primarily an indication of the articles/authors/journals that influence JRI articles rather
than the influence of JRI articles/authors upon scholarly journals in general, including
the JRI. In other words, the focus of our study is on the influence of JRI authors both
on the JRI and on other WOS journals. As noted earlier, our approach is more analogous
to the construction of WOS journal impact factors than existing studies. Further, as noted
in our discussion of Colquitt et al. (2009) below, the majority of citations to JRI articles
appear outside the JRI. Browne (2003) reports the 10 journals most frequently cited in
the JRI and the 10 authors most frequently cited in the JRI. The JRI is the most frequently
journal cited in the JRI and is the only insurance journal among the 10 most frequently
cited in the JRI. Lastly, Browne (2003) reports the 10 articles most frequently cited in the
JRI.
Weiss and Qiu (2008) present an analysis of JRI articles based on following metrics over a
75-year period and for different decades: number of articles, pages per article, number of
authors, authors per article, North American author affiliation versus not, international
content, topical content, empirical versus theoretical versus both types, and academic
versus industry-affiliated authors. The 14 leading authors based on number of articles
or number of pages is identified. A particular interesting finding related to our study is
that the average number of authors per article has increased over time, consistent with
increasing collaboration and increasing level of sophistication and analysis requiring
more collaboration.
Colquitt et al. (2009) present an analysis of citations to articles published in 17 risk
journals, which includes 7 that are principally actuarial. The authors analyze citations
appearing in the 17 risk journals and 16 finance journals (33 sample journals). This study
updates Colquitt (2003) to examine citations appearing between 2001 and 2005 in the
33 sample journals. They rank the journals based on citations and “insurance impact
factors” (Colquitt, 1997). They also report the most cited JRI articles for selected time pe-
riods. In regards to the present study, it is important to note that Colquitt (1997), Colquitt
(2003), and Colquitt et al. (2009), which together examine citations appearing over the
1991–2005 period, leave out approximately 78 percent of citations (i.e., references) ap-
pearing in JRI articles because most citations appearing in JRI articles are to journals that
are not among their 33 sample journals. That is, their sample journals account for only
about 22 percent of citations appearing in JRI articles over this 15-year period.8 In effect,
most of the journals that have an influence on the JRI as measured by reference lists in JRI
articles are not among the sample journals in the three preceding studies. On the other
hand, an important advantage of the three Colquitt studies is its fairly comprehensive
nature relative to insurance journals, both WOS and non-WOS. The results of our analy-
sis suggest that the Colquitt studies are likely to significantly understate the influence of
JRI articles and authors on noninsurance journals because in their study only 8.8 percent
of citations to JRI articles appear in their 16 noninsurance journals, while in our study
the 46.8 percent of citations to JRI articles appear in noninsurance journals.9
8 This percentage (22) is calculated using data reported in table 2 of Colquitt (1997) and (2003),
and table 1 of Colquitt et al. (2009).
9 Table 1 of Colquitt et al. (2009) shows that out of 1,222 citations to JRI articles, 107 appear in
noninsurance journals. Of the 13,921 citations to JRI articles in our study, 6,516 are to noninsur-
ance journals. If we collect and include citations to JRI article that appear in non-WOS insurance.344 RISK MANAGEMENT AND INSURANCE REVIEW
TABLE 1
JRI Publication Counts (1989–2010)
Panel A: Not Adjusted for Number of Coauthors Panel B: Adjusted for Number of Coauthors
N (Number) Authors Cumulative N (Weighted) Authors Cumulative
14 2 2 Over 5.0 5 5
11 3 5 4.51–5.0 8 13
10 3 8 4.01–4.50 3 16
9 5 13 3.51–4.00 4 20
8 4 17 3.01–3.50 6 26
7 7 24 2.51–3.00 8 34
6 7 31 2.01–2.50 14 48
5 11 42 1.76–2.00 26 74
4 25 67 1.51–1.75 10 84
3 47 114 1.26–1.50 30 114
2 117 231 1.01–1.25 16 130
1 655 886 0.76–1.00 142 272
0.51–0.75 37 309
0.125–0.50 577 886
Note: Coauthorship adjustment is strictly proportional to the number of authors on an article;
thus, an author receives a “credit” (or count) of 1 divided by the number of authors on each article
published in the JRI during the publication period.
Danielson and Heck (2014) identify the “most prolific contributing authors” to “seven
high-impact RMI journals” over the 1984–2013 period. The seven high-impact journals,
as identified by the authors based on presence in the SSCI and Colquitt et al. (2009),
include the JRI, JRU, Geneva Review, Geneva Papers on Risk and Insurance – Issues and
Practices (Geneva Issues), JII, JIR, and Risk Management and Insurance Review (RMIR).
These authors report that 72.4 percent of authors appear in these seven journals only
once. Author rankings are reported both adjusted and unadjusted for coauthorship both
for the seven high-impact journals and for the three “elite” journals, where the elite
journals include only the JRI, JRU, and Geneva Review.
SAMPLE AND DATA
We begin by collecting the following information for every article published in the
JRI, other than book reviews, from March 1989 through December 2010: (1) full name
journals without also doing so for non-WOS noninsurance journals, then a bias against JRI arti-
cles and authors with relatively more cites in non-WOS noninsurance journals is created. Our
use of WOS journals is motivated by the WOS consisting of a well-defined set of high-quality
insurance and noninsurance journals; in addition, we do not have to make a subjective choice
of sample journals among the hundreds of non-WOS journals.TABLE 2
Citations to Journal of Risk and Insurance Articles (1989–2014) Summary Statistics on
Authors
Variables Authors Mean Standard Deviation Minimum Maximum
cites total 886 15.71 27.14 0 295.00
cites ins 886 8.36 17.61 0 214.00
cites nonins 886 7.35 12.08 0 98.00
cites article 886 8.69 9.92 0 84.00
cites article ins 886 4.31 5.77 0 50.00
cites article nonins 886 4.38 6.16 0 69.00
cites weighted 886 7.32 13.09 0 158.04
cites year 886 1.18 2.00 0 27.13
Note: cites total = citations in Web of Science (WOS) journals; cites ins = citations in seven WOS in-
surance journals (see Sample and Data Section); cites nonins = citations in WOS journals other than
seven insurance journals; cites article = author’s citations in WOS journals divided by author’s
number of articles; cites article ins = author’s cites in seven WOS insurance journals divided by
author’s number of articles; cites article nonins = author’s citations in WOS journals other than
seven insurance journals divided by author’s number of articles; cites weighted = author’s cita-
tions divided by author’s average number of authors per article; cites year = author’s citations
divided by (2014 minus year of author’s first JRI article in publication period).
of each author; (2) number of authors on article; (3) month and year of publication;
and (4) article type. The article types are as follows: feature, shorter, invited, panel
discussion, comment, reply, presidential address, and editor report.10 Comments, replies,
presidential addresses, and editor reports are not included in our analysis of publication
counts or citations due to the nature of the articles and for consistency with prior
literature. In total, we include 696 articles and 886 authors, which correspond to an
author–article total of 1,444. Among the 696 articles, 201 (28.9 percent) have one author,
304 (43.7 percent) have two authors, 147 (21.1 percent) have three authors, and 44 (6.3
percent) have four or more authors.11 The median (mean) number of authors per article
during the publication period is 2 (2.07).
Next, citations to JRI articles of the article types indicated earlier are collected by author
for the period 1989–2014, that is, 4 years past the last JRI publication in our sample,
from Thomson Reuters WOS. Two databases within WOS are used—the SSCI and Science
Citation Index. First, citations are classified as self or nonself. Then, nonself-citations
are classified based on the journal where the citation appears as either insurance or
noninsurance citations. The following seven journals indexed by WOS are classified as
10 We do not distinguish between “feature articles” and “shorter articles.” The designation “shorter
article” has not been used since 1998. 11 Thus, the one author articles relates to 201 author–article observations, the two author articles
relate to 608 (2 × 304), and so forth, to reach the total number of author–article observations of
1,444.