949). The sociology of sociability. The American Journal of Sociology, 55(3), 254261.
Thornton, S. (1996). Club cultures: Music, media, and subcultural capital. Wesleyan University Press.
Tomlinson, M. (2003). Lifestyle and social class. European Sociological Review, 19(1), 97111.
Turco, C. J. (2010). Cultural foundations of tokenism evidence from the leveraged buyout industry. American Sociological Review, 75(6), 894913.
Warde, A., Wright, D., & Gayo-Cal, M. (2008). The omnivorous orientation in the UK. Poetics, 36(2), 148165.
Weber, W. (1977). Mass culture and the reshaping of european musical taste, 17701870. International Review of the Aesthetics and Sociology of Music, 8(1), 5
22.
Weber, M. (1994). Sociological writings. (W.V. heydebrand, trans.). Continuum International.
Weininger, E. B., Lareau, A., & Conley, D. (2015). What money doesnt buy: Class resources and children's participation in organized extracurricular activities.
Social Forces, 94(2), 479503.
Omar Lizardo is an Associate Professor of Sociology at the University of Notre Dame and a member of the Interdisciplinary Center for Network Science and
Applications. His research deals with various topics in sociology, social psychology, cultural sociology, network theory, and cognitive science. His work has
appeared in such venues as American Sociological Review, Theory and Society, Sociological Theory, Poetics, Cultural Sociology, and Social Forces. With Jessica
Collett, he was the guest-editor of the June 2014 special issue of Social Psychology Quarterly dedicated to advancing the connections between Social
Psychology and Cultural Sociology. He is currently a member of the editorial advisory board of more than six journals, including Social Forces, Theory and
Society, Poetics, and Sociological Forum, and with Rory McVeigh and Sarah Mustillo he is the one of the current co-editors of American Sociological Review.

Lizardo, 2016

Uploaded by

Copyright:

Available Formats

Lizardo, 2016

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lizardo, 2016

Uploaded by

Copyright:

Available Formats

Poetics 58 (2016) 117

Contents lists available at ScienceDirect

Why cultural matters matter: Culture talk as the

E-mail address: olizardo@nd.edu (O. Lizardo).

2. Culture talk as sociability

2.1. The importance of aesthetic discourse as conversational content

2.2. Culture industries as the source of aesthetic content

2.3. Cultural content as a strategic interactional resource

3. The stratication of culture talk

3.1. Culture talk and socioeconomic status

3.2. Cumulative advantage processes

3.3. Cross-domain differences in the stratication of culture talk

4. Data and variables

4.1. Data source

4.2. Outcome variables

4.3. Main predictors

4.3.1. Indicators of culture consumption frequency

4.3.2. Indicators of socioeconomic status

4.4. Additional predictors

4.5. Analytic strategy

4.5.1. The additive model

4.5.2. The multiplicative model

Movies Music Books Arts Plays

Family Income (centered) 0.08 0.17* 0.10* 0.07 0.18*

Age (centered) 0.45* 0.27* 0.25* 0.16* 0.16*

Gender (Woman = 1) 0.23* 0.06 0.33* 0.19* 0.27*

Urban Location (Yes = 1) 0.05 0.19* 0.10* 0.08 0.23*

Consumption Freq. 1.04* 0.62* 1.08* 1.28* 1.05*

Tau1 1.15* 0.70* 0.91* 0.13* 0.36*

Tau2 0.31* 0.13* 0.04 1.05* 0.76*

Tau3 0.71* 1.00* 0.79* 1.85* 1.61*

5.1. Status-linked differences in the frequency of culture-talk

5.2. Heterogeneity of socioeconomic status effects by levels of engagement

Factor 1 Factor 2 Uniqueness

5.3. Heterogeneity of socioeconomic status effects across cultural domains

Movies Music Books Arts Plays

5.4. Culture talk differences across commercial and ne arts audiences

6.1. Summary of the argument and ndings

6.2. Limitations and suggestions for future research

6.3. Implications for culture and inequality studies

See Tables A1A3

Movies Music Books Arts Plays

Family Income (centered) 0.28* 0.01 0.05 0.21* 0.24*

Age (centered) 0.77* 0.60* 0.10* 0.10* 0.22*

Gender (Woman = 1) 0.12* 0.17* 0.36* 0.10* 0.04

Urban Location (Yes = 1) 0.23* 0.04 0.10* 0.18* 0.23*

Tau1 0.51* 1.70* 0.90* 0.14* 0.34*

Tau2 0.59* 0.99* 0.18* 1.08* 0.60*

Tau3 1.33* 0.36* 0.43* 1.88* 1.40*

N 1792 1798 1795 1792 1794

Movies Music Books Arts Plays

Education (centered) 0.12* 0.08 0.19* 0.22* 0.14*

Family Income (centered) 0.06 0.03 0.01 0.02 0.13*

Age (centered) 0.60* 0.21* 0.00 0.04 0.00

Gender (Woman = 1) 0.20* 0.00 0.41* 0.13* 0.20*

Urban Location (Yes = 1) 0.03 0.06 0.03 0.03 0.19*

Tau1 1.14* 0.79* 0.89* 0.10* 0.39*

Tau2 0.31* 0.14* 0.04 1.03* 0.80*

Tau3 0.71* 1.10* 0.77* 1.79* 1.69*