Dispersion of ICT-related subject terms in information and knowledge management publications: A Bradford analysis

The study investigates whether Bradford's law of dispersion can be applied to identify core ICT-related subject terms within the information and knowledge management (IKM) literature. Motivated by the growth of ICT and knowledge management research in LIS and the importance of content analyses for delineating subject concepts, the paper aims to determine core subject terms, understand their dispersion over time, and assess the applicability of Bradford’s law to subject terms rather than journals. This is important for clarifying evolving research foci in IKM and guiding indexing, retrieval, and collection practices.

Bradford’s law (Bradford, 1934) posits that when journals are ordered by decreasing productivity on a topic, they can be divided into a core and subsequent zones in the ratio 1:n:n²:n³. It has commonly been used to identify core journals and to test scatter (Diodato, 1994). Scholars have refined, interpreted, critiqued, and extended the law (e.g., Brookes, 1968; Bailón-Moreno et al., 2005; Hjørland and Nicolaisen, 2005; Yatsko, 2012). Recent applications include delineating core journals in various fields and analyzing citation distributions (e.g., Singh and Bebi, 2014; Desai et al., 2018; Wahid and Idrees, 2017; Yang et al., 2016). Extensions beyond journals include subject changes in citing vs. cited literature (Tsay, 2008), web “sitation” distributions (Faba-Pérez et al., 2003), and evaluation of book collections by subject headings (Girap et al., 2014), which reported approximate but not perfect Bradford-type distributions. Building on suggestions to apply the law to various data types (Yatsko, 2012), this study explores its use for identifying core ICT subject terms within IKM literature.

Data sources: EBSCO Discovery and EBSCOhost’s LISTA and LISS databases were used, focusing on LIS where KM is a sub-field. Searches were conducted in the Subject (SU) field for peer-reviewed scholarly articles published 1998–2017. Search strategy: Advanced search combining ICT and IKM terms. Exact query: (SU ("information technology" OR "information technologies" OR "information systems" OR "information communication technologies" OR "information and communication technologies" OR "information & communication technology" OR "information & communication technologies")) AND SU ("information management" OR "information resources management" OR "information services management" OR "knowledge management" OR "knowledge organization"). The search yielded 9479 articles. Data processing: Subject terms from each record were exported as text and analyzed with Bibexcel to generate term occurrence frequencies. Subject terms were ranked by frequency for four periods: 1998–2002, 2003–2007, 2008–2012, 2013–2017. Bradford analysis: Following Andrés (2009) and Singh and Bebi (2014), Bradford’s constant κ was computed for each period using κ = (e^γ Ym)/p with Euler’s constant γ = 0.5772, Ym = max frequency of the top subject term, and p = number of zones (set to 4 due to large T). Reported κ values: 1998–2002: 4.738191 (Ym=283); 2003–2007: 3.839334 (Ym=122); 2008–2012: 3.201688 (Ym=59); 2013–2017: 3.937762 (Ym=135). The core size r0 was computed by r0 = T(κ−1)/(κ^p−1), and subsequent zones by r1 = r0 κ¹, r2 = r0 κ², r3 = r0 κ³. Numbers of subject terms and associated article counts per zone were tabulated (Table 1). Core subject terms for each period were listed (Tables 2–5). Fit to Bradford’s law was assessed by: (1) comparing within-period multipliers r_n/r_{n−1} to κ; and (2) evaluating proportional patterns against 1:n:n²:n³.

• Dispersion and core sizes (Table 1):
  • 1998–2002: Core contained 7 subject terms associated with 776 articles; subsequent zones (terms/articles): 34/572, 160/728, 759/857.
  • 2003–2007: Core 36 terms with 1380 articles; zones: 139/1237, 532/1592, 2043/2160.
  • 2008–2012: Core 86 terms with 1902 articles; zones: 274/1898, 878/2255, 2811/2819.
  • 2013–2017: Core 39 terms with 959 articles; zones: 151/1295, 595/1728, 2344/2509.
• Core subject terms by period:
  • 1998–2002 (N=1143): Technology–Information Services (283; 24.76%), Information Technology (263; 23.01%), Information Society (58; 5.07%), Internet (53; 4.64%), Libraries (50; 4.37%), Library Science (39; 3.41%), Electronic Publications (30; 2.62%). Emphasis on IT in libraries and the information society, with internet and e-publications prominent.
  • 2003–2007 (N=2813; 30 of 36 shown): Focus shifted to management and commerce-related terms: Technology (122; 4.34%), High Technology (106; 3.77%), Management (83; 2.95%), Information Systems (75; 2.67%), Business Enterprises (72; 2.56%), Electronic Commerce (40; 1.42%), Medical Care (39; 1.39%), Industrial Management (33; 1.17%), Intellectual Capital (32; 1.14%), Public Administration (31; 1.10%), Information Technology (31; 1.10%), Decision Making (29; 1.03%), etc. Management Information Systems was the principal ICT-specific term.
  • 2008–2012 (N=3265; 30 of 86 shown): Top terms included Management Information Systems (52; 1.59%), Decision Making (49; 1.50%), Online Social Networks (48; 1.47%), Electronic Commerce (44; 1.35%), Business Enterprises (42; 1.29%), Social Networks (42; 1.29%), Industrial Management (39; 1.19%), High Technology (35; 1.07%), Corporate Culture (32; 0.98%), Technology (31; 0.95%), Data Security (28; 0.86%), Management (27; 0.83%), Organizational Learning (27; 0.83%), Theory of Knowledge (26; 0.80%), Intellectual Capital (25; 0.77%), Educational Innovations (25; 0.77%), Information Technology (24; 0.74%), Public Administration (24; 0.74%), Technology & Society (22; 0.67%), Project Management (22; 0.67%), Technology Acceptance Model (21; 0.64%), Organizational Change (21; 0.64%), etc. ICT acceptance and diffusion themes (TAM, Diffusion of Innovations) appeared.
  • 2013–2017 (30 of 39 shown): Information Technology in Medicine (135; 5.98%), Decision Making (45; 1.99%), Diffusion of Innovations (43; 1.90%), Higher Education (41; 1.82%), Information Technology Industry (32; 1.42%), Organizational Performance (30; 1.33%), Medical Care (30; 1.33%), Telemedicine (26; 1.15%), Business Enterprises (26; 1.15%), Structural Equation Modelling (25; 1.11%), Public Sector (23; 1.02%), Public Administration (22; 0.97%), Project Management (21; 0.93%), Electronic Commerce (21; 0.93%), Social Networks (21; 0.93%), Online Social Networks (19; 0.84%), Technology (18; 0.80%), Data Analysis Software (17; 0.75%), Industrial Management (17; 0.75%), Theory of Knowledge (16; 0.71%), Learning Strategies (16; 0.71%), Information Technology Periodicals (16; 0.71%), Cell Phones (15; 0.66%), Data Security (15; 0.66%), Medical Communication (15; 0.66%). New ICT enablers included social media and cell phones.
• Thematic areas of ICT application in IKM across periods: medicine/health, business and commerce, education and training (higher education), decision sciences, industrial management, and information resources management.
• Temporal dynamics: Core subjects varied substantially between periods; many subject terms appeared only once overall; cumulative subject terms rose from 960/2933 articles (1998–2002) to 4049/8874 (2008–2012), then decreased to 3129/653 (2013–2017), with the late decrease attributed to retrospective indexing lag.
• Fit to Bradford’s law (Tables 6–7): Within-period multipliers r_n/r_{n−1} closely matched calculated κ for each zone. Proportional patterns approximated 1:n:n²:n³ with multipliers n by period: 1998–2002 ≈ 5 (1:4.86:22.86:108.43), 2003–2007 ≈ 4 (1:3.86:14.78:56.75), 2008–2012 ≈ 3 (1:3.19:10.21:32.69), 2013–2017 ≈ 4 (1:3.87:15.26:60.10). Authors noted that while subject-term dispersion fits Bradford’s pattern, dispersion of articles across zones was not perfectly representative of the law.

The study addressed whether Bradford’s law can be used to identify core ICT-related subject terms in IKM and to characterize their dispersion. By ranking subject terms from LISTA/LISS and partitioning them into Bradford zones, the authors identified evolving cores across four periods, revealing shifts from IT-in-libraries themes (late 1990s) to organizational management and commerce (2003–2007), and later to diversified ICT use in management, education, and especially medicine/health, with social media and mobile technologies emerging in 2013–2017. The two criteria for Bradford fit—zone multipliers approximating κ and proportional distributions approximating 1:n:n²:n³—were satisfied for subject terms, supporting the applicability of Bradford’s law to subject analysis beyond journals. These findings are significant for IKM and LIS because they provide an empirical basis for delineating core concepts over time, informing indexing and retrieval strategies, guiding collection development toward high-yield subject areas, and supporting curriculum and thesaurus development around demonstrably core ICT topics. The recognition that article counts per zone did not perfectly follow Bradford’s law highlights nuances in applying Bradford analysis to units other than journals, suggesting careful interpretation when moving from term-level to document-level dispersion. The observed temporal shifts underscore the dynamic, interdisciplinary nature of ICT in IKM and the role of emerging technologies (e.g., social media, telemedicine, mobile devices) in shaping the field’s core concerns.

The study determined that core ICT subject terms within IKM vary by period, with 7 core terms (1998–2002), 36 (2003–2007), 86 (2008–2012), and 39 (2013–2017). Core compositions changed across periods, with 2008–2012 and 2013–2017 showing greater similarity. ICT applications in IKM concentrate in medicine, business/commerce, education and training, decision sciences, industrial management, and information resources management. The proportional distribution of subject terms across Bradford zones approximated 1:n:n²:n³, and zone multipliers matched calculated κ values, indicating that Bradford’s law can be used to identify core ICT subject terms in IKM. Future work should replicate the approach in other subject domains to further validate the law’s applicability for units beyond journal titles.

• Indexing time lag and retrospective indexing likely reduced the number of captured subject terms in 2013–2017, potentially affecting core composition.
• The authors note that dispersion of articles across Bradford zones did not accurately represent the principles of the law, even though subject-term dispersion did, indicating caution when interpreting article-level zone distributions.
• Results depend on subject indexing in LISTA and LISS; dynamic indexing practices may influence term frequencies and core identification.