The Turfgrass Information File (TGIF):
Database Characteristics and Content Evaluation

Carrie Preston

August 2000

The turfgrass sector comprises a large number of businesses, organizations, educational programs and research programs related to the use of grass in lawns, landscapes, golf courses, athletic fields, parks, and other non-agricultural contexts. Turf-related education programs currently exist at over 100 colleges and universities in North America, with over 30 related graduate education/research programs (Golf Course Superintendents Association of America, 1997). However, turfgrass researchers and professionals may have difficulty finding bibliographic access to published information in their field. Turfgrass literature, being unrelated to food production, may be outside the scope of bibliographic databases focused on agriculture or may be processed on a low-priority basis. Turfgrass research and practice is also heavily dependent on a literature of professional, trade, and university publications, many distributed on a local or regional basis, which are not peer-reviewed and so do not appear in scientific journal databases, while being too specialized to appear in general-interest nonscientific databases.

The Turfgrass Information File (TGIF) is a bibliographic database produced by the Turfgrass Information Center at Michigan State University. Created in 1984, TGIF is solely dedicated to indexing turfgrass-related literature, turfgrass being broadly defined as any grass cultivated in a non-agricultural context (Cookingham, 1999; Turfgrass Information Center, 2000a, 2000b). TGIF is intended to provide turfgrass researchers and professionals with a mode of access to bibliographic information in their field, by providing a large proportion of unique content, bibliographic information on turf-focused publications not indexed by other online databases. TGIF is also assumed to provide for more efficient searching of turf-related topics by reducing the need to eliminate turf-irrelevant content from search result sets.

Bibliographic databases can be evaluated using a number of criteria related to software, hardware, and content or "dataware" (Jacso, 1992). Even when evaluation is limited to database content, a number of factors may be taken into account, including general database scope, composition, source coverage, time period coverage; the range of serials indexed by the database; and the consistency and completeness of coverage of individual serials (Jacso, 1997, 1998). This report focuses on the content of TGIF, both in itself and in comparison to the general agricultural databases that are the main alternative for users seeking access to current, turfgrass-related bibliographic information. The general characteristics of TGIF will be summarized and its content evaluated on a variety of dimensions. TGIF will be also compared to three major agricultural databases: AGRICOLA, produced by the U.S. National Agricultural Library; AGRIS, produced by the Food and Agriculture Organization of the United Nations; and CAB ABSTRACTS, produced by CAB International in the United Kingdom (Thomas, 1990). TGIF will be compared to the agricultural databases in terms of utility for turf-related subject searching. The proportion of unique versus overlapping content between TGIF and each of the major agricultural databases will be examined.

As of June 2000, TGIF contained over 65,000 bibliographic records, of which about 60,000 represent articles from print serials. The remaining records represent a range of other materials including books and book chapters, theses and dissertations, booklets and brochures, university agricultural extension service fact sheets, plant cultivar registrations, audiovisual materials, and electronic materials. Electronic materials indexed in TGIF include over 3,000 World Wide Web documents, of which 90% may be accessed by direct links from their TGIF records when the records are viewed in the Turfgrass Information Center's Web interface. In addition to documents available in full text on the Web, full text articles from certain print serials are currently being added to TGIF; as of June 2000 these totaled about 500 records. Material formats indexed by TGIF are summarized in Table 1.

Serials indexed by TGIF are published by a variety of organizations for scientific, professional, and popular audiences. Serials indexed by TGIF include peer-reviewed journals; scientific research report sources such as Agronomy Abstracts and university turfgrass department research reports; national and international-level professional publications such as USGA Green Section Record, Golf Course Management, and International Turfgrass Bulletin; local professional newsletters; proceedings of local, national and international turf-related conferences; and popular magazines such as Golf Digest. TGIF currently includes records for articles drawn from over 1800 serials. As of August 2000, serials regularly scanned by TGIF included 129 scientific publications, 159 professional and trade publications, and 25 other publications. Over 50% of TGIF records represent content from professional and trade magazines, newsletters and conference proceedings; these publications are widely utilized within the turfgrass profession and most are not found in other indexes and databases. Major types of publications indexed by TGIF are summarized in Tables 2 and 3.

Serials are selectively indexed according to subject relevance; most or all articles from turfgrass-focused professional and trade publications are indexed, while only turf-related articles are selected from broader agricultural journals. Book reviews and extremely brief "news item" type articles are typically not indexed. Efforts are made to restrict items published in more than one serial to a single TGIF record containing information on all known appearances of the item.

TGIF is primarily focused on indexing current materials, with ongoing secondary efforts to index older materials. Over 90% of the materials indexed by TGIF were published in or after 1970, and 60% were published in or after 1990. However, the database does include about 4,000 records for materials published before 1970, including all United States Golf Association Green Section serials published since 1921. Time period coverage of TGIF is summarized in Table 4.

Currently, TGIF is composed primarily of English-language materials, with a limited number of non-English materials represented, particularly articles from the German professional journal Rasen. Over 70% of the publications regularly indexed by TGIF have a North American geographic focus. However, TGIF does index a selection of professional and trade publications from the United Kingdom, continental Europe, Australia and New Zealand and a limited number from Asia, Africa and South America, as well as internationally focused peer-reviewed journals and conference proceedings. Language and geographic coverage of TGIF are summarized in Tables 5 and 6.

Items listed in TGIF are indexed using the Turfgrass Information Center's Turfgrass Thesaurus, a highly specific controlled vocabulary focused on turfgrass science. About 70% percent of existing records also include abstracts (provided either by the publisher or by the Turfgrass Information Center), full text within TGIF records, or full text on separate web pages directly linked to their TGIF records, and all records being added to the database on an ongoing database are provided with abstracts. TGIF is thus designed to facilitate both keyword (title, abstract and full text) and controlled vocabulary searching.

Method

General search. In an attempt to roughly estimate the number of turf-related articles in the agricultural databases, a search for the keywords turf* or lawn* or golf was performed in each database. (The search term grass was not used in order to avoid confusion with materials on forage grasses, as the term turf(grass) tends to be preferred by authors writing about grasses used in non-agricultural contexts.) This search was intended as a very rough estimate using the means available to researchers in turfgrass science who would be unable to individually examine large numbers of articles.

Subject searches. TGIF and the three agricultural databases were also compared using five turf-related subject searches. In addition to TGIF, the Turfgrass Information Center produces Hot Turf TOPICs, a series of prepared subject bibliographies on turf-related topics of current interest. The numbers of online user requests for each subject bibliography were tabulated and the five most popular subjects selected as test searches, since these were topics of demonstrated interest among real searchers interested in turfgrass science. Keyword searches were composed in an attempt to isolate subject-relevant items in each of the four databases. Attempts were made to keep the searches used in TGIF and those used in the agricultural databases as similar as possible while attempting to eliminate non-turf-related materials from the agricultural database result sets. The subjects and search syntax used in the test searches are listed in Table 7.

Results of each search were evaluated for relevance to the original subject, using the titles, abstracts and index terms provided by the databases. Items that appeared to mention the subject only in passing and those that were completely irrelevant to turf contexts (e.g. articles about marine algae rather than algae found on turfgrass) were classified as irrelevant. Uncertain cases (e.g., those in the agricultural databases where abstracts were not provided) were decided in favor of relevance. No clear cases of duplicate records within the same result set were found; procedures for dealing with duplicate records (e.g. Hood & Wilson, 1999) were therefore unnecessary. Search precision was measured as the ratio of relevant results to total results from each keyword search (Lancaster & Warner, 1993). Overlap between relevant result sets from TGIF and each of the other databases was also determined, defining overlap as follows (Gluck, 1990):

percent overlap	=	number of documents in the intersection of the two result sets	X	100
		number of documents in the union of the two result sets

Indexing of Major Turf-Related Publications. Databases were also compared for their indexing of major turf-related publications, particularly professional and trade publications published by national and international turf organizations. Databases were searched by publication title (for the agricultural databases) or TGIF serial code (for TGIF) for records drawn from 16 major publications.

All TGIF searches were performed using the Turfgrass Information Center's Power Search web-based search interface. All AGRICOLA, AGRIS and CAB searches were performed using Silver Platter's WebSPIRS web-based search interface as made available to the public at Michigan State University.

Results and Discussion

Results of rough general searches in each of the three agricultural databases are presented in Table 8. Though these searches are not meant to be compared directly to the full range of items available through TGIF, the results of the searches demonstrate the relatively small number of articles available to searchers who attempt to limit their agricultural database search results to turf-related items.

Subject searches. As shown in Table 9, TGIF tended to produce considerably larger result sets than the agricultural databases when searching for turf-related topics, while maintaining high precision. TGIF search precision was comparatively high for all searches except the dollar spot disease search ("dollar spot" is a phrase not commonly found outside the turfgrass literature and so produced high precision in all databases). The algae and effluent water searches produced relatively low precision in the agricultural databases due to common non-turf-related uses of these terms (e.g., use of "turf" and lawn" for growths of marine algae and "effluent" for the outflow of water that occurs during various laboratory procedures). The biological control search tended to produce low precision due to variation in usage of the term (e.g., for "natural" pest control methods other than the use of living organisms) and a general tendency to return results focused primarily on chemical, rather than biological, pest control. Overall, TGIF allowed better isolation of relevant, turf-related items, even when using keyword searches rather than TGIF's turf-based controlled vocabulary.

As shown in Tables 10-12, overlap between TGIF and the major agricultural databases was low for all searches, even taking into account the relatively large result sets produced by TGIF. Items commonly returned by both the agricultural databases and TGIF included scientific journal articles and, for AGRICOLA and AGRIS only, USGA Green Section Record articles. Items commonly returned by the agricultural databases only included non-U.S. publications (particularly for AGRIS and CAB) and agricultural extension service publications from U.S. universities (particularly for AGRICOLA). Items commonly returned by TGIF alone included articles from professional and trade publications, many of which are indexed only by TGIF among the four databases.

Indexing of Major Turf-Related Publications. As shown in Table 13, TGIF has considerably greater coverage of the sixteen major professional and trade publications than the major agricultural databases. Of the sixteen publications studied, only four had significant coverage in any of the agricultural databases.

The Turfgrass Information File (TGIF) appears to offer bibliographic access to a variety of turfgrass-related literature not indexed by other major databases. For researchers interested specifically in turf-related literature, TGIF offers advantages over general agricultural databases both in the amount of information available and in relevance of typical search results. A major advantage of TGIF include its coverage of professional and trade journals, conference proceedings, and university turfgrass research reports, all of which are widely utilized by turfgrass researchers and professionals and largely ignored by agricultural and general-interest databases. TGIF's sole focus on turf-related literature also allows for more precise searching while requiring less effort to produce focused, relevant search results.

One weakness of TGIF in comparison to the major agricultural databases is its relatively limited coverage of non-U.S. publications. However, the number of turf-related, non-U.S. publications found in AGRIS and CAB is so small that international researchers might want to supplement information found in those databases with U.S.-focused information found in TGIF. TGIF also lists relatively few agricultural extension publications from U.S. universities when compared to AGRICOLA. The Turfgrass Information Center does index agricultural extension publications but currently lacks a systematic mechanism for receiving and processing these publications; with greater availability of these publications over the World Wide Web, increasing numbers of these publications may be included in TGIF. Furthermore, the total number of results for turf-related searches in AGRICOLA was not large enough to suggest complete or near-complete coverage of agricultural extension publications in this database. Even taking into account some degree of advantage in the area of agricultural extension publications, the total number of turf-related publications unique to AGRICOLA is small enough that searchers might wish to use it as a supplementary information source when information provided by TGIF is inadequate.

Overall, TGIF provides greater access to turfgrass science literature than other available resources and should be of value to turfgrass researchers and professionals.

Cookingham, P. O. (1999). The Turfgrass Information File: Monitoring the turf science literature. In American Society of Agronomy, Crop Science Society of America, & Soil Science Society of America, 1999 Annual Meeting Abstracts. Madison, WI: American Society of Agronomy, 1999.

Gluck, M. (1990). A review of journal coverage overlap with an extension to the definition of overlap. Journal of the American Society for Information Science, 41(1), 43-60.

Golf Course Superintendents Association of America (1997). GCSAA college guide to the golf course management profession. Lawrence, KS: Golf Course Superintendents Association of America.

Hood, W. W., & Wilson, C. S. (1999). The distribution of bibliographic records in databases using different counting methods for duplicate records. Scientometrics, 46(3), 473-486.

Jacso, P. (1992). CD-ROM software, dataware, and hardware: Evaluation, selection, and installation. Englewood, CO: Libraries Unlimited.

Jacso, P. (1997). Content evaluation of databases. Annual Review of Information Science and Technology, 32, 231-267.

Jacso, P. (1998). Analyzing the journal coverage of abstracting/indexing databases at variable aggregate and analytic levels. Library & Information Science Research, 20(2), 133-151.

Lancaster, F. W., & Warner, A. J. (1993). Information retrieval today. Arlington, VA: Information Resources Press.

Thomas, S. E. (1990). Bibliographic control and agriculture. Library Trends, 38(3), 542-561.

Turfgrass Information Center (2000a). Turfgrass Information File (TGIF). http://www.lib.msu.edu/tgif/tgifda.htm (viewed August 22, 2000).

Turfgrass Information Center (2000b). Turfgrass Information File (TGIF) database specifications. http://www.lib.msu.edu/tgif/tgifspecs.htm (viewed August 22, 2000).