May 25, 1993
Presented to the Information Systems Division Conference of the International Communication Association, Washington, DC on May, 1993. Submitted to the Journal of Communication, August, 1993
But television is also an appliance, a physical object, a box that comes in various colors, shapes, and sizes. A particular television, especially when we own more than one, is often dedicated to a particular room, a particular type of content, or a particular reason to view. The question for this study concerns the possible influence of television as a box on responses to the screen. Could differences between television sets change responses to television content, especially when we assign a screen to one particular role among its many possible uses and functions?
What could it mean to respond to television as a box? There are three possibilities, and for two of them there is already evidence that the television set can change responses to identical content. First, responses to television can change in relation to physical attributes of the box, especially its size and shape. Experiments have shown that screens of different sizes, each displaying identical content, can change attention to, memory of, and evaluations of people shown in news (Lombard, 1992), people presented in a neutral context (Reeves, Lombard & Melwani, 1992), and reactions to people and action in segments from feature films (Reeves, Detenber & Steuer, 1993).
A second response to television as a box considers television as a place, a place where things happen and where people exist. As a place, television becomes a part, but only one part, of the spatial arrangements in a room. This means that when a television set moves or when we move in relation to it, everything in the box (or on the screen) moves as well. Research suggests, for example, that if individuals move close to the screen, they feel that they are physically closer to the people displayed on the screen (Lombard, 1992; Reeves, Lombard & Melwani, 1992). In this sense, television images are evaluated as merely one part of the viewing environment which surrounds the box; the images are not processed solely in relation to the cues within the frame of the screen.
Two quite different experimental approaches demonstrate the influence of spatial arrangements between the box and viewer. The experiments share the theoretical notion that physical space is a critical cue that guides attention, memory, and behavior. First, in two studies, manipulations of viewing distance (close versus far) were treated as changes in the arrangements of physical space, and because faces were shown on the screen, viewing distance represented differences in the arrangement of personal space (Reeves, Lombard & Melwani, 1991; Lombard, 1992). Both studies showed that closer viewing distances (just as closer interpersonal distances) encouraged the evaluation of people on the screen to be more extreme.
Second, an unrelated experiment demonstrated the importance of physical space as a memory cue. Rothkopf, Dixon & Billington (1986) had subjects watch a discussion on the MacNeil-Lehrer News Hour either on one television monitor, with six discussants portrayed within the frame of a single screen, or on six different monitors, arranged from left to right in front of the viewers with each discussant assigned to a separate monitor. Subjects who watched on multiple monitors attributed statements to their proper source more accurately than subjects who watched a single monitor, presumably because they had a spatial cue to aid memory.
These experiments suggest that the television box can define the place where people and action exist. Thus, the placement of the box determines the physical relationship of people and action within it to other people and things in the viewing environment (cf., Nass & Steuer, 1993). Most notably, the box can determine the spatial relationship between the content of the screen and the viewer.
A third response to television as a box (and the subject of the present experiment) considers the assignment of particular roles to different television sets, regardless of whether the assignment is made consciously or not. Imagine a household where sports is normally watched on the large TV in the family room, news and soap operas on the small portable in the kitchen, movies on the set attached to the VCR, and late night talk shows in the bedroom. When we own more than one television or when we have a special need for one form of content, it is common to watch a particular program genre on one particular screen. It is not necessary that the screens be the exclusive location for specific content (the news will certainly be seen, on occasion, on the large screen in the living room). But regardless of slippage around a special role, we can reasonably question whether the different roles assigned to television sets can shape viewer's expectations about the content that will appear, and consequently, viewer's evaluations of the programs that are shown.
Admittedly, this proposal may seem far-fetched. Could people really think that a role assigned to a television set should change their thoughts about a television program? We think there are two explanations that suggest such a link is plausible, even though people are likely not aware of the influence, and even though they would label the confusion between television sets and television programs as absurd if questioned about it after the fact.
First, the assignment of technologies to specific roles is quite familiar. It is not unusual for a bedroom radio to be tuned to one radio station and the kitchen radio tuned to another. The computer in the study runs different software than the computer in the children's room. And the telephone in the dining room is reserved for different calls than the one in the bedroom or car. It isn't necessary that these roles be consistent without exception. Rather, the examples demonstrate that technologies can become specialized because we assign them different roles; technologies are not unique only because they represent a particular collection of hardware, because they are made by different companies, because they are programmed by different people, or because they are sold in different stores. And conversely, technologies are not similar just because they look alike. Identical hardware can be used in quite different ways.
Second, the assignment of roles to technologies shares many of the same qualities as the assignment of roles to people. To the extent that people respond to technologies (especially televisions and computers) as if the technologies were people--a confusion that comes naturally and one for which Nass and colleagues have provided ample evidence (e.g., Nass & Steuer, 1993; Nass, Steuer, Henriksen, & Dryer, 1993; Nass, Steuer, Tauber, & Reeder, 1993)--then the concept of role becomes a reasonable way to summarized people's expectations about constancy. Once we learn to expect consistent behavior form a person, that person becomes specialized; that is, that person is assigned a role (Parsons, 1962; Wallace, 1983). Since people have considerable experience with roles assigned to other people, and because communication technologies often behave as other people do, then we can reasonably predict that viewers will judge and expect content from a role-specific television in the same fashion they would judge and expect behavior from another human assigned to the same specialized role.
The expectation that the specialist television would encourage evaluations of content in favor of its role was based on two characteristics of specialization. First, agents (whether people or televisions) are expected to behave consistently within their specialties, and this should make specialists appear more representative of their roles. In contrast, generalist behavior is diverse. Hence, any particular instance of generalist behavior should not be prototypic of any particular role.
A second rule is that specialists are better than generalists, at least in the realm of their specialization. Just as brain surgeons are perceived as better at brain surgery than are general practitioners, entertainment or news televisions should show better entertainment or news, respectively.
The specific expectations for this experiment then were that people who watch a news-only television set should perceive the news segments to better reflect fundamental characteristics of news than people who watch the identical content on a television used to display several different types of content (generalist television). Similarly, people who watch an entertainment-only television set should perceive the entertainment segments to better reflect fundamental characteristics of entertainment than people who watch the generalist television. Furthermore, the perceived quality of content should be superior for the specialist televisions than for the generalist television.
Subjects were randomly assigned to either the specialist TV (two TVs) condition or the generalist TV (one TV) condition. Random assignment was also made for TVs serving as the "news TV" and the "entertainment TV" in the specialist TV condition, and for the TV serving as the "news/entertainment TV" in the generalist TV condition.
After entering the laboratory, subjects were told that they would participate in a research project that involved watching short segments from several television news programs and from several entertainment programs. The purpose of the research, they were told, was to examine how people respond to different types of news and entertainment programs. Each subject was instructed on how to use the remote control for the VCR (two remotes in the specialist TV condition).
In the specialist TV condition, a sign was placed on the TVs which indicated the content type that would be shown. Also, the TVs showed only the content that was indicated by the sign. While physical characteristics of the box are self-evident (e.g., screen-size and placement), content characteristics are not. Without a sign, subjects would need to watch a TV for an extended period of time before they could perceive that the TV showed only one type of content. During the instructions, the experimenter referred to the TV in the generalist TV condition as the "news/entertainment TV" and to the TVs in the specialist TV condition as either the "news TV" or the "entertainment TV."
The experimenter told each subject that he or she would watch news and entertainment programs, would be asked to answer a few questions about each program segment, and then would answer a few more questions at the end of the session. All answers were recorded on a self-administered paper-and-pencil questionnaire.
Subjects were instructed to watch two segments, pause the VCR, and answer questions about each segment. In the generalist TV condition, subjects watched two more segments, answered more questions, and continued until all eight segments were watched and until all questions were answered. In the specialist TV condition, subjects were instructed to watch the first two segments and answer the corresponding questions. The segments they watched, either news or entertainment, corresponded to the particular role that was assigned to the TV they were watching. They were told to then get up and move to another chair in front of the other TV. (The chairs were next to each other but faced in opposite directions.) After watching two more segments, they were to answer the corresponding questions. They continued this procedure until all eight segments were watched and until all questions were answered. After all instructions were given, the experimenter left the room.
The four entertainment segments came from network situation comedies: 1) Who's the Boss?, 2) Roseanne, 3) Cosby, and 4) Cheers.
Each segment was recorded on VHS videotape. Eight segment orders were chosen at random from all possible order combinations of like-segment pairs. That is, each subject saw either two news segments followed by two entertainment segments, then two more news segments and two more entertainment segments, or vice versa. Each experimental session consisted of all eight video segments.
Based on these criteria, five "news" variables were chosen: important, informative, disturbing, serious, and interesting. Two "entertainment" attributes were chosen: funny and relaxing.[1]
The means across condition for content representativeness are shown in Table 1. A higher mean indicates that subjects perceived the particular program content as being more representative of a news or entertainment attribute. Consistent with the hypothesis that specialist TVs provide more representative content than do generalist TVs, subjects who watched the specialist TVs rated the news to be significantly more important (t = 2.36, p = < .05), more informative (t = 1.76, p < .05), more disturbing (t = 2.06, p < .05), and more interesting (t = 1.84, p < .05) than subjects who watched the same content on the generalist TV. The result for the "serious" attribute, although not significant (t = 0.94, n.s.), was in the predicted direction.
As predicted, subjects who watched the specialist TVs rated the entertainment to be funnier (t = 1.42, p < .1) and more relaxing (t = 2.07, p < .05) than subjects who watched the generalist TV.
To demonstrate the consistency of the effect, we disaggregated the representativeness indices. Twenty measures of news representativeness were examined (five news characteristics x four program segments). As predicted, subjects who watched the specialist TVs rated the news segments to be more representative of news than the subjects who watched the generalist TV for 18 of the 20 items, a clearly significant difference by sign-test (p < .002).
For all eight entertainment representativeness items (two entertainment characteristics x four program segments), content on the specialist TVs was perceived as more representative of entertainment that was content on the generalist TV. A sign-test shows that this pattern is significant (p < .005).
Table 1:
Mean Representativeness for Specialist and Generalist TVs
Specialist TVs
(2 Generalist TV
(1 t TV) TV)
News
Important 7.9 7.0 2.36*
Informative 7.6 6.7 1.76*
Disturbing 7.8 6.7 2.06*
Serious 8.0 7.6 0.94
Interesting 7.5 6.5 1.84*
Entertainment
Funny 6.5 5.5 1.42+
Relaxing 7.5 6.1 2.07*
+
= p < .1; * = p < .05
Subjects who watched the specialist TVs rated the quality of both the news (t = 3.50, p < .01) and entertainment segments (t = 2.05, p < .05) significantly higher than subjects who watched the generalist TV (as shown in Table 2). Also, subjects who watched the specialist TVs liked both the news (t = 1.98, p < .05) and the entertainment (t = 2.89, p < .01) segments better than subjects who watched the generalist TV.
Table 2:
Mean Specialist Superiority for Specialist and Generalist TVs
Specialist TVs
(2 Generalist TV
(1 t
TV) TV)
News
Quality 8.1 5.6 3.50*
Liking 7.2 5.7 1.98*
Entertainment
Quality 7.4 5.8 2.05*
Liking 8.3 6.2 2.89**
+
= p < .1; * = p < .05; ** = p <
.01.
The traditional stance in mass communication research is to treat television as simply the means of presenting audio/visual material--books and newspapers as means of presenting text, radio as the means of presenting audio material, and computers as the means of facilitating interactivity. This study demonstrates that this approach, at least for television, is incomplete: The physical embodiment that we call "television" has a psychological reality separate from its content and modality.
Nass, C., & Green, N. (1993, March). Stereotyping of computers with gendered voices: Computers are social actors IV. Paper presented at the annual meeting of the Western Academy of Management, San Jose, CA.
Nass, C., & Steuer, J. (1993). Voices, boxes and sources of messages: Computers as social actors, Human Communication Research, 19(4), 504-527.
Nass, C., Steuer, J., Henriksen, L., & Dryer, C. (1992). Machines and social attributions: Performance assessments of computers subsequent to "self-" or "other-" evaluations. Unpublished manuscript, Stanford University.
Nass, C., Steuer, J., Tauber, E., & Reeder, H. (1993, April). Anthropomorphism, agency, & ethopoeia: Computers as social actors. Paper presented at INTERCHI `93, conference of the ACM/SIGCHI. Amsterdam, the Netherlands.
Parsons, T. (1961). An outline of the social system. In T. Parsons, E. Shils, K. D.. Naegele, & J. R. Pitts (Eds.), Theories of Society. New York: Free Press, pp. 120-153.
Reeves, B., Lombard, M., & Melwani, G. (1992). Faces on the screen: Pictures or natural experience? Manuscript submitted for publication.
Reeves, B., Detenber, B, & Steuer, J. (1993). Effects of enhanced spatial context on television message retention. Paper presented at the annual meeting of the International Communication Association, Washington, D.C.
Rothkopf, E. Z., Dixon, P., & Billington, M. J. (1986). Effects of enhanced spatial context on television message retention, Communication Research, 13(1), 55-69.
Wallace, W. L. (1983). Principles of scientific sociology. New York: Aldine.