Search Concentration, Bias & Parochialism: Comparative Study of Baidu, Google, Jike
1. Search Concentration, Bias & Parochialism:
A Comparative Study of Baidu, Google &
Jike’s Search Results from China
Min
Jiang
(Ph.D.)
Associate
Professor
Of
Communica:on,
UNC
Charlo?e
Affiliate
Researcher
Center
for
Global
Communica:on
Studies,
Univ.
of
Pennsylvania
TPRC
Conference
|
George
Mason
Law
School
|
September
2013
6. Chinese Search Engine Market
• 591 million Internet users (CNNIC, 2013)
• 450 million search engine users (CNNIC, 2013)
• Market size: $1.5 billion (iResearch, 2013)
• Google exit from mainland China in 2010
11. Search Concentration, Bias, Parochialism
• Search concentration
• Tendency to drive traffic to well-established sites
• “Rich-get-richer,” popularity over quality
• Search bias
• Tendency to favor their own content over others’
• “Own-content bias” & “other-content bias”
• Search Parochialism
• Tendency to direct traffic to domestic sites
• Parochialism vs. cosmopolitanism
12. Operationalization
• Search concentration
• Percentage of search results concentrated in top 5 and
top 10 websites (e.g. 50% results from top 5 sites)
• Search bias
• Own-content bias: favored inclusion and ranking of
search engine’s own content
• Other-content bias: exclusion and lowered ranking of
rivals’ content
• Search Parochialism
• Percentage of search results from domestic sites
• Presence of operating license issued by Chinese authorities
14. Data Collection
• Information retrieval
• Infer search characteristics from search results patterns
• Challenge of search personalization
• Machine interprets user relationship to reality & select info
• Time
• 2011 August & 2012 August
• Procedure
• Two researchers in southern China city
• Same location, same router, different laptops
• IE browser, disable cookie (except Google CN)
• High overlaps in search results between researchers
• Google 88%, Baidu 90%, Jike 98%
15. Query Sample
• Top 20 Chinese Internet events of 2010
• Tencent vs. 360 Dispute
• Shanghai Expo
• Internet celebrity “Sister Feng”
• Li Gang’s son’s drunken hit run kills student on campus
• Foxconn suicides …
• 15 general terms
• Transportation, military, medicine, blog, entertainment,
school, government, news, tourism, fashion, plane
ticket, car, law, economy, music
16. Distribution of Top 20 Events Search Results
Among Popular Chinese Websites (2011)
Baidu
2011
Google
2011
Jike
2011
1 Baidu
(22)
Sina
(13)
Xinhua
(13)
2 Sina
(21)
Sohu
(13)
QQ
(12)
3 Sohu
(18)
QQ
(10)
Sohu
(11)
4 QQ
(15)
NetEase
(10)
NetEase
(10)
5 NetEase
(15)
ifeng
(9)
Sina
(9)
6 ifeng
(9)
Baidu
(6)
360doc
(4)
7 Xinhua
(6)
People’s
Net
(6)
People’s
Net
(4)
8 Ku6
(5)
Huanqiu
(5)
Huanqiu
(3)
9 Huanqiu
(4)
Ku6
(4)
ifeng
(3)
10 Youku
(2)
Youku
(4)
Gov.cn
(2)
Top 5 Sites 45.5%
30.6%
21.0%
Top 10 Sites 58.5%
44.4%
35.5%
17. Distribution of Top 20 Events Search Results
Among Popular Chinese Websites (2012)
Baidu
2012
Google
2012
Jike
2012
1 Baidu
(32)
Sina
(14)
QQ
(21)
2 Sina
(24)
ifeng
(11)
Sina
(20)
3 Sohu
(18)
QQ
(10)
Sohu
(13)
4 NetEase
(16)
NetEase
(10)
Xinhua
(9)
5 ifeng
(11)
Baidu
(8)
NetEase
(8)
6 QQ
(9)
Google
(8)
Youku
(8)
7 Ku6
(5)
Wikipedia
(8)
ifeng
(5)
8 Huanqiu
(4)
Sohu
(7)
Baidu
(5)
9 People’s
Net
(3)
People’s
Net
(7)
Huanqiu
(4)
10 Xinhua
(3)
Xinhua
(5)
Ku6
(4)
Top 5 Sites 50.5%
29.4%
35.5%
Top 10 Sites 62.5%
44.0%
47.5%
18. Distribution of 15 General Terms Search Results
Among Popular Chinese Websites (2011)
Baidu
2011
Google
2011
Jike
2011
1 Baidu
(43)
NetEase
(9)
Gov.cn
(9)
2 Sina
(6)
Baidu
(8)
QQ
(8)
3 NetEase
(5)
People’s
Net
(6)
Sohu
(8)
4 Hao123
(5)
Sina
((7)
ifeng
(8)
5 Sohu
(5)
Sohu
(7)
Trends
(6)
6 QQ
(4)
Huanqiu
(5)
NetEase
(5)
7 ifeng
(4)
Hexun
(5)
Sina
(5)
8 Hexun
(2)
ifeng
(4)
Baidu
(4)
9 Huanqiu
(2)
Google
(4)
Hao123
(4)
10 CNTV
(2)
Xinhua
(2)
ifeng
(3)
Top 5 Sites 42.7%
24.7%
29.3%
Top 10 Sites 52.0%
37.3%
43.3%
19. Distribution of 15 General Terms Search Results
Among Popular Chinese Websites (2012)
Baidu
2012
Google
2012
Jike
2012
1 Baidu
(40)
Google
(8)
Baidu
(18)
2 Gov.cn
(11)
Sina
(7)
Sina
(8)
3 Sina
(6)
QQ
(7)
QQ
(7)
4 Huanqiu
(5)
NetEase
(7)
Hudong
(7)
5 Hao123
(4)
Huanqiu
(6)
Sohu
(6)
6 Huanqiu
(4)
Sohu
(6)
NetEase
(5)
7 ifeng
(3)
Baidu
(5)
Hexun
(4)
8 QQ
(3)
ifeng
(4)
Ifeng
(3)
9 People’s
Net
(3)
Wikipedia
(4)
Gov.cn
(3)
10 Xinhua
(3)
Hexun
(3)
Huanqiu
(2)
Top 5 Sites 44.0%
23.3%
30.7%
Top 10 Sites 54.7%
38.0%
42.0%
20. Online Encyclopedia Services In Search Results (2012)
Top
20
Events
Baidu
Google
Jike
Google
(0)
Google
(8)
Google
(0)
Baidu
Baike
(7)
Baidu
Baike
(5)
Baidu
Baike
(2)
Wikipedia
(0)
Wikipedia
(5)
Wikipedia
(1)
Hudong
(0)
Hudong
(4)
Hudong
(3)
15
General
Terms
Baidu
Google
Jike
Google
(0)
Google
(8)
Google
(1)
Baidu
Baike
(16)
Baidu
Baike
(2)
Baidu
Baike
(13)
Wikipedia
(0)
Wikipedia
(4)
Wikipedia
(0)
Hudong
(0)
Hudong
(4)
Hudong
(7)
21. Distribution of Overseas Search Results from Baidu, Google, Jike
Top
20
Events
(2011)
Baidu
Google
Jike
WSJ
CN
(2)
Google
(2),
Wikipedia
(2)
Zaobao
(2)
Epoch
Times
(2),
Taiwan
(2)
WSJ
CN
(1)
Google
HK
(1),
Blogspot
(1)
Google
(1)
Youtube
(1),
Zaobao
(1)
German
(1),
Canadian
(1)
1% 7%
2%
Top
20
Events
(2012)
Baidu
Google
Jike
WSJ
CN
(1)
Wikipedia
(8)
Zaobao
(1)
Google
(1)
Wikipedia
(1)
Canadian
(1)
FT
CN
(1)
0.5%
5%
1.5%
22. Search Concentration, Bias, Parochialism
• Search concentration
• More visible in Baidu; in news queries than general terms
• Chinese commercial Internet giants; state websites
• Search bias
• Most palpable in Baidu, less so in Google or Jike
• Online encyclopedia: Baidu Baike, Hudong, Wikipedia
• Search Parochialism
• Google more likely to direct traffic to overseas sites
• Decreasing diversity of overseas sites in Google
23. Make Our Means Match Our Ends
“Our inventions are wont to be pretty toys,
which distract our attention from serious
things. They are but improved means to an
unimproved end.”
- Henry David Thoreau