Canadian traditions and directions in information systems research

Grant, Rebecca

As a field of study matures, interest in its evolution and future increases. Individuals begin to study the choice of research topics and methods. Frameworks for research emerge, authors call for overarching theory to guide research, and articles appear analyzing the work done to date. In the information systems (IS) field, this interest surfaced most clearly in 1980, at the First International Conference on Information Systems (ICIS), where leaders in the field presented papers and led discussions on the evolution of IS research and the directions it should take in the future.

Since the 1980 ICIS, we have seen conferences, colloquia, and journals that concentrated on research methodology and paradigms. In 1984, the Harvard Business School sponsored a Research Colloquium on Information Systems to consider research history and needs in the field. Nineteen eighty-five brought the publication of Research Methods in Information Systems (Mumford, 1985). More recently, ICIS 1989 and 1990 included paper tracks on research methodology and direction, the IFIP WG 8.2 1990 Working Conference chose “The Information Systems Research Arena of the 90s” as its theme, the 1992 Hawaii International Conference on Systems Sciences (HICSS-25) included sessions on methodology and measurement issues, and the ICIS 1993 included a track on relevant IS research methods.

The development of thought in a particular field can be traced in two ways. First, surveying and analyzing published research demonstrate historic topics of interest, paradigms, and issues. There are numerous examples of such reviews in IS, including work by Alavi et al. (1989), Culnan (1986), Elam et al. (1986), and Swanson and Ramiller (1993). These studies examine the field’s publications and categorize its historical evolution. Such work tells us where we’ve come from and where we are now.

One shortcoming in studying publications is the inevitable lag between the research project and publication of its results. From January, 1990, through December, 1993, for example, the average time from submission of an IS-related paper to its publication in Management Science ranged from 15 to 44 months, with a mean lag of 27.2 months. In Information Systems Research, time to acceptance during the same period ranged from 8 to 34 months, with a mean of 22.4 months. Conference proceedings fare better, but are rarely included in publication studies. Thus, studies that look at published papers run the risk of describing the field as it existed two or more years earlier. Such studies also often use only a fragment of published research, such as that appearing in top journals or by the best-known researchers. As a result, they report on only a part of the field and may ignore significant research by newcomers.

A second shortcoming is the publication bias (Light & Pillemer, 1984) or file drawer problem (Rosenthal, 1978) arising in studies of publications. Some research fails to demonstrate significant effects, and thus may not be considered worth publishing. In other cases, the research may not explore a “new” or “exciting” issue, and thus be less apt to find an outlet. As a result, counts or analyses of publications can underestimate the relative amount of research in a particular area of IS.

There are alternatives to publication analyses. One is to track the topics being studied that have yet to lead to published papers. Notable among sources for such research are doctoral student research and works in progress by established researchers. By studying the topics investigated in these works, we can anticipate where future research is heading, and what the favored topics, issues, and paradigms may become. This approach looks forward from the present.

Teng and Galletta (1991) surveyed 397 MIS researchers to assay the quality of historical work and research-in-progress, They identified three vital issues that would help demonstrate IS’s progress as a discipline, or movement out of a preparadigmatic state. Those issues included the emergence of a cumulative tradition advocated by Keen (1980), identification of dependent variables, and identification of appropriate reference disciplines. They concluded that the field was still highly fragmented, with significant differences in the research interests of faculty in graduate programs versus those in predominantly undergraduate programs. The bulk of their analysis concentrated on issues of research quality and maturity, leading them to conclude that IS research quality was improving and would continue to do so.

This paper describes a survey study and report on the history of IS among Canadian faculty and researchers, as well as its future. It differs from prior work in five ways. First, it uses the Barki, Rivard, and Talbot (1988, 1993) Information Systems Research Literature (ISRL) taxonomy as a tool for individuals to report their own interests. Other studies have relied on external raters to categorize the work. Second, it concentrates on work in progress, rather than published research. Thus, it focuses on the current rather than historic state of the field. Third, the study uses a larger base of informants to identify research in IS and captures the work of doctoral students completing their degrees as recently as 1990. Fourth, this work considers the genealogy or intellectual lineage of Canadian IS educators and researchers. Finally, it examines Canadian research interests separate from those of researchers in other countries. The Canadian IS community is a close-knit group, as evidenced by its visibility and activity at the Administrative Sciences Association of Canada Annual Conference and the extensive publication record of many of its researchers. While this group does not isolate itself from the U.S. research community (and, indeed, maintains a high profile there as well), its numbers are small compared to that larger domain. The Canadian IS community comprises approximately 100 individuals, out of more than 2000 names appearing on the MISRC (1990) database. Thus, trends among Canadian researchers disappear when aggregated in studies that span North America. Never before have such studies looked specifically at the work going on in Canada.

This study identifies crucial areas of current research interest, as well as research areas that receive little attention. By considering these research streams, individuals could direct their efforts toward areas that may emerge as distinctive competencies for Canadian IS faculty.

The paper has three parts. We first present the survey research design and methodology. Second, we discuss the results of the survey in three areas: (1) the major categories in which current IS researchers conducted their doctoral research and the contribution of reference disciplines to IS research over time; (2) the breakdown of current research interests and works-in-progress according to the ISRL; and (3) significant families’ of researchers. The paper concludes by looking at the implications of the study for those trying to understand the field and its research future.

Survey Design and Methodology

Sample

This study used the Management Information Systems Research Center (MISRC) database of IS faculty. This database, updated weekly, draws on a number of sources to produce an extensive directory of individuals in IS departments or teaching IS courses. Foremost among those sources are the annual registrations for the International Conference on Information Systems. As of November, 1990, the database included the names and addresses of 2361 individuals who classified themselves as being in IS.

From this database, we identified 101 IS faculty at Canadian universities. Twelve individuals, known not to have a doctorate, were removed from the list. In December, 1990, we sent each of the remaining 89 faculty a survey (shown in Appendix A), instructions on how to complete it, and a stamped reply envelope.(Appendix A omitted) The survey was precoded with a four-digit number tying the survey to the individual’s name and employer in our database. We needed to know the name of each respondent in order to develop the genealogies discussed later in this paper.

Table 1 shows the breakdown of data used in our analyses.(Table 1 omitted) Three waves of survey distribution produced a total of 52 responses, for a 58.4% response rate. These 52 surveys (Group A) provided data for analyses of graduation trends, dissertation trends, lineage, and research interests. We used three secondary sources to gather a limited amount of information on non-respondents. First, dissertation advisors provided secondary data about five non-respondents (Group B). This secondary data was combined with data drawn from other secondary sources, to allow us to include the five faculty members in analyses of graduation trends, dissertations, and lineage. Second, the 1989 MISRC directory (MISRC, 1989) provided the year a faculty member’s degree was granted, his or her major and the school granting the degree. Third, UMI’s Dissertation Abstracts International (UMI, 1986 and 1991) also provides this information for faculty members who had not completed their dissertations in time for the data to appear in the 1989 MISRC directory. These two sources gave us basic information on 23 additional non-respondents (Group C). We could then include those 23 faculty members in our analyses of graduation trends. Thus, our discussion of graduations, majors, and granting schools includes 80 of the 89 Canadian faculty in our original population. Table 2 provides a breakdown of respondents by the school at which they were employed.(Table 2 omitted)

The Survey

The one-page survey comprised three sections. The first asked the respondent to supply the title of his/her dissertation, the year completed, the school granting the degree, the name of his/her dissertation advisor, the name of an equally or more significant mentor (if any), and up to two ISRL codes classifying the dissertation. This data was used to analyze trends in doctoral research over time and, working backward, to develop lineage profiles. The second section asked for current research interests, in terms of a brief topic description and an ISRL code. Each respondent could provide up to three such areas of interest. The third section asked for a list of doctoral students whom the respondent had supervised. It provided space for the student’s name, dissertation title, ISRL code(s), and year completed. The form allowed for up to four advisees. Respondents were instructed to attach another sheet with the requested data, if they had advised more than four students.

In prior studies (e.g., Alavi et al., 1989; Culnan & Swanson, 1986; Ives & Olson, 1984), expert raters classified research. Such raters were used even when those doing the research supplied a description of it. We deliberately asked respondents to categorize their own work. Many topics are very broad and might fit under several ISRL codes. For example, if someone reported that they were interested in DSS, would their specific interest be in DSS adoption (ISRL code EM05), interface design (code FB04), justification (code EJO1), or DSS applications (code HA03)? Without a lengthy description of the work, it is unlikely that different individuals assigning a code to the topic would all make the same assignment. Thus, the reliability of the coding would be suspect. It could be a particular problem if the coders’ interpretation differed from that of the researcher.

The Barki, Rivard, and Talbot (1988, 1993) taxonomy is a five-level hierarchy, with each level providing more detailed classifications. We used the first two levels of the taxonomy (see Appendix B).(Appendix B omitted) The first or top level (Barki et al., 1988, 1993) consists of nine broad classes, such as Reference Disciplines, Technological Environment, IS Management, and Information Systems. Each top-level class is represented by a single character code. The second level breaks the nine classes into 57 subcategories, each represented by a two-character code. We limited the survey to these two levels in order to simplify the instructions, reduce the effort required to classify a research topic and limit the time needed to complete the survey. With only 89 names on the mailing list, it was also likely that responses would be scattered across a range of third-level categories and we would have to combine them into second-level groups for meaningful analysis.

Results and Discussion

The survey provided two sources for data about doctoral dissertations. First, individuals who returned a survey reported their own dissertation title, advisor’s name, and ISRL code(s) for the dissertation. Second, dissertation advisors provided the name, title, and code(s) of the research done by their doctoral students. In some cases, therefore, we could use the advisor’s report to add a non-respondent to the base of doctoral research data. We derived information about five non-respondents from surveys completed by their dissertation advisors.

Doctoral Trends–Canadian Roots

Analysis of doctoral trends is based on primary sources (surveys completed by the faculty members themselves) and secondary sources described above, for an n of 80. Discussion of dissertation topics is based on self-reports of 52 faculty (Group A) and the reports of advisors for five faculty (Group B), producing an n of 57.

Graduates over time. The survey and secondary data demonstrated a dramatic increase in the number of IS faculty graduating since 1974 (see Figure 1).(Figure 1 omitted) Sixty-three of the 80 individuals (78.8%) received their degrees between 1974 and 1990. That increase coincided with growth in IS programs (and hence hiring requirements) and the appearance of the first Ph.D. programs in IS.

The continued attraction of students to IS graduate programs also indicates the outside community’s needs and interests. Continuing growth of the field is driven in part by practitioner interest in our undergraduate and graduate students, as well as interest in our research results. This continued interest is a further indication of the emergence of IS as a separate and distinct field of study.

The Canadian IS community is, like that in the U.S., relatively young. Predicting retirement patterns is difficult unless one knows the precise age distribution of the faculty population. In their analysis of academic supply and demand, Jarvenpaa et al. (1991) assumed an academic career span of 30-35 years. Even using the conservative figure of 30 years, one would project only ten retirements in Canada in the next ten years. Thus, new faces in the academic community will probably fill positions generated in one of three ways: growth in faculty at existing programs, replacement of faculty who leave Canada or academia, and death. Estimates forecast that supply will outpace projected demand from these three sources throughout North America for years to come. At the same time, the hiring bulge of the late 1970s and early 1980s suggests that attrition will produce another period of heavy demand in the long term, as faculty hired in those decades retire.

Major programs and dissertation topics. The two most popular Ph.D. majors among Canadian faculty were IS and management science/operations research (Figure 2).(Figure 2 omitted) These accounted for 28 and 15 of the 80 faculty respectively. An additional five people did a joint accounting/IS major. Together, these three majors reflect 60% of the Canadian IS faculty considered in this study. Other faculty are scattered across a number of majors, most notably the reference disciplines of computer science and management.

Teng and Galletta (1991), in a recent study capturing dissertation information, reported that 39% of their respondents (drawn from the an earlier version of the MISRC database) wrote dissertations in MIS. By contrast, 41.5% of Canadian faculty in our study majored or joint majored in IS. This is the result of the high proportion of faculty (33, or 41.5%) teaching at schools with IS majors or joint majors at the doctoral level. Teng and Galletta reported that “the MIS PhD-granting institutions are more likely to hire faculty members with MIS as a PhD thesis area” (p. 55). Our study confirmed this trend in Canada (see Table 3).(Table 3 omitted)

Our survey produced dissertation titles and up to two ISRL categories for each of 57 faculty. Classifying dissertations according to ISRL codes offers some advantages over simply looking at the names of the graduates’ major programs. First, the codes allow finer distinctions of doctoral research topics than do names of major programs. Second, there are overlaps among topics considered appropriate in different major programs. For example, an IS major might do a dissertation on artificial intelligence (AL), but so might a student majoring in computer science. Similarly, students majoring in management science and IS might both classify their DSS dissertations as decision theory (AC), IS characteristics (HD), or software (CB). Looking at topics according to their ISRL classifications makes some of these common interests more explicit.

The 57 dissertations were scattered across 37 second level ISRL codes, with only four codes accounting for five or more dissertation topics. Together these four codes (shown in Table 4) represented 42.1% of the dissertations captured by the survey.(Table 4 omitted) These categories included FC (Development Methods and Tools), AK (Management Science), FA (Development Strategies), and GA (Organizational Use of IS). Thirty-eight dissertations fell into 32 other IS-related categories. Five dissertations were classified as ZZ; that is, they did not fall into any specific reference discipline or IS category. Examples of such ZZ dissertations included those in pure math, physics, and chemistry. The diversity of dissertation research topics demonstrates the wide variety of research backgrounds within Canada.

Because of the wide dispersion of categories, analysis of trends over time would be meaningless at the second level. However, Figure 3 shows how doctoral background in different major categories has changed over time.(Figure 3 omitted) For example, the number of candidates hired from outside IS or its reference disciplines has declined dramatically, while hires of graduates studying IS Usage (code G) and IS Management (code E) have increased in recent years. The number of researchers with dissertations in reference disciplines (codes AA-AZ) has varied over time. However, they remain a significant influence in the research community. All told, 13 dissertations (22.8%) fell exclusively into one or more references disciplines. Another 11 (19.2%) were classified in both a reference discipline and one of the explicitly IS categories. The most prevalent reference discipline was management science, followed by behavioral science and psychology. Since the Canadian faculty base is relatively stable, the significant percentage of individuals with backgrounds in IS’s reference disciplines will continue to influence research trends for many years.

Current Research Interests

While dissertations tell us where our training and early interests lie, they may bear little relation to current research interests. This is increasingly true as one’s tenure in the field increases. Respondents were asked to describe up to three current research interests and classify them as primary, secondary, and tertiary interests. Fifty respondents supplied at least one area of research interest. We compared the three current interests to dissertation topics for the 50 individuals for whom we had data. (Analyses are based on primary data only–that is, the 52 surveys returned by Group A respondents.)

Breaking faculty into three groups (degrees granted 1954-1970, 1971-1980, and 1981-1990) demonstrated significant patterns. As shown in Table 5, no one who graduated before 1970 was currently doing research in his or her dissertation area.(Table 5 omitted) Nine individuals who graduated between 1971 and 1980 (50% of the graduates of that period) were pursuing research in their dissertation area, as a primary, secondary or tertiary interest. Finally, 26 post-1980 graduates (93%) were researching in their dissertation area, generally as their primary or secondary research interest. Thus it is important to consider current interests separately from dissertation interests.

The current interests of Canadian researchers diverge from findings of other studies. Alavi et al.’s (1989) analysis of published articles from 1968 to 1988 indicated IS Management (E), Information Systems (H) and IS Development and Operations (F) as the three most prevalent major ISRL codes among articles published in seven key journals. As shown in Table 6, IS Management was the most popular topic among Canadian researchers, chosen by 29 (58%) of the researchers as one of their three interest areas.(Table 6 omitted) Development/Operations was a close second (compared to its more distant third place position in Alavi et al.’s study). It was chosen by 28 (56%) researchers. However, Reference Disciplines (A) placed third among Canadian choices, with 24 individuals (48%) indicating that one of their three areas of current interest fell into a reference discipline. Among reference disciplines, the most popular areas were Artificial Intelligence (AL), Decision Theory (AC), and Management Science (AK). Artificial Intelligence is an evolving area of interest in the U.S. as well, with 22% of Teng and Galletta’s respondents choosing it as an area of current research. Alavi et al.’s second code, H (Information Systems) ranked a distant fourth among current Canadian research interests. One other marked difference between the current Canadian interests and Alavi et al.’s findings appeared with codes I(IS Education and Research) and C (Technical Environment). These two codes were reversed in the Canadian rankings.

When we look at the second level codes (in Table 7), the differences between the current interests and the publication history of the field become more apparent (Table 6).(Table 7 omitted) Most notably, the single most popular code for current research interests among Canadian faculty was FC–Development Methods and Tools. Fifteen individuals (30%) chose this classification as one of their three areas of interest; eight chose it as their primary area of interest. This code did not appear in Alavi et al.’s second level breakdown of the top three ISRL codes. Teng and Galletta reported that systems development, which may include both Strategies or Approaches (code FA) and Methods/Tools (FC), was mentioned by 19% of their North American respondents as a current area of interest. Programming/Software Engineering (one subset of Methods and Tools) was mentioned by 7% of their researchers. Thus, there appears to be considerably more widespread interest in development methods and tools among Canadian researchers. We interpret this to reflect interest in CASE, model management and group support system tools, which has only recently begun to appear in IS research journals.

The second most popular areas of current research were Organizational Use of IS (GA) and IS Evaluation (EJ). Organizational Use of IS did not appear on Teng and Galletta’s list of the thirteen most often-mentioned areas of interest (Table 8), although their End-User Computing is a subcategory of GA in the ISRL scheme. Furthermore, IS Evaluation was well down their list, cited by only 7% of respondents, compared to 12.3% of Canadians. By contrast, the Alavi et al. study ranked IS Evaluation as the second most prevalent code, primarily in the areas of user satisfaction and cost-benefit analysis.

Researchers in IS must contend with a technological environment that changes rapidly. The Canadian emphasis on management issues, specifically organizational use of IS and IS evaluation, may indicate an emphasis and recognition of the need to study issues that transcend the technology. This survey shows that, in emphasizing topics beyond the current technology fads, Canadian faculty are researching those issues Teng and Galletta’s (1991) respondents considered difficult to study.

Faculty in Canada are pursuing research in a broad range of areas. Thirty-nine of the 57 possible categories were mentioned as an area of interest by at least one respondent. Furthermore, there were few areas being explored by many individuals: only the top five codes were chosen by more than 10% of respondents. Radical changes in technology or targeted research funding (such as SSHRC’s strategic research grant themes in science and technology) for a specific area of emphasis may encourage more research toward particular research arenas. However, given the prediction of a relatively stable group of IS researchers in Canada for the next 10 to 15 years, the variety of research efforts among many different topics is likely to continue.

In this regard, Canada is much like the rest of the IS community. Teng and Galletta reported similar diversity in their survey results, with respondents choosing 38 different categories as areas of interest and two-thirds choosing areas that fell outside the top 13 (Table 8). Teng and Galletta, however, reported significant differences between the research interests of faculty at schools with graduate IS programs and those at undergraduate institutions. Respondents at graduate degree schools were more likely to be studying managerial issues, while those at undergraduates were looking at technical areas. This is not the case in Canada. Chi-square tests of research interests by graduate versus undergraduate teaching schools were not significant (alpha > 0.10).

Keen (1980) called on IS researchers to build a cumulative research tradition. Continued fragmentation of research interests and efforts, shown here among Canadian faculty and in Teng and Galletta among the U.S. faculty, would obstruct the progress of building cumulative IS theory. Researchers must collaborate or build on one another’s work in common areas of interest to advance knowledge in the field.

The survey demonstrated the diversity of interests within schools, with no school solely managerial or technical in its focus. For example, faculty from the University of Calgary listed IS evaluation, organizational use, and IS research among their interests; those from HEC included IS implementation, task characteristics and management science; and UWO faculty included IS management and technological environment among other interests.

Interests are also spread across the country. Researchers at UBC, University of Alberta and McMaster (among others) are looking at development methods/tools. Individuals at Dalhousie and UQAM (again, among others) are studying AI issues. The same lack of regional concentration applies to all of the leading areas of research. Even schools with an explicit focus or primary research area include faculty interests outside the focus area. There are advantages to the wide diversity across the country. It suggests that there are numerous opportunities to extend one’s research and build partnerships within the country and to tap funding sources that limit grants to Canadian participants. This is especially relevant in light of SSHRC’s new funding for research networks and research partnership development. In addition, faculty to add to one school’s research strength in a particular area may be readily available from other Canadian schools.

The Intellectual Lineage of the Canadian IS Community

The choice of doctoral program and dissertation advisor can be a major determinant of one’s perspective on research and the field of IS. The backgrounds of Canadian faculty are diverse. Table 9 indicates that slightly over one-half received their doctoral training in the U.S., one-third in Canada, and the remainder in Europe and the Middle East.(Table 9 omitted) Only five schools accounted for more than 5% each of the current, doctorally qualified faculty in Canada. Of these five, only the University of Waterloo and the University of British Columbia are Canadian.

The survey asked respondents to indicate their dissertation advisor and an equally (or more) significant mentor. Using information provided by graduates and their advisors (for the 57 Group A and Group B faculty members), we linked members of the IS community through a network of intellectual genealogies. For most individuals, this was a short genealogy. The U.S. data gathering is still in the early stages, so many grandparent and more distant relationships cannot be constructed with the data available. Moreover, the scattering of sources and recency of major IS programs suggests that few individuals can trace their IS roots back more than one or two generations.

Nonetheless, four distinct families emerged in this look at the history and influences on the community. Each of these families is headed by an individual (two, in the case of Minnesota) from a different major university. The most prolific of these families, from the view of populating the Canadian community, was the Minnesota family (Figure 4).(Figure 4 omitted) Headed by Gordon Davis and Gary Dickson, this family produced five of the 57 faculty analyzed in this portion of the study. The mentoring relationship between Gary Dickson and Izak Benbasat meant that the family also influenced another four Canadian faculty.

The second most prolific family was that of Waterloo, headed by David Conrath (Figure 5).(Figure 5 omitted) Waterloo provided more Canadian IS faculty than any other school, and Conrath was the advisor or mentor for five of the seven Waterloo graduates. This is particularly interesting because the University of Waterloo does not have an IS major in its Ph.D. program. Instead, candidates major in Management Science with a concentration in IS. Both the Davis/Dickson and the Conrath family lines were scattered across the Canadian IS community, although the faculty at Queen’s University was predominantly from the Davis/Dickson line.

The Harvard family ranked third (Figure 6).(Figure 6 omitted) Directly responsible for two Canadian faculty, James McKenney’s lineage also included five individuals through the mentoring relationship between Sid Huff and Michael Scott Morton. The McKenney family line was, not surprisingly, most influential on the University of Western Ontario’s faculty and graduates.

Finally, the MIT family produced two Canadian faculty (Figure 7).(Figure 7 omitted) While there are other advisors not represented in our figures who are responsible for two graduates, the Rockart line incorporated Ephraim McLean of UCLA. UCLA was, in turn, responsible for five Canadian faculty and thus the advising or mentoring lineage is undoubtedly broader than the survey responses would indicate. The Rockart line was most apparent among the faculty of Universite de Laval.

Ideas represented by key figures in Canadian research families, specifically Gordon Davis, Gary Dickson, Izak Benbasat, James McKenney and Malcolm Munro, were shown to support and define the intellectual structure of IS (Culnan, 1987). As thought passes from teacher or mentor to doctoral student, these lineages demonstrate a direct influence of the IS intellectual founders on the Canadian research faculty.(1)

Limitations

Our discussion is clearly limited by the extent to which respondents are representative of the Canadian research community as a whole. One factor limiting its generality might be the distribution of non-respondents. Notably, rates were low from Laval (25%) and the University of Waterloo (20%). The most obvious effect of the low rates would be in the discussion of research lineage. There is a strong link between the faculty at Laval and the doctoral program at UCLA, and while we can use secondary sources to establish that link, those sources do not indicate advisor/student or mentor relationships. Similarly, the University of Waterloo has produced many of the faculty currently teaching in Canadian schools. Its programs and faculty clearly play a significant role in the lineage of Canadian faculty.

A second limitation of the study is the relatively small population of IS faculty in Canada. Our discussion of research interests reports results from a sample of 52 individuals within a population of 89 Canadian faculty. Research interest area and doctoral dissertation area distributions in the Canadian milieu are more likely to be influenced by the topics of a few individuals than those of a larger population such as the U.S. Indeed, the scattered distribution of research interests among second level ISRL codes precluded a detailed analysis at that level. However, 58% of researchers indicated that IS management was one of their interest areas, and 56% chose development/operations, When viewed at this level, the numbers show an emphasis not likely to be influenced by a few researchers.

Another area of concern is the potential difficulty of using ISRL codes, particularly by nonexperts. Some individuals classified their research interests as UF, the shorthand used by Barki, Rivard, and Talbot to indicate “use for”. In most such cases, the surveys also included descriptions that paralleled the labels in the ISRL instructions, so we were able to recode those specific items. In others, individuals circled the code on the attached ISRL code description sheet, and we could directly transcribe their intended code.

Restricting choices to only the first two levels reduced the precision of classification. For some respondents, the lack of precision was a problem: Seeing the lower levels would have made their choices easier. While the survey instructions gave the reference for the full, multilevel taxonomy, individuals with questions may not have looked it up. We chose to limit the levels to reduce the length and complexity of classification. However, this seems to have increased the difficulty for some respondents. The ISRL has been widely tested and used, but rarely by anyone other than expert raters with immediate access to the complete taxonomy. It is our conclusion that instructions for the taxonomy must be explicit if it is to be used by novices; the taxonomy as it appears in the Appendix to Barki et al. (1988, 1993) must be supplemented by examples and expanded instructions for wide-scale use.

Finally, this research is based on the population of Canadian faculty in 1991. While relatively stable, this population does change: Faculty leave Canada to teach in the U.S., return to Canada from the U.S., or leave academia altogether. Resources prevented us from distributing another survey to capture movement in faculty while this article was in production. In the interest of using primary and secondary data of the same vintage, we have not supplemented our discussion of graduation or employment trends with more recent secondary data. Therefore, interpretation of these findings must acknowledge that some such changes have occurred since data collection.

Conclusion

The Canadian IS community is often merged into that of the U.S. when researchers study trends and areas of interest. Certainly there are more researchers in the U.S., and including the 89 doctorally qualified Canadian faculty in analyses does not distort visible U.S. patterns or trends. However, there is a strong community of IS researchers and faculty within Canada, which deserves to be considered on its own as well as in partnership with the U.S. There are significant areas of shared research interest between the two communities. There is also a shared intellectual heritage by virtue of Canadian faculty who received training in the U.S. and Canadian graduates who subsequently teach across the border. Nonetheless, this study identified some distinct elements to the research and heritage of the Canadian arena. These unique elements suggest that there are streams of research emerging in Canada that may become distinctive areas of competence unmatched in the U.S. These primary areas of Canadian research interest are development methods and tools, organizational use of IS, and IS evaluation. Individuals interested in these popular topics are spread across the country, rather than concentrated in a particular school or region. The widespread popularity of research in development methods/tools and artificial intelligence point to opportunities for research teams that cross university boundaries and researcher exchanges among Canadian schools.

References

Alavi, M., Carlson, P., & Brooke, G. (1989). The ecology of MIS research: A twenty year status review. Proceedings of the 10th Annual International Conference on Information Systems, Boston, MA, 363-375.

Barki, H., Rivard, S., & Talbot, J. (1989) An information systems keyword classification scheme. MIS Quarterly, 12, 299-322.

Barki, H., Rivard, S., & Talbot, J. (1993). A keyword classification scheme for IS research literature: An update. MIS Quarterly, 17, 209-226.

Culnan, M. (1986). The intellectual development of management information systems, 1972-1982: A co-citation analysis. Management Science, 32, 156-172.

Culnan, M. (1987). Mapping the intellectual structure of MIS, 1980-1985: A co-citation analysis. MIS Quarterly, 11, 341-353.

Culnan, M., & Swanson, E.B. (1986). Research in management information systems, 1980-1984: Points of work and reference. MIS Quarterly, 10, 289-301.

Elam, J.J., Huber, G.P., & Hurt, M.E. (1986). An examination of the DSS literature (1975-1985). In E.R. McLean & H.G. Sol (Eds.), Proceedings of the IFIP 8.3 working conference on decision support systems: A decade in perspective (pp. 1-17). Noordwijkerhout, The Netherlands. Amsterdam: Elsevier Science.

Ives, B., & Olson, M. H. (1984). User involvement and MIS success. Management Science, 30, 586-603.

Jarvenpaa, S.L., Ives, B., & Davis, G.B. (1991). Supply/demand of IS doctorates in the 1990s. Communications of the ACM, 34, 86-99.

Keen, P.G.W. (1980). MIS research: Reference disciplines and cumulative tradition. Proceedings of the First International Conference on Information Systems, Philadelphia, PA, 9-18.

Light, R.J., & Pillemer, D.B. (1984). Summing up: The science of reviewing research. Cambridge, MA: Harvard University Press.

MISRC (1991). MISRC/McGraw-Hill directory of management information systems faculty database (MIS Research Center). Minneapolis, MN: Carlson School of Management, University of Minnesota.

Mumford, E. (1985). Researching people problems: Some advice to a student. In E. Mumford et al. (Eds.), Proceedings of IFIP WG 8.2 Colloquium: Research methods in information systems. Manchester, England. Amsterdam: North-Holland.

Rosenthal, R. (1978). Combining results of independent studies. Psychological Bulletin, 85, 195-193.

Swanson, E.B., & Ramiller, N. C. (1993). Information systems research thematics: Submissions to a new journal, 1987-1992. Information Systems Research, 4, 299-330.

Teng, J., & Galletta, D. (1991). MIS research directions: A survey of researchers’ views. Data Base, 22, 53-62.

UMI (1986). Doctoral Abstracts International Online–ProQuest. Ann Arbor, MI: UMI.

UMI (1991). Doctoral Abstracts International Online–ProQuest. Ann Arbor, MI: UMI.

Note

1. We received survey responses from four individuals who earned their PhD. Degrees in Canada, but were teaching at U.S. Schools at the time of the study. In the interest of confining our study to faculty at Canadian schools, we did not include these responses in this analysis. The primary impact would have been to strengthen the Corath and McKenney lines: Higgins was the advisor and Huff the mentor for two of the respondents, while Goldstein was the advisor for the other two.

Table 8

Most-Mentioned Current Areas of Research Interest (from Teng & Galletta, 1991)

Research Area–Percent Choosing–32

Decision Support Systems–22

Artificial Intelligence–21

Database Management Systems–19

System Development–15

End-User Computing–10

Human Factors–9

Telecommunications–7

MIS Evaluation–7

Organizational Impact of MIS–7

MIS Education–7

MIS Management–7

Programming/Software Engineering–7

Other (25 items)–33

Note: Responses total more than 100% due to multiple responses.

The authors are indebted to the Management Information Systems Research Center of the University of Minnesota, and particularly Ms. Jan deGross, for providing the mailing lists used in this research. Address all correspondence to Rebecca Grant, School of Business, University of Victoria, Victoria, BC, Canada, V8W 3P1.

An analysis of its determinants among male and female teachers

EFFECT OF THE INTERACTIVE SMALL GROUP STUDY METHOD ON DEMOCRATIC ATTITUDES AND SELF-EXPRESSION ABILITY, THE