Channel repertoires: using peoplemeter data in Beijing.

Every year, television viewers around the world have more channels from which to choose. In the United States, for example, the average household receives more than 100 channels of programming--a threefold increase since 1990 (Nielsen Media Research, 2004). In China, the world's largest television audience has seen a fourfold increase in less than a decade (CVSC-Sofres Media [CSM], 2004). We have known for some time that Americans cope with this abundance by winnowing the field to a smaller channel repertoire within which regular viewing occurs (Ferguson, 1992; Ferguson & Perse, 1993; Heeter, 1985; Heeter, D'Alessio, Greenberg, & McVoy, 1983; Neuendorf, Atkin, & Jeffres, 2001). The precision of that information, however, often leaves something to be desired. To date, no study has extended this line of research beyond the U.S. marketplace. This research adds to that literature by (a) investigating channel repertoires using peoplemeter data, thus affording a more finely calibrated look at channel use; and (b) documenting the use of channel repertoires in Beijing, suggesting that this behavior is characteristic of audiences in channel-rich environments worldwide. The authors find that although a large number of channels are sampled each week, a small number account for the lion's share of viewing. Variation in those repertoires is most easily explained by structural factors (i.e., viewer and channel availability), as specified in the theoretical framework developed by Webster and Phalen (1997).

Channel Repertoires

Heeter et al. (1983) coined the term channel repertoire to describe "the set of channels watched regularly by an individual or household" (Heeter, 1985, p. 133). Using household-tuning data collected at a cable headend, Heeter et al. (1983) found that although the cable system offered subscribers 34 channels, the average home watched fewer than 10 a week. These repertoires were conceptualized as a mechanism that viewers used to cope with an increasingly abundant and complex media environment (Heeter, 1985). Early studies (Heeter & Greenberg, 1988) further established that repertoires varied in size (with cable subscribers watching more than nonsubscribers) and composition (with major broadcast networks common to most repertoires, but dissimilar combinations beyond that).

Subsequent research has elaborated on definitions of channel repertoire and sought to further explain variation in repertoire size and composition. Some studies have continued to define repertoires as the total number of all channels watched over a certain period of time--usually a week (Heeter, 1985). Others have drawn a distinction between total channel repertoires (TCRs) and "mindful" channel repertoires--those that come to mind without aided recall (Ferguson & Perse, 1993). Neuendorf et al. (2001) grouped channels into "sets," which were summed to create repertoires, and they attempted to weight channels or sets by the frequency of viewing. Regardless of the definition, the overall pattern is clear: Viewers with abundant choices watch far fewer than the total number of available channels. This is consistent with a recent industry estimate that the average U.S. household watched only 14.8 channels in the course of a week (Nielsen Media Research, 2004).

Researchers have tried to explain variation in the size of repertoires using a range of predictor variables. Webster and Phalen (1997) offered a useful theoretical framework for summarizing these results that draws a distinction between microlevel and macrolevel structural determinants. Some studies favor the microlevel determinants by hypothesizing the individual viewers' media use and demographic characteristics as the primary predictors of the channel repertoires. Heeter (1985) found that the viewers who had an exhaustive channel-search pattern had larger channel repertoires, and that education was a positive predictor of repertoire size. Neuendorf et al. (2001) found that the use of other mass media explained a small portion of the variance of channel repertoire.

A smaller number of studies have considered variation in what Webster and Phalen (1997) identified as structural variables. These include audience availability and the number of choices in the viewing environment. Ferguson and Perse (1993) operationalized availability as the time spent watching television and the number of choices as a cable-no-cable dichotomy. They concluded that "audience behavior can be explained well without considering individual audience characteristics. The findings show in a powerful way that TCR is a function of audience availability as it interacts with media structure" (p. 42). Table 1 summarizes the principal academic studies of channel repertoires, their operationalizations of the construct, methods, and key findings.

Two limitations in this literature are apparent. First, in the wake of Heeter et al.'s (1983) groundbreaking study, investigators have relied on some form of recall to assess the size and composition of repertoires. In the increasingly complex television viewing environment of the 21st century, which features dozens of channels and near universal penetration of remote control devices, such methods produce suspect results (Webster, Phalen, & Lichty, 2006). Second, the findings are based exclusively on U.S. viewers. It seems likely that viewers in similarly complex media environments would adopt similar coping mechanisms, but that is yet to be demonstrated. This study takes advantage of peoplemeter data collected in Beijing to address both shortcomings.

Chinese Television

A few words about the nature of Chinese television and its audience may be useful. Chinese television has undergone significant growth since the 1980s (Chang, Wang, & Chen, 2002). Currently, with some 1.2 billion viewers, it has the world's largest audience. In some metropolitan areas such as Beijing, audiences have an ever-growing number of channels delivered by cable and satellite systems. Figure 1 shows the dramatic increase in channel availability for Beijing's audience in the last 8 years.

[FIGURE 1 OMITTED]

In Beijing, 95% of the households have at least one TV set with a remote control. The availability of increased programming coupled with low subscription fees has resulted in a combined cable and satellite penetration rate of almost 90%, a level slightly higher than the comparable U.S. national average (Nielsen Media Research, 2004). An average adult viewer spends about 200 minutes a day watching television, or about 80 minutes less than a typical American (Veronis Suhler Stevenson, 2004).

There are three categories of television in the Beijing market. First, China Central Television (CCTV), the only national television service, has 12 channels. About half of those channels are broadcast over the air. These have a longer history and enjoy higher audience shares than the newer cable channels. Except for CCTV-1, which has a breadth of programming comparable to a traditional U.S. broadcast network, each of the other CCTV channels specializes in one or two specific program categories such as news, sports, music, lifestyle, and so on. Second, the local service (i.e., Beijing Television [BTV]) has 9 channels, most of which are distributed via cable. Among these channels, BTV-1 has comprehensive broadcast programming, and the rest of the channels are more or less specialized. Third, there are approximately 50 distant channels from other provinces and cities that are brought to the Beijing audience by cable. Similar to "superstations" in the United States, these distant channels offer a broad range of content.

Today, most Chinese television is advertiser supported, so, much like their U.S. counterparts, Chinese broadcasters need data on the size and composition of audiences at both the national and local levels. Audience viewing data in China are collected by CSM using peoplemeter panels. These panels are created through a process of multistage area probability sampling, in which each stage is stratified and sample elements are drawn in proportion to their incidence in the population. Similar to the meters Nielsen Media Research uses to produce national audience ratings in the United States, the CSM peoplemeter is an electronic device attached to the TV set that automatically records the minute-by-minute viewing behavior of all the members of the household. Such meters are known to produce a much more precise record of viewing behavior than either diaries or telephone recall techniques and have become the preferred method for measuring television audiences worldwide (Webster et al., 2006).

Hypotheses and Questions