The goal of this study is to test predictions derived from both implicit assumptions made by professional news producers and theoretical predictions derived from the limited capacity model of mediated message processing (LCMMMP) about how the effects of production pacing and the length of television news stories affect news viewers' channel changing behavior and their processing of the news in a free choice viewing context.
Production Changes in a Multichannel Remote Control Environment
In recent years, in order to keep viewers on channel in the multichannel remote control environment, television producers have changed the structure and content of their messages (Bellamy & Walker, 1996). One common change has been to make messages shorter and faster (Bellamy & Walker, 1996; Eastman & Newton, 1995). As Bollier (1989) and Eastman and Neal-Lunsford (1993) noted, practitioners are using more cutting, shorter scenes, faster-paced shows, and more shorthand visual techniques. "Pregrazed" programs like Short Attention Span Theater (Comedy Central) and The Edge (FOX) were designed assuming that if the program itself is changing, the viewer need not change the channel (Bellamy & Walker, 1996).
Changes in Audience Viewing Behavior
Research on channel changing behavior has provided descriptive data on who changes, how often, when, and why. Some viewers rarely change while others change constantly. Ferguson (1994) measured channel changing with college students using an electronic counting device and found that the number of changes ranged from 3 to 396 times per hour, with a mean of 107. Kaye and Sapolsky (1997), also using a mechanical counter, reported a channel changing range of 1.23 to 178, with an average of 36.6. When self-report methods are used, the average frequency of channel changing is much lower but still suggests that many viewers change channels frequently and that long periods of single channel viewing is not the norm (Ferguson, 1992).
Channel changing behavior differs as a function of age and sex. Previous studies have found that younger viewers change channels more often than older viewers (Eastman & Newton, 1995; Greenberg, Heeter, & Sipes, 1988) and males change more than females (Copeland & Schweitzer, 1993; Eastman & Newton, 1995; Heeter, 1985). Ferguson and Perse (1993) found that age and gender interact. Older women change less than older men but younger men and women do not differ significantly.
Research suggests that viewers change channels for many reasons, and the reasons may be related to channel changing frequency (Ainslie, 1988; Ferguson, 1992; Walker & Bellamy, 1991). For example, Heeter and Greenberg (1985a, 1985b, 1988) identified four types of channel changers: those who change rarely, those who change between programs, those who change during commercials, and those who change at all times. Perse (1990) found that 1.8% of adults say they never change, 23.5% report changing between programs, 44.3% change during commercials, and 30.4% change during programs. Moriarty and Everett (1994) also found that most viewers change during program and commercial breaks, but some change at all times.
Walker and Bellamy (1991) report that viewers change channels to see what is on other channels and to avoid commercials. Perse (1998) found channel changing is associated with ritualistic viewing, low attention, and engagement. Eastman and Newton (1995) found that viewers change most during sports and the least during pay-cable movies but that some viewers change channels regardless of genres or content.
Production Pacing, Story Length, and Program Choice
Research in this area has been primarily descriptive, asking what people do and when they do it. Theories about how television's structure and content affect channel changing behavior or about why people change channels have rarely been tested and often reflect the implicit theories or assumptions that appear to be operating in a professional world. The literature argues that producers assume that younger viewers prefer faster-paced television production and that fast pacing will hold viewers on channel. Alfstad (1991) argues that young viewers have been "programmed" to switch their attention rapidly from topic to topic and image to image (p. 20), and Bellamy and Walker (1996) characterize younger viewers as "raised on a collage of rapidly shifting images, able to absorb visual information quickly, fascinated with new technology, and easily bored" (p. 96). Greenberg et al. (1988) found that younger viewers are more likely to watch short segments, change channels frequently, view multiple programs simultaneously, and carry out orienting searches.
Research looking at the older viewers, on the other hand, suggests that older viewers may have more trouble processing fast-paced messages and be turned off by fast-paced production (Lang, Schwartz, & Snyder, 1999). The first hypothesis tests this prediction:
[H.sub.1]: Younger viewers will prefer fast-paced programming and short stories, whereas older viewers will prefer slow-paced programming and long stories.
Production Pacing and Cognitive Effort
Although both research and practice suggest that increasing production pacing may affect channel preference, we know that, at least in the laboratory, in a forced viewing situation, production pacing has a significant effect on how viewers process the information presented in a news story. A great deal of research has been done--in laboratories--to learn how increased production pacing affects the information processing of mediated messages (e.g., Geiger & Reeves, 1993; Lang, 1990, 1991,1994; Lang, Bolls, Potter, & Kawahara, 1999; Lang, Geiger, Strickwerda, & Sumner, ]993; Lang, Zhou, Schwartz, Bolls, & Potter, 2000). One theory that has frequently been used to explicate how production pacing affects how viewers process messages is the LCMMMP (Lang, 2000). The hypotheses put forward in this article about how production pacing will affect channel changing behavior and the processing of mediated messages are derived from that model. Previous tests of these hypotheses have all been done in a forced viewing environment where participants were instructed to play close attention to a message and were not given the opportunity to change channels or do anything else. In this experiment, participants were instructed to watch television as they would at home, were provided with a remote control device, and were told they could change channels at will. Hence, if hypotheses derived from the LCMMMP are supported, it will be possible to begin the process of generalizing predictions made by the theory from the forced choice viewing, high attention environment to a free choice environment.
LCMMMP defines attention as the allocation of processing resources to a message. How many resources are allocated to a message is determined by a combination of controlled and automatic allocation mechanisms (Shiffrin & Schneider, 1977). Television viewers can allocate their processing resources to a message in response to their goals and motivations using controlled allocation mechanisms. Viewers allocate more resources to relevant, interesting, and involving messages. If viewers do not like a message, they are likely to decrease their cognitive effort. Thus, if older viewers prefer slow-paced messages to fast-paced messages, as predicted, they should exert less cognitive effort during fast-paced messages compared to slow-paced messages, whereas younger viewers should show the reverse pattern. Thus:
[H.sub.2]: Fast pacing will elicit greater cognitive effort in younger viewers and less cognitive effort in older viewers.
Production Pacing and Memory
Of course, controlled resource allocation does not tell the whole story about how messages are being processed and whether they will be remembered because, according to LCMMMP, resources are also being automatically allocated to processing in response to structural and content features of the message (Lang, 2000). Production techniques like fast pacing, novel visuals, and video graphics elicit automatic attention in television viewers (Fox et al., 2002; Lang, 2000). The combination of controlled and automatic resource allocation determines overall resource allocation. How well the message is processed depends on whether sufficient resources are allocated to the message processing task.
Prior research demonstrates that creating messages that increase resource allocation (either automatic or controlled) does not automatically increase memory for the message. Instead, research shows that introducing structural features that automatically elicit processing resources will increase memory up to a point. After that point, however, memory has been shown to decline. LCMMMP defines that point as the point of cognitive overload (Lang, 2000), which occurs when the viewer's limited capacity of processing resources has been completely allocated and that allocation is insufficient to fully process the message. When resource requirements exceed the viewer's capacity, then fewer resources will be allocated than are required, and the job of processing the message will be performed less thoroughly. Content difficulty, familiarity, and viewers' age have...