Measuring Cultural Capital: Taste and Legitimate Culture of Czech Youth

Introduction

1.1 The flourishing field of empirical research into the relationship between social stratification and taste originated with Bourdieu's Distinction (1984). This seminal work drew attention to workings of cultural capital, resource accumulated in the milieu of well-off families and elite educational institutions. Its nature was mostly connected with acquaintance of so-called highbrow culture – classical music, literature and theatre. Despite its influence, this pivotal concept was not spared of objections and critical reappraisals (e.g. Bennett & Silva 2011; Prieur & Savage 2011; Sullivan 2007; Warde 2008). One of the first critical discussions of cultural capital involved the notion of omnivorousness derived from the empirical discovery that the upper classes are distinctive by wide ranging tastes rather than focusing solely on highbrow culture. This evidence was used to contest the relevance of cultural hierarchies in contemporary society (see e.g., Lizardo & Skiles 2016; Peterson 2005). Recently, this objection was superseded by a more sympathetic inquiry into new emerging forms of cultural capital (Prieur & Savage 2013; Roose 2014). Despite the new development of methodological tools employed in the study of cultural capital (Friedman et al. 2015; see also Beer & Taylor 2012), there remains a major gap between accounts of institutions legitimizing cultural products (e.g. Allen & Lincoln 2004; Kersten & Verboord 2014; Verboord et al. 2015) and survey-based studies of the distribution of taste. The aim of this paper is to explore the possibility of incorporating institutional evaluation of cultural taste into survey-based data.

1.2 When examining the relationship between the differentiation of taste and social inequalities, the notion of legitimate culture plays a crucial role (Lamont & Lareau 1988; Lareau & Weininger 2003). Institutions that establish cultural legitimacy (e. g. universities, museums or elite media) imbue unequal value into cultural goods. This cultural hierarchy allows us not only to speak about differences in taste, but also about unequal tastes, eventually considering aesthetic judgment to be a key part of the accumulation of cultural capital (Warde 2008).

1.3 The role of legitimate culture in present day society is widely disputed in the current research. The crumbling of the privileged structure of the fine arts caused some to maintain that in the light of the dispersion of legitimate forms of cultural capital, we ought to search for its new forms (DiMaggio & Mukhtar 2004; Prieur & Savage 2011, 2013; Roose 2014). The competition from popular forms of culture, the de-institutionalisation of high culture itself, and the emerging omnivorous taste (DiMaggio & Mukhtar 2004) have been causing a slow but steady decline of its importance for social distinction. Prieur and Savage (2011) equate highbrow culture with legitimate culture, and blurring boundaries of the arts in this way encourage them to jettison the concept of legitimate culture as an integral part of cultural capital. Instead, what becomes important is a search for new dimensions that can define differences in taste more effectively, such as participation versus disengagement, or the cosmopolitan versus the national.

1.4 Recent methodological innovations in this field of research, such as the implementation of Multiple Correspondence Analysis (MCA), one of Bourdieu's authentic analytical tools (Bennett et al. 2009; Gripsrud et al. 2011; Hanquinet et al. 2014; Roose 2014, 2016), generate a more imaginative picture of contested social fields from quantitative data. However, despite these developments, studies integrating the institutional embeddedness of cultural capital into accounts of the distribution of individual tastes in the general population are still lacking.^[1] The intrinsic tendency of survey methodology to fall back on the paradigm of methodological individualism complicates the empirical incorporation of the institutional dimension of cultural production into the survey data.

1.5 Studies focusing on the institutional dimension of cultural production show that the actual content of legitimate taste could be prone to significant change, as is the case with the shifting tastes of the general public. Through the content analysis of movie reviews, Baumann (2001) shows how cinematography was institutionalised as a legitimate field of art, adopting the fine arts' criteria of evaluation. Along similar lines, Fishman and Lizardo (2013) show the impact of different educational systems on the complex composition of the aesthetic field. Therefore, discussion of emerging forms of cultural capital should take into account how current institutions, such as the state, the educational system, the media, professional criticism, cultural organizations, or artists' celebrities, imbue cultural products with legitimacy.

1.6 This paper explores the possibility of integrating empirically-based measures to determine the institutional evaluation of cultural products into traditional survey data of the distribution of taste. Rather than dismissing the role of legitimate culture on the basis of an a priori assumption about its content, it is more appropriate to address this issue empirically. A survey of the cultural consumption is used to tackle this issue in a limited social context of the Czech younger generation.^[2] If the proposition about the crumbling of highbrow culture legitimacy is true (DiMaggio & Mukhtar 2004), it should probably be most expressed in cultural consumption of youth.

1.7 This study aims to minimize a priori assumptions about the structure of taste. Open-ended questions are used to gather data from the respondents about their favourite cultural products in the fields of music, cinema, and literature. Using this innovative approach, the legitimacy of the respondents' tastes is then assessed by comparing their answers with the professional critics' judgement of respective cultural items. The results of this approach are then juxtaposed with the results of other standard methods of measuring cultural capital.

1.8 The next section discusses some of the major approaches to the measurement of cultural capital. The third section discusses the integration of the empirical study of cultural legitimation into the survey data, which develops into the formulation of a specific research question. The fourth section is dedicated to a description of the data and the methods of measurement. Section 5 presents the results of the analysis and is followed by the concluding discussion.

Measuring cultural capital

2.1 Various methods are used to measure the volume of cultural capital in current sociological research, ranging from the simple indicators as achieved educational level to the more elaborate examinations of various components, such as taste, practice, and knowledge, in several cultural fields. There are a few general strategies how to operationalise cultural capital in quantitative research, guided by important but often unquestioned assumptions with significant empirical consequences (see also Robette & Roueff 2014).

2.2 These assumptions usually underlie studies not directly engaging the nature of cultural capital, which often employ widely available but simplified ways of measurement. Firstly, it is quite commonplace to equate cultural capital and formal indications of its institutionalised form, e.g. achieved level of education. Such approaches are supported by the homology thesis which associates tastes and cultural practices with social position. Assuming the homology between social space and the space of judgement, this approach completely bypasses the issue of taste and its content. People in a privileged position are thus by definition assumed to have both accumulated cultural capital and acquired legitimate taste.

2.3 The second common strategy identifies several cultural practices traditionally understood as highbrow culture, such as visiting a museum or attending an opera or classical music performance. On the basis of an a priori assumption about the essential nature of these activities as somehow necessarily being a part of legitimate culture, such practices are used as an indication of cultural capital (van Eijck & Bargeman 2004; van Hek & Kraaykamp 2013). These assumptions are easily contested by the changing nature of legitimate culture over time and across geographical space (Baumann 2001; Bourdieu 1998). Such an a priori approach can be criticised on the grounds of its excessive reliance on an intuitive understanding of legitimate culture, without an empirical inquiry into the actual cultural content appreciated by current cultural institutions (Warde 2008; Warde & Gayo-Cal 2009).

2.4 Studies exploring the content of cultural capital are more sensitive to its contextual nature. The first approach relies on the researcher's a priori understanding of the legitimacy of taste in order to confine the scope of legitimate culture to the traditional definitions of fine arts and highbrow culture (Daenekindt & Roose 2014; Gripsrud et al. 2011). Again, determining the content of legitimate culture via a priori assumptions often leads to results showing its decreasing importance and the blurring of cultural boundaries, resulting in the search for new forms of cultural capital (Gripsrud et al. 2011; Prieur & Savage 2013). Little attention is paid to the actual institutional mechanisms of ascribing legitimacy (Allen & Lincoln 2004;Lizardo 2008; Verboord et al. 2015; Warde 2008), whether through the evaluative work of critics' in the form of reviews (Baumann 2001), canonical pedagogical practices (Fishman & Lizardo 2013), or the activities of cultural intermediaries in general (Maguire & Matthews 2014).

2.5 The second branch of research determines legitimacy based on the assumption of homology thesis mentioned above. The legitimacy of cultural practices is derived from its disproportional distribution among dominant groups, measured in terms of various indicators of social position, such as education (Warde & Gayo-Cal 2009) or amalgamation of socioeconomic variables (Prieur et al. 2008, similar logic is employed in Bennett et al. 2009 or Savage et al. 2013). As an example, Warde and Gayo-Cal's (2009) innovative, straightforward, and transparent assessment of the legitimacy of various cultural practices is based on a comparison of the positive and negative evaluation of these practices by university graduates and respondents without qualification. While this logic faithfully mirrors the intersubjective nature of cultural capital conceptualized in terms of 'widely shared, high status cultural signals' (Lamont & Lareau 1988 p. 156), it doesn't directly acknowledge the relevance of the evaluative and legitimizing functions of cultural institutions, and from a methodological point of view, it is at the edge of circular reasoning. The dominant class necessarily becomes the wielder of legitimate culture if the legitimacy of cultural practice is defined by the social status of its practitioner.

Institutional legitimacy and survey data

3.1 The existing quantitative research of cultural capital often neglects the institutional dimension of legitimising taste and withdraws from an attempt to determine legitimacy in an empirically-based manner. In fact, it appears as though there are no studies which attempt to connect the respondents' judgements of taste with other empirical measurement of the cultural product's legitimacy. The enactment of cultural legitimacy itself is, on the other hand, a well-studied process.

3.2 The principles of evaluation and consecration are studied especially in the context of film production (Allen & Lincoln 2004; Kersten & Verboord 2014), but also literature (Verboord et al. 2015) and specific niche genres and fields (Allington 2015; Dubois & Méon 2013). They show the nature of criteria that leads to cultural recognition, or even consecration in 'halls of fame' of the cultural world. The evaluation of cultural products is often divided into popular recognition, peer recognition and critics' recognition. Popular recognition is shown in sales volume, audience awards or vote ratings on websites. It is linked to commercial success and closer to economic capital, rather than cultural capital. Legitimate culture is more likely to be established by critical and peer recognition. These types of appreciation are empirically embodied in prestigious newspapers' top ten lists and reviews, or critical and peer awards. Critical recognition should be understood as an institutional delineation of legitimate culture, a sign of good taste sanctioned by a social group endowed with cultural capital.

3.3 The abovementioned studies employ well-founded methodological approaches of empirical inquiry into cultural legitimation, but their techniques have only limited application in the context of survey data. They often focus on delineated segment of cultural production, mostly defined by sample of time span or a set of previously consecrated cultural works. Effort to integrate institutional legitimation of cultural capital into survey analysis is thus constrained to two feasible alternatives. First, it is possible to limit the dimensions of taste to several a priori defined cultural items, which open the risk of biased selection. Second, the expression of respondents' taste could be unrestricted, but classification has to be based on a broad source of critical evaluation.

3.4 The principal aim of this study is to explore the possibilities of the second approach. It develops a method to measure respondents' cultural capital by empirical evaluation of legitimacy of their taste (institutional measurement), a method that doesn't confine preferences to a predefined set of cultural items or genres, but evaluates the answers to open-ended questions. The specific focus is to answer how this methodologically innovative approach relates to the prevalent instruments for the measurement of cultural capital and how it could inform the study of change of legitimate culture. The key research questions are:

• How does institutional measurement of cultural capital relates to other often employed methods? Are these different ways of measurement empirically equivalent or divergent?
• To what extent is institutional measurement of cultural capital feasible and effective for survey research?
• What implications do the methods of measurement have for the substantial results of survey research of cultural capital? Specifically, does the institutional measurement confirm the decline of legitimate culture, in other words the dissolution of social status, judgements of taste and structures of cultural legitimacy?

3.5 The analysis opens up possibilities for the further development of ways to empirically measure the legitimacy of taste in survey research without recursion to data itself, or without basing one's determination on an unfounded assumption about the content of legitimate culture. The wide variety of respondents' favourite cultural goods, despite the focus on a demographically homogeneous population, brings enormous demands of data processing, and the scope of institutional evaluation. Employing open-ended questions to measure the legitimacy of taste mitigates the risk of inappropriate assumptions about taste differences, but leads to many methodological difficulties and limitations in the analysis of data.^[3]

Data and measurement

Data

4.1 The study uses data from a survey, 'Volný 99as a kulturní zájmy maturantů 2015'^[4], which was administered to a representative sample of high school students in Prague in the final grade. Respondents were 18 to 20 years old, and were drawn from high schools with either academic or vocational specialisations. The questionnaire was administrated through web-based application during class time, and therefore the survey's response rate was very high. The final sample size consisted of 524 respondents, which is quite sufficient, considering the homogeneity of the sample with respect to the respondent's ages and geographical location.

Types of measurements

4.2 In order to answer the research question, three different methods of measuring cultural capital were used: (1) homology measurement, which uses indicators of social position as a proxy for accumulated cultural capital;^[5] (2) a priori measurement, which determines cultural capital from a traditional definition of highbrow culture; and (3) institutional measurement, which evaluates respondents' tastes in light of institutionalised criticism.

Homology measurements

4.3 Two indicators of social position were used in this analysis. First, the parents' education is a common indicator of a family's cultural capital. It is differentiated into four categories: 'both parents without GCSE', 'at least one parent with GCSE', 'at least one parent with tertiary education' and 'both parents with tertiary education'.

4.4 The second indicator more accurately reflects the respondents' current social positions. It arises from the fundamental division of Czech secondary high schools into two distinctive types: the school type 'academic' represents the more prestigious general education with the assumption that students at these schools will continue on to university, while 'vocational' schools cover more practice-oriented education and are generally considered to be of a lower status, as its curricula make transition to tertiary level education more difficult.

A priori measurements

4.5 A widely applied, and also disputed, indicator of cultural capital is linked to consumption of highbrow culture. In line with this type of measurement the scale was computed as a summary of three variables, including frequency of attendance at theatres, art exhibitions, and museums (0 = never; 4 = at least once a month). The scale ranged from 0 to 12 and had sufficient statistical properties (Cronbach's alpha = 0,791).

4.6 A more elaborated approach adopts the Bourdieusian method of Multiple Correspondence Analysis (see Roose 2016; Roux & Rouanet 2004) to decompose the complex set of cultural tastes and behaviours into fundamental dimensions. MCA uses a set of variables to construct a multidimensional space of differences between individual respondents. The further apart the two respondents are in this space, the more dissimilar their taste. The most important of these dimensions is then interpreted as a volume of cultural capital.

4.7 In this analysis, a large number of variables (177) covering areas of cultural consumption such as music, movies, TV series, and literature, were dichotomously coded either positively (liking, doing, agreeing) or negatively (neutral stance or disliking, not doing, disagreeing).^[6] The first dimension (axis) that showed to be highly correlated with both highbrow cultural items and high social positions has been interpreted as a second way to a priori measurement of cultural capital.

Institutional measurements

4.8 The institutional measurement of cultural capital determines the legitimacy of one's tastes from an empirical assessment of preferred cultural products. In the three fields under study – movies, music and literature – respondents were asked open-ended questions for which they were to provide their three favourite movies, bands, and books. The open-ended form of the questions limits the risk of imposing genre categories and/or preconceived assumptions about taste differences upon the respondents. Two sources of institutionalised judgements of taste were used to classify the respondents' selections. The cultural legitimacy of the movies and music artists was based on the aggregate score of critical reviews from the Metacritic website^[7]; the cultural legitimacy of the books was determined based on the official assessment of the Municipal Library of Prague.

4.9 The Metacritic website publishes a so-called 'metascore', which represents the weighted average of at least four critics' reviews from carefully selected publications.^[8] Unlike widespread popularity (e.g. voted ratings on imdb.com or box-office revenue), which doesn't necessarily reflect critical appraisal (Allen & Lincoln 2004), critical recognition of cultural product is a solid base for assessment of its legitimacy,. The Metacritic website aggregates a wide range of reviews, while having sufficient criteria of quality applied on the tracked publications. It standardizes different expressions of critics' evaluations on a hundred-point scale. Last, but not least, Metacritic is a popular reference guide for the classification of cultural production, thus, in itself, it represents one of the contemporary institutions of cultural legitimation.

4.10 On the other hand, the partially opaque Metacritic aggregation system doesn't allow its methodological procedures to be further elaborated or put under scrutiny, e.g. weighting of different review sources. Its strong focus on recent Anglo-Saxon production also leads to unwanted bias in the scope of available ratings. These shortcomings are outweighed by other unique features of the Metacritic website. First, as this study doesn't engage the issues of the dimensionality and dynamics of the cultural field itself (cf. Allen & Lincoln 2004; Kersten & Verboord 2014; Verboord et al. 2015), the primary concern of the methodological approach lies in encompassing as wide range of cultural products mentioned by respondents as possible. Despite its limitation, the Metacritic database offers the reliable aggregation of critical evaluation of not only recognized works, but also those less common and unconsecrated. Other sources common to the studies of cultural legitimation, such as peer awards or 'best of the year' lists, would result in rather inadequate coverage of respondents' answers. Second, the actual cultural orientation of the youth substantially reduces the issue of geographical and novelty bias, especially in the field of movies. Of all movie titles mentioned by respondents, 86% have been produced in the United States or the United Kingdom, and 75% were shot in the year 1999 or later (the same year that the Metacritic site was launched). Of the all movie titles in the respondents' answers only 15% have no metascore assigned in the database. The field of music was significantly more affected by this issue. Almost half of the mentioned artists have no rating of their albums in the database, many of them being artists beyond general recognition.

4.11 The recognition of book titles was based on the Municipal Library of Prague's catalogue where each book is assigned a tag describing its aesthetic quality. Each new book acquisition is assessed by a qualified librarian who reads through the book and fills out the book record with a description of the general genre categories as well as one of the available evaluation ratings: classic, very good, good, popular, average, low quality. The last two categories, average and low-quality, are not publicly visible in the library's catalogue, and therefore, were not available for the analysis. The book evaluation system used here is not as comprehensive as the Metacritic website, as there is only one evaluator and the evaluations are mostly support material for the library's acquisition department, rather than actual critical reviews. Still, the organizational system, assigning two working hours of a university-educated librarian's time to make each evaluation, should ensure the review is of adequate quality.

4.12 Each cultural item (movie title, book title or music artist) mentioned by the respondents was identified in the respective database and its rating, either a metascore or a librarian's evaluation, was assigned to it. In each cultural field, respondents could state up to three favourite items, so the mean value of all available ratings was used to determinate the legitimacy of the taste.^[9]

Results

5.1 This section first brings in a detailed description of a priori and institutional measurements, as they are pivotal to this study. The following subsection then shows the interrelationship between the various measurements of cultural capital.

Multiple Correspondence Analysis

5.2 The rich descriptive content, composed of a wide range of variables concerning respondents taste and cultural practices, was sufficiently replicated by the four major axes of the MCA. The axes were able to account for 80% of the variance (the sum of the modified rates), which is, considering the number of variables, a generally acceptable rate. The 45% share of the first axis indicates dominance, followed by only a 21% share of the second axis. The cultural practices most contributing to the first axis are also in line with the traditional understanding of legitimate culture (see Figure 1). On the right side, there is a higher concentration of respondents who visit museums and galleries, like to read classical literature and authors such as Bukowski, Goethe, Havel, or Kafka, and appreciate jazz music. On the left side, there are several indicators of illegitimate cultural taste, e.g. not liking to be in a museum or a theatre, or dislike of Poe and Orwell.^[10]

5.3 It should be mentioned that some non-traditional aspects of cultural taste also contributed significantly to the first axis. Especially remarkable is the wider scope of authors outside of the usual literary canon, including writers of recent bestsellers (e.g., Larsson and Coelho) and various authors of the fantasy genre (e.g., Tolkien, Paolini, Martin, and Pratchett). Even though traditional highbrow culture is an important structuring dimension even for today's youth's cultural consumption, it is definitely not established in its purified form, and is instead enriched by the current subculture genre of fantasy literature.

Figure 1. Plane 1-2: Modalities most contributing to the first axis

5.4 The second axis is significantly connected with gender differences in consumption practices (see Figure 2). At the top of the plot, there is a concentration of modalities expressing tastes for romantic movies (e.g., Love Actually, The Fault in Our Stars, Twilight), soap operas (Ošklivka Betty, Ulice) and shopping at the mall. Dislike of these cultural practices is represented at the bottom, together with a like of computer gaming and G. R. R. Martin novels. In addition, the second axis also structures the preferences of popular music and nightlife, which are more liked on the feminine side and disliked on the masculine side. The gender interpretation of the second axis is confirmed by its high correlation coefficient with gender variable (0.55).

Figure 2. Plane 1-2: Modalities most contributing to the second axis

5.5 While the other axis of the MCA opens up new dimensions of cultural taste that are certainly interesting, their description is out of the scope of this study. In fact, among the first four dimensions of the MCA which account for 80% of the variance, all except for the first one have only weak or insignificant association with other indicators of cultural capital (see Table 1). In the following analysis, the first axis of the MCA is used as one of the measurements of cultural capital. This interpretation is supported by both homology argument – the axis is strongly correlated with the type of school – and the traditional highbrow rank of many cultural products contributing to this axis.

Table 1: The result of the MCA - correlation between the axis and the traditional indicators of cultural capital (Spearman's correlation coefficient)
		correlation coefficients
	modified rate of variance	parents' education	school type	highbrow culture
Axis 1	45.4	0.41**	0.49**	0.66**
Axis 2	20.6	0.18**	-0.16**	0.14**
Axis 3	8.2	0.07	-0.24**	0.04
Axis 4	6.3	0.07	-0.01	0.05
N = 519

Institutional measurements

5.6 Processing answers to open-ended questions is a potentially difficult procedure, especially when the responses consist of a wide variety of cultural items. The actual distribution patterns of favourite cultural products are very similar in all three of the fields covered (see Table 2). In each field, 1100-1300 responses were gathered, with references being made to approximately 500 different cultural products. Thus, despite the survey's focus on a demographically and geographically homogenous population (mostly 18- to 19-year-old students from Prague) with supposedly deep immersion into mainstream culture, open-ended questions concerning favourite cultural products resulted in a vast variety of answers. Only about one fifth of all responses concentrated on the ten most frequently mentioned items, while about one third of responses consisted of items mentioned only once in the whole survey. Even among movies, which could be considered as a product of more centralised culture industry than music or literature, we could see high variability among the mentioned items. The research-tailored evaluation of such vast variety of cultural products (e.g. by cultural critics as evaluators) would be demanding on resources. Utilising existing critical reviews is desirable in terms of resource effectiveness, but also because of its non-intrusive character and thus presumably higher validity.

Table 2: Favourite cultural products - distribution of responses
	Movies (titles)	Music (artists)	Literature (books)
No. of responses	1 303	1 153	1 169
No. of different items	517	498	516
Share of (%)
- ten most frequent items	22.1	16.7	21.9
- items mentioned once	33.0	32.0	38.2

5.7 The critics' rating of movie and music cultural items was based on the Metacritic database. Tables 3 and 4 show the distribution of these ratings of all of the responses. A metascore of 81 or higher generally means that movie title or music artist is 'universally acclaimed'.^[11] Nearly one fifth of movie responses (18.1%) and 9.1% of music artist responses belong to this category. Examples of some of the cultural products often mentioned in this category include two successful movies from 1994, Forrest Gump and Pulp Fiction, both of them winning Academy Awards in the United States. On the other end of the scale is the action film franchise The Fast and the Furious or the unfavourably-reviewed science fiction film Divergent.

Table 3: Movie legitimacy - the distribution of Metacritic's scores
Metacritic's score	% of responses	Example of frequently mentioned items (no. of references)
81 or higher	18.1	Forrest Gump (45), Lord of the Rings (43), Pulp Fiction (36)
71-80	14.5	The Shawshank Redemption (20), Inception (17), The Wolf of Wall Street (9)
61-70	21.9	Harry Potter (28), The Hunger Games (26), The Fault in Our Stars (15)
51-60	15.2	The Fast and the Furious (24), The Intouchables (22), The Hobbit (21)
50 or lower	15.9	Divergent (12), Just Go with It (10), The Expandables (7), Grown Ups (7), Pearl Harbor (7)
N/Aa	14.4	Pelíšky (19), Some Like It Hot (7), Babov114esky (6)
N = 1 303
a) No metascore value in the database (not enough reviews or not included at all)

Table 4: Music artist legitimacy - the distribution of Metacritic's scores
Metacritic's score for highest-rated album	% of responses	Example of frequently mentioned items (no. of references)
81 or higher	9.1	The Beatles (12), Pink Floyd (8), System of a Down (7)
71-80	17.4	Taylor Swift (17), Coldplay (16), Eminem (16)
61-70	22.1	Linkin Park (25), One Republic (23), Ed Sheeran (22)
60 or lower	4.4	Imagine Dragons (15), Afrojack (3), Bastille (3), Enrique Iglesias (3), Skrillex (3)
N/Aa	47.1	Kabát (23), Ektor (16), Kontrafakt (12)
N = 1 153
a) No metascore value in the database (not enough reviews or not included at all)

5.8 The evaluation of book items was derived from the classification of the municipal library. Over one third of the books mentioned by respondents belong to the category of 'literary classics'. This large proportion could be a result of the guiding effect of the maturita exam. Each adept of this exam has to read at least 20 book titles from the school's approved reading list. These lists include both Czech and foreign books, and in itself this list could be understood as a quintessential legitimacy-defining tool. It could be assumed that, for many students, compulsory maturita reading is the only experience that they have with literature.^[12] In other words, it may be quite common for respondents not to have read any other book aside from those on the compulsory canonical list, and therefore, not be able to mention any books other than literary classics.

5.9 The 'not available' category is not present in Table 5, since every mentioned book title was found in library's catalogue. Due to public unavailability of 'low quality' tag in the database (see above), it was impossible to distinguish between average and lower evaluation.

Table 5: Book legitimacy - the distribution of the municipal library evaluations
Library evaluation	% of responses	Example of frequently mentioned items (no. of references)
classics	35.6	All Quiet on the Western Front (32), The Lord of the Rings (23), The Hobbit (18)
very good	14.4	Harry Potter (60), The Alchemist (9), The Catcher in the Rye (8)
good	17.9	Zoo Station: The Story of Christiane F. (30), The Fault in Our Stars (21), The Godfather (11)
average or lower	32.1	Fifty Shades of Grey (24), Hunger Games (15), Inheritance (13)
N = 1 169

5.10 The institutional measurement is based on a mean rating of favourite cultural products in the respective field.^[13] Items without an available rating were excluded from the computation. In each field, between 11-18% of respondents did not mention any cultural products and therefore the legitimacy of their tastes in that field could not be measured (see Appendix A1). In the field of music, there was a substantial percentage of respondents who mentioned only music artists who did not have any rating in the Metacritic database. Most of them mentioned either Czech music artists or artists from niche genres such as dance music or metal, enjoying various musicians under the radar of international recognition.

Comparison of measurements

5.11 A comparison between the institutional measurements and the other ways of measuring cultural capital examines the empirical support for the assumptions ascribing legitimacy of taste based on a priori intuitions or homology hypothesis. The correlation matrix (Table 6) shows that all three types of institutional measurements are significantly correlated with the a priori and homology measurements of cultural capital. However, the magnitude of correlation is weak to moderate, and shows substantial variation.

5.12 First of all, the first axis of the MCA is most closely associated with all three measurements. Its correlation coefficient is highest for the legitimacy of taste for movies (0.33), followed by music (0.25), and literature (0.22).^[14] On the other hand, the scale of highbrow culture shows very weak association with empirical measurements of legitimacy, ranging from 0.17-0.18. Homology measurements associate strongly in the case of movie tastes (0.31), but poorly in the other two fields (0.14 and 0.10).

5.13 This leads to the conclusion that the main differences in taste represented by the first principal axis of the MCA are supported by institutional evaluation. People who are located on the right side of Figure 1 are more prone to mention as favourites cultural products that have higher ratings among critics' reviews, and vice versa. Nonetheless, taste for these consecrated cultural goods is not homologous to participation in traditional highbrow culture. There are significant deviations that make the results of the MCA different from highbrow culture scale (see the discussion in the next section).

Table 6: Correlations between the measurements of legitimate taste (Spearman's correlation coefficient)
	institutional measurement
	movies	music	literature
a priori measurement
- highbrow culture	0.18**	0.17**	0.17**
- MCA - axis 1	0.33**	0.25**	0.22**
homology measurement
- parents' education	0.31**	0.14**	0.10*
- school type	0.28**	0.14*	0.14**
N	450	337	422

5.14 Graphical representation of the correlated variables in the MCA space further confirms and qualifies their relationship. Figures 3a to 3c show the distribution of the legitimised movies, music, and literary tastes across the plane of the MCA's first and second axes. Despite small deviations, a substantial increase in values from left to right is clearly visible. In the movie and music fields, there is also a visible trend of an increasing mean rating in the bottom part of the plot. The middle location of the categories omitted in the construction of the mean values ('no answer' and 'no rating', see above) in the context of the first axis confirms that they should be interpreted as a methodological issue, in terms of missing values, rather than as an indication of deficient mastery of legitimate culture.

5.15 The shape of the plot in Figure 3d reflects shared variables used for the construction of the a priori scale of highbrow culture and the variables which are major contributors to the first axis. The highbrow scale develops in full length along the first axis, with wide and almost regular spacing. Both of these types of measurements are very closely correlated (see Table 1).

5.16 Both indicators of social position (Figures 3e and 3f) are in line with the homology hypothesis, horizontally aligned with the first axis and leaning subtly towards the bottom right quadrant. Supposedly, highbrow cultural practices are more prevalent among students with an educated family background or who are studying at academically-oriented schools. The diagonal skew indicates that the homology effect is related to the institutional measurement of the movie field and the location of non-traditional cultural products (like fantasy literature) on the bottom right area of Figure 1.

Figure 3. Measures of cultural capital on plane 1-2 of MCA

For description of the horizontal and vertical axes see Figures 1 and 2. 'No answer' and 'No rating' are plotted only if these categories included more than 30 respondents.

Discussion and conclusion

6.1 Various measurements of cultural capital show much common variance as an expression of their singular conceptual origin, on the other hand they also manifest significant divergence, as they tap different facet of cultural capital. Bivariate correlations between assorted measurements were of weak to moderate magnitude. The indicators of highbrow culture had a very weak connection to the measurement based on taste for institutionally legitimate cultural goods. This means there was a significant number of respondents, whose taste was classified as not being highbrow, but in the open-ended questions about their preferences they wrote down cultural goods that are in fact critically praised. In contrast, the first axis of the multidimensional cultural space derived from the Multiple Correspondence Analysis was the most correlated with institutional measurements in all three fields of culture. Thus, it could be argued that this method is most faithful in capturing the aspect of legitimacy in culture consumption (see below).

6.2 The proposed method of institutional measurement of cultural capital has two major downsides. First, its feasibility is affected by the demanding procedure of coding and evaluating respondents' answers. Second, its applicability is dependent on the existence of an extensive list of quantifiable reviews of cultural items. While careful employment of automated coding scripts could deal with the first obstacle, the latter issue remains a serious limitation for its applicability in different areas of cultural consumption and the wider population. Nonetheless, the innovative incorporation of potential data sources capturing the content of current 'legitimate' culture should improve the precision of further measurement possibilities. Other possible data sources beyond critics' reviews, such as radio playlists, pedagogical canons, or lifestyle media experts, could be incorporated into further research on this topic.

6.3 Considering the above mentioned technical difficulty, it is useful to assess how this measurement relates to other, perhaps more viable, types of measurement. The principal axis of the MCA seems to be a promising instrument, as the results indicate its central position among other operationalisations of cultural capital. It has strong correlations with both the institutional measurements and all other types. It corresponds with Bourdieu's preference of this method and the recent resurgence of its application in the field of cultural stratification (e.g. Bennett et al. 2009; Hanquinet et al. 2014; Roose 2014, 2016). Its ability to disentangle the geometry of cultural space is probably one of its main assets, making it more sensitive to the changing constellations of cultural space, and less vulnerable to bias of strictly defined a priori measurements. This study results support the assumption that the Multiple Correspondence Analysis of cultural space is able to retrieve relevant quantification of legitimate taste, i.e. cultural capital, as one of its axis.

6.4 The data reveals the continuing social significance of legitimate culture, even among the younger generation. The first principal axis of the MCA is closely related to the traditional attributes of highbrow culture. Still, its content is not limited to these items but also includes preferences for non-traditional cultural items such as fantasy literature. The empirical measurement of institutionally legitimized taste is more similar to the MCA's explorative identification of the principal axis of taste differences than it is to the traditional scale of highbrow cultural participation. It is thus premature to dismiss the connection between cultural capital, taste and legitimate culture on the basis of (a) a decreasing association between social status and highbrow participation; or (b) an intuitive interpretation of the changes in cultural space. Without looking at what is considered as legitimate in the contemporary field of culture, 'it remains imperative to avoid using everyday impressions, or recollections of past circumstances, precisely because the content of high culture changes.' (Warde & Gayo-Cal 2009 p. 122-123). While considering Pulp Fiction or The Lord of the Rings as legitimate culture products could cause some unease for the researcher, it could also indicate his or her lack of familiarity with actual illegitimate culture – not surprisingly, bearing in mind his or her social position – rather than being a sign of the fall of legitimate culture.

6.5 Despite the well-documented breach of boundaries between popular and fine arts, it seems that the legacies of autonomous artistic fields have been objectified and transported across genres (Lizardo 2008). The capacity to discriminate within any field of culture could be understood as a sign of acquiring cultural capital, as 'fields of "mass produced" popular art themselves come to acquire and institute the same schemes of perception, appreciation and historical canonization distinctive of the restricted field of cultural production' (Lizardo 2008p. 18; see also Lizardo & Skiles 2016). Further research into the institutional appreciation of cultural products is still needed, though, to make more reliable accounts about the fate of legitimate culture (Warde 2008).

---

Appendix 1

---

Notes

¹ Inventive studies which capture the institutional establishment of cultural hierarchies and the subjects' individual taste in a single analysis focuses on niche cultural subfields like wind bands (Dubois & Méon 2013) or the interactive fiction retro-gaming scene (Allington 2015). However insightful and inspiring, their methods are hardly transferable to an analysis of the general population.

² It is well-documented, that the expression of cultural capital can be identified even in the social milieu of youth (see e.g. McCulloch et al. 2006, Abeele & Roe 2013). The comprehensive analysis of taste and cultural practices of Sweden's higher education students (Börjesson et al. 2016) show the pronounced origins of future differences in the volume of cultural capital.

³ The coding of a wide range of varied answers is a foremost resource-demanding procedure. However, using an open-ended question also results in substantial analytical constraints. For example, the majority of cultural products are mentioned only once or twice in whole sample, which substantially limits the possibility to use this type of questions to measure similarity or dissimilarity of respondents' taste.

⁴"Leisure and Cultural Interests of Maturants 2015". A 'maturant' is a student in the final grade of high school, who will take the 'maturita' exam at the end of school year (comparable to GCSE).

⁵ Lizardo and Skiles (2016) make thoughtful critique of common misunderstanding of what Bourdieu meant by 'homology'. Homology measurement in this paper represents this common research practice misconception, rather than genuine interpretation of Bourdieu's theory.

⁶ A detailed description of variables and results of the MCA are available from the author upon request.

⁷ http://www.metacritic.com.

⁸ See http://www.metacritic.com/faq and http://www.metacritic.com/about-metascores for a detailed description of the aggregation and selection process.

⁹ Cultural items without a metascore pose serious methodological issues for the construction of the aggregate scale. They could be understood as (a) unrecognised, and thus, not legitimate cultural products; therefore, a low rating would be appropriate; or (b) not reviewed, therefore understood as a 'missing value'. Likewise, when a respondent did not write down three favourite items, it could be interpreted both as (a) an indication of unfamiliarity of the respective cultural field; (b) a respondent's satisfaction with a response consisting of only one or two items; or (c) unwillingness to bother with open-ended questions at all. The presented construction of the legitimacy of taste scale has been based on (b) both in the case of cultural items without a rating and missing answers. Therefore, missing values were not taken into account in any way during the construction of this scale of measurement.

¹⁰ Considering the range of variables used, only a narrow selection of contributing modalities is plotted in Figure 1. A common criterion of more than an average contribution to a given axis was applied in a more restrictive magnitude, filtering these modalities whose which contribution was higher than triple the average. The full description of modalities contributing to axis 1 is available from the author upon request.

¹¹ See http://www.metacritic.com/about-metascores.

¹² According to the PISA 2009 survey, as much as 43% of 15-year-old Czechs 'never read for enjoyment'.

¹³ Numerical ratings were assigned to library categories to achieve a scale comparable to that of the Metacritic database ('classics' = 85, 'very good' = 75, 'good' = 65, 'average or lower' = 55).

¹⁴ As a reminder, open-ended variables used for the construction of institutional measurements were not included in the set of input variables in the MCA.

---

References