E-Journal User Study
Report of First Survey
March 2002
Footnotes
1
The Stanford E-Journal User Study includes quantitative research (several online surveys and a data mining effort), qualitative interviews (ethnographic interviews with users of e-journals), and experts' workshops.
2
A journal is defined here as a peer-reviewed serial publication for scientists or medical professionals. An e-journal is a journal available online (e.g., through the Internet). A printed journal is a journal available in a printed format and delivered to subscribers by mail.
3
We created a variable defining a person as a "high collegial communicator" if the person had communicated with colleagues on scientific matters through "conferences/seminars/workshops" within the last month, through "e-mail" that day or one day before, by "phone calls" that day or one day before, and at "personal meetings" that day or one day before filling out the survey. We did not use this variable in the regression equations because it had inconsistent and sometimes conflicting results and because causality was difficult to determine; instead, we tested for correlation using one-way ANOVA analysis.
4
The two variables are highly correlated (.83). Ninety-one percent of the 2,234 "daily average" users also said they had used e-journals for article access "yesterday or today" (9% said "within the last week," and less than 1% said any of the other categories). Forty-four percent of the 4,643 respondents who said they had used e-journals for article access "yesterday or today" also said they had been daily users on average over the past year, and 52% said they were weekly users; only 4% said they were monthly users, and less than 1% said they "seldom" used e-journals for article access. Similarly, 52% of the 5,757 "weekly average" users also said they had used e-journals for article access "within the last week," whereas 42% said "yesterday or today." Seventy-three percent of the 4,103 respondents who said they had used e-journals for article access "within the last week" also said they had been weekly users on average over the journals for article access "within the last week" also said they had been weekly users on average over the past year, with 20% saying they were monthly users, 5% saying they were daily users, and 2% saying they "seldom" used e-journals for article access. Most recent usage thus appears to be a reasonable proxy for usage frequency and is the most accurate indicator we have in this survey.
5
The equation for this regression was as follows:
P[EJU] = F (DEMO1, INTERNET, SUBS, PAPER1, FACILITY, ACCESS)
where EJU is frequency of e-journal use, indicated by how recently the respondent used e-journals (measured as: 1=yesterday or today; 2=within the last week; 3=within the last month; 4=longer than a month ago; 5=never ); DEMO1 is a set of demographic characteristics — AGE, MD, BIOLOGY, MALE, and USA-CAN — which represent respondents' age, whether a respondent is a medical doctor, whether a respondent's field of research is biology, whether a respondent is male, and whether a respondent resides in either the United States or Canada or in another country (refer to Appendix II, Table 1); INTERNET measures weekly hours spent on the Internet (excluding use of e-mail) — an indicator of Internet familiarity; SUBS measures the number of personal journal subscriptions; FACILITY measures whether or not a respondent has a PC, printer, and Internet access at both home and work — an indicator of available basic IT resources; ACCESS measures scientific journal access through institutions; PAPER1 measures the number of papers submitted in the past year — an indicator of respondents' research productivity; PAPER2 measures the number of papers accepted/published in the past year — indicators of respondents' research productivity. Appendix II, Table 2, shows the EJU question and responses by category. Appendix II, Table 8, shows the results of the regression analysis.
6
Ceteris paribus means "other things being equal." The term is used in economic analysis when the analyst wants to focus on explaining the effect of changes in one (independent) variable on changes in another (dependent) variable without having to worry about the possible offsetting effects of still other independent variables on the dependent variable under examination. See Appendix II for full equations for all analyses. All regression results in this report may be assumed to be controlling for all other variables in their respective equations.
7
There is some support for this interpretation. Numbers of biology articles published in referred journals have decreased compared to numbers of articles in other life sciences (such as medical research) last decades. The proportion of biology to all life science R&D funding in U.S. academic institutions decreased by 2% from 1993 to 1997, whereas R&D funding in the medical sciences increased by 2% during that period (NSF Science and Engineering Indicators 2000). This indicates that biology may be becoming more competitive in terms of grants than are other fields (such as medical sciences) in the life sciences.
8
We refer to the number of personal subscriptions in our tables as "SUBS," to access to institutional subscriptions (for most journals) as "ACCESS," and to our index of access to PCs, printers, and the Internet as "FACILITY."
9
We ran a probit analysis that included having DSL at home as an independent variable predicting e-journal usage frequency and found a statistically positive impact on e-journal usage frequency. We encountered a serious multicollinearity problem caused by the inclusion of the DSL factor in the analysis, however, precluding us from using DSL as an independent variable in the final analyses. The other results reported in this section thus do not control for having DSL at home.
10
PAPER1 in our tables in the Appendix II.1.
11
See Part II-A for the topline statistics on usage frequency, and note in the Methodology section in Appendix I.B. how usage frequency was included in the equations.
12
The U.S. share in the world's article output in major scientific engineering journals decreased by approximately 5% and Canada's share by .5% between 1989 and 1997; the U.S. share of the world's biology article output decreased by approximately 7%, health and professional by 6%, clinical medicine by 2.5%, and biomedical research by less than 1%.
13
See http://ejust.stanford.edu for background and details. A grant from the Andrew W. Mellon Foundation supports this project.
14
This respondent is working with a definition of e-journals as nonpeer-reviewed literature, which is at odds with the definition used by this study. We define e-journals as peer-reviewed journals available online, whether or not they are also available in conventional, printed form. The confusion about what the term means was not uncommon among those interviewed for the qualitative research and among several scholars who sent unsolicited comments about the survey, which suggests that survey respondents may have been using a range of definitions.
Feedback:
Please share your reactions to the Survey Findings by filling out a simple form. Your feedback is a valuable component of the Electronic Journal User Study.
Last updated: 03-29-02
| 
