User Perceptions of the Characteristics of Media

One such perspective can be found in the perceptions of highly experienced users of a variety of media. A series of questions in the 1988 survey of IBMPC participants allow us to test this perspective. Participants were asked, in these questions, to rate computer conferencing and six other media on three broad characteristics of media. These three characteristics were selected to represent the three dimensions of media identified in the previous chapter. The six media were selected to represent both the extremes of these dimensions and a sampling of traditional, technological, and computer media.


The three characteristics surveyed were:

Media Speed
Respondents were asked, for each of the seven media, "How fast is the medium. (How quickly can a message travel? How quickly must/will the message be received? How dynamic is the content?)" Answers were selected from among the following choices:
  1. instantaneous
  2. almost instantaneous
  3. very quick
  4. moderately quick
  5. fairly quick
  6. somewhat slow
  7. very slow
  8. extremely slow

This question attempts to simulate the dynamism factor in the typology. The simulation is not, of course, perfect. Dynamism, in the factor analysis, entails more than just the raw speed with which a message travels from sender to receiver. Other important components of this factor which are not adequately represented in this measure include the persistence of the message (how long it continues to exist after it is created), and at least one measure of bandwidth (negatively correlated). The result is a pure measure of speed that is probably somewhat correlated with the second measure, media interactiveness.

Media Interactiveness
Respondents were asked, for each of the seven media, "How interactive' is medium x? (How easy is it to reply to messages? How quickly can you reply? How quickly will the reply be read? Does it allow conversation?)" This question attempts to simulate the interpersonal versus mass media factor in the typology. Again, the simulation is less than perfect. In the typology, this factor represents both interactiveness and audience size, with larger audience size generally correlated with less interactiveness. Here, by contrast, we are only concerned with interactiveness, and the result is probably at least somewhat correlated with speed.

Users were asked to select from among the following choices:

  1. totally interactive
  2. extremely interactive
  3. very interactive
  4. moderately interactive
  5. somewhat interactive
  6. a little bit interactive
  7. not at all interactive
Media Bandwidth
Respondents were asked, for each of the seven media, "How rich is the content of medium x? (How much information must be processed all at once? How much information is available?)" This question attempts to simulate the bandwidth factor in the typology. Note that the question is phrased not in terms of the number or potential information content of the channels of communication available, but in the relative information content of messages transmitted. This differs somewhat from the hard measures of bandwidth used to form the typology. Indeed, it is more similar, in many ways, to the concept of "media richness" (Trevino, Lengel, and Daft, 1987).

Users were asked to select from among the following choices:

  1. extremely rich in content
  2. very rich in content
  3. moderately rich in content
  4. somewhat rich in content
  5. somewhat poor in content
  6. moderately poor in content
  7. very poor in content
  8. extremely poor in content


The seven media surveyed included:

A test of the typology

It was expected, on the basis of the formal typology, that the media would line up along the three dimensions from highest to lowest as follows (note that some media have similar expectations associated with them, and are therefore ranked at the same level):

When tested by Spearman rank correlation of these expectations with the results of the survey data (the ranks of significantly different means), the expectations are only partially supported. Media speed compares well with expectations (r(s)=.99, n=7, p<.01). Media interactiveness also compares well (r(s)=.86, n=7, p<.05). Media bandwidth does not compare well with expectations, however. Indeed, there appears to be no relationship between the media bandwidth survey variables and the bandwidth factor in the typology, at least within the narrow range of media surveyed. Rank correlations between the three measures support their independence from each other.

Speed and Interactiveness

The success of the speed and interactiveness variables speaks well for the reasonableness of these two factors in the typology. People seem to have a sense of these dimensions and the way they apply to the media they use that squares up with the way these media classify themselves when their characteristics are compared in a numerical taxonomy.

Both measures attract a fairly wide range of response. The mean estimates for the speed of individual media ranges from very slow to almost instantaneous, and the seven variables differentiate themselves at five clearly differentiated levels. The mean estimates for interactiveness are even more various, and range from extremely interactive to not at all interactive. All media measured differ significantly in interactiveness from all others.

Respondents quite reasonably regarded face to face interaction as the fastest, most interactive, and highest bandwidth (richest) medium surveyed. It is broadly regarded as being almost instantaneous (the median and mode), totally interactive (the response selected by two-thirds of respondents), and very rich (the median and mode). The telephone trailed closely on both interactiveness and speed. It was regarded as both very quick and extremely interactive (each the median and mode for its measure).

Publications and correspondence, by contrast, were estimated to be both slow and non-interactive. Publications were seen as being significantly slower and less interactive than correspondence, as might be expected. Correspondence was perceived as being somewhat (the mode) to moderately (the median) interactive and somewhat slow (the mode and median). Publications were perceived as being very slow and not at all interactive (the median and mode for each measure).

It is television that, more than anything, assures the independence of the speed and interaction measures. Respondents classified it as one of the faster media. Its mean for media speed does not differ significantly from that of the telephone, with a mode and median of very quick. Respondents also classify it as the least interactive of the media surveyed, with 83% of respondents declaring it "not at all interactive."

These results agree with common sense expectations. Direct interaction, whether face to face or on the telephone, is both faster and more interactive than either correspondence or publications. Broadcast television is fast, but not interactive. The placement of these media on these measures, and the relative value of the estimates, only acts to give a common sense affirmation that these measures are, in fact, measuring what they are supposed to be measuring. The results for electronic mail and computer conferencing should be more interesting, however, as the reader may not have the experience with these media that would be required to make similar common sense judgements.

One would expect, on the basis of the typology, that these media would tend toward the center. When dynamism (speed) and interpersonal versus mass communication (interactiveness) factors were juxtaposed, the result was a "hidden" seventh cluster of computer media, which seemed to reach across the middle ground of these two dimensions. This expectation is not disappointed.

Respondents classify both computer conferencing and electronic mail as being moderately quick, significantly slower than the telephone or television, but significantly faster than either correspondence or publications. Although, in terms of pure rank, respondents rated computer conferencing as being the faster of the two, the difference in speed was not significant.

The same difference is significant when these media are compared on interactiveness, however. Computer conferencing was estimated to be very interactive (the median and mode), while electronic mail was seen to be moderately interactive (the median and mode). In both cases, the interactiveness of these media are well behind face to face communication or telephone conversations. They are, on the other hand, seen as being well ahead of correspondence, publications, or television on the same measure.

Media Richness

The failure of the bandwidth variable makes it difficult to make reasonable comparisons with the bandwidth dimension from the typology, but does not preclude use of the data to make some interesting observations about the media measured, including computer media. The failure itself is probably due to the difference between channel bandwidth and message content. Increased bandwidth creates the potential for the transmission of more information, but does not preclude the transmission of large amounts of redundancy. Decreased bandwidth, by contrast, may limit the amount of information that can be transmitted, but may well challenge the creators of messages to use that bandwidth well.

Media richness, which is what, realistically, this question measures ("How rich is the content of ...") is not a measure of potential. It is a measure of realization, how well people use the bandwidth a medium offers to create messages. This contrast between potential and reality is apparent in several of the measures used in the typology, including persistence, speed, and audience size. No such measure was used, however, for resolving bandwidth to the reality of message content. In the typology's defense, the author is aware of no unambiguous measure of the real message content associated with various media that could have been used in the typology. Still, the significant difference between the potential measured in the typology's bandwidth variable and the reality measured in the survey question raise serious questions about the adequacy of the typology's bandwidth dimension.

The reality observed in this question is quite interesting, however, in part because the range of that reality is so constrained. No medium surveyed is seen as being "extremely rich". By the same token, no medium is seen as being less than "somewhat rich". The bulk of responses then, for all media, are within the range of "somewhat", "moderately", and "very" rich in content.

This constraint should not, perhaps, be unexpected. All of the media surveyed are commonly used (and by extension useful) media which all respondents should be very familiar with. It is hard to imagine that many members of this survey group (selected based on their apparent use of electronic media) use any of the media surveyed less than once a week. It is expected that most respondents use all considerably more often. It does raise some interesting questions, however.

Trevise, Lengel, and Daft (1987), when making a series of hypotheses concerning media richness, refer to written correspondence as a lean medium. It is clear, however, from the responses to this question, that a lean medium is somewhat akin to lean meat or lowfat milk. Lean meat and lowfat milk may both have substantial fat content, perhaps 50% or more of the fat content of "fatty" meat and milk. It is only by comparison that they are lean or lowfat. So it is with media richness. Correspondence is assessed, in this study, to be, along with television, the least rich medium, with a mean and median of somewhat rich. At worst, correspondence is a moderate content medium.

One suspects that the truly "lean" or low content media are the ones people don't use very often. Hence studies of media richness might do well to include some of these media as a base for comparison. One suspects that one would find that telegraph messages, telex, tape letters, C.B. Radio, filmstrips, and even pop records would all be classified as truly lean media by at least some people.

Given the standard of the media richness literature, the classification of face to face interaction as high content ("very rich") and of correspondence as moderate content ("somewhat rich") is probably expected. The same cannot be said of the ratings of some of the other media. Publications, for instance, are classified as being fairly rich in content; significantly richer than either telephone conversations or television. The median for publications is, like face to face interaction, "very rich in content". The mode, however, is "moderately rich", and the mean is closer to that standard, significantly less rich than face to face interaction.

Television, moreover, despite its multichannel (sight and sound) presentation, is seen as only moderate in content. Although its median is "moderately rich", its mode is "somewhat rich", and it is seen as being no richer in content than correspondence. Telephone conversations, by contrast, inhabit a middle ground. The median and mode of this medium for media richness is "moderately rich in content".

On reflection, these results seem very sensible. Television generally wastes its potential by broadcasting a high level of redundancy. Shows are repeated exactly as they were when first broadcast. Commercials may be repeated several times during the course of the same show. Innovation is unusual in either, and it is a rare series that doesn't duplicate the same plots and themes over and over again; borrow the formulas that have been successful on other shows. News broadcasts are very dependent on similar formulas. This does not necessarily diminish its entertainment value. If it did, the redundancy would be reduced. In any case, the same has been said, at different times, of all of television's entertainment predecessors, including movies, plays, circuses, and vaudeville. Indeed, these media invented most of the formulas that television makes continued use of.

Publications, by contrast, are frequently used to communicate detail, sometimes at levels exceeding those apparent in face to face interaction. A novel may give details of a face to face interaction, including the thoughts and feelings of the participants, that may be hard to get at during interaction. A text will frequently give details of a topic that will be difficult to reproduce in a video. A newspaper can frequently afford to publish details of an event that won't fit within the format of a television news show. Hence the relative richness of publications is not entirely surprising.

Telephone conversations occupy a middle ground, short on the detail inherent in publications or the high non-verbal bandwidth inherent to face to face interaction, but with more varied content than television and more non-verbal detail than correspondence. Hence, at least among the traditional and technological media surveyed, this measure of media richness is sensible. The question, of course, is where the computer media, computer conferencing and electronic mail, fit in.

Computer conferencing turns out to be the richer of the two. Its median is "very rich in content". Its mode is "moderately rich in content". Its mean is not significantly different from, that of publications. Electronic mail, by contrast, is rated as similar in content to (its mean is not significantly different than) the telephone. Its mean and mode are somewhat rich in content, but its mean lies directly between somewhat and moderately rich in content. Several contrasts stand out in these results, and deserve more detailed exploration.

Electronic Mail and Correspondence

Electronic mail is not, as is suggested by Trevino, Lengel, and Daft's (1987, p. 560) second proposition, as lean a medium as written mail, at least in the view of highly experienced users of both media. This difference is probably related to several elements, including the speed and increased interactiveness of electronic mail. Other factors that may be important include the relative ease with which electronic mail can be saved for future reference and referenced once saved, and the ease with which electronic correspondence can be edited. The electronic mail user can edit the mail he or she receives in creating new mail, and readily revise mail before sending it.

In fairness, it should also be noted that this difference may also be a response effect. Electronic mail is new, and although widely used in IBM, is relatively new to most users. Hence respondents may be attributing an increased richness where there is none. One notes, however, that the effect seems to be real enough. Electronic mail use now exceeds that of hard copy mail in IBM, at least among users of IBMPC, by a considerable margin. Harvard Business School (1988) reports that the number of envelopes physically mailed in IBM starts to decline in 1986. It is increasingly unusual for letters, memos, software, or even publications to be distributed via conventional hard copy mail.

This radical shift is readily observable throughout IBM's business and has been clear in measures that have been documented elsewhere in this study. It does not appear to be due to any management decision. Stationary closets remain full. It does not appear to be a fad change. Electronic mail use has steadily increased in IBM over a long period of time, and the increase is continuing.

The more reasonable explanation is that electronic mail does its job better than conventional mail does. It was already clear that some of this superiority was related to its speed and relatively greater interactiveness. Now it appears that, at least in the eyes of those that use it, it may also be richer in content.

Electronic Mail and Telephone Conversations

The comparative richness of content inherent to electronic mail when compared with telephone conversations is also interesting, especially given the contention, made elsewhere in this study, that electronic mail was replacing telephone use. Respondents do not regard electronic mail as being either as speedy or as interactive as the telephone. It is clear that the telephone is richer of the two in non-verbal content. Yet respondents rate the richness of these two media very closely (no significant difference between means).

One suspects that electronic mail makes up for its disadvantages in several ways. First, electronic mail is written, and allows the writer to take time to think out a position before stating it. When sent, moreover, the response will often contain greater detail, and can easily be copied to others. Second, electronic mail can be more readily saved, reaccessed, and, if necessary, edited. Third, because electronic mail is asynchronous, one can frequently exchange several notes during a period in which one might otherwise play telephone tag. All of these factors act to enrich the content of electronic mail, and make it an effective substitute for the telephone under many conditions.

Computer Conferencing and Publications

The lack of significant difference in the richness of content attributed to computer conferencing and publications is also interesting, as it suggests that computer conferencing can be used to replace publications under some circumstances. Given the clear superiority of computer conferencing to publications in terms of both speed and interactiveness, moreover, it is easy to imagine circumstances under which such replacement would be desirable. Some of these circumstances have already revealed themselves in the genres of communication that have revealed themselves on IBMPC.

For a more global statement, however, one might think of a computer conference as an interactive publication in which large numbers of users educate each other on the details of a particular topic. The result is compariable to neither magazine, newspaper, nor book. None of these publication types are currently threatened by computer conferencing. Yet the function of each, the dissemination of information, is gradually being absorbed by computer conferences, and one imagines that there are already cases of publications that haven't been written because a computer conference did the job better. An example of such can be found in the yet to be written tutorial for the TOOLS software that is used to run IBMPC, PCTOOLS, and most computer conferences in IBM. It has been replaced by a forum, shared jointly between the IBMPC and IBMVM conferencing facilities, called TOOLSRUN LESSONS.

Once more into the Middle Ground

From the standpoint of the typology, the most important result of this analysis is the confirmation that computer media do form the hidden "seventh cluster" of media that was apparent in the juxtaposition of the interpersonal versus mass communication and dynamism dimensions. Indeed, the results seem to strengthen this notion that computer media are creating a middle ground that bridges otherwise polar dimensions of media.

This is particularly apparent in examining the speed and interactiveness variables, where electronic mail and computer conferencing take sole possession of the middle ground, creating continua in both measures where there were once only extremes. These continua may well be the great social contribution of computer conferencing and electronic mail. Where we once had fast media in which the movement of messages was almost instantaneous and slow media in which the movement of messages was measured in days, we now have a continuum of media, including computer media that deliver messages in minutes or hours. Where we once had very non-interactive media that were studied in departments of mass communication and interactive media that were studied in departments of rhetoric and communication, we now have a continuum of media including computer media whose status as mass or interpersonal media is indistinct.

This affirms what was apparent on close inspection of the typology: Electronic media do form a new cluster of media. The most important characteristics of these media is their ability to bridge the gap between mass media and interpersonal media. It is apparent in the results, moreover, that this centrality also seems to hold for media richness, although in the context of an existing continuum which computer media supplement, often providing higher levels of speed and interactiveness than media with equivalent richness.

3D view of User Descriptions of Media

This general trend is clearly apparent in the figure above, which shows, in three dimensions (media richness is seen as depth, with the richer media forced to the foreground as larger objects). The computer media, electronic mail and computer conferencing, stand alone in this middle ground.

The insufficiency of mediator based characteristics

Another important result, from the perspective of the present theory of media as process, is that media often are not fully defined by their mediators and mediator-based characteristics. A medium with fairly high bandwidth can be very lean on content. A medium with a fairly small bandwidth can, by contrast, be moderately rich in content. This variation, whether viewed in the programming of television or the detailed content of publications, is related to the way people use media; to the subject of the next chapter -- media practices.