Online First

The journals have a total of 468 EBMs, these representing 404 distinct individuals. Of these, 358 individuals serve on a single board, a percentage (76.5%) that is very similar to the figure of 79% quoted by Cronin (2009) for a set of ten LIS journals, seven of which are included in those studied here. Of the remainder, 32 individuals serve on two boards, ten on three boards and four (Judit Bar-Ilan, Ian Ruthven, Amanda Spink and Mike Thelwall) on four boards. There is hence a fair degree of inter-connectedness between the boards of the chosen sample of journals, as noted in several previous studies (Baccini & Barabesi, 2010, 2011; Malin & Carley, 2007; Rosenstreich & Wooliscroft, 2006).

The gender could be identified for 401 of the 404 distinct EBMs: 238 were male and 163 female, i.e., there was a slight (59.4%) male majority in the sample. The 404 EBMs were based at institutions in 40 different countries. Of these, the five most frequently occurring nations were the USA (198 EBMs), the UK (46), Canada (15), and the People’s Republic of China and the Netherlands (both with 10). A total of 231 different institutions was identified, the overwhelming majority of which were universities and the majority of which (166 of them) provided just a single EBM. The most frequently occurring institutions were Illinois at Urbana-Champaign and Indiana with 15 and 13 EBMs, respectively, followed by Maryland (9) and Florida State, Loughborough, Syracuse, California at Los Angeles and Washington (all of which had 6 EBMs); Loughborough was hence the only highly productive institution from outside the USA. In some cases, an institution was strongly represented on a particular journal’s board, e.g., Illinois at Urbana-Champaign has eight board members for LT, and all four members of the Hungarian Academy of Sciences in the sample were on the board of SCI. For comparison, 48,849 different authors were associated with the 39,932 journal articles, proceedings papers and review articles in the Web of Science Information Science & Library Science category published in the period 2001-2011. Even allowing for variant forms of author names, making the figure of 48,849 probably a slight over-estimate, it is clear that the 404 distinct EBMs are a very small subset of the current population of LIS authors, comprising just 0.83% of the latter. It has been noted in the literature review that, in some cases, EBMs do not exhibit the research track record that might have been expected, and it is hence of interest to see whether this is the case for LIS.

The first article was hence chosen from the final issue in 2011 of each of the 16 journals, these 16 articles being associated with a total of 31 different authors. After eliminating seven of these who were included in the set of 404 EBMs (with six of this seven publishing in a journal for which they were an EBM) there was thus a total of 24 authors typical of non-EBMs contributing to the 16 journals. Their publication and citation counts were collected as for the EBMs above, and the resulting counts for the sets of EBMs and non-EBMs compared as summarized in Table 2. A χ² test using Yates’ correction shows that the EBM and non-EBM publication counts are significantly different (χ² = 16.02, υ = 2, p <= 0.005) and that this is also the case for the citation counts (χ² = 15.05, υ = 2, p <= 0.005). The differences are, of course, hardly surprising in that individuals are approached to become EBMs largely on the basis of their research contributions, and for some of the EBMs these contributions were very large: 10 had more than 100 publications and six had more than 2000 citations, the EBM median values for these two variables being 9.5 publications and 39 citations, respectively. However, the contributions are small for some of the other EBMs, as is clear from the bottom row of Table 2: indeed, there were 41 EBMs with no publications and 69 EBMs with no citations. In saying that, it is most important to remember that Web of Science has been used to obtain the publication and citation data. Many of the EBMs of some LIS journals are practicing library or information professionals, with strong publication records in the professional literature rather than in the high-impact academic journals that form the principal input to Web of Science, and the boards of journals such as CRL, LRTS or RUSQ contain many such practitioners (as is entirely appropriate given the aims of these journals).

At least in part, the large contributions that the EBMs make to the literature reflect the fact that many of them have been working in the field for an extended period. The median date of first publication for the EBMs (for those with publications) was 1995 whilst that for the set of non-EBMs was 2007, with the largest single number of the latter making their first contribution as recently as 2011. At the other extreme, the first publications in Web of Science for Gene Garfield and for Tom Wilson date from 1954 and 1963, respectively; and the presence of several very highly experienced researchers on the board of the ARIST results in it having by far the earliest median date of first publication in Table 1.

Country data was available for 36,141 of the set of 39,932 LIS articles mentioned above. These records represent a total of 145 different countries, of which the four most productive were the USA (15,076 publications), the UK (3,731), the People’s Republic of China (2,958), and Canada (1,626) (note that there is some duplication here since the authors of a publication may come from more than one country); the USA’s dominance is also clear in a study of LIS journals by Davarpanah and Aslekia (2008). Whilst there is clearly a fair degree of match between numbers of EBMs and contributions to the LIS literature (as reflected in Web of Science), it is also clear that the People’s Republic of China is markedly under-represented at board level, and this is also the case for other countries. Table 3 lists the country data for the EBMs and for LIS publications in general, and a χ² test data shows that these two distributions are indeed significantly different (χ² = 33.64, υ = 4, p ≤ 0.0001). There is a strong bias in the data since three of the journals - CRL, LRTS and RUSQ - are published by divisions of the American Library Association and have editorial boards that are overwhelmingly composed of people from the USA. However a significant difference (χ² = 16.44, υ = 4, p ≤ 0.005) in the distribution of nationalities remains if the analysis in Table 3 is restricted to the other 13 journals.

In brief, it would seem that the set of EBMs considered here is not totally dissimilar in character to the respondents to a questionnaire survey of the EBMs of 56 accounting journals, where Brinn and Jones (2007) noted that the average respondent “was a US, male professor with a PhD and professional qualification who was over 50 and had been in service for over 10 years [and] who was active and successful in research”.

Inspection of the summary data for the 16 journals listed in Table 1 will show that the boards vary considerably in size, from RUSQ at one extreme to SCI at the other, with a median of 22 EBMs. Their composition is equally variable, in terms of both gender and nationality. The median male fraction is 0.59, with the lowest values being for two of the more professional journals, RUSQ and LRTS, as might be expected given the predominance of women in the LIS workforce. These two journals are the only ones in the sample of 16 that have female editors, and they (together with CRL) also have the smallest spread of nationalities. JDOC and LT have the most geographically diverse boards using the chosen statistic, with SCI (25) and IR (19) having the largest number of distinct nationalities represented on the board. It is interesting to compare these gender and nationality results with previous studies in other disciplines (though it should be noted that the figures quoted below are taken at a specific moment of time, and some longitudinal studies suggest that the fraction of male and USA representation are both declining).

While men clearly dominate the editorial boards of the LIS journals considered here, this is far less marked than in previous studies of other disciplines: for example, the median male fraction in 2004 for the editorial boards of 57 management journals surveyed by Metz and Harzing (2009) was no less than 0.83, a figure that is identical to that obtained from an analysis by Morton and Sonnad (2007) of 3,473 individuals on 39 professional society boards and 54 journal editorial boards in medicine. A very similar fraction, of 0.82, was obtained for the fraction of male responses to a questionnaire distributed to the editorial boards of 56 accounting journals (Brinn & Jones, 2007) and for the median male fraction of the boards of 14 leading medical journals (Jagsi, Tarbell, Henault, Chang & Hylek, 2008). Fong et al. (2009) quote a much lower value of 0.53 for the editorial boards of five educational psychology journals, a figure that is broadly comparable to the 0.58 that may be estimated from inspection of Figure 6 in a survey of four school psychology journals by Roberts et al. (2006).

Turning now to the international spread, most previous studies in the literature have focused on the fraction of EBMs that come from the USA. In our sample of 16 LIS journals, 198 EBMs were from the USA, i.e., 0.49 of the set of 404 individual EBMs, with the individual journals having fractions ranging from 0.05 (JDOC) to 1.00 (LRTS and RUSQ) and with a median of 0.62 (or of 0.56 if the three American Library Association journals are excluded). The median figure can be compared with those reported in studies of the nationalities of editorial boards for leading journals in other disciplines: for example, 0.86 for ten marketing journals (Rosenstreich & Wooliscroft, 2006), 0.75 for 16 emergency medicine and critical care journals and 0.59 for 18 anaesthesiology journals (Boldt & Maleck, 2000), 0.57 for 57 management journals (Harzing & Merz, 2012), 0.57 for 281 journals across 15 scientific and engineering disciplines (Garcia-Carpintero, Granadino & Plaza, 2010), and 0.42 for 20 leading chemistry journals (Braun & Diospatonyi, 2006).

The early date of first publication for the ARIST EBMs has been noted above; other journals in Table 1 that have a significant number of such experienced contributors include IPM, JDOC, LT and SCI, all of which have median first publication dates prior to 1992. There are, as would be expected, strong correlations between the publication and citation data in Table 1, with the five journals having the largest median numbers of publications (ARIST, JDOC, JIS, JASIST and SCI) also having the largest median numbers of citations; and all of these (bar JIS, which is replaced by IPM) are included in the five journals with the largest 2-year impact factor.

The data in Table 1 have been used to generate a classification of the 16 journals, with each of the seven columns being Z-standardized so that each variable had a zero mean and a unit standard deviation. The resulting 16×7 data matrix was then submitted to the group-average hierarchic agglomerative clustering routine in SPSS, with the inter-journal distances calculated using the squared Euclidean distance.

The resulting classification contained three well-marked clusters: ARIST on its own in a singleton cluster; then a six-member cluster (CRL, JAL, LISR, LQ, LRTS and RUSQ) consisting of, predominantly, library-focused journals with a strong professional slant, and then the remaining nine, research-focused journals (IPM, IR, IS, JDOC, JIS, JASIST, LT, OIR and SCI). Exactly the same partition is obtained if one takes the three-cluster solution resulting from the widely-used Ward’s minimum variance clustering method, instead of using the group-average method as here. The same set of 16 journals has been clustered by Milojević et al. (2011) on the basis of common title words using the Jaccard coefficient and the group-average method. The dendrogram shown in Figure 5 of their publication again reveals a well-marked, three-cluster solution: one cluster consisting of SCI on its own; a cluster of library-focused journals (CRL, JAL, LISR, LQ, LRTS, LT and RUSQ); and then a cluster containing the remaining, more information-focused journals. It will be realized that this is very similar indeed to the three-cluster solution obtained in the present work, despite the very different variables used to characterize the journals in the two studies. The singleton status of SCI in the earlier study is likely to arise from the bibliometrics-focused vocabulary employed in the titles of the articles published therein, while the comparable status of ARIST in the present study arises from its EBMs’ very strong research profiles (the journal is also, of course, unique in containing only review articles but this is not a characteristic coded explicitly in Table 1).

The three-cluster solution obtained here is based on the use of all of the seven variables in Table 1 to characterize each of the journals and, as noted previously, there are two rather different types of variable considered in this study: those that characterize each journal as a whole (i.e., number of EBMs, the male fraction, the international spread and the IF) and those that characterize the individual EBMs (i.e., their date of first publication, and their publication and non-self citation counts). Use of the journal-specific variables yielded a two-cluster solution: one consisting of CRL, JAL, LQ, LRTS and RUSQ, i.e., the library-focused cluster obtained previously with the exception of the more research-focused LISR; and one consisting of the other eleven journals. Use of the EBM-specific variables yielded a three-cluster solution: ARIST on its own; a group of five high-impact research journals (IPM, JDOC, JIS, JASIST and SCI); and finally the remaining ten journals. Both of these classifications are entirely understandable given the natures of the journals that are being clustered. Alternative classifications of LIS journals are discussed by Ni et al. (2013), who have used distances between pairs of journals based on numbers of common authors, of common EBMs, of co-citations in reference lists, or of common title and abstract words.