(2) With the goal of generating a testable hypothesis for the place of architectural research in this array, the present study investigates the sources cited in nine journals of architectural research Ref.1 published in the year 1990.
(3) These nine "source journals" are: Architecture et Comportement, Architectural Science Review, Assemblage, Built Environment, Building and Environment, Daidalos, Design Methods and Theories, the Journal of Architectural Education, and the Journal of Architectural and Planning Research. Each is published either quarterly or three times a year. The population comprises a convenience sample drawn from among the titles likely to be found in architecture-school libraries and self-evidently covering the chief departments (empirical research, theoretical and critical investigation, historical inquiry, pedagogy) and foci (building use, building technology, building design, building environments, architectural pedagogy) of architectural research.
METHODOLOGY
(4) The nine source journals together published a total of 247 articles in the year 1990. These articles included 3560 footnoted citations to various antecedent works. A data base was created in which each cited work was classified into four publication types (standard monograph; periodical; reference or gray literature; non-book material). This data base identified each cited work by author name(s), title, and publication date.
FINDINGS FOR THE SOURCE JOURNALS AS A GROUP
(5) Basic analysis of the data reveals that 54.5 percent of the citations were to standard monographs and 29.6 percent were to periodicals. This tracks with the percentages for these two categories found in studies of arts and humanities research. Such studies have found a range of 49.9 to 71.4 percent of citations to monographs, 25.3 to 45.7 percent of citations to periodicals, and 5.2 to 16.7 percent of citations to other sources. Ref.2
CHART1
|
(6) Analysis of the data base for age of the cited material at the time of citation shows that 91.2 percent of the standard monographs, 92.2 percent of the periodicals, and 95.3 percent of the reference and gray literature cited had been published in the past forty years (i.e., since 1950). These statistics do not parallel the findings of other studies of arts and humanities research literatures. Studies of citation patterns that have examined the rate of obsolescence of sources generally agree that information in these fields "ages" more slowly (i.e., remains useful for scholarly purposes longer) than that in the natural sciences or social sciences. Ref.3 The data presented in Chart 1 indicate the nine architecture journals studied here as a group use newer material than would normally be expected of arts and humanities research journals.
FINDINGS FOR THE SOURCE JOURNALS INDIVIDUALLY
(7) Analysis at the source-journal level reveals differences among them which may help to explain why as a group they do not clearly adhere to a single paradigm.
CHART 2
 |
(8) Chart 2 shows the distribution by type of works cited across the source journals. By inspection, it is evident there are marked differences from journal to journal. The use of a pair-wise test of proportions objectively identifies statistically significant differences between the journals. This analysis was done at the 95 percent confidence level using the Student-t statistic.
(9) Authors in Daidalos (a German journal of architectural history and theory) cited a significantly higher percentage of standard monographs than authors in any of the other journals. Likewise, they cited a statistically lower percentage of periodicals than all but one (Architecture & Comportement) of the other journals. In fact, only 10 percent of the citations in Daidalos were to periodicals. This extraordinarily low use of periodicals marks the 1990 volume of this journal as unusual among architectural research journals. Daidalos falls at the upper limits of the arts-and-humanities paradigm for type characterization.
(10) Authors in Assemblage (an American journal of architectural history and theory) used a significantly larger percentage of standard monographs than authors of all but three of the other journals (Daidalos, Journal of Architectural Education, or Design Methods and Theories). Yet Assemblage still appears to conform rather faithfully to the arts-and-humanities paradigm.
(11) Authors in Building and Environment cited a significantly lower percentage of standard monographs than authors in any of the other journals except Architectural Science Review. Authors in this journal also cited a significantly higher percentage of reference materials and gray literature than did authors contributing to Design Methods and Theories, Journal of Architectural Education, or the Journal of Architectural and Planning Research. The more even distribution of references to monographic, periodical, and reference literature by authors in Building & Environment reveals the essentially social-science nature of this journal.
(12) Similar to those publishing in Building and Environment, Architectural Science Review authors also cite materials in a pattern much like that typical of social-science researchers. Monograph citations in this journal's sample were significantly fewer than those in five of the other journals (Architecture & Comportement, Assemblage, Daidalos, Design Methods and Theories, and the Journal of Architectural Education). The more even distribution of citations among source types matches closely the pattern characteristic of social-science research.
(13) Of the other journals in the sample, only the citations found in the Journal of Architectural and Planning Research exhibits any significant differences in type characteristics. This journal cited significantly fewer reference and gray-literature titles than three of the other journals (Architecture & Comportement, Built Environment, and Building and Environment). However, this journal still exhibits the characteristics of the arts-and-humanities paradigm for type distribution, as indeed do most of the journals in the sample.
CHART 3
 |
(14) Chart 3 presents the distribution by age of standard monographs across the nine source journals. Two of these journals, Built Environment and Building and Environment, use significantly newer monographs than the other journals in the sample. From the point of view of the age of their cited material, these two journals appear to obey the social science paradigm. In fact, of 512 citations, only 17 (3.3 percent) were to items published prior to 1950.
(15) Two of the journals studied, Assemblage and Design Methods and Theories, cited significantly fewer monographs published in the 1980s than did the other seven. There is, however, a special consideration in the case of Design Methods and Theories. During the year sampled, Design Methods and Theories printed a translated Polish book as a series of articles. This book had been produced earlier and this fact may have skewed the results of the present analysis. Yet in spite of their differences from the other journals in this sample for the most recently published materials cited, these two journals do not exhibit significant differences in older materials. In fact, in only one of these journals, Daidalos, was more than ten percent of the material cited by authors published prior to the twentieth century. If the nine journals studied were arts and humanities research journals, all would be expected to have about ten percent of their citations refer to material from early years. Therefore, only Assemblage and Daidalos come close to conforming to the arts-and-humanities paradigm with respect to the age of cited monographs.
CHART 4
 |
(16) Chart 4 presents the distribution by age of periodicals cited in the nine source journals. Authors in two of the journals, Assemblage and Design Methods and Theories, used significantly fewer periodicals published in the 1980s than did those writing in five of the other journals (Architecture & Comportement, Built Environment, Building and Environment, the Journal of Architectural Education, and the Journal of Architectural and Planning Research). (The possible skewing of the data for Design Methods and Theories must be borne in mind.) Again, however, only Assemblage and Daidalos come close to conforming to the arts-and-humanities paradigm for age of cited periodicals.
(17) Only the Journal of Architectural Education uses significantly newer material than other journals (Architectural Science Review, Assemblage, Design Methods and Theories, and the Journal of Architectural and Planning Research). In the light of studies of citation patterns in social-science literature, this fact suggests that the Journal of Architectural Education obeys the social-science paradigm more strongly than some of the other journals studied.
CHART 5
 |
(18) Statistical comparisons of the age of reference and gray literature cited by authors are difficult because of the small sample sizes. Notable, however, is the relative age of this material used by authors in Daidalos and Assemblage. Perhaps this is a further indication of conformity to the arts-and-humanities paradigm by researchers writing for those two journals.
CONCLUSIONS
(19) Because the sample upon which this study is based was a convenience sample rather than a random sample, the conclusions drawn may not be generalizable to the entire journal literature of architectural research. Nevertheless, they have implications for the management of library services to architectural researchers significant enough to warrant further investigation.
(20) Judging from the patterns of citation found in the nine journals included in this study, the literature of architectural research does not fall neatly into any one of the paradigmatic patterns characteristic of scholarly writing in the sciences, the social sciences, or the arts and humanities. Rather, this literature seems to be a collection of more or less separate literatures, each adhering more or less closely to one of those established patterns. In a sense, the literature of architectural research lies at the intersection of the social-science and arts-and-humanities paradigms of documented scholarship.
(21) Assemblage and Daidalos are the only two of the nine journals studied here whose citation patterns could be said to follow the arts-and-humanities paradigm of higher use of monographs and lesser use of periodicals along with a higher percentage of older material (although in this sample the higher percentage is not statistically significant). The citation patterns of Architectural Science Review and Building and Environment clearly follow the social-science paradigm (even distribution of type citations across categories and the use of quite recent sources of information).
(22) It could be said that the Journal of Architectural Education joins these two by virtue of the strong significance of modern emphasis in periodicals cited by its authors. The evidence for the proper model for the remaining journals in this sample is ambivalent. They display the characteristics for the arts-and-humanities model for citation type while displaying the citation age characteristics of the social-science model.
(23) The chief lesson of these conclusions for managers of libraries supporting architectural research is this: large retrospective collections are not important to the architectural researchers who wrote for the 1990 issues of the nine journals included in this study. If their writings turn out to be representative of current research in architecture, over 90 percent of the demand for architectural research materials could be met by a library collection whose oldest holdings date back only to about 1950, some forty years ago.
(24) For librarians and library-collection developers, the broadness of the literature cited by authors in the nine journals here considered has other implications. To truly support such research a library must be capable of providing both important architectural information and also information from other fields with (apparently) only tenuous or peripheral connections to architecture. Once the entire list of cited materials is examined, it becomes clear that only a large, full-service academic or public research library could support such research as was done to create the 247 articles found in these nine journals. This finding does not bode well for the concept of branch libraries in academic settings (where much architectural research is now carried out). Such libraries -- even if large -- may be too specialized for the contemporary architectural researcher.
REFERENCES
Ref. 1: For an overview of "architectural research" see: James C. Snyder, ed. Architectural Research (New York: Van Nostrand Reinhold, 1984), esp. pp. 1-8. See also: Mary Vance, Architectural Research (Architecture Series: Bibliography A- 974) (Monticello, IL: Vance Bibliographies, 1983).
Ref. 2: Among the most important of such studies are: Wesley Clark Simonton, "Characteristics of the Research Literature of the Fine Arts During the Period 1948-1957" (Ph.D. dissertation, University of Illinois, 1960); Lois Bebout, Donald Davis, Jr., and Donald Oehlerts, "User Studies in the Humanities," RQ 15 (Fall 1975): 40-44; and Diane M. Nelson, "Methods of Citation Analysis in the Fine Arts," Special Libraries 68 (November 1977): 390-395. See also Jean-Pierre V. M. Herubel, "Materials Used in Historical Scholarship: A Limited Citation Analysis of the Journal of Garden History," Collection Management 14 (1991): 155-162.
Ref. 3: The paradigmatic patterns are best set forth comparatively in: George A. Barnett, Edward L. Fink, and Mary Beth Eckert, "The Diffusion of Academic Information: A Mathematical Model of Citations in the Sciences, Social Sciences, and Arts and Humanities," paper presented at the annual meeting of the International communication Association, Chicago, IL, May 1986 (ERIC Document 275328).
Material appearing in Architronic may be distributed freely by electronic or any other means, providing that any such distribution is without charge (unless for purposes of cost recovery by interlibrary loan services) and that Architronic is acknowledged as the source. However, no article may be reprinted in any publication without the explicit written permission of the author(s). This statement must accompany all distributions of Architronic, whether complete or partial.