Image needs on social Q&A sites: a comparison of Zhihu and Baidu Zhidao
| Date | 03 June 2019 |
| Pages | 454-473 |
| Published date | 03 June 2019 |
| DOI | https://doi.org/10.1108/EL-09-2018-0192 |
| Author | Shengli Deng,Anqi Zhao,Ruhua Huang,Haiping Zhao |
| Subject Matter | Information & knowledge management |
Image needs on social Q&A sites:
a comparison of Zhihu and
Baidu Zhidao
Shengli Deng
The Center for the Studies of Information Resources of Wuhan University,
Wuhan, China, and
Anqi Zhao,Ruhua Huang and Haiping Zhao
School of Information Management, Wuhan University, Wuhan, China
Abstract
Purpose –This study aims to examine why users searchfor images, how users describe their image needs
and what the images areused for by analysing questions obtained from two Chinesesocial Q&A sites, Zhihu
and Baidu Zhidao.
Design/methodology/approach –A total of 1,402 image questions were collected from Zhihu and
Baidu Zhidao. Both quantitative analysis and qualitative content analysis were performed to identify user
image needsand the potential differences on the two social Q&A sites.
Findings –Question-asker’s intention varies in different platforms. Zhihu users asked questions mainly
aiming at a promotion of subsequent discussion, whereas users of Baidu Zhidao often did so to seek
information. Syntacticattributes were not frequently used in both two sites. Zhihu users were more likely to
express subjectiveevaluations on images (concept, emotion, theme andstyle) in their questions than users of
Baidu Zhidao. In contrast, questions from Baidu Zhidao showed a tendency to more frequently include
descriptive metadata(rights, format, size, quality and authenticity) andsemantic attributes (generic activity,
specific people,fashion and text) of the images than questions fromZhihu. Learning was an important use on
social Q&A sites, especiallyon Baidu Zhidao. In addition, the images were primarilyused to trigger emotion
or served a persuasivepurpose in Zhihu.
Practical implications –This study contributes to a better understanding of user image search behaviour,
and the findings could be used to develop better image services on social Q&A sites. Meanwhile, the image
attributes extracted from the questions are conducive to the improvement of image retrieval systems.
Originality/value –This study explored the features of image needs on social Q&A sites, especially
considering image use specified in the question. The difference of image needs between two Chinese social
Q&A sites (Zhihuand Baidu Zhidao) was identified.
Keywords Social Q&A sites, Image use, Image demands, Image description, Image needs
Paper type Research paper
Introduction
Image, compared to text, is a more efficient medium for informationcommunication. Along
with rapid advances in internet-based services, an increasing amount of people engage in
image searching via the internet in their daily life. Image searching has switched from an
expert-oriented task to an everyday activity for normal users (Cunningham, 2006). By
September 2018, there were more than 51,000 image-related questions posed in Zhihu,
followed by 38,738 users (Zhihu, 2018a). A survey revealed that the number of users who
preferred to use both texts and images in their posts had reached 81.5 per cent on Sina
Weibo and 50.7 per centon WeChat by 2016 (Chyxx, 2017).
EL
37,3
454
Received30 September 2018
Revised16 December 2018
Accepted10 April 2019
TheElectronic Library
Vol.37 No. 3, 2019
pp. 454-473
© Emerald Publishing Limited
0264-0473
DOI 10.1108/EL-09-2018-0192
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0264-0473.htm
Significantly, image needs are complex and diversified in nature. Social question and
answer (Q&A) sites, such as Zhihu, Quora, Baidu Zhidao, and Yahoo Answers, provide a
new method for users to search for images which enable users to express detailed image
needs and turn to other users by asking natural-language questions in a network (Oh et al.,
2009). Albeit there have beennumerous studies on image search behaviour, only a few have
placed a focus on image description or image use on social Q&A sites (Cunningham et al.,
2004;Yoon and Chung, 2011). Moreover,little is known about the features of image needs on
social Q&A sites and whether a difference exists between Zhihu and Baidu Zhidao in the
context of image-relatedbehaviour.
Image needs on social Q&A sites embodies question-asker’s intentions (why users ask
questions on image searches), image description (how users describe the image features in
their questions), and image use (what the images will be used for). Understanding user
image needs could add new insights to user image search behaviour and offer a useful
reference for practitioners to improve image indexing and retrieval systems. In the present
study, questions were manually categorised, 701 collected from Zhihu and 701 questions
collected from Baidu Zhidao, characteristics of image needs were identified, and the
differences evidenced in those questions were explored. The work was guided by the
following researchquestions:
RQ1. What are the characteristics of image needs according to the image questions
posted on social Q&A sites?
RQ2. What differences are present between Zhihu and Baidu Zhidao with regard to
question-asker’sintention, image description, and image use?
The remainder of this paper is structuredas follows. A literature review is first provided on
image needs on social Q&A sites,followed by discussions on the research methodology, and
the results of data analysis. Finally, the practical implications, limitations, and possible
trends for further research are discussed.
Literature review
Image search behaviour
Similar to other information behaviours,image search behaviour reflects user image needs.
Research into image search behaviour has produced a significant body of literature. The
literature on online image search gained interestmainly on four topics: search tools, search
intentions, search queries,and search tactics. Researchers working in the first of these areas
have identified that Google or Google Image were the most frequently used tools to search
for images among college students (Yoon, 2011). Some researchers worked on designing a
high-density image search interface for creative professionals, such as journalists and
copywriters (Göker et al., 2016).There are three types of user intentions –including explore/
learn, entertain, and locate/acquire –dealingwith image search behaviour (Xie et al.,2018).
Soleymani et al. (2017) found thateye gaze and implicit user interactions (mouse movements
and keystrokes) were the most informative features to predict user search intentions. For
search queries, prior studies showed that most image queries were described with person,
thing, event or condition depending on location or time (Choi and Rasmussen, 2003). Based
on 1,000 queries on Yahoo Image Search,Park et al. (2015) found that queries of arts, culture,
and entertainment and queries for specific entities (especially people) were dominant.
Researchers also reported that image queries with different search intentions could be
diverse (Soleymani et al., 2018). In addition, studies related to contextual factors that can
affect image search tactics have also been performed. For instance, Choi (2010) found that
Zhihu and
Baidu Zhidao
455
To continue reading
Request your trial