What do Chinese-language microblog users do with Baidu Baike and Chinese Wikipedia? A case study of information

Note. A pre-print version accepted for presentation at Opensym 2014. It may be different from the final version for the conference proceeding.
What do Chinese-language microblog users do with Baidu
Baike and Chinese Wikipedia? A case study of information
engagement
Han-Teng Liao
Oxford Internet Institute
University of Oxford
Oxford, United Kingdom
hanteng@gmail.com
ABSTRACT
This paper presents a case study of information engagement based
on microblog posts gathered from Sina Weibo and Twitter that
mentioned the two major Chinese-language user-generated
encyclopaedias. The content analysis shows that microblog users
not only engaged in public discussions by using and citing both
encyclopaedias, but also shared their perceptions and experiences
more generally with various online platforms and China’s
filtering/censorship regime to which user-generated content and
activities are subjected. This exploratory study thus raises several
research and practice questions on the links between public
discussions and information engagement on user-generated
platforms.
Categories and Subject Descriptors
[Human-centered computing]: Collaborative and social
computing systems and tools—Blogs, Wikis, Empirical studies in
collaborative and social computing
Keywords
Censorship, Chinese Internet, Information use, Information
engagement, Internet filtering.
1. INTRODUCTION
The microblogging services Sina Weibo and Tencent Weibo were
launched in 2009 and 2010 respectively and have contributed to
the rapid growth of online advertising markets in China [5, 12].
The term “Weibo” ( 微博 ) is the transliteration of the term
“microblog” in Mandarin Chinese. Weibo allow users to publish
short messages like tweets with a 140-character limit per post.
Like other major companies in China and elsewhere whose
business is user-generated content (hereafter UGC), the amount of
content and traffic a platform can garner is crucial for its survival.
Microblogs can be seen as part of the historical development of
the rise of user-generated content since 2005[19]. A recent survey
has shown that user-generated content is more popular in Asia
than in the West [3]. By the first half year of 2010, the content
produced by amateur Chinese Internet users surpassed that
produced by professional websites [26]. According to a survey
report based on 2012 data, around 66.1% of mainland Chinese
Internet users used a blog/personal space, 54.7% used microblogs,
and 48.8% used social networking websites [5]. One US market
report found that 47% of Chinese broadband users had
contributed to UGC in China, including online review sites,
forums, blogs, etc., and that this UGC content had influenced
58% of purchase decisions in China. Both numbers are much
higher than the corresponding statistics in the US [24].
The growth of UGC in mainland China may have contributed to
the creation of a so-called “public opinion monitoring sector” (舆
情监测服務市場), where managers (mostly from the public
sector) buy up-to-date information about the latest online public
opinion in mainland China. The online public opinion data relies
heavily on processing and mining UGC, especially from
microblogs. One of the most successful example is the Online
Public Opinion Monitoring and Measuring Unit of the People's
Net (人民网舆情监测室), which alone has an annual revenue of
close to 200 million RMB [21, 28], or about £20 million. It is
noted that the Chinese government and corporations have been
developing and using “Internet public opinion monitoring
systems,” which involve intensive collection and analysis of web
data [21, 28]. As of 2014, there is little Internet research in
English about this development of public opinion monitoring,
except for a panel discussion [15], which I organized during the
2013 Chinese Internet Research Conference. Further work will be
needed to compare the political and commercial uses of
microblog data in China and elsewhere.
As expected, microblog platforms that are hosted in mainland
China are subject to Beijing’s filtering and censorship regime.
Several recent research papers have reverse-engineered the
censorship mechanism, mostly based on the selection of keywords
to understand how censorship works on Weibo platforms.
Analysing sensitive terms on Sina Weibo and Twitter, researchers
found initial evidence that content is filtered or censored on Sina
Weibo[1]. Researchers at Hong Kong University, who
systematically collect Weibo data on a website called Weiboscope,
have also conducted censorship research [9, 10]. Another group of
researchers found that individual criticisms of the government are
permitted but that collective expression (which might generate
protest against the government) is censored [13]. Yet another
study tracked and quantified sensitive topics on Weibo and found
that the posts were often short-lived and confined to a small core
of users who posted sensitive content, suggesting that multiple
layers of filtering produced effective censorship to contain the
spread of such messages [36].
According to an industry report, in 2012, about 88% of Internet
users in mainland China are Sina Weibo users [7]. Weibo
platforms are thus important sites of observation to see how user-
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. To copy
otherwise, or republish, to post on servers or to redistribute to lists,
requires prior specific permission and/or a fee.
OpenSym '14 August 27 - 29 2014, Berlin, Germany
Copyright 2014 ACM xxx-x-xxxx-xxxx-x/xx/xx ...$xx.xx.

2
generated encyclopaedias like Chinese Wikipedia, Baidu Baike
and Hudong Baike[19] are mentioned and used. (Baike means
“encyclopedia” in Chinese. Baidu is the name of China’s major
search engine company. Hudong means “interactive” or
“interaction” in Chinese.) First, the microblog posts are likely to
contain the views and experiences of Chinese Internet users with
the encyclopaedias. Second, since microblog platforms are also
user-generated content platforms, it is possible that the users of
microblogs are more likely to use user-generated encyclopaedias.
Third, researchers can observe how Beijing’s censorship
influences the use or perception of user-generated encyclopaedias.
Especially in the case of Sina Weibo, which reaches a large
enough number of mainland Chinese users, the posts can provide
important indications about how mainland Chinese users are
receptive to the two online encyclopaedias. Fourth, users are
likely to use microblogs (both reading and writing, which can be
regarded as one amongst several online information engagement
practices [2]) in conjunction with search engines and user-
generated encyclopaedias. Researchers can thus observe how
users use different platforms in combination.
The concept of “information engagement” is used in two
disciplines with different audience and research concerns. In the
context of civic engagement including youth engagement, the
focus is to engage citizens for more and better political
participation, as proposed by W. Lance Bennett and Chris Wells
to build “civic learning repertoires” for “information
engagement” [2]. In the context of warfare, including information
warfare, the practical concern is to provide actionable guidelines
to inform foreign and domestic audiences [30], which has been
discussed as part of the redefinition of information warfare [25].
In the Chinese context, the Internet is central for civic
engagement [6, 35] and the state’s cyberwarfare doctrine [11],
and the study of the use and reception of user-generated
encyclopaedias expressed on microblogs can be conducted as a
case study of information engagement in both senses.
This paper aims to explore what Chinese-language microblog
users do with major user-generated encyclopedias such as Baidu
Baike (Baidu’s encyclopedia website) and Chinese Wikipedia[19]
by analysing the content of their posts. As user-generated
encyclopaedias are expected to solve information problems for
users, microblog content about users’ experiences and perceptions
should indicate the status (and thus also the gaps) of information
engagement experiences with the two major online Chinese
encyclopaedias as expressed in a major type of platform that
constitute a major part of Chinese public opinion.
2. METHODS
Computer-assisted quantitative methods (including several
Chinese natural language processing and text-mining techniques)
are first used to identify patterns and posts relevant to the research
question, and then qualitative methods are used to interpret the
findings. Similar to how Zhu, Phipps, Pridgen, Crandall, &
Wallach [36] tracked and quantified censorship by selecting
topics for research, I developed a means of filtering relevant posts
that mention Baidu Baike or Chinese Wikipedia, which produced
a dataset that combines data from Sina Weibo and Twitter.
This exploratory method mimics the tools that are likely used for
public opinion monitoring. Relevant posts are first screened from
designated online platforms using keywords based on the
analyst’s judgement. Although there are no guidelines for how
many and which terms yield sufficiently comprehensive results, it
can be expected a few iterations will yield certain heuristics for
gauging the results. As an exploratory study, the combination of
qualitative and quantitative methods is bound to be refined in the
future as the analysis of microblogs advances. For example, one
must consider the issues of sampling when claiming the
observations on microblog platforms are valid public opinion
data, which is outside the scope of this exploratory study.
Nevertheless, this research should provide indications for future
research and practices on monitoring and analysing microblog
content. I now examine how the Chinese term ‘wiki’ in a number
of variants, which are mentioned in microblog posts.
2.1 Data selection and data sets
For data collection, the Chinese term (both simplified and
traditional characters) “wiki” was used instead of the more
specific term “Wikipedia”, in addition to specific Chinese terms
such as “Baidu Baike” and “Chinese Wikipedia” (see Table 1).
Table 1. Wikipedia-related and Wiki-similar Chinese terms
Simplified
Chinese
Traditional
Chinese
Notes
Baidu Baike 百度百科百度百科 unambiguous term
Chinese Wikipedia 维基百科維基百科 unambiguous term
Hudong Baike 互动百科互動百科 unambiguous term
wiki 维基維基 ambiguous term
wikileak 维基解密維基解密
wikileak 维基揭密維基揭密 alternative name
wikileak 维基泄密維基泄密 alternative name
"Wong Wiki" 王维基王維基 Ricky Wong, unrelated to Wiki
Two unrelated entities were expected to yield false positives since
they contain exactly the same characters of popular Chinese term
of “wiki”: One is “WikiLeaks” with its variant Chinese terms and
another is a Hong Kong businessperson Ricky Wong, whose first
name happens to share the exact same Chinese characters as
“wiki” (see the bottom row in Table 1). The keyword filter used
here can also introduce other false positive outcomes since the
Chinese term “wiki” can also refer to general wiki technologies
and practices, yet in this case the resulting dataset can be used to
see how dominant Wikipedia or wiki-implemented
encyclopaedias are on the subject of wiki.
Based on the keyword list, two datasets were retrieved from two
data intermediaries. The WeiboScope dataset, provided by Dr. Fu
at the Hong Kong University, randomly sampled results from Sina
Weibo between January 2011 and April 2012 (see Table 2 for
details). Second, the DiscoverText dataset contains publicly
available microblog posts from Sina Weibo and Twitter for a time
period of 23 days in 2012. Table 2 summarizes the number (N)
and proportion (percentage) of posts from the two platforms and
the two intermediary sources, showing a disproportionally large
dataset from Weibo (81%).
Table 2 Posts collected: including false positives
Microblog platform N % N % N %
Twitter 0 0% 11640 63% 11640 19%
Weibo 42992 100% 6865 37% 49857 81%
Total 42992 18505 61497
Timestamps
Earliest
Latest
Time Duration
DiscoverTextWeiboScope Total
03 Jan 2011 00:21
458 days
20 Mar 2012 18:08
23 days
03 Jan 2011 00:21
465 days
05 Apr 2012 13:05 12 Apr 2012 12:09 12 Apr 2012 12:09

3
The first step then was to remove the false positive data (unrelated
to the research topic) included in Table 2, and then also to remove
another set of “jamming” posts targeted at Wen Yunchao,
resulting in a reduced dataset as summarized in Table 3. Wen
Yunchao is a mainland Chinese activist who launched a series of
online campaigns against Internet censorship in China, through
writing messages on blogs and other social media platforms.
Table 3 Posts collected: false positives removed
Microblog platform N % N % N %
Twitter 0 0% 1152 15% 1152 3%
Weibo 37185 100% 6330 85% 43515 97%
Total 37185 7482 44667
Timestamps
Earliest
Latest
Time Duration
05 Apr 2012 13:05 12 Apr 2012 12:09 12 Apr 2012 12:09
458 days 23 days 465 days
WeiboScope DiscoverText Total
03 Jan 2011 00:21 20 Mar 2012 18:08 03 Jan 2011 00:21
The removal of posts targeted at Wen Yunchao resulted in a
notable reduction (more than 10,000 posts) in the Twitter dataset;
so, it is worth explaining why these were removed. I examined
these posts and noticed that almost all of them followed the
format: “@wenyunchao Some Entry or Web Page Title – Baidu
Baike or Wikipedia” (for example, “@wenyunchao 李明博 - 维
基百科” and “@wenyunchao 山海关_百度百科”). These posts
were often written by users with ID names that are likely to be
randomly generated by machines (e.g. 5sx63cyd, bd8wwkp). All
these machine-generated posts were posted to Twitter within the
23 days covered by DiscoverText as shown in the first row of
Table 4. These posts mostly cited Baidu Baike, but a few posts
cited Chinese Wikipedia. The two posts containing
“wenyunchao” from the WeiboScope dataset, in contrast, were
sent by human users.
Table 4. Posts that mention wenyunchao (the set "WYC")
Microblog
platform
Dataset BB CW HD "wiki" wikileak "Wong"
Twitter DiscoverText 8755 1426 0 1524 10 4
Weibo HKU 0 0 0 2 2 0
8755 1426 0 1526 12 4Total
As these posts would skew any content analysis, I decide not to
include them, but rather to consider how the cyber-attacks against
Wen Yunchao relate to the research here.
The remaining 44667 posts are then processed for Chinese-
language word-tokenization, a necessary procedure to analyse
Chinese-language texts because, unlike languages such as
English, Chinese is written without space.
2.2 From overview to detailed analysis.
I first ran a word frequency analysis to obtain an overview about
hot topics and frequently associated words. Then I compared the
posts that mention the two encyclopaedias individually. In the
process, I grouped and differentiated microblog posts using
computer-assisted techniques (including Latent Semantic
Analysis) so as to break down the large number of posts into
manageable clusters. Originally developed for word-document
relationship analysis in the field of natural language processing
and information retrieval, these techniques proved suitable for the
following reasons. First, they can process a large amount of
documents, assisting researchers in clustering and differentiating
documents and words for further analysis. Second, they mimic or
even simulate the way the online texts are processed by search
engines and social media analysts (including China’s online
public opinion monitors). Third, techniques such as Latent
Semantic Analysis has been applied to Chinese-language texts
with noticeable success [4, 33], solving effectively the matching
problems caused by synonymy or polysemy [34]. Finally, by
repurposing the quantitative techniques for cultural research,
researchers can ground their analyses not only on specific
relevant words, but also on the latent semantic relationships
amongst them that are not immediately apparent to casual readers.
For this study, I implemented such analysis using a combination
of NLTK (Natural Language Toolkit), scikit-learn (a set of
machine learning tools) and gensim (“generate similar”
documents).
After the posts were clustered, I categorized the relevant posts
under the two main themes (filtering/censorship regime and
information engagement) to elicit how the two encyclopaedias
were discussed by microblog users.
3. RESULTS
The findings are now presented, proceeding from overall patterns
to specific individual posts. Table 5Error! Reference source not
found. lists the most frequently occurring words. The most
mentioned term is “encyclopaedia” ( 百科 ), after which the
Chinese names of the three encyclopaedias follow in this order:
Baidu Baike, Wikipedia and Hudong Baike. This paper focuses
on the posts mentioning the first two of these.
Table 5. Most frequently-occurring words
1 Encyclopedia 百科 39951 31 Company 公司 1601 61 Sprinkle over 撒过 1346
2 Baidu 百度 22672 32 Harm 危害 1600 62 Test 检验 1344
3 Wiki 维基 21495 33 (English) wikipedia 1570 63 Initiation 引发 1342
4 Baidu Baike 百度百科 18623 34 Addition 增加 1569 64 Kai-fu Lee 李开复 1338
5 Wikipedia 维基百科 10780 35 Ha ha ha 哈哈哈 1561 65 Nerve 神经 1331
6 Interactive 互动 6948 36 Cause 导致 1536 66 Times 万倍 1324
7 Hudong Baike 互动百科 6583 37 Reference 参考 1522 67 Notes 笔记 1312
8 China 中国 4428 38 Life 生命 1518 68 Jokes 恶搞 1308
9 Floating Spirit 浮灵 3954 39 The book 全书 1513 69 100,000 times 十万倍 1307
10 Share 分享 3291 40 One kind 一种 1495 70 Bad 不良 1290
11 Reply 回复 3023 41 Not work 不行 1474 71 Necessary 就要 1286
12 Activity 活动 2732 42 Network 网络 1453 72 Search 搜索 1283
13 Excessive 超标 2682 43 Occur 发生 1444 73 Scientific name 学名 1268
14 Wallet 钱包 2608 44 Community 社会 1427 74 Serious harm 严重危害 1265
15 Wiki 維基 2310 45 In vivo 体内 1422 75 Ridicule 调侃 1264
16 Microblogging 微博 2290 46 Similar 类似 1421 76 Notebook 笔记本 1263
17 United States 美国 2239 47 University 大学 1418 77 Syndrome 综合症 1247
18 Recommend 推荐 2222 48 Knowledge 知识 1411 78 Sitting 坐姿 1241
19 Discovery 发现 2217 49 Oxygenated 含氧 1400 79 (English) syndrome 1235
20 World 世界 2181 50 Oxygen 含氧量 1399 80 (English) piriformis 1232
21 Introduction 介绍 2130 51 Temporarily 暂时 1399 81 Ischial 坐骨 1225
22 Website 网站 2008 52 Oxygen 氧量 1399 82 Intelligence 智力 1225
23 Forwarding 转发 2008 53Canadian fish 加鱼 1395 83 Sciatic 坐骨神经 1222
24 Free 自由 1938 54 Originally 原本 1380 84 Country 国家 1211
25 Health 健康 1795 55 Complex 综合 1367 85 History 历史 1208
26 First 第一 1752 56 Time 时间 1366 86 Information 信息 1197
27 Shanghai 上海 1714 57 Soon 眼看 1365 87 Beijing 北京 1193
28 Click 点击 1701 58 Encyclopedia 百科全书 1353 88 Face 正视 1191
29 Entry 词条 1669 59 Hypoxia 缺氧 1352 89 Because fish 因鱼 1181
30 Call 称为 1642 60 Extend 延长 1350 90 Really 真的 1177
Ranking Word CountsRanking Word CountsRanking Word Counts
3.1 When mentioning either
Table 6 provides an overview for each of the two portions of posts
mentioning the two encyclopaedias, and ranks the most frequently
co-occurring words. Instead of commenting on the two
encyclopaedias per se, many posts use or cite the encyclopaedias
to comment on certain popular incidents. The left-hand side of the
table shows the predominance of the Yúfúlíng (鱼浮灵) incident
for posts mentioning Baidu Baike, whereas the right-hand side
shows the predominance of the ‘Wallet’ incident for posts
mentioning Chinese Wikipedia. The findings mainly demonstrate
the topicality of microblogs in responding to the latest events.
Baidu Baike was mentioned because of the background

4
information it provided concerning an online rumour that a
chemical used to revive dying fish called Yúfúlíng is
carcinogenic. Chinese Wikipedia was referred to for its content on
piriformis syndrome, a topic around which the ex-Google China
Chief Lee Kaifu was ridiculed by many and defended by some for
recommending that people not sit with their wallets in their rear
pockets. To gauge their impact of these events on the datasets, I
used clustering and keyword search independently, and found
about 1,400 posts for the Yúfúlíng incident and about 1,200 posts
for the ‘Wallet’ incident. I also identified the words
predominantly associated with the incidents (see the “Memo”
columns in Table 6), which included ‘Shanghai.’ The findings
highlight the potential of using how online encyclopaedias are
referenced and linked in topical microblog public discussions
when authoritative information is needed.
Table 6. Most frequently-occurring words when mentioning…
Baidu Baike Chinese Wikipedia
Counts Memo Counts Memo
1 Floating Spirit 浮灵 39951 Yúfúlíng Wallet 钱包 2600 Wallet
2 Excessive 超标 22672 Yúfúlíng Introduction 介绍 1451 Wallet
3 Baidu 百度 21495 Reference 参考 1317 Wallet
4 Share 分享 18623 Call 称为 1307 Wallet
5 China 中国 10780 Kai-fu Lee 李开复 1297 Wallet
6 Discovery 发现 6948 Yúfúlíng Health 健康 1293 Wallet
7 Addition 增加 6583 Yúfúlíng Similar 类似 1292 Wallet
8 Life 生命 4428 Yúfúlíng Kuso 恶搞 1258 Wallet
9 In vivo 体内 3954 Yúfúlíng Cause 引发 1251 Wallet
10 Oxygen level 含氧量 3291 Yúfúlíng Scientific name 学名 1239 Wallet
11 Canadian fish 加鱼 2732 Yúfúlíng Notebook 笔记本 1238 Wallet
12 Temporarily 暂时 2682 Yúfúlíng Sitting 坐姿 1237 Wallet
13 Hypoxia 缺氧 2608 Yúfúlíng piriformis piriformis 1232 Wallet
14 Extend 延长 2321 Yúfúlíng syndrome syndrome 1228 Wallet
15 Sprinkle over 撒过 2310 Yúfúlíng Ridicule 调侃 1224 Wallet
16 Originally 原本 2290 Yúfúlíng Sciatic 坐骨神经 1222 Wallet
17 Test 检验 2239 Yúfúlíng Syndrome 综合症 1217 Wallet
18 Shanghai 上海 2222 Yúfúlíng Face 正视 1184 Wallet
19 Serious harm 严重危害 2217 Yúfúlíng Wikipedia 維基百科 1096
20 Necessary 就要 2181 Yúfúlíng Website 网站 768
Word Word
Note that some word segmentation errors are present in Table 6:
The term Yúfúlíng (鱼浮灵) was segmented into two separate
words: Yú (fish) and fúlíng (floating spirit). Also, the first two
characters from the Chinese term for using Yúfúlíng was wrongly
segmented into “Canadian fish”. This reveals the challenges in
dealing with Chinese texts containing seldom-used terms. At the
same time, these can be used to detect new terms.
While the above findings indicate the “hot topics” on which
microblogs cite the two encyclopaedias, they say little about the
comments about the encyclopaedias. Thus the posts about the two
incidents are removed, producing the results listed in Table 7.
Table 7. Most frequently-occurring words when mentioning…
(two major incidents removed)
Counts Counts
1 Baidu 百度 2552 Website 网站 768
2 Share 分享 2418 Free(dom) 自由 669
3 China 中国 1784 China 中国 663
4 〝a bit〞一下 1135 United States 美国 660
5 Entry 词条 983 Forwarding 转发 608
6 Microblogging 微博 930 Encyclopedia 百科全书 533
7 Ency. 百科 683 Wiki 维基 498
8 One kind 一种 624 Google 谷歌 447
9 World 世界 618 〝a bit〞一下 444
10 Quote 引用 587 Baidu 百度 430
11 United States 美国 567 Microblogging 微博 423
12 Forwarding 转发 562 Foundation 基金会 401
13 Company 公司 551 google google 394
14 Search 搜索 536 Event 事件 382
15 Beijing 北京 507 Community 社会 377
16 Support 支持 501 Entry 词条 374
17 Introduction 介绍 494 World 世界 374
18 Activity 活动 485 Knowledge 知识 371
19 Network 网络 453 English 英文 366
20 Website 网站 441 Data 资料 359
Word
Baidu Baike Chinese Wikipedia
Word
Some of the words most commonly associated with Baidu Baike
included the verbs “share” and “quote,” the nouns “China,”
“entry,” and “Weibo,” and some geographic entities such as the
“United States,” “Beijing,” and “Japan.” With Chinese Wikipedia,
the list of words included “website,” “free(dom),” “China,” the
“United States,” “wiki,” “Google” (both in Chinese and in
English), “a bit,” “Baidu,” and “Weibo.” As will be discussed
later, the term “a bit” (一下) points to the phenomena where the
term “wiki” and other wiki-based encyclopaedias become verbs
that describe the action of using online encyclopaedias (e.g. to
wiki a bit”).
The same procedure that was used to extract the major discussion
threads and “hot topics” could be deployed again for obtaining
high-frequency words listed in Table 7. The words on the left
provide clues that lead to a series of prize-winning campaigns
initiated by Baidu Baike to promote itself on Weibo, accounting
for the high frequency of the term “share”. The words on the
right, including the terms “website” and “free(dom),” point to
Internet filtering and censorship in China.
The high frequency of the verb “share” for Baidu Baike is the
direct result of a 5-year anniversary campaign by Baidu Baike
that started in early April 2011. More than 85% of the instances
were simply sharing a certain entry article on Sina Weibo, with an
explicit note: “shared from @Baidu Baike”. Weibo users were
incentivized to promote Baidu Baike in this way by a chance to
win a prize in addition to gaining points in Baidu Baike. With the
help of machine-assisted clustering of the texts, I also identified
another set of more than 2600 posts that promoted Hudong Baike
and Baidu Baike with mentions of awards. (Like Baidu Baike,
Hudong Baike is another commercial user-generated
encyclopedia website hosted in mainland China[19]). In fact,
more than 1800 posts that came from Hudong Baike constituted
more than 30 different prizes from smartphones to chocolate. This
phenomenon also indicates certain forms of cross-spherical
activities between the platforms of Weibo and those of other user-
generated encyclopaedias.
The high frequency of the terms “free(dom)” and “websites” for
Chinese Wikipedia, in contrast, is largely the outcome of the
filtering and censorship experienced by Weibo users across
several websites or topics. Using terms such as “sensitive” (敏感),
“the Great Firewall” (GFW), “walled” (被墙), “climbed over the
wall” ( 翻墙 ), I found that many Weibo users share their
experiences on a range of websites and topics to gauge the
boundaries of filtering/censorship
Table 8 lists some of these posts. (Each ID number refers to a
particular post; negative numbers signal that the post is collected
from the WeiboScope; positive ones from the DiscoverText).
Hereafter I refer to a post by using the format of ID123 or ID-45.
For instance, one Weibo user questioned why her/his post from
Wikipedia to the Sina blog was deleted but survived in Sina
Weibo (ID4414) and another user claimed, “Wikipedia has
become classified as sensitive words in Sina Weibo” (ID-29805).
For others, Wikipedia become one of the tests of censorship, with
various access issues to its website and specific pages. In post
ID2667, a Weibo user shared her/his experience of being blocked
from visiting Wikipedia. Post ID-5667 described another users’
observation that the article on the Chinese Cultural Revolution
was unblocked but the article on a Chinese official was not. Post
ID1303 even suggested that, her/his experience of being blocked
probably because of several visits to sensitive articles on
Wikipedia, could somehow be resolved by visiting Baidu.

5
Table 8. Selected posts that mentioned CW in relation to
blockages or censorship
Category ID Main text translated
4414 ․Sina blog deleted my post copied from Wikipedia; is Sina blog more sensitive
than Sina Weibo??? The same post survive Weibo.
15243 ․[citing CW] … after the Wang Lijun incident, the name "Jiang Weiping" was no
longer blocked by Baidu or Sina Weibo.
-18987 ․Well, examples such as the "June Fourth" is blocked by Sina and Baidu as
sensitive historical events, but Wikipedia dares to tell the truth on historical facts.
-29805 ․Wikipedia has been listed as sensitive words in Sina Weibo [emoticon: pity]
Access
Issues
2667 ․Can any one open Wikipedia websites? Why after several pages of visit, I saw
"this page cannot be displayed". Nothing sensitive really ah.
-5667 ․The ban on Wikipedia is lifted, and the term "mosquito cells" [referring to
China's Cultural Revolution] is also removed from the list of sensitive terms. But
the recent ex-leader of Chongqing remains inaccessible.
1303 ․On Wikipedia, I visited a relatively sensitive topic, ended up with no access to
all pages. Later I entered Baidu's address by accident. As a result, I can go online
and search for information. Thank all-powerful "Baidu", my grateful tears of
i2609 ․Just finished the exam on Marxism, and then you will not let me to visit wiki
now?!! You crazy ah! I hate the Great Firewall!!
2997 ․After searching some politically sensitive terms as well as several major events
using Wikipedia, it feel really good, and should be closer to the truth. I have a new
perspective on the historical events. The same words on Baidu Search I find
"Search results are not displayable because they may not comply with relevant
laws, regulations and policies." Better use foreign websites, for they may be closer
to the truth. My mourning for the innocent people who were persecuted.
3270 ․What triggers the sensitive nerve system as Wikipedia becomes inaccessible?
Life is like a farce at times. Can never get the intention of the "superstructure"
[high officials]. Better to live in the moment and avoid commenting. Time is the
best encyclopaedia, older when we revisit, we'll realize other stories exist.
Sina as
censors
(blog or
Weibo)
Views
on the
isssue of
access
Still other users voiced their opinions on their encounters with
blockages or censorship when using Wikipedia. Post ID2609
seemed to express a student’s angry response for not being able to
access Wikipedia after a Marxism exam. Posts ID 2997 and
ID-18987 expressed their belief that Wikipedia was more likely to
be true for sensitive topics, whereas post ID3270 expressed a
more passive attitude about the futility of commenting.
This way of using high frequency words to identify major
discussion threads and “hot topics” can be iterated after removing
found sets of posts. The procedure can be useful to identify major
topics one by one. We can thus now move to the posts that
explicitly mention both Baidu Baike and Chinese Wikipedia.
3.2 When mentioning both
To zero in on the posts that explicitly compare Baidu Baike and
Chinese Wikipedia, we can now examine the posts that mention
both Baidu Baike (BB) and Chinese Wikipedia (CW). Figure 1
shows that there are 281 posts mention both.
Figure 1. The mentions of the three encyclopaedias
Users’ comparisons are most likely to occur in these posts, which
can highlight contrasting experiences and attitudes about the two
encyclopaedias Thus, focusing on the 281 posts that mention both
Baidu Baike and Chinese Wikipedia, two set of analysis can be
conducted to see (1) the relationship of users to the
censorship/filtering regime and (2) how Weibo and Twitter users
compare. I address these two avenues of analysis separately
because it is likely that the differences in censorship/filtering
between Weibo and Twitter influences users. To do so, I used
Chinese terms such as “wall,” “harmony,” and “river crabs” that
refer to the filtering and censorship regime to derive the first
sample. Then I manually coded the remaining posts for the second
analysis.
At least 21 posts specifically mentioned the keywords referring to
the censorship/filtering regime, of which 15 posts are listed in
Table 9 (the unlisted ones contained irrelevant content).
Table 9. Selected posts that mentioned both encyclopaedias (in
relation to filtering/censorship)
Category ID Main text translated
Baidu
Baike is
better
-13491 ․On "Socialism with Chinese characteristics", I responsibly vouch for the correct
version in Baidu Baike, Wikipedia's entry is simply nonsense! How could such an
unreliable site have a place in China? It should be thrown over the Wall.
-26194 ․For the people of China, visiting Baidu Baike is enough. Consider the risk of being
"Walled" when visiting Wikipedia.
Baidu
Baike as
alternative
-28952 ․Having no choice, Wikipedia is "Walled". Please use Baidu Baike as the
foundation.
4388 ․Wikipedia is more fair and comprehensive! I advise everyone not to use that
"harmonized" Baidu Baike!
5323 ․Some people died on foreign soils, but remained alive in the country. Some died in
Google, but still alive in Baidu. Some died in Weibo, but still alive on the CCTV
[Chinese Central TV]. Some died in Wikipedia, but still alive in Baidu Baike. Some
are dead outside the Wall, but inside the Wall still alive.
-5667 ․Some people died in a foreign country, but still alive in the country. Some died in
Google, but still alive in Baidu. Some died in Weibo, but still alive on the CCTV.
Some died in Wikipedia, but still alive in Baidu Baike. Some is dead outside the Wall,
but still alive inside.
1303 ․[On certain botany entry..] Opposite plants are treated as the same species in
Baidu Baike. What a miracle in this harmonious society ah! While accumulating
your wealth, Du-girl [feminized name for Baidu], would you please do something
substantial a bit? Access to Wikipedia blocked, Baidu Baike cannot be trusted. No
wonder many brain-damaged in this Heavenly Dynasty!
2609 ․Ever since on the other side of the Wall, I become used to using Google and
clicking on Wikipedia.. [@other user]:To be honest, I have now subconsciously use
Baidu because Google is often off due to some unknown forces. Besides, Baidu
Baike, Baidu Zhidao and Baidu Wenku indeed make a great contribution to the
diffusion of knowledge to the ignorant Netizens like me. [@yet another user] I
choose to wait another ten minutes when Google is not accessible.
2997 ․Baidu Baike not only copied but also harmonized the content from Wikipedia.
Much to see in Wikipedia once over the Wall.
3270 ․I always thought Baidu Baike is the best, because it understand Chinese language
very well! I was wrong! Completely wrong! In fact, Wikipedia is the best! And truly
the best when over the Wall!
3301 ․Better to visit Wikipedia over the wall . . . All lies in Baidu Baike.
3403 ․This "river-crabbed"[meaning censored] entry existed in Baidu Baike long time
ago, which tells us nothing about the Great Firewall. Whether we can visit certain
entries in Wikipedia is the standard test whether the "river-crab"(censorship) is
loosening.
Chinese
Wikipedia
is better
From the top to the bottom, the first category of posts showed
minority opinions of the selected dataset: one claimed that
Wikipedia’s specific entry on “Socialism with Chinese
characteristics” is nonsense when compared to the Baidu Baike’s
correct version (ID-13491). The other argued that Baidu Baike
was enough for people of China, citing the risk of being “Walled”
(ID-26194). The post in the second category shared a similar
opinion that Baidu Baike is an alternative when Chinese
Wikipedia is blocked. Of the majority of the posts that mention
both encyclopaedias in relation to filtering/censorship, the third
and most representative category consisted of posts arguing that
Chinese Wikipedia is better because it is not censored: Wikipedia

6
is "more fair and comprehensive" (ID4388), "the best" (ID-33992),
"have much to see" (ID-33991) and can serve as "the standard
test" of censorship (ID-40579), whereas Baidu Baike is censored
(ID5323) or "harmonized"(ID4388, ID-33991 and ID-40579) or
full of “lies” (ID-40570), repetition (ID2790) and errors (ID-
3473). Among them, clear anger was also expressed in posts
ID2790 and ID-24937 against access blockages.
Note that post ID-32406 discussed the decisions users face when
they are blocked. One user reflected on the gradual change to the
use of Google and Wikipedia that occurred when living outside
the Wall, and another comment reflected on the subconscious
shift to Baidu because of the Wall, while another made the
conscious decision to wait for ten minutes when “Walled”. These
comments clearly suggest that not all users of mainland China are
oblivious to the existence of Wikipedia and that their use and
reception of the websites inside and outside the Wall, while
diverse, is more or less conditioned by it. Concerning Baidu
Baike and Chinese Wikipedia, users clearly expressed contrasting
attitudes to the Great Wall, which also provides additional user
testimony on the impacts of the censorship/filtering regime
imposed on the two: Baidu Baike is often censored and Chinese
Wikipedia is blocked.
To compare how user opinions differ apart from in relation to the
censorship/filtering regime, I examined the remaining posts for
users’ evaluative terms about the two encyclopaedias, and found
39 posts (listed in Table 10). Of these, the majority found Chinese
Wikipedia to be better. Baidu Baike was reported to be less
reliable (e.g. post ID3502, 5733, 5917, 6210, -7323, -7371), less
objective (e.g. post ID3570, 6190, -1457, -10079, -13965, -19690,
-25687, -34458) and prone to government’s influence or
censorship (e.g. post ID5587, -19547, -19690, -21512, -25182,
-25221). In contrast, Chinese Wikipedia was perceived to be
more useful (e.g. post ID6306), better referenced (e.g. post
ID5733, -3755) and to have an elegant interface and style (e.g.
post ID3664). Nonetheless, some posts exist stated Baidu Baike
was better. Post ID-14431, for example, claims that Baidu Baike
is more detailed on R&B (rhythm and blues music). Post
ID-26631 suggested that Baidu Baike was richer in content
because it could “freely” copy content from other websites, and
thus had more comprehensive content that suits the needs of users
in China. Post ID-19336 seems to be ironic by agreeing that
Baidu Baike is thus “freerer” (I am not be certain whether it is
meant to be ironic).
Table 10. Selected posts that mentioned both (apart from
those concerning filtering/censorship)
Main text translated
-14431 ․On R&B, Baidu Baike is more detailed than Wikipedia.
-19336 ․It turns out Baidu Baike is freerer than Baidu Baike [Haha]
-26631 ․Users of China do not like Wikipedia. Why? For users who only read without
editing, they do not care whether the content is copied or original. All they want is
plenty and complete information. Because Wikipedia cannot copy but Baidu can
copy Wikipedia and others, Baidu Baike's content is richer than Wikipedia's.
-27322 ․You are welcome to read the essay penned by "Between Men&God", an
administrator of Chinese Wikipedia, published in Douban.com: "Why is (Chinese)
Wikipedia encyclopaedia worse than Baidu Baike?" (Chinese Wikipedia indeed
needs constant self-inspections. Wiki-girl approves this article [emotion:greivance])
16123 ․Chinese Wikipedia recently did something stupid. Competing with Vietnamese
Wikipedia on the number of entries. It appears that, outnumbered by Baidu Baike
and Hudong Baike, Chinese Wikipedia starts to enjoy itself by comparing with
minor languages. After Chinese Wikipedia criticized its competitors’ value of
quantity over quality, a rather ironic development that it enters the number game.
Both good -5667 ․Occasionally Baidu Baike is useful and very comprehensive. Wikipedia is better
when looking up English words.
Category ID
Baidu
Baike is
better
(Table 10 continued)
Chinese
Wikipedia
is better
1303 ․[On "Great Chinese Famine"] Better read more detailed Wikipedia. … Just also
checked Baidu Baike's, which is almost consistent with Wikipedia's. It is a rare
situation where Baidu Baike shows some progress.
2609 ․[On Zhang Zhixin] Huge difference between Baidu Baike and Chinese
Wikipedia. Which lies?
2997 ․The difference between Wikipedia and Baidu Baike? Can tell from one simple
entry on Fang Zhouzi..
3270 ․Stronger than Baidu Baike: Wikipedia
3301 ․My foreign teacher asked us to consult Wikipedia and share found information
in class. Finally realized that Baidu Baike is the epitome of the new online
"Sakoku" (isolationism). Although having less entries, Chinese Wikipedia has
absolute essentials, totally unlike Baidu Baike which copies and pastes stuff
everywhere. Better learn English well to enjoy the fun reading English Wikipedia!
3403 ․Wikipedia better than Baidu Baike. Checked the history of Haifeng, Wikipedia
has the same info as the county government website, but Baidu Baike is way off!
3502 ․Baidu Baike's content unreliable. Contributors often promote companies, books,
and so on. The entries on the Internet of things, Smart Industries, Industry clusters,
etc. mislead readers. I checked; none is comparable to English Wikipedia.
3570 ․For "Wukan incident", now we can visit Wikipedia's description, which is
objective, very detailed, encyclopaedic, and thus much better than Baidu Baike's
3664 ․Wow the Web interface is elegant and language is stylish in Wikipedia. No more
Baidu Baike for me from now on.
4021 ․Our history class assignment has the Red Guards as its theme, a bit heavy and
difficult. Everyone is advised to take a look at Wikipedia's and Baidu Baike's entry.
Chinese
Wikipedia
is better
5587 ․Very fun contrasting Wikipedia and Baidu Baike. Cannot open Wikipedia's
page because it contains the phrases such as "... regarded as a dissident for her
opposition to Mao's personality cult and extreme leftism". Can open it in Baidu
Baike, because its description goes like this. ...
5733 ․Baidu Baike not reliable. It appears to have complete information but all of
them unreferenced or self-referenced... Better use Wikipedia. At very least, each
Wikipedia's entry has clear references, respecting original authors' copyright.
5917 ․After reading the linked article, I search for information on "Pan Jinfu". With
the help of @hischild, I found this article. Baidu Baike's entry proved to be
"watery"... Wikipedia is more reliable. Of course, what is the most reliable is our
habits to look for original sources, which can help understand the facts.
6190 ․The description on Milton Friedman in Baidu Baike, in contrast to Chinese
Wikipedia, removed the parts where he was challenged and questioned by others.
6210 ․Scholars with expertise have conducted research and found the precision of
Wikipedia is higher than encyclopaedia Britannica, and thus can be treated as
reliable sources. Noted: Do not ever use Baidu Baike and Hudong Baike, where
false information is distributed, otherwise you will hurt yourself and others.
6306 ․Now I realized that Wikipedia is more useful than Baidu Baike.
-1457 ․Better use Baidu Baike less. Like Baidu search, it is castrated and inflicted by
viruses. ...Wikipedia has verification and coordination mechanisms and Baidu
Baike has not. With a political leaning, Baidu Baike is wildly unfair and chaotic.
-3755 ․Wikipedia first demands info sources, second requires footnotes in scholarly
formats, third prohibits [editors'] original research, and fourth says no to vandals
and ads. If Baidu Baike can do any of the above, it is no longer Baidu Baike
-7323 ․[Comments on the errors in Baidu Baike's prazepam entry...] Everyone please
use Wikipedia when consulting entries in professional fields such as medicine,
architecture, law, etc. Baidu Baike is far behind Wikipedia in terms of
professionalism and precision.
-7371 ․[On making a mistake regarding the age of an actor.] I recognized my mistakes
after googling. ... I realized this: Wikipedia is indeed better than Baidu Baike. It is
not about the word count, but about correct information
-7622 ․Fortunately Wikipedia is not wrong on this. When can Baidu Baike correct itself
to avoid repeating baseless assertions?
-10079 ․Today I bumped into a new term "neoliberalism". Not knowing what it is, I
looked it up in Baidu Baike and Wikipedia. After comparison, the orientation of
the two sites is clearly different. The former contains largely negative comments,
with emphasis on the opposition of socialism and capitalism... The latter does not
contain comments with ideological orientations, except for explaining all
knowledge backgrounds. ... From now on wiki only .... Baidu you garbage.
Chinese
Wikipedia
is better
-13965 ․Encyclopaedias are supposed to document facts. Have a look at the content in
Wikipedia and Baidu Baike on TVXQ (a South Korean pop group) and its five
members. I seldom visit Baidu, and today at a casual look I found it contains weird
stuff in it. Speechless. An encyclopaedia is used as a fan site, with all its biased
comments. Only Baidu is brain-damaged enough to allow such content. No chance
for such content to survive in Wikipedia's standard.
-16600 ․In practice, Wikipedia is proved to be more cultured than Baidu Baike.
-19547 ․Wikipedia is better than Baidu Baike. Baidu is a dog for the government.
-19690 ․On many topics, the results out of using Wikipedia is more objective than Baidu
Baike.
-21512 ․Both as encyclopaedias, why such a big different between the content from
Baidu Baike and that from Wikipedia? ...[sh!] [emoticon:depressed]
-25182 ․Baidu Baike does not have and Chinese Wikipedia has it. Immediately I know
why.
-25221 ․No such paragraph in Baidu Baike. So much content is different from
Wikipedia.
-25470 ․Ha ha ha ha ha Said the teacher. Wikipedia is more authoritative than Baidu
Baike!
-25687 ․Consult Wikipedia and Baidu Baike respectively on the topic of the "White Sea -
Baltic Canal". Contrast and see the obvious evil of Baidu.

7
Still, these comments do not contradict the observation that
overall, Chinese Wikipedia values content quality and respects
copyright more [17]. In fact, the essay mentioned by post
ID-237322 is originally titled: “What kind of improvements does
(Chinese) Wikipedia need when compared with Baidu Baike.” In
the essay mentioned, before summarizing his opinions that
Chinese Wikipedia’s content is too limited, the page editor too
complex to use, and it is too unfriendly for newcomers, the writer
(an administrator of Chinese Wikipedia) acknowledged that much
has already been said about where Chinese Wikipedia is better.
Altogether, these posts represent a sample of opinions voiced in
both Weibo and Twitter, suggesting that Chinese Wikipedia is
perceived as more reliable and less censored than Baidu Baike. In
particular, posts ID3301 and ID-25470 use teachers’ opinions, and
the former shares her/his class experience in using Wikipedia as
that “Baidu Baike is the epitome of the new online ‘Sakoku’”
(isolationism). The two posts are amongst many posts that
mentioned Baidu Baike or Chinese Wikipedia as one of the
important digital tools for improving one’s understanding of the
world. The idea that online encyclopaedias are important tools for
gaining information is further supported by the fact that the
Chinese term “wiki” has become a verb similar to the verb “to
google”.
3.3 When “wiki” becomes a verb.
Indeed in the dataset, like “google” in English, “Baidu" and the
Chinese term “wiki” (维基) are used as verbs that refer to visiting
these encyclopaedias for information. I found nearly 200 posts for
each of the Chinese-language phrases “Baidu a bit” and “Wiki a
bit.” To find out what kinds of “verbification” are expressed, I
categorize these posts according to the preceding words because
they provide contexts for such verbification. Table 11 lists the
categorized results.
Table 11 Baidu/Wiki a bit
Precedent Word context Precedent Word context
Action done 刚(2) 刚(11) ...just (recently) Wiki a bit…
剛(3) ...just (recently) Wiki a bit…
刚/剛才(2) ...just (recently) Wiki a bit…
于是(4) ...then Baidu a bit… 顺手(2) ...casually Wiki a bit…
特意(3) ...purposely Baidu a bit… 重新(2) ...again Wiki a bit…
再(2) ...again Wiki a bit…
Action 去(17) ...go Baidu a bit… 去(10) ...go Wiki a bit…
可(10) ...can Wiki a bit…
能(10) ...can Wiki a bit…
可以(8) ...can Baidu a bit… 可以(3) ...can Wiki a bit…
先(6) ...first Baidu a bit… 不会(4) ...why not Wiki a bit?…
要(2) ...must Baidu a bit… 为啥不(2) ...why not Wiki a bit?…
只要(2) ...simply Baidu a bit… 最好是(1) ...better Wiki a bit…
请(2) ...please Baidu a bit…
大家(4) ...everyone Baidu a bit… 大家(2) ...everyone Wiki a bit…
我(3) ...I Baidu a bit…
你(3) ...you Baidu a bit… 你(3) ...you Wiki a bit…
同学(2) ...classmate, Baidu a bit… @(5) ...@[a Weibo and Twitter user]…
Promotion 都(23) ...all Baidu a bit…
百度(8) ...Baidu/Wiki a bit…
或者(6) ...or Wiki a bit…
和(3) ...Baidu and Wiki a bit…
Action
combined
Category
Action
better
required
"Baidu a bit" (百度一下) "Wiki a bit" (维基一下)
...just (recently) Baidu a
bit…
Action
possible
Pronoun
and noun
The four categories in Table 11 distinguish between when an
action is done, will be done, can be done, or should be done.
More than 20 posts refer to the posters’ past actions (the category
“Action done” in the table) in using one of the encyclopaedias for
information, and there are more such posts for Chinese
Wikipedia: e.g. “… just recently wikied a bit”. This suggests that
some users share their experiences using online encyclopaedias
on microblogs. In addition, the categories “Action possible” and
“Action better required” refers to when users make the suggestion
that online encyclopaedias should be consulted.
Note that there are eight posts that use the combined verb
“Baidu/Wiki a bit” (百度维基一下), with an example post quoted
below (ID-31873):
The masses should take advantage of this opportunity to
popularize this universal knowledge ah! If one does not believe
in experts, one can Baidu/Wiki a bit.
One instance that distinguishes Baidu Baike and Chinese
Wikipedia is when the verb “Baidu a bit” is used by Baidu to
promote itself in Weibo. At least 23 posts, most likely generated
by some Weibo users who participated in promotion campaigns,
followed the format:
Got news? Brush [meaning "refresh" or “update” in English]
Weibo. Got new knowledge? Brush Weibo too. Because what
matters most for communication is brush something anew. It is
most foreign flavour (meaning “exotically fashionable” in
English) to all Baidu a bit on everything. What you do not know,
@BaiduZhidao let you know. I am @UserID [Weibo user] at
Weibo.com. Got something to ask? Don't ask me. Ask @Baidu
Baike.
What the above post demonstrates is exactly the crossover of
communicative and information spaces, where the information
spaces of Baidu Baike and Baidu Zhidao [a Baidu user-generated
questions and answers website,] can be used to update new
information in the communicative spaces of Weibo. Although
such cross-over or cross-spherical activities may in this case not
be completely voluntary, there is ample evidence that Weibo or
Twitter users share their experiences in using search engines,
user-generated encyclopaedias and microblogs across platforms.
Some of them have also noticed the relationship between the
SERPs and UGEs:
…As matter of fact, you will know by simply Baidu a bit. Key
in three-character “Yinbingshi” to search, the first item should
be Baidu Baike’s entry on Yinbingshi…. (ID-33642)
Google is sensitive….search “lie group” one can find Lie
group-Wikipedia, the free encyclopaedia as the top result.
(ID-23435)
…Simply Baidu a bit, Baidu Baike has a specific entry on him….
(ID-34850)
…Amazing that through Baidu I found the “sensitive”-men (i.e.
Tiananmen) incident twenty years ago on a sensitive day and
sensitive month in a sensitive year. The first SERP item is
Wikipedia’s entry page on this incident, and it can be opened.
(ID-42437)
Thus in general, what is shared between “Baidu a bit” and “Wiki
a bit” as a verb is the specific use of Baidu Baike and Chinese
Wikipedia: looking for information. I found no cases where the
verb is used to refer to writing or contributing content to them.

8
Thus, in the Chinese-language context, as shown by the text
analysed here, when used as verbs, the terms “Baidu a bit” and
“Wiki a bit” refer to general information-seeking actions, not to
the original meaning of wiki for collaborative writing. The posts
clearly indicate the prominence of using Wikipedia and/or Baidu
Baike as information practices on par with Google or Baidu.
4. DISCUSSION
Ample evidence exists to show the existence of cross-platform
activities. Weibo and Twitter users are often aware of both
encyclopaedias. Research must take such cross-platform activities
into account, as suggested by some digital method researchers
[27]. Such cross-platform activities also include promotion
campaigns initiated by Baidu Baike and Hudong Baike, and even
cyber-attacks against Wen Yunchao’s Twitter account by flooding
tweets using Baidu Baike’s entry, as also testified by Wen
Yunchao himself at a US Senate hearing:
….For about a year starting April, 2011, unidentified persons
“tweet bombed” me on Twitter with trash information. Using a
software [program] called Tween to filter the trash, I found the
heaviest attack took place on April 25, 2012 – with a staggering
590,000 spam posts within 24 hours. Unidentified persons also
posted viciously defaming information about me online at the
rate of over 10,000 times per day. As far as I know, artist Ai
Weiwei has been similarly attacked ….
Thus, the Twitter data collected for this research (from March 20,
2012 to April 12, 2012) contains the early phase of the spam post
attack against Wen Yunchao on Twitter. Encyclopaedia pages’
titles and links, mostly from Baidu Baike, can be used to jam a
Chinese activist’s account in Twitter that is outside Beijing’s
filtering and censorship regime. This is evidence of information
engagement in the sense of information warfare. The availability
of user-generated encyclopaedia data is repurposed to jam a
microblog account by introduced introducing noise to keep
certain Chinese activists from engaging their intended audience.
While this exploratory study had not yet employ a method to
detect and then remove such jamming data points, it is feasible for
future research to identify the potential abuse of user-generated
encyclopaedias on microblogs. Posts that keep mentioning user-
generated encyclopaedias’ different articles in a short timeframe
targeting individual accounts are likely to be jamming,
Also, the systematic method of first clustering posts (by
computers programs) and then categorizing them (by analysts)
one by one through identifying and then removing the hot topics
should also help detect other spamming or en-masse promoting
activities, as demonstrated by the self-promotion campaign
initiated by Baidu Baike and Hudong Baike. Although the
approach of the study may be questioned on the grounds of
accuracy and validity, it has several merits that traditional
methods such as user interviews cannot rival. First, this approach
is feasible and reproducible enough for constant monitoring.
Dynamic reports could be quickly generated with a workflow
consisting of computer programs and social media analysts.
Second, this approach may be better to anticipate or even predict
the hot topics of public discussions that use or comment on user-
generated encyclopaedias. Third, the accuracy and validity of
findings can be further enhanced through better programs and
more experienced analysts after several iterations.
The use and reception of the two major user-generated
encyclopaedias on Weibo and Twitter yielded several findings.
First, it confirmed that some Weibo and Twitter users use Chinese
Wikipedia and/or Baidu Baike, and that Beijing’s filtering and
censorship regime has different effects on these uses, thereby
contributing to the existing research on online Chinese-language
encyclopaedias [16, 20, 22]. Several users were also aware of the
differences between the two encyclopaedias. In fact, several posts
suggest the contrast between Baidu Baike and Chinese Wikipedia
as a test for tracking what counts as a sensitive topic. While users
did sometimes get frustrated when they were denied access to
Chinese Wikipedia, many of them preferred Chinese Wikipedia
precisely because it was not censored. Second, overall, users had
positive experiences using online encyclopaedias, and more so in
using Chinese Wikipedia than in using Baidu Baike. For example,
Chinese Wikipedia was perceived to value content quality and
respect copyright more, in contrast to Baidu Baike, which was
known for copying and pasting content from other websites.
Explicit comparisons were also made by Weibo and Twitter users
in short posts, and more users perceived Chinese Wikipedia to be
of higher quality. Third, like the new English word “google”
being used as verb, the Chinese terms “Baidu” and “Wiki” are
also used as verbs to describe looking up content in Baidu Baike
or Wikipedia. This suggests that the use of online user-generated
encyclopaedias have become an important part of users’ online
experience, including when users cite information to discuss
topical issues, request that others find information, or verify
information seen elsewhere. For these users, user-generated
encyclopaedias were resources they relied on.
Moreover, the findings show that many users view the two
resources as an established part of everyday information
activities, like using the search engines Google and Baidu. These
activities are often linked with information or digital literacy, as
exemplified by comments from students using Wikipedia in
classroom settings. Here we can see the expectation that user-
generated encyclopaedias lead to the betterment of self and
society through gaining knowledge, partly expressed by a post
suggesting that “[t]he masses should take advantage of this
opportunity to popularize this universal knowledge ah! If one
does not believe in experts, one can Baidu/Wiki a bit”
(ID-31873). The new experiences with information and
communication spaces are increasingly associated with these
websites, as are cultural and social norms about information
access, knowledge sharing, and content filtering and censorship.
The online practices of information engagement are thus expected
to be influenced by users’ experiences with different platforms,
which are of cultural-political significance for the Chinese-
language Internet [18].
The findings also have implications for youth engagement in
mainland China since Weibo users in mainland China are more
likely to be younger. More than 50% of Sina Weibo users were
26-35 years old according to an industry report in 2012 [7] and
more than 80% of the visiting traffic came from 10-39 year old
users according to an annual report released by Chinese Academy
of Social Sciences in 2014 [29]. In light of the rough user profile,
it can be argued that state and market players are expected to
engage the younger generation in mainland China through
microblog platforms, which may explain why the information
warfare kind of information engagement is observed in the case of
jamming microblog accounts of Chinese activists such as Wen
Yunchao. It may also explain why the learning aspect of citing,
using, and sharing experiences about user-generated
encyclopaedias is also important for several microblog users.

9
5. CONCLUSIONS AND IMPLICATIONS
In terms of information engagement, this study finds that
microblog users cite major user-generated encyclopaedias to
engage in information-critical or information-intensive public
discussions. In a broadened sense of open collaboration, the open
discussion on microblog platforms use information resources from
user-generated encyclopaedia platforms.
The dynamics of the information engagement concerns the largest
non-English segment of the world’s Internet population, and this
paper has examined how it is practised and expressed on two
major Chinese-language microblog platforms about two major
Chinese-language encyclopaedias. The findings demonstrate the
potentials and gaps towards more active and sound collaborative
information engagement beyond mere information seeking and
consumption. The microblog posts were still predominantly about
using encyclopaedias (in a way contributing to better informed
users and public discussions), not about contributing back to the
content of user-generated encyclopaedias.
Still, both Baidu Baike and Chinese Wikipedia enjoy mentions
from Twitter and Weibo users, and both have become verbs in the
Chinese language just like the new English verb “google.” The
Chinese verbification of the two encyclopaedias thus indicates
what has been achieved and what has yet to be achieved. More
research and practice is needed to advance the understanding of
the potentials and the challenges to improve information
engagement on both platforms, possibly using microblog
discussion to improve encyclopaedia platforms and their content
as well as other open collaboration activities that are more
“egalitarian,” “meritocratic” and “self-organizing” [8].
With regards to the recent efforts by Chinese Wikipedians to
engage users on Sina Weibo [31], the findings point to a direction
in which users engage in public discussions on microblog using
and enhancing encyclopaedia articles. The research methods here
can be repurposed and streamlined to build a public opinion
monitoring system. First, robust jam- or spam-detection software
can be developed and implemented to highlight the abuse and/or
misuse of user-generated encyclopaedias. In addition, it might be
possible to show (or even predict) which incidents may become
hot discussion topics that require quality and authoritative
information sources that may or may not be found in major user-
generated encyclopaedias. Future work can be conducted to
explain, predict, or even plan interventions to the information-
related activities that can be publicly observed across the
microblog platforms and user-generated encyclopaedia platforms.
Overall, the development of certain event monitoring devices,
information engagement rules and new encyclopaedia articles
may result in better public discussion and better human
knowledge. This may in turn translate into more kinds of civic
information engagement and less information warfare.
A division of users and user experiences currently exists in
mainland China characterized by the choice between Baidu Baike
and Chinese Wikipedia. While the frequent occurrences of the
verbs such as “Google a bit”, “Baidu a bit”, and “Wiki a bit”
suggest that user-generated encyclopaedias are part of users’
overall information-seeking activities, the choice between Baidu
Baike and Chinese Wikipedia entails different norms and
attitudes. Baidu Baike was often associated with the Beijing-
censored version of information sources and Beijing-sanctioned
ways of information engagement. On the other hand, Chinese
Wikipedia is not censored and thus was perceived by some to be
more reliable and was associated with the freedom of information.
Access blockages to Chinese Wikipedia were also perceived as
signals of Beijing’s political “sensitivity” towards information.
The perspective of learning to find and cite information in order
to participate in public discussions shows the potential of social
media platforms as “civic learning repertoires” for “information
engagement” [2]. For users in mainland China, at least two major
learning strategies can be identified: one takes the censorship and
filtering regime for granted, and the other seeks information and
communication platforms that are outside the influence of the
regime. It is likely that the “civic learning repertoires” for
“information engagement” are limited to the bounds of the regime
(i.e. mainland China) for those who use Baidu and Baidu Baike.
Indeed, both Baidu Baike and Chinese Wikipedia become
important reference points and even everyday communicative
practices (e.g. “Baidu a bit” or “Wiki a bit”). Still, the different
uses and receptions suggest two set of learning experiences and
information choices. For contributors and readers, the choice of
one over the other reflects the different experiences in civic
learning made possible by the two websites.
The findings add to the cross-spherical analysis of web spaces
proposed by Richard Rogers[27]. Because of the citing and
commenting activities, Web spaces are indeed porous between
Chinese-language microblog and encyclopaedia platforms. On the
other hand, the distinct choice between Baidu Baike and Chinese
Wikipedia suggests not only different information engagement
practices because of the censorship/filtering regime, but also
different civic learning repertoires. The censorship/filtering
regime clearly has impacts on shaping web spaces and their
information engagement activities.
The findings also update our understanding of the online “Chinese
cultural sphere” of which Chinese-language resources are
important[32]. Chinese-language platforms such as microblogs
and user-generated encyclopaedias constitute major instances
where such resources can be indexed, searched, retrieved and
viewed amongst mostly Chinese-language users. Such openly
available resources permit potentially new information
engagement and even open collaboration. However, such
potentials may be limited by censorship (e.g. Baidu Baike) or
blockage (e.g. Chinese Wikipedia), shaping the engagement
dynamics with different sets of norms and knowledge sources.
Future research may also consider how the concept of information
engagement intersect with the official and popular discourse of
the Chinese term “suzhi" (quality) of people, a discourse that
demands individual competition, responsibility and self-
improvement—including the ability to use technology to do so. In
the dataset, several posts discussing the lack of “quality”
information and people can be further analysed. Since the notion
of information engagement is relatively unknown to the general
public in China, the notion of suzhi can be used as potential
bridging concept to discuss which kind of information
engagement is better in promoting knowledge for the betterment
of Chinese individuals and societies. Some questions can also be
formulated based on the existing Chinese studies literature on
“suzhi” for online platforms. For instance, Suzhi discourse may
have aimed to turn peasants into modern Chinese citizens [23],
but how does it work, say, to turn migrant workers into modern
online Chinese citizens? If suzhi discourse justified social
hierarchies [14], how is it related to the choices of information
platforms and the practices of information engagement? More
empirical research is thus needed to examine and compare across
different information platforms and groups of users.

10
Important links are shown between public discussions and
information engagement in this exploratory case study. Since both
platforms of microblogs and encyclopaedias depend on users and
their contributed content, users learn to use and choose platforms
while platforms learn from users. Future research and practices
are needed, with better information engagement and content, to
advance the quality of public discussions.
6. ACKNOWLEDGMENTS
I acknowledge the WeiboScope team and Dr. Fu King-wa at the
University of Hong Kong for the provision of the WeiboScope
dataset. I also acknowledge Texifter.com for the trial of their
platform that allowed me create the DiscoverText-Weibo dataset.
7. REFERENCES
[1] Bamman, D. et al. 2012. Censorship and deletion practices in
Chinese social media. First Monday. 17, 3-5 (2012).
[2] Bennett, W.L. and Wells, C. 2009. Civic Engagement: Bridging
Differences to Build a Field of Civic Learning. International
Journal of Learning and Media. 1, 3 (Aug. 2009), 1–10.
[3] Bolsover, G. et al. 2013. Social Foundations of the Internet in
China and the New Internet World: A Cross-National
Comparative Perspective. (Oxford Internet Institute, University
of Oxford, Jun. 2013).
[4] Chen, B. et al. 2006. Chinese Spoken Document Summarization
Using Probabilistic Latent Topical Information. 2006 IEEE
International Conference on Acoustics, Speech and Signal
Processing, 2006. ICASSP 2006 Proceedings (May 2006), I–I.
[5] CNNIC 2013. Statistical Report on Internet Development in
China (January 2013). Technical Report #31st. China Internet
Network Information Center.
[6] Damm, J. 2008. Technological Empowerment: The Internet,
State, and Society in China. By Yongnian Zheng. Stanford,
Calif.: Stanford University Press, 2008. xviii, 272 pp. $50.00
(cloth). The Journal of Asian Studies. 67, 03 (2008), 1080–1081.
[7] DCCI 2012. Blue Book of China MicroBlog. Data Center of
China Internet (DCCI).
[8] Definition of Open Collaboration: 2012.
http://www.wikisym.org/2012/09/28/definition-of-open-
collaboration/. Accessed: 2013-03-26.
[9] Fu, K. et al. 2013. Assessing Censorship on Microblogs in
China: Discriminatory Keyword Analysis and Impact
Evaluation of the “Real Name Registration” Policy. IEEE
Internet Computing. 17, 3 (2013), 42–50.
[10] Fu, K. and Chau, M. 2013. Reality Check for the Chinese
Microblog Space: a random sampling approach. PLOS ONE. 8,
3 (2013), e58356.
[11] Hagestad II, W. 2012. 21st Century Chinese Cyberwarfare. IT
Governance Ltd.
[12] iResearch 2013. Sina’s Weibo New Advertising Model Marks
Monetization of the Long-tail Marketing Strategy. iResearch
Consulting Group.
[13] King, G. et al. 2012. How censorship in China allows
government criticism but silences collective expression. APSA
2012 Annual Meeting Paper (2012).
[14] Kipnis, A. 2006. Suzhi: A Keyword Approach. The China
Quarterly. 186, (Jun. 2006), 295.
[15] Liao, H.-T. et al. 2013. Chinese Web Data: Definition, Uses,
and Scholarship.
[16] Liao, H.-T. 2009. Conflict and Consensus in the Chinese
version of Wikipedia. IEEE Technology and Society Magazine.
28, 2 (2009), 49–56.
[17] Liao, H.-T. 2013. How do Baidu Baike and Chinese Wikipedia
filter contribution? A case study of network gatekeeping.
Proceedings of the 9th International Symposium on Open
Collaboration (WikiSym + OpenSym 2013) (Hong Kong, Aug.
2013).
[18] Liao, H.-T. 2014. The Cultural Politics of User-Generated
Encyclopedias: Comparing Chinese Wikipedia and Baidu.
University of Oxford.
[19] Liao, H.-T. 2014. “User Generated Content” (用户生成内容).
The Internet in China. Berkshire Publishing.
[20] Lou Z.-C. and Fu Z. 2008. The Impact Analyze of Some
External Forces on the Wikipedia Order Process (外部因素对
维基百科序化过程的影响分析). Document, Information &
Knowledge (圖書情報知識). 123, (May 2008), 28–33.
[21] Luo, T.(罗婷) and Li, C.(李成) 2010. How to develop public
opinion monitoring products? - The modes of operation of the
Public Opinion Monitoring Unit, People’s Net. [如何开发舆情
监测产品?——人民网舆情监测室的运作模式]. Chinese
Journalist (中国记者). 6 (2010).
[22] Ma, F. and Xia, Y. 2008. Mechanism of ordering in Wikipedia
based on the CAS theory. Library Tribune. 28, 6 (2008), 85–92.
[23] Murphy, R. 2004. Turning Peasants into Modern Chinese
Citizens: “Population Quality” Discourse, Demographic
Transition and Primary Education. The China Quarterly. 177,
(Mar. 2004), 1–20.
[24] Netpop Research 2007. Chinese consumers have dramatically
surpassed Americans in adopting Web 2.0 behavior, relying
heavily on social media for guidance in purchase decisions.
Netpop Research.
[25] Porche, I. et al. 2013. Redefining Information Warfare
Boundaries for an Army in a Wireless World. Rand Corporation.
[26] Qiang, X. 2010. User-generated content online now 50.7% of
total. China Daily.
[27] Rogers, R. 2013. Digital methods. The MIT Press.
[28] Tan, Y.F. 2012. Online public opinion monitoring: an
exploratory report [探照网络舆情监测]. CAIJING (财经).
[29] Tang, X. 2014. Annual Report on Development of New Media
in China (2014) 《中国新媒体发展报告》. Technical Report
#5. Institute of Journalism and Media, Chinese Academy of
Social Sciences (CASS).
[30] United States Department of the Army 2009. Tactics in
Counterinsurgency. Mundus Publishing.
[31] Wang, A. 2013. Using social media to engage Wikipedia
readers and editors in China. Wikimedia blog.
[32] Yang, G. 2003. The Internet and the Rise of a Transnational
Chinese Cultural Sphere. Media Culture Society. 25, 4 (Jul.
2003), 469–490.
[33] Yang, Y. and Xie, L. 2008. Subword Latent Semantic Analysis
for Texttiling-Based Automatic Story Segmentation of Chinese
Broadcast News. 6th International Symposium on Chinese
Spoken Language Processing, 2008. (Dec. 2008), 1–4.
[34] Yu, Z.-T. et al. 2006. Answer Extracting for Chinese Question-
Answering System Based on Latent Semantic Analysis. Chinese
Journal of Computers 《计算机学报》. 29, 10 (2006), 1889.
[35] Zheng, Y. 2007. The Internet, Civic Engagement, and Public
Distrust. Technological Empowerment: The Internet, State, and
Society in China. Stanford University Press.
[36] Zhu, T. et al. 2012. Tracking and Quantifying Censorship on a
Chinese Microblogging Site. arXiv:1211.6166. (Nov. 2012).

What do Chinese-language microblog users do with Baidu Baike and Chinese Wikipedia? A case study of information

More Related Content

What's hot (20)

Similar to What do Chinese-language microblog users do with Baidu Baike and Chinese Wikipedia? A case study of information (20)

What do Chinese-language microblog users do with Baidu Baike and Chinese Wikipedia? A case study of information