|
|
ABOUT DoIT
Research1999 Student Computing Survey - TrendsJune 3, 1999 written by Marketing Communications BackgroundThe 1999 DoIT Student Computing Survey was designed to provide DoIT guidance on how well it is doing and how it might better serve student information technology needs. This year, the Student Information Technology Initiative (SITI) committee authorized a methodological comparison between the use of traditional mail questionnaires versus new web-based questionnaires. The justification for this study was the potential cost savings available using in-house DoIT resources for surveying. The current cost to DoIT for conducting a web survey is between a quarter and a third of the cost incurred by contracting outside sources to conduct a mail survey. MethodologyThe Wisconsin Survey Research Laboratory (WSRL) picked two random samples of 1,000 UW-Madison students from Registrar records. Past samples of students indicate that about 4% do not have a preferred email address. Thus we could draw two valid, representative samples from the student population, one of mailing addresses and one of email addresses, making this an excellent population for which to draw conclusions about the differences and/or similarities between mail and online research. A general questionnaire was developed to fulfill SITI requirements and to meet the research needs of various DoIT groups. Using Active Server Pages software, DoIT’s Business and Financial Applications group put the questionnaire online. This website was programmed so that skip-and-fill sequences were automatic and appeared seamless to the respondent. WSRL printed a booklet-size questionnaire as has been used in previous years survey research. Both the mail and online questionnaires were virtually identical regarding question wording and sequence. A cover letter was sent to both groups. For the mail survey, this was sent along with the questionnaire in each mailing. For the online research the cover letter took the form of an email. This email explained the purpose of the survey and how to connect to the survey web site. The web site’s Universal Resource Locator (URL) was listed in the email. A copy of the email with the referring URL can be found here: http://magicbus.doit.wisc.edu/studentsurvey/ While all respondents were referred to the same web site, each individual URL address had a unique random number attached to it. This ensured that only those respondents with the correct random number or "password" could access the site. This also operated as a check against respondents filling out the survey more than once since they were only given one "spot" that corresponded with their random number "password." An Access database recorded survey responses. One initial mailing/emailing was sent to all respondents in both groups on February 18. Two follow-up emails were sent to non-respondents on March 15 and April 5. Every effort was made to make sure both groups received the survey materials at approximately the same time. The online survey garnered 270 completed questionnaires out of a non-redundant, cleaned (13 email addresses were judged to be undeliverable) sample of 987. This corresponds to a 27.3% response rate and a margin of error of +/- 6.0%. The mail survey garnered 415 completed questionnaires out of a total sample of 990 (10 questionnaires were undeliverable). This corresponds to a 41.9% response rate and a margin of error of +/- 4.8%. Overall, with the data of both modalities combined (n=685), this survey has a margin of error of +/- 3.7%. The data reported below are from the merged dataset, containing the responses of students in both groups. The distribution of respondents' year in school was compared with the UW's actual distribution of year in school. A weight which took into account the slight disparity between the two distributions was computed and applied to the data for both modalities and to the combined dataset. An Overview: Survey HighlightsThe 1999 Student Survey found a jump in overall computer ownership from 65% last year to 77% this year. Much of the increase can be attributed to undergraduates, with freshmen having the highest percentage of computer ownership (79%) among undergraduates. Windows 95/98 continues to increase its market share among students, with Mac OS dropping from 22% to 18% this year, and Windows 95/98 increasing substantially from 59% to 81% this year. Given a choice, more students prefer Mac (22%) than use it (18%), while less students prefer Windows 95/98 (67%) than actually use it (81%). Notably, many more students prefer to use Windows NT (23%) than use it (8%). Overall, modes of access to the Internet remained very stable, with the most popular method of connecting being the WiscWorld dial-in modem pool with 63% using this. There was a large drop in the percentage of students using InfoLabs or other computer labs to connect to the Internet, with 46% using it this year versus 60% last year. Other measures indicate that general use of InfoLabs is decreasing along with an increase in computer ownership (see Questions 9 and 22). Commercial ISP use leveled off at 13%. When asked why they used a commercial ISP, the most popular responses were that they already had this service where they live (family already had it, it’s a long distance charge to dial-in to the WiscWorld modem pool, etc.). About 27% of ISP users cited that they used these services because of busy signals with the modem pool, or that the connection to their ISP was faster. Connection time to the Internet almost doubled this year from 8 hours to 14 hours. Much of this increase was attributable to about 26% of student who connect to the Internet for 15 or more hours per week. When these individuals are controlled for, we see a much more graduated increase in connection time the past three years (see Question 5). These power users differed significantly from their moderate counterparts on a number of behaviors. There were slight increases in awareness across the board for most general DoIT services. The largest increases occurred for free Student Peer Computer Training and TechNews. While many services’ awareness increased, there were decreases in use for many of these same services. The largest drops occurred for InfoLabs and ARCH labs. Use of TechNews more than doubled from 8% to 20%. Among WiscWorld services, Eudora use dropped from 93% to 81%. Even though use was generally down, satisfaction ratings were generally up, some dramatically, although this could partly be a function of the small number of users rating these services. Interest in computer training continued to trend downward in 1999, with 45% expressing interest. Internet tools and web site development were the first and third most popular training areas, respectively. Use of spreadsheets and presentation graphics programs increased and this is reflected in their second and fourth ranking, respectively. The top barrier to training continued to be that students are too busy with other things (72%). More than a third (36%) said they were not aware of it. When asked to allocate $100 across a number of computing services, the top three areas were continued unlimited dial-in access, online course registration, and additional computers in computer labs. There was a drop in the percentage of students regularly using an InfoLab (from 47% to 39%). Checking email continued to be the most used activity while in the InfoLabs. Far fewer computer owners regularly use an InfoLab than non-owners (see Question 22). Available Trend DataFrequency RunsSection 1 - Computer and Internet UseThis section presents the frequency counts in percentages for the questions asked on the survey. Each question appears in the same sequence as it did on the actual survey.
Overall computer ownership increased by about 12% this year. Interestingly, freshmen had the highest computer ownership of all but graduate students (see Figure 1). A pattern has emerged during the past few years--shown in Figure 2--in which the growth in computer ownership levels off every other year, followed by a spurt in ownership. If this pattern follows next year, we can expect computer ownership to level off once again. At some point overall ownership will not be able to increase anymore (ceiling effect).
Figure 1: Computer ownership by class.
Figure 2: Overall, undergraduate and graduate computer ownership.
[as a percentage of computer owners] Linux appears to be gaining popularity among students (it was not an offered response option last year). Of other particular note in this table is the substantial drop in use of Windows 3.x and a corresponding jump in the use of Windows 95/98.
[as a percentage of computer owners] There were large drops in preference for Windows 3.x as well as Macintosh. Also, far more students prefer to use Windows NT (23%) than actually use it (8%) on the computers they own.
Modes of access to the Internet remained steady with the exception of InfoLabs or other computer labs, which experienced a sizable drop is use for accessing the Internet.
We asked Commercial ISP users why they were using these services. A plurality of the reasons offered were that they already had the service--for a variety of reasons--where they currently lived. Twenty-seven percent said they were using ISP’s because of busy signals with the WiscWorld modem pool, and that these services had faster connections. Thirteen percent volunteered that the cost was expensive or that they had unlimited access with their ISP’s, which is confusing considering that the WiscWorld modem pool is free and offers virtually unlimited access. Unlike the 1999 UW Online Computing Survey, there was a paucity of respondents who wished to keep their private use separate from their academic use.
In 1999, respondent spent an average 14.2 (25.4 s.d.) hours per week connected to the Internet. In 1998, students reported spending an average of 7.7 hours on the Internet. In 1997, student reported spending an average 6.5 hours per week connected to the WiscWorld dial-in modem pool. As with the past two years, a close look at the numbers reveals that the increase in average connection time per week is due to a small group of users. Mathematical averages are sensitive to extreme values and a small percentage of people using the Internet for relatively extreme lengths of time can inflate average Internet use. In 1999, 26% of Internet users spent more than 15 hours per week on the Internet. When these extreme users are taken into account (or held constant), average Internet use in 1999 was 5.4 hours. The proportion of extreme users has been growing at an almost exponential rate during the last three years. It could be that our rather arbitrary designation of extreme use (15 hours + per week) is no longer a good delineation of extreme use as more people spend more time online. Figure 3: Internet use overall and without power users
Figure 4: Percentage of students connected to the Internet 15 or more hours per week.
As with faculty and staff power users, student power users differed from their more moderate counterparts on a number of dimensions:
Somewhat less than half of students have made at least one purchase over the Internet in the past year, and these individuals were likely to have made more than just one purchase.
Section 2 - Awareness and Use of Computing Services
There were slight increases in awareness across the board for most general DoIT services. With the exception of FTP/Telnet, awareness of most of the main components of the WiscWorld suite has remained quite stable over the past three years. Changes in awareness of 10 percentage points or more are shaded.
While awareness of most general DoIT services increased slightly, here we see the exact opposite; most general DoIT services experienced slight to moderate declines in use during 1999. Only one service, TechNews, bucked this trend and increased its usage substantially. Among WiscWorld services there was a rather large drop in the use of Eudora, with the rest of WiscWorld components remaining relatively stable in usage. As with awareness, changes of 10 percentage points or more are highlighted.
Only respondents who indicated they used a service were allowed to rate their satisfaction with it. Interpretation of the following ratings should be tempered by the fact that the number of respondents providing ratings for each of the services below varies widely. Satisfaction ratings followed a pattern similar to awareness, with most general DoIT services posting impressive increases in satisfaction. WiscWorld ratings remained positive, and for the most part, stable. The largest increase in satisfaction among these services was FTP/Telnet.
Overall satisfaction remains very high, with 89% of students "very satisfied" or "satisfied," the same figure as in 1998. Figure 5: Overall satisfaction with computing resources, 1997 through 1999.
[valid %]
The following table shows a substantial drop in the percentage of students ratings of the availability of computing resources at UW-Madison. However, this could be an effect of a change in the response options. This year, we included a five-point scale with a neutral position (neither good nor poor), and changed the scale itself. Last year’s scale is in parentheses in the table below. While this methodological change can explain away some of the ratings drop, it is unlikely to account for the whole of the drop.
Word processors, Internet tools, and email continue to be the top three software used on campus. Use of spreadsheets and presentation programs increased substantially over the past two years. Use of specific software assigned in class also increased from last year, but has varied widely since 1997.
Figure 6: Interest in training from 1994 through 1999.
This year, instead of providing response categories for respondents this question was left open-ended. Not surprisingly, the results are different from prior years' findings. The table below presents the answers provided by students in rank order. The last column in the table shows the percentage of respondents volunteering interest in each training area. Last year's rank for each service is provided in parentheses. Not surprisingly, interest in most areas was muted in comparison with 1997 and 1998. Yet last year's top ten areas are well represented in this year's top ten. Desire for training in spreadsheets and presentation graphics seems appropriate considering the increase in their use documented above. There were several general or vague areas respondents expressed interest in, such as "general use of computer," "software," and "troubleshooting."
The top two barriers to receiving training continue to be that students are too busy, and that they are not aware of training.
Section 3 - Other Services
Continued unlimited dial-in access remains a very popular service, along with the addition of computers in computer labs. On-line course registration was allocated substantially more money this year, indicating an increase in its perceived importance.
This year, fewer respondents said they regularly use an InfoLab. When broken out by computer ownership, we find that those who regularly use an InfoLab are less likely to own a computer (62%) than those who do not use an InfoLab regularly (87%). A similar relationship held last year with 56% of InfoLab users owning computers compared with 74% of non-users owning computers. Thus, an increase in computer ownership might be leading to a decline in the use of InfoLabs.
[average rank] Email is by far the most popular use of the InfoLabs. Also, computer owners are less likely than non-owners to use email and more likely to use printing facilities when in the InfoLabs.
This is an excellent example of following up an information campaign with a measure of success. Since posters were DoIT’s main publicity vehicle for this campaign, it is reassuring that they were the most visible source of information on the "Rules of the Road" campaign.
Section 4 - About Yourself
The second and fifth items have been reverse-coded so that they are on the same scale polarity as the affirmative questions. This series of questions was included as a measure of respondent "need for cognition," and as a guide in how to differentially appeal to different types of customers.
Overall, there were significant differences between web and mail respondents regarding where they live. Web respondents were more likely than mail respondents to live in on- or off-campus residence halls or dormitories
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
