Characteristics of the IBMPC User

IBMPC User Demographics

User's of IBMPC include a broad spectrum of IBM employees, but that spectrum may not be entirely representative, either of IBM or the world at large. The two surveys measured a broad range of IBMPC user demographics, including:

These demographics are explored, in part, in methodology, the development of IBMPC, and other chapters. This appendix provides some additional demographic information.

General Demographics

Several questions in the surveys concerned general demographics, starting with question 1, which in both 1986 and 1988 asked "How old are you?" There were five possible answers, including '0-25', '26-35', '36-45', '46-59', '60+'. In both surveys, the largest share of respondents belonged to the '26-35' and '36-45' age groups. 66% of respondents belong to one or the other of these groups in 1986. 70% of respondents belong to one or the other of these groups in 1988.

Although the average age of respondents creeps up slightly between the two surveys, with the median shifting from the 26-35 year old group to the 36-45 year old group, a t-test comparison shows no significant difference, indicating that IBMPC has roughly the same age profile in 1988 that it had in 1986. These results are hardly surprising, and although I don't have company wide statistics to compare them with, it would be surprising if company-wide age statistics differed markedly from those observed among participants in the IBMPC computer conferencing facility.

Three of these general demographic questions explore language skills. One of these, typing skill, was only included in the 1986 survey. Most respondents to this question indicated that they were average to quite experienced typists, and only 10% considered themselves below average typists. Another, fluency in the English language, did not significantly change (although the average did decline; t=1.57, .2 < p < .1 ) between 1986 and 1988. In 1986, 83% of respondents considered themselves extremely fluent in English, and another 16% considered themselves somewhat fluent. In 1988, 75% or respondents considered themselves extremely fluent in English, with 24% considering themselves somewhat fluent. A third skill, or at least inclination, was perceived verbalness. There is effectively no change in the mean for this measure between the two surveys. Most respondents (53% in 1986; 49% in 1988) consider themselves quite verbal. Roughly 32% of respondents in both surveys consider themselves about average.

Together, these three measures paint a picture of the typical IBMPC participant as an average to excellent typist, as might be expected of experienced computer users; generally highly fluent in the English language, as might be expected, even in an international computer company where English is often used in international technical discussions.

Only the last of these language measures might be considered surprising. Hiltz and Turoff (1978, p.101) suggest that people "who dislike face-to-face interaction, or who are not skilled at interaction rituals," may find computer conferencing an important new communication outlet. The results of the two surveys of IBMPC participants does not seem to support this position. Although it cannot be asserted that the aggressively verbal don't find a home in computer conferencing, it seems clear that people who aren't likely to speak up in the course of normal interaction may be similarly unlikely to speak up on a computer conferencing facility.

More surprising, however, are the results of the second question in both the 1986 and 1988 surveys, which queried the gender of respondent. In 1986, only 3 of 122 respondents said they were female (116 claimed to be male; 3 respondents did not answer the question). In 1988, only 6 of 177 respondents claimed to be female. (169 claimed to be mail; 2 respondents did not answer the question). These results are clearly at variance with expectations, as a large percentage of IBM employees are female; certainly a much larger percentage than the roughly 3% female participation in IBMPC that is indicated in these surveys (there is no significant difference between the results in 1986 and 1988), are female.

One hesitates to interpret this statistic, as almost any interpretation might be considered sexist. Still, it is both surprising and, in an age where women are striving for equality in all parts of the workforce, disturbing. While there is no evidence that IBMPC participation is prerequisite to success within IBM, there is evidence, presented in the chapter on the benefits of computer conferencing, that participation in IBMPC provides tools for success, and may, indeed, contribute to success. Regardless of such evidence, moreover, it seems clear that IBMPC, as by far the largest lateral communication vehicle within the company, is a valuable place to meet people. It seems equally clear that, because a computer conference reveals a persons words without presenting the distractions of their appearance, that it is a good place to present self free of gender based stereotypes and expectations.

One would like to better understand why women in IBM are not, in general, participating in the IBMPC computer conferencing facility. With adequate participation in the survey, one might find clues that would help in understanding it. As things stand, however, the survey participation is so low (3 women in 1986 and 6 in 1988) that reliable male/female statistical comparisons cannot be made. Still one wonders. Is IBMPC simply a conferencing facility for technical discussions that fall outside the work scope and interests of women in IBM? Does voluntary participation in a conferencing facility that is not a part of ones job requirements conflict with a gender based view of what work is? Is there an element of computer conferencing interaction that women find threatening? The data only raise questions. Neither the surveys nor the authors participation in IBMPC provides answers to such questions.

Demographics related to computer usage

Consider that the IBMPC Computer Conferencing facility is:

It seems reasonable to expect that the people who use the facility would have considerable experience with, and feel fairly comfortable using, computers. Another set of questions directly addresses this expectation by asking respondents about their computer use. The results, which with two interesting exceptions, do not change between 1986 and 1988, confirm this general expectation. Specifically:

These responses paint a picture of the typical IBMPC participant as a fairly knowledgeable computer user who likes using computers and has been doing so for a long time. Indeed, many IBMPC participants feel sufficient commitment to use computers at home as well as at work, and a full 30% claim to have accessed the IBMPC conferencing facility from home.

Hours spent computing

A fuller picture of this commitment is seen in a series of questions, posed in both 1986 and 1988, which ask users to estimate the number of hours they spend at the computer doing various tasks. Each of these questions was phrased as a multiple choice question, with the answers incremented logarithmically (following evidence from the ratio scaling literature that the accuracy of ratio answers is logarithmic). Hence the possible answers were "less than 1 hour", "1-5 hours", "6-16 hours", "17-35 hours", and "greater than 35 hours". Two questions looked at work related and personal use of computers. Six looked at various applications. It was possible, in both surveys, to rank the applications into three groups based on the significance of differences between means. Each group represents a set in which estimates of use are not significantly different from the others. The results:

  • Over two-thirds of the respondents to both surveys indicated that they used a computer for over 17 hours a week in the course of doing their IBM work. The median (48% in both surveys) was 17-35 hours. Another 30% (in both surveys) used the computer for more than 35 hours each week in the process of doing IBM work. A modest upward drift in the number of hours computers were used for work was not statistically significant.
  • Roughly three-quarters of the respondents to both surveys indicated that they made personal (non-work related) use of a computer for at least 1 hour a week. Although the median shifts from 1-5 hours in 1986 to 6-16 hours in 1988, the shift did not prove statistically significant. At the extreme, roughly 9% of respondents to both surveys indicated that they made personal use of a computer for more than 17 hours a week.
  • In both surveys, 65% of respondents claim to use their computer for calculation less than an hour a week, and at least 89% calculate with the computer less than 5 hours a week. In both surveys calculations stand alone as the least used application among IBMPC participants, significantly less used than any other application.
  • The most used application among IBMPC participants in both surveys is programming. The estimate for hours spent programming rises between 1986 and 1988, but not significantly so. In 1986, 43% of IBMPC participants spend six or more hours a week programming. In 1988, 50% of IBMPC participants spend six or more hours a week programming. But while programming may be the number one application on IBMPC, technical forums concerned with programming are only one of its attractions. Indeed, roughly a third of IBMPC participants in both surveys program less than one hour a week. Most of these people probably never program at all.
  • In 1986 Word Processing was the second most important application among IBMPC participants, with estimated usage levels that were not significantly different than programming. By the 1988 survey, the relative (but non-significant) rise in programming combined with a small (but non-significant) drop in word processing estimates to drop word processing to the second group, significantly less important to IBMPC users than programming. Some of this drop may be due to the rise of IBMTEXT, an increasingly popular IBM conference facility concerned with word processing and writing tools. Still, both surveys indicate that roughly a third of all IBMPC participants spent more than six hours a week writing documents with word processing software.
  • The median estimate (46% of respondents) for use of other non-communications applications of computers (not programming, word processing, calculation, or communication) is one to five hours. Roughly three-quarters of respondents in both surveys claim to use such applications (which presumably include games, graphics, and database applications) less than 6 hours a week. There was no significant change in this measure, which ranked in the second group of applications in both surveys, between 1986 and 1988.
  • In both surveys, 57% of the survey respondents indicated that they spent between 1 and 5 hours a week on computer conferencing. This strong median masks, however, a significant shift (t=3.38, df=298; p<.0005) in computer conferencing usage between 1986 and 1988. 29% of the 1986 respondents spent less than an hour a week engaged in computer conferencing (14% spent 6 or more hours a week on conferencing). 27% of the 1988 respondents spent 6 or more hours a week on conferencing (only 15% spent less than an hour a week conferencing). In both surveys, only 2% of respondents spent 17 or more hours a week engaged in conferencing. This shift is not sufficient to move conferencing out of its position in the second tier of applications, but its raw rank (based strictly on the mean, without regard to the significance of its difference with other application estimates, moves from fifth in 1986 to third in 1988.
  • The same general pattern is seen more strongly for non-conference electronic communications. The median, once again, is 1 to 5 hours (56% in 1986 and 50% in 1988). Respondents who communicate electronically for less than one hour a week decline, however, from 23% in 1986 to 9% in 1988 while those spending six or more hours a week engaged in non-conferencing electronic communication expands from 21% in 1986 to 39% in 1988. The resulting shift is strongly significant (t=4.72; df=298; p<.00001). As a result, the rank of non-conference electronic communication shifts as well. In 1986 this application was one of the second tier of applications. In 1988, this application was a strong member of the first tier, only marginally (and non-significantly) behind programming.

    The last of these results is, perhaps, the most interesting, as they indicate substantial growth, not only in the use of computer conferencing, but of computer mediated communication in general. Non-conference communication is now one of the leading applications of computers among IBMPC computer conference participants, and computer conferencing is clearly at least as important an application as word processing.