Page 139 - JOURNAL OF LIBRARY SCIENCE IN CHINA 2018 Vol. 43

ZHAO Xing / Exploring the measurement features of usage data for academic literature 139

accumulation, besides measuring the statistical features of usage, this paper calculates the citation
data of the sample literature set for a comparative study. See the descriptive statistics of the citation
and usage data on the four disciplines in Table 2.

Table 2. Descriptive statistics of the citation count and usage count of literature in four disciplines in 2013 (TC: total
citation; U: total usage)
Discipline Index Max. Min. Ava. Std. K-S Z K-S P
TC 539 0 5.30 10.36 110.53 ＜0.001
Physics
U 1568 0 26.38 36.44 85.13 ＜0.001
TC 719 0 3.08 6.79 67.77 ＜0.001
Computer Science
U 830 0 15.81 20.00 44.71 ＜0.001
TC 82 0 2.25 3.74 36.66 ＜0.001
Economics
U 707 0 15.64 16.30 23.94 ＜0.001
Library and Information TC 48 0 2.24 3.542 15.85 ＜0.001
Science U 392 0 28.24 27.888 9.35 ＜0.001

As shown in Table 2, no matter what the maximum or mean values are, the value of usage is
always greater than that of citation, fully implying that usage is more discriminative and sensitive.
Relative to extreme values, the mean value can more directly reflect the mean level of the data
set. When comparing the mean levels of the usage of various disciplines, the data reveal a very
interesting phenomenon. The mean usage count of Library and Information Science is even higher
than that of physics, possibly because the scholars of Library and Information Science are more
accustomed to and skilled at using WoS and other relevant platforms. This speculation can also be
demonstrated by the conversion rate of “usage-citation”. The literature of Library and Information
Science has a high usage count but a low citation count. The usage count is about 13 times greater
than the citation count. For the literature of physics, the mean citation count is 5.3, and the mean
usage count is only about five times the mean citation count, so the conversion rate is far higher
than the rate of Library and Information Science. Assuming that the ratio of the number of acquired
pieces of literature to the number of cited pieces of literature is basically fixed, the scholars of
Library and Information Science largely acquire literature from the WoS platform whilst the
scholars of physics probably tend to rely more on other channels. In addition, the mean usage count
of computer science is also worthy of close attention, as its value is basically equal to the mean
usage count of economics. In other words, although computer science is one of the disciplines
attaching great importance to conference articles, its usage data suggest that the discipline also has
a rather high usage count of journal articles. These results combine to reveal that in the statistical
sense, the value feature of usage is not consistent with that of citation, but has its own feature or
even shows its unique discipline difference in comparison to citation. The literature usage behavior
of scholars from different disciplines will provide a topic for a future detailed comparison.