Barriers to live participation

On the other hand, Tables 9.3 and 9.4 show barriers to participation for both classical and popular live attendance. As with motives, respondents provide two different barriers (first and second). Looking at classical music, lack of interest stands out as the main barrier to attendance, which, in turn, can be equated to non-participation. Once this is accounted for, three barriers emerge as major obstacles to participation: time, price, and supply availability. Overall, 83% of the survey-takers choose any of these as their first or second barrier to attend more to classical music concerts. Indeed, this seems plausible as live participation is both time intensive and geographically predetermined. As for the former, attendees need not only to allocate time for cultural consumption but also the coincidence of an individual’s time availability with the time slot when the show is programmed. Regarding geographical constraints, the concentration of supply in specific urban areas means that even if one has (time and financial) resources the spatial distribution of supply may make attendance non-feasible. Therefore, time and supply availability are active hurdles for most individuals. Add to that the existence of budget constraints, in this case a trait not necessarily specific of live performances, and we obtain a detailed depiction of the main restrictions individuals face.

The lowbrow nature of popular live music makes it more accessible to the general public and therefore makes lack of interest a less relevant barrier. Apart from this, the same live-specific barriers emerge (time, price, and supply restrictions) and in the same order as in the case of highbrow participation.

Table 9.3 Barriers to attendance: classical music

Table 9.4 Barriers to attendance: popular music

Overcoming live-attendance barriers through digital consumption

The coincidence of barriers for both highbrow and lowbrow live music engagement suggests audiences face similar attendance problems and, to enlarge them, similar strategies could be addressed. Specifically, it could be argued that digital consumption, inasmuch as it is unaffected by time and

Classical and popular music in Spain 147 geographical restrictions, could cater to the segments of the market mostly comprised of occasional and non-participants and, in so doing, raise participation. Additionally, and given that minimal technological requirements apply to the online consumption of classical and popular music concerts, digital participation may prove to be an inexpensive means of accessing cultural consumption (overcoming financial constraints) and an alternative way of taste formation, therefore enlarging audiences. Of course, one should also account for the possibility of digital consumption cannibalizing live participation.

The above discussion brings about some research questions regarding digital participation that this study aims to address. First, we pursue evidence regarding the composition of audiences brought about by digital participation. Specifically, we are interested in knowing whether it incorporates new cultural consumers (enlarging audiences) or it adds a new channel to current participants (deepening or increasing consumption of individuals that already participate). Second, and linked with the previous question, we examine the issue of whether online participation reproduces those barriers identified in live consumption or if, on the other hand, creates new ones. This, as suggested, is linked to the enlargement of audiences through digitization and its potential to lifting structural constraints enabling the incorporation of new consumers, not only as digital consumers, but also as live consumers through learning and taste shaping. Third, we aim at identifying under what circumstances digital consumption substitutes, that is, cannibalizes, live participation or, on the contrary, it complements it. Note that the three research specific goals we have spelled out are interwoven in that audience enlargement comes with effortless digital access and, possibly, the accumulation of competences and specific knowledge that leads to taste formation.

To understand the role of digital participation, we start by looking at Tables 9.5 and 9.6, which cross-tabulate live and digital participation for classical and popular music respectively. The latter is the binary answer to the question “Have you watched a music concert through the Internet in the last year?” Individuals answer twice (classical/popular) to this question.

From these, it is apparent that digital participation represents less than half of the audience of live music. Data show that 4.8% of the population engages in classical music online (Table 9.5), while 13.5% engage in popular music (Table 9.6). Compare these with offline participation rates of roughly 10% and 30% respectively, and one sees what could be deemed as a complementary contribution to the overall audience. However, it is true that one sees some marginal evidence of non-participants of live music accessing to music performances online: 2.3% for classical music and 5.7% for popular music. These shares could be considered net gains in terms of audience. Yet the pattern of participation shows that the probability of going digital increases as we move up in Tables 9.5 and 9.6, which means that the more frequent a live cultural consumer is, the more likely she is to use additional channels.

Table 9.5 Classical music: live and digital participation (row percentages)

Table 9.6 Popular music: live and digital participation (row percentages)

If we look now at the effect of digital consumption on live participation (Figure 9.2) we see that, in general, digital participation enhances both classical and popular live participation. In pop music, roughly 52% of digital audiences are also live audiences, compared to merely 27% of non-digital ones. This effect is even more intense in classical music: 31% of digital audiences are also live audiences, while only 8% of non-digital audiences are live participants. As a consequence, digital participation reduces the probability of live non-participation in all cases: only 15% of popular online music audiences and 32% of classical online audiences are only digital audiences. Furthermore, non-participation in the digital arena also increases the likelihood of nonparticipating live: 67% and 40% for classical and popular music respectively.

At this point, we acknowledge that the descriptive evidence is somehow mixed. We see that some individuals that do not attend live are attracted to online consumption, a finding that is consistent with an audienceenlargement effect. However, we also find that cultural consumers seem to have a propensity to participate through different channels, which pinpoints the reproduction of consumption patterns and barriers in the digital world.

Methodology

The foregoing descriptive findings are suggestive of a non-trivial relation between offline and online music concerts consumption. To identify it, a

Figure 9.2 Frequency of live classical and popular music conditional on digital participation (D: digital; ND: non-digital).

bivariate probit model (Marra & Radice, 2017) using R is estimated (R Core Team, 2017). Specifically, we aim to single out: (1) the link between online and offline engagement and the extent to which distribution channels are complements or substitutes; (2) the barriers audiences face when consuming live/online music; (3) the contribution of online engagement to the democratization of culture by increasing the diversity of audiences.

Live and digital participation are simultaneously determined in a model that we separately estimate for each cultural manifestation. Let Y,j, a binary indicator for individual i whose value determines whether she participates in manifestation (classical/popular) through channel k (live/online). Then, we estimate the following specifications:

where two binary dependent variables (digital/live participation) are explained in terms of a set of covariates. These are classified as follows:

• 1 Socio-demographics (sociodemo¡). We include age, gender, labor-market, and family situation of the survey-taker. Expression (1) also includes the square of age to account for nonlinearities.
• 2 Educational attainment is measured as a set of dummy variables: secondary, vocational, and university. The reference value is an individual that has completed at least primary education (or possibly less).
• 3 Labor market position is a set of dummies that control for being employed, unemployed, student, househusband, retired, and other, with the reference value being disabled.
• 4 We include marital status (being married) and having children as two dummies. We also include the number of members of the household as a trait to describe family situation.
• 5 Household equipment (phisicalk¡) is included to control for the availability and variety of physical capital which, we assume, facilitates cultural consumption and participation. Furthermore, and given the survey does not include information about income, this will allow us to control for it. Here we use information about reading, music, and other audiovisual equipment as well as computer and other Internet-related hardware and online and TV subscriptions which a household has. Overall there are 39 variables describing a household’s equipment that we reduce, using principal components analysis, to nine scores that account for over 50% of the total variance.
• 6 Geographical covariates (geo J are included in the model to control for supply and demand spatial structural traits, such as the geographical distribution of cultural supply and regional income differences to name two. To this end a dummy for each region in Spain is included. Furthermore, a variable controlling for the city of residence of the survey-taker (province capital; more than 50,000 inhabitants; less than 50,000 inhabitants) has also been included.
• 7 A measure of the digital expertise of the survey-taker (internet,} has been included to account for the skills and competences needed to participate online. In this respect, two variables are included. First, we use a binary variable to account for whether the survey-taker uses the Internet on a regular basis at work. Second, we also control for the weekly number of hours the respondent uses the Internet at work.

Additionally, we control for the term when the survey-taker was interviewed. This is relevant as the supply of live performances exhibits a strong seasonal pattern. This is specially the case of popular music, in which big events (music festivals) are concentrated in the summer months (see Montoro-Pons & Cuadrado-García, 2016).

Table 9.7 provides a summary of the main explanatory variables used. Note that it provides the overall mean for the whole sample (or proportion

Table 9.7 Description of the sample

for categorical variables) as well as the summaries across the categories of the dependent variables.