Chinese Television Audience Behavior in a Multi-Channel Environment:

　　A case of Beijing

　 Elaine J.Y. Yuan　　Ph. D Student　　Communication Studies

　　School of Communication　　Northwestern University

　　When summarized in a general way, the figures associated with China’s television market present us a picture of a fast evolving mass medium with an audience of enormous size and potential. As of 1999, 94% of its 1.2 billion population (CSM, 2001) was able to be reached by the television signals sent from its some 600 local and national stations. The number of television sets owned by the country was over 300 million. An average person spent more than 3 hours per day watching television. In some cities such as Beijing and Shanghai, the number of channels that could be received was more than 70. Cable penetration reached 88 million households.

　　Chinese television is not only catching up in terms of sheer quantities but also rapidly opening up to the advertising industry in a much fuller capacity. And consequently its audience is being recognized as pivotal impetus to the operation of all related businesses. There are many aspects of both the television medium and the audience that are unique and call for special attention to the inquiries into the audience behaviors in China.

　　However, little has been done to further examine how this mass audience actually watches television. Does it exhibit any similar patterns of regularity such as those found among the television audience in other countries? While the advancement of the new transmission technology has now made more channels available to the Chinese audience than ever before, is there also a changing trend of audience fragmentation? Do people distribute their viewing evenly across the many available choices or is the audience splitting into small groups only loyal to their favorites and little else?

　　This study assesses the current state of mass audience behaviors across 50 of the most prominent television channels in China and compares the results with American audience data of the same kind in an effort not only to present a comprehensive picture but also to take a look at the reasons behind.

　　Regularities in Mass Audience Behavior and Structural Determinants

　　 When examined in a gross manner, television audience behavior has been shown to possess some consistent patterns. In fact, some of those patterns are found to be so consistent that they are sometimes referred to as “laws”. They are channel loyalty, inheritance effects, double jeopardy, repeat viewing (Goodhardt & Ehrenberg, 1969; Goodhardt, Ehrenberg & Collins, 1987), and the popularity-weighted heterogeneity of consumption (Comstock & Scharrer, 1999)

　　 Researchers have identified that media structural factors such as the number of channels available to the audience, the scheduling of programs within and across channels, as well as other relevant market characteristics are responsible for these mass behavioral regularities (Webster & Lichty, 2000). For example, research on television audience flow has found program scheduling and audience availabilities account for a large portion of the variance in audience duplication (Goodhardt, Ehrenberg & Collins, 1987; Webster 1985). Further more Webster (1988) has grouped the structural factors into two categories: a) long term market characteristics that are usually beyond the control of broadcasters, b) relatively transitory program scheduling characteristics.

　　 The number of channels available to viewers is one of the key long term market characteristics that have significant effects on mass audience behavioral patterns. As the number of channels increases, we would expect greater audience fragmentation, i.e. a lower average ratings for any one channel, which is one of the two features of macro-level audience behavior in the new media environment suggested by Webster (1986).

　　 The juxtaposition of programs also affects the size and composition of audiences. Audience flow research found that more people will watch a program if it follows a show with a high ratings (Webster, 1985). The phenomenon is called “inheritance effect” or “lead-in” effect which has long be recognized and employed by television programmers in their effort to keep audience tuned from one program to the next.

　　 In recent years, China has witnesses a period of rapid increase in the number of channels available to viewers due to both the advancement of satellite and cable systems in the transmission technology and the favorable regulatory police changes. As of 2003, all the 31 provincial television channels, whose broadcasting signal could only cover their own province and surrounding areas in the past, are now made available nationwide via satellite. Meanwhile China Central Television, the only state-owned station, has expanded its number of channels to 12, all of which can reach a national audience. The overall picture of Chinese television market is appearing to resemble that of America in its early years of expansion of the “new media environment”.

　　New Television Environment and Audience Fragmentation

　　 Webster (1986) described this “new media environment” by highlighting three ways in which it differed from the old. First, in the new environment, programming is more diverse. Second, in the new environment, content is correlated with channel. Third, channels in the new environment are differentially available. He further suggests that these changes would produce two features in macro-level audience behavior; fragmentation and polarization. Fragmentation, the phenomenon in which each of many channels takes a relatively small share of the audience, is well illustrated by the on-going battle imposed onto the older, mainstream broadcast channels (i.e., ABC, CBS, and NBC) in U.S. by their newer cable channels in competing for audience share. The obvious trend seems to be that as the average number of channels receivable increases, the share of each channel inevitably decreases. There are two reasons for polarization: the correlation of content and channels, and the differential availability of those channels (Webster, 2003). The former is another conventional practice of the television industry nowadays in U.S. where some of the newly established networks try to appeal to a particular audience demographic that can be sold to advertisers. The latter is related to “de facto polarization” (Webster, 1986) and “channel repertoires”, i.e. a limited and relatively stable number of channels usually watched by viewers even when channel capacities are unconstrained (Ferguson & Perse, 1993).

　　 All of those new environmental characteristics have been observed in China. While diversity is subject to many definitions (Webster, 2003), a telling example is the increasing amount of television programs imported from other countries. Wang and Chang (1996) reported that as of 1990, 30% of the total television programming in China were imported, of which 73.6% were from developed countries such as U.S. and Japan. The programs imported included movies, dram series, children’s programs, sports, and documentaries etc. Although the 31 provincial television channels are now supposedly made available by satellites to the whole nation, in practice the reception of the satellite signals is not guaranteed to reach 100% of the nation’s television households. There are three conceivable reasons: a) the uneven development of the local cable system that are used to receive the satellite television signals, b) the different policies employed by local governments towards the satellite reception, and c) the limited capacities of some old television sets that are still in use in many rural households. While some households in metropolitan area can receive as many as 70 channels, the average number of channels available nationwide is around 20 to 30 in cities and 10 to 20 in rural areas (CSM, 2001). As a result, it is reasonable to speculate that de facto polarization has some impact on audience viewing behavior. Meanwhile most multi-channel stations in China have started to “specialize” their channels, i.e. to designate each of their channels to a special program genre such as movie, news and sports, etc. The most prominent example would be China Central Television. All but one of its eleven channels are exclusively dedicated to one or a couple of program genres. So are the eight Beijing TV channels. All the above stated current environmental features give us reasons to expect some degree of audience fragmentation and polarization. But how do audiences actually view these specialized channels? Are there groups of small but loyal audience who devoted much of its time exclusively to its channel of choice and nothing else? Barwise and Ehrenberg, however, argued that specialized channels that usually attract a relatively small audience would suffer from a “double jeopardy.”

　　“Double jeopardy” (McPhee, 1963; Goodhardt, et.al., 1987) denotes the phenomenon in mass audience behavior that unpopular programs have not only small but also disloyal audiences. Although there have been exceptional cases found for some local markets, e.g. Barwise and Ehrenberg (1984) demonstrated minority-language and religious stations enjoyed inordinately high time-spent-viewing (TSV) levels despite their limited reach; Webster (1986) found that the Spanish-language station had an unusually loyal audience, more general studies of audience behavior which have operationalized loyalty as either TSV (Barwise & Ehrenberg, 1984) or repeat-viewing (e.g., Ehrenberg & Wakshlag, 1987; Webster & Wang, 1992) indicate that double jeopardy is the dominant rule.

　　 The “double jeopardy” is directly at odds with the notion of “small-but-loyal” audience that is the rationale behind the undergoing channel specialization in China. To what extent does this phenomenon hold up in today’s multi-channel environment? Is television still a broadcasting medium or as is it increasingly fragmented and we are already entering a “narrowcasting” era? The answer to these questions has pragmatic implications for China’s television industry.

　　Method

　　 The study is a re-analysis of the ratings data collected by CVSC-Sofre-Media (CSM) in Beijing during the week of March 3 to 9, 2001 using its peoplemeter panel. As one of the most developed cities in China, Beijing often represents the trend of future socioeconomic development for the rest of the country. Therefore it is an ideal locale for the current study. The total sample is 300 households projecting to 7.2 million people. The number of channels that can be received by an average household in Beijing is 33 while some households in the metropolitan area may receive as many as 77 channels (CSM, 2001). The study includes 50 major channels available in Beijing which account for 94% of the total viewing during the period of time in question.

　　The results of the analysis are presented in a seven column table, the same format as use by Webster (2003) in a study on the same issue. The first column lists all the 50 channels in an alphabetic order. The second is the universe of households in which each channel can be received, expressed as a percent of total TV households (TVHH). The next two columns are the measures of the size of each channel’s audience. The first is the channel’s cumulative audience (or reach) expressed as the percent of all persons 4+ who have viewed for at least one minute during the week. The second is the channel’s overall share of audience, i.e. the percent of the total number of “man-hours” that is attributable to that particular channel. The remaining three columns of data are various measures of polarization.

　　The first is “time-spent-viewing” (TSV), the average number of minutes per week spent viewing the channel in question, among those who tuned in for at least a minute (i.e., those in the cume). It is also a common measure of loyalty in studies assessing the double jeopardy effect (Barwise & Ehrenberg, 1984)

　　The second measure is “share-within-cume.” This measure adjusts for the total amount of time a channel’s users spend watching television by representing TSV as the percent of total TV viewing time.

　　The third is measure is a “within-channel duplication index”, the ratio of share-within-cume to total market share (Webster, 198? & 2003). It is a rough approximation of the constant in the duplication of viewing law (Goodhardt, et. al., 1987). An index of 1.6 would mean that a channel’s audience is 60% more likely than the general population to watch that channel.