Page 101 - Journal of Library Science in China, Vol.45, 2019

100 Journal of Library Science in China, Vol.11, 2019

selected as research objects. Usage data (HTML views and PDF downloads) and citations from
both Chinese journal official websites and CNKI platform was crawled and used to compare
their user platform preference and user interest preference, then summarize laws and contributory
factors of usage patterns, providing references for theoretical research and application of Chinese
usage metrics under new technology background.

1 Data and method

1.1 Data processing

Since January 2014, usage data of open-access CSSCI and CSCD Chinese journals (including
journal lists of CSSCI and CDCD extended versions) have been tracked and investigated. Finally,
academic papers from 61 Chinese open-access journals in the fields of “Library, Information and
Archival Science”, “Management Science”, “Economics”, “Pedagogy”, “Computer Science”,
“Earth Science”, “Math” and “Biology” (the first four belong to social sciences and the final four
belong to natural sciences), published during 2014-2015 and indexed by CSSCI (2017-2018
version) and CSCD (2017-2018 version), are selected as research objects. The specific selection
criteria are as follows: 1) All sample papers are open-access on their official websites and are
also full-text indexed by CNKI platform to ensure that users can browse or download academic
papers on both journal official websites and CNKI platform. 2) All sample papers were published
from January 2014 to December 2015 to ensure that browsing, downloading and citation data of
each paper were accumulated to a stable level (usage counts, especially citations, accumulate to
a steady level in the first 2 or 3 years after publication (Lippi & Favaloro, 2013)). 3) Browsing,
downloading and citation data of each paper must be complete and real-time, which means that
each paper can be browsed or downloaded right after its official publication. If any journal official
website is launched after February 2014 or browsing or downloading data of any papers are
missing, they will not be included in this study.
The data processing steps are as follows: 1) Usage data (full-text downloads) along with other
metadata (e.g., article titles, authors, total citations, keywords, etc.) of each paper were crawled
from CNKI platform. 2) Usage data (full-text and abstract views of HTML formats, full-text
downloads of PDF format) along with other metadata (e.g., article titles, authors, etc.) of each
paper were crawled from journal official websites. 3) Usage data from both journal official
websites and CNKI platform were merged by “title”, “author” or “DOI”. Finally, editorials and
letters were excluded, only articles, proceedings papers and reviews were kept, and the final data
set are shown in Table 1. Usage data of papers from “Library, Information and Archival Science”,
“Management Science”, “Economics” and “Pedagogy” fields were collected and processed
between March 14th and March 21st, 2018. Usage data of papers from “Computer Science”,
“Earth Science”, “Math” and “Biology” fields were collected and processed between April 1st and