Professional Documents
Culture Documents
Preview
Preview
EV
IE
W
APPROVAL SHEET
Online Searching
W
Dissertation and Abstract Approved: Dr. Anita H. Komlodi
IEAssociate Professor
W
University of Maryland Baltimore County
IE
Intercultural Communication, M.A., 2009
Publications:
Wang, J., Komlodi, A., & Ka, O. (2018). Understanding Multilingual Web Users’ Code-
Switching Behaviors in Online Searching. In Proc. of ACM ASIST’18. Vancouver, BC,
Canada.
Wang, J., Komlodi, A. (2018). Switching Languages in Online Searching: A Qualitative Study
of Web Users’ Code-Switching Search Behaviors. In Proc. of ACM CHIIR’18, New
Brunswick, NJ, USA.
Wang, J., Komlodi, A. (2016). Understanding Users' Language Selection: Code-Switching in
Online Searches. In Proc. of ACM CHIIR’16. Chapel Hill, North Carolina, USA.
Wang, J., Joardar, S. (2015). Cultural capital at work in Facebook users' selection of different
languages. In P. L. P. Rau (Ed.), Cross-Cultural Design: Applications in Mobile
Interaction, Education, Health, Transport and Cultural Heritage (pp. 101-109). New
York, NY: Springer.
Weaver, K., Komlodi, A., Wang, J., Joshi, K., Seller, B. (2012). Parents Views and Rules about
Technology: As told by their middle school children in Hungary and India. In Proc. CA
TaC2012. Aarhus, Denmark, 2012.
Wang, J., Komlodi, A. (2012). Children’s Formal and Informal Definition of Technology, In
Proc. of iConference 2012, Toronto, ON, Canada.
Ahmad, R., Wang, J., Hercegfi, K., & Komlodi, A. (2011). Different people Different Styles:
W
Impact of Personality in Web Sites Credibility Judgment. In M. J. Smith & G. Salvendy
(Ed.), Human Interface and the Management of Information (pp. 521-527). New York,
NY: Springer.
IE
Ahmad, R., Komlodi, A., Wang, J., & Hercegfi, K. (2010). The Impact of User Experience
Levels on Web Credibility Judgments, In Proc. of ASIS&T 2010, Pittsburgh, PA, USA.
EV
Online Searching
W
The availability of information in different languages on the Internet allows multilingual web
users to search in multiple languages for the same search task. When web users can search in
IE
multiple languages, they may combine these languages in various ways during searching. Few
studies have examined multilingual web users’ code-switching search behaviors in two or more
EV
languages online. This work bridges this gap, aiming to investigate the factors that impact code-
switching search behaviors of native Chinese web users who speak English as their second or
PR
foreign language (L2 users) and explore the L2 web users’ code-switching search patterns and
This dissertation consists of two phases. The Phase I study confirms that code-switching search
behaviors naturally occur when L2 web users search for information online by using the methods
of diaries and interviews. Findings highlight that code-switching occurred in a wide range of
search topics in L2 web users’ daily searching. The factors, such as language proficiency and
translation, information sufficiency, cultural preferences, and feelings and attitudes, impacted L2
web users’ situational and metaphorical code-switching search behaviors. In the Phase II study,
controlled lab experiments and follow-up interviews were conducted so as to study L2 web
users’ code-switching information seeking behaviors in more detail during searching. This in-
depth study revealed the L2 web users’ code-switching search models, strategies, and more
factors such as context, information verification, language image, and website credibility that
The purpose of this work is to describe code-switching search behaviors and to provide design
implications for multilingual search systems. It contributes to a detailed study of L2 web users’
W
language selection search behaviors from an interdisciplinary perspective using sociolinguistics
ONLINE SEARCHING
By
W
Jieyu Wang
IE
EV
Human-Centered Computing
2018
ProQuest Number: 10846657
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
W
IE
EV
ProQuest 10846657
Published by ProQuest LLC (2018 ). Copyright of the Dissertation is held by the Author.
All rights reserved.
PR
This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106 - 1346
© Copyright by
Jieyu Wang
2018
W
IE
EV
PR
PR
EV
IE
W
Acknowledgements
My committee members: Dr. Omar Ka, Dr. Lina Zhou, Dr. Kuber, and Dr. Jinie Pak.
W
IE
EV
PR
ii
Table of Contents
Chapter I Introduction…………………………………………………………………...1
1.1 Background……………………………………………………………………….1
1.2 Contributions……………………………………………………………………...3
1.3 Dissertation Summary ………………………………………...………………….4
Chapter II Literature Review…………………………………………………………….8
2.1Second Language Acquisition…………………………………………………….8
2.2Users’ Second Language Behaviors in Technology……………………………...12
2.3 Code-switching……………………………………………………………….….16
2.3.1 Definition………………………………………………………………16
2.3.2 Discourse-related Code-switching……………………………………..17
2.3.3 Code-switching in Technology………………………………………...19
Chapter III Previous data collection………………………………………………………24
3.1 Second language online searching: User behaviors and challenges……………..24
3.1.1 Introduction………………………………………………………….....24
3.1.2 Participants……………………………………………………………..24
W
3.1.3 Methods……………………………………………………………...…25
3.1.4 Data Analysis……………………………………………………..…....26
3.1.5 Results/Discussion…………………………………………………..…27
3.1.6 Future Work……………………………………………………………36
IE
3.2 Code-switching on Facebook………………………………………………….....36
3.2.1 Introduction…………………………………………...………………..37
3.2.2 Methods………………………………………………………………...38
EV
3.2.3 Results…………………………………………………………...……..40
3.2.4 Discussion…………………………………………………...…………41
3.2.5 Conclusion and Future Work………………………...………………...48
3.3 Code-switching within a U. S. University Classroom: When, how and why it happens
with heritage students………………………………………………………………...49
PR
3.3.1 Introduction…………………………………………………………….49
3.3.2 Participants……………………………………………………………..50
3.3.3 Methods………………………………………………………………...51
3.3.4 Results and Discussion………………………………………...……….52
3.3.5 Conclusion……………………………………………………………...56
3.4 Outline of Previous Studies’ Results……………………………………………..57
Chapter IV Research Questions and Methodology……..……………………………….59
4.1 Research Questions…………………………........................................................59
4.2 Methodology…………………………..................................................................60
4.2.1 Phase I Diary Studies and In-Context Interviews………………………............62
4.2.1.1 Participants…………………………………………………………...63
4.2.1.2 Demographic Questionnaire………………………………………….63
4.2.1.3 Diary Studies…………………………………………………...…….65
4.2.1.4 In-context interviews...……………………………………………….66
4.2.1.5 Data Analysis………………………………………………………....68
4.2.1.5.1 Descriptive Data Analysis…………………………………..68
4.2.1.5.2 Qualitative Data Analysis: Open-Coding……………….…..68
4.2.1.5.3 Qualitative Data Analysis: Axial-Coding………………..….68
iii
4.2.1.6 Results of Pilot Testing……………………………………………..…69
4.2.2 Phase II Controlled Lab-based Experiments………….….……………………...69
4.2.2.1 Participants……………………………………….……………………70
4.2.2.2 Procedures……………………………………………………………..70
4.2.2.3 Instruments..……………………………………………………………71
4.2.2.3.1 Tasks…………………………………………………………71
4.2.2.3.2 Protocol………………………………………………………73
4.2.2.3.3 Experiment Observation……………………………………..73
4.2.2.3.4 Screen Video Analysis……………………………………….73
4.2.2.3.5 Follow-up Interviews………………….……………………..73
4.2.2.4 Analytical Approach…...……………………………………………....74
4.2.2.4.1 Descriptive Data Analysis……………………………………74
4.2.2.4.2 Observation and Video Data Analysis……………………….74
4.2.2.4.3 Interview Data Analysis……………………………………...76
4.2.2.5 The Pilot Study…………………………………………………………77
4.2.2.6 Limitations……………………………………………………………...78
W
Chapter V Results……..……………………………………………………………………..79
5.1 Findings from Phase I……………………………………………………………....79
5.1.1 Participants’ Demographics………...……………………………………..80
5.1.1.1 Gender…………………………………………………………...80
IE
5.1.1.2 Age……………………………………………………………....81
5.1.1.3 Major and Degree………………………………………………..81
5.1.1.4 English Language Proficiency, Code-Switching Frequency, and
EV
International Experiences………………………………………………..82
5.1.1.5 Participant Limitation……………………………………………84
5.1.2 Topics of Code-Switching Episodes………………………………………85
5.1.3 Functions of Code-Switching……………………………………………...90
5.1.3.1 Situational Code-Switching in Language and Information
PR
Sufficiency………………………………………………………………91
5.1.3.1.1 Situational Code-Switching in Language and
Translation………………………………………………………92
5.1.3.1.2 Situational Code-Switching in Information
Sufficiency……...........................................................................94
5.1.3.1.2.1 Situational Code-Switching in Information Sufficiency:
Quality of Information………………………………………….94
5.1.3.1.2.2 Situational Code-Switching in Information Sufficiency:
Quantity of Information………………………………………...95
5.1.3.1.2.3 Situational Code-Switching in Information Sufficiency:
Both Cultural Perspectives of Information………………...…...97
5.1.3.2 Metaphorical Code-Switching in Cultural Preferences, Feelings
and Attitudes……………………………………………………………98
5.1.3.2.1 Metaphorical Code-Switching in Cultural Preferences of
Mixed Language Queries……………………………………….98
5.1.3.2.2 Metaphorical Code-Switching in Feelings and
Attitudes………………………………………………………...100
5.2 Finding from Phase II…………………………….………….….……………….…102
iv
5.2.1 Participants’ Demographics……………………..…………………...…...102
5.2.1.1 Gender…………………………………………………………..102
5.2.1.2 Age…………………………………………………...………....103
5.2.1.3 English Language Proficiency, Major, Degree, and Code-
Switching Frequency …………………………………………………..103
5.2.2 Search Engines…………………………………..................................….104
5.2.3 Queries and Query Reformulations…………………………………...….105
5.2.4 Factors Impacted Users’ Situational and Metaphorical Code-Switching...107
5.2.4.1 Situational Code-Switching…………………………...………..108
5.2.4.1.1 Situational Code-Switching due to Information
Sufficiency…………………………………………………...…108
5.2.4.1.1.1 Situational Code-Switching due to Information
Sufficiency: Quantity and Quality of the Results…………...….108
5.2.4.1.1.2. Situational Code-Switching due to Information
Sufficiency: Information in Both Cultural Perspectives………..109
5.2.4.1.2 Situational Code-Switching due to Language Proficiency
W
and Translation………………………………………………….109
5.2.4.1.2.1 Situational Code-Switching due to Language
Proficiency………..…………………………………………….110
5.2.4.1.2.2 Situational Code-Switching due to Translation….....111
IE
5.2.4.1.3 Situational Code-Switching due to Information
Verification……………………………………………………..111
5.2.4.1.4 Situational Code-Switching due to Context…………..113
EV
5.2.4.2 Metaphorical Code-Switching………………………………….114
5.2.4.2.1 Metaphorical Code-Switching due to Language
Image……………………………………………………………114
5.2.4.2.2 Metaphorical Code-Switching due to Feelings and
Attitudes…………………………………………………...……115
PR
v
6.6.1.1 Simple Vocabulary…………………………………………......143
6.6.1.2 Checklists………………………………………………………143
6.6.1.3 Query Recommendations…….………………………………...144
6.6.2 Design Implications: Translation Tools………………………………….145
6.6.3 Design Implications: Websites’ Different Language Versions…………..147
6.6.4 Design Implications: Providing Sufficient Information………………….150
6.6.5 Design Implications: Support for Comparing Information from Multiple
Sources…………………………………………………………………………151
6.6.6 Design Implications: Culturally adaptable UI design……………………153
6.6.7 Design Implications: Enhance Design Features to Build L2 User Trust...155
Chapter VII Conclusions…………………………………………..………………………..157
7.1 Contributions………………….……………………………………………………157
7.2 Limitations and Challenges………………………………………………………...159
7.3 Future Work…………………………………………………………………….….160
7.4 Related Publications………………………………………………………………..161
Appendices……………………………………………………………………………...……. 162
W
Appendix 1 UMBC IRB Approvals…………………………………………................162
Appendix 2 Recruitment Description…………………………………………………..167
Appendix 3 Demographic Questionnaire………………………………………………168
Appendix 4 In-Context Interview Guide……………………………………………….179
IE
Appendix 5 Controlled Lab Experiments….…………………………………………...180
Appendix 6 Topics of Code-Switching in Phase I……………………………………..183
Appendix 7 The Factors and Reasons for Code-Switching in the Phase I Episodes…..185
EV
Appendix 8 The Factors and Reasons for Code-Switching in Phase II………………..186
References………………………………………………………………………...…………...188
PR
vi
List of Tables
W
Table 13: Language Proficiency and CS Frequency in Phase II……………………………….104
Table 14: Search Engines the Participants Used in Phase II…………………………………...105
Table 15: Numbers of Queries and Query Reformulations in Phase II……………….………..106
Table 16: T-Test of Queries and Query Reformulations in Phase II……………………...……107
IE
Table 17: Numbers of Situational and Metaphorical Code-Switching with Factors and Reasons in
Phase II………………………………………………………………………………………….108
Table 18: Comparison of the Factors and Reasons of CS in Phase I and II……………………124
EV
Table 19: Languages the Participants Started with in Phase II………………………………....125
Table 20: L2 web users’ Search Strategies…………………………………………...………...139
Table 21: Results and Design Implication……………………………………………………...142
PR
vii
List of Figures
W
Figure 12: Hierarchy of Code-Switching and Reasons and Factors in Phase II………………..123
Figure 13: L2 Web Users’ Code-Switching Search Model for the News Task Part I……….....128
Figure 14: L2 Web Users’ Code-Switching Search Model for the News Task Part II…………130
Figure 15: L2 Web Users’ Code-Switching Search Model for the Travel Task Part I………....132
IE
Figure 16: L2 Web Users’ Code-Switching Search Model for the Travel Task Part II………..134
Figure 17: L2 Web Users’ Code-Switching Search Model for the Medical Task Part I…….....136
Figure 18: L2 Web Users’ Code-Switching Search Model for the Medical Task Part II……....137
EV
Figure 19: Design Example Mayo Clinic Symptom Checker………………………………….144
Figure 20: Design Example Google Translate………………………………………………….146
Figure 21: Design Example Baidu Translate…………………………………………………...147
Figure 22: Design Example Website’s Language Option……………………………………....148
Figure 23: Design Example BBC Different Language Versions……………………………….148
PR
viii
Chapter I Introduction
1.1 Background
Since the Internet was introduced, the number of web users who search for information on it has
increased to 3.366 billion (Internet World Stats, 2015). In 2012 26.8% of internet users were
English language users and 56.6% of the information on the web was in English (Internet World
Stats, 2012). However, as the Internet has developed and the number of web users who are
speakers of other languages has increased, the role of languages on the web is becoming more
W
important.
IE
According to the U.S. Census Bureau (2003), in the United States 18% of the population is
bilingual or multilingual. While in Europe, a European Commission survey suggested that 56%
EV
of the respondents speak two languages (Marian & Shook, 2012). As a whole, only 40% of the
world population is monolingual, whereas 60% of the world population is bilingual, trilingual, or
PR
multilingual (ILanguages, 2015). Though it is hard to investigate how many non-native English-
speaking web users use English websites (L2 users), Kachru (2005) estimated there were over
Researchers in the field of technology have been studying how these L2 users interact with the
information on English websites. Recent studies have focused on different factors that influence
users’ second language web surfing such as language proficiency and culture, domain
knowledge, web accessibility in other languages, and language translation (Kralisch & Berendt,
2005; Nzomo, 2012; Marlow et al., 2008). However, little research has been conducted on the
1
phenomena of users with multilingual backgrounds searching for information by switching
between their native languages and second/foreign languages in order to get more satisfying
results.
defined and analyzed code-switching in different social contexts (Gumperz, 1982; Myers-
Scotton, 1993; Auer, 1998; Woolard, 2004). Here we define users’ native/first language (L1) as
the language that a speaker acquires first chronologically and speaks most fluently as a result of
W
other speakers in his environment speaking the same language. Second language (L2) is the
language that a speaker speaks other than his native language. Meanwhile, this language is
IE
spoken as a native language in his immediate environment (Ringbom, 1980). As a comparison, a
foreign language refers to the language other than a speaker’s native language. This language is
EV
hardly spoken in his immediate environment (Ringbom, 1980). Discourse-related code-
switching between one’s native languages and second/foreign languages has been extensively
PR
studied in sociolinguistics. Few studies have been done on web users’ code-switching behaviors
users’ code-switching behaviors between their native language and their second/foreign language
when they search online. Ringbom (1980) pointed out that the difference between a second
language and a foreign language is that the second language is spoken in the immediate
environment of the speaker while the foreign language is not spoken in that environment. For
2
example, in this study 25 participants whose native language is Chinese and whose
second/foreign language is English have been recruited. Participants who live in China are
considered web users who search in a foreign language. Participants who stay in the US within
an English environment are considered web users who search in their second language. Overall,
the research investigates users’ language switching between their native language (Chinese) and
their second/foreign language (English) when they search for information online.
1.2 Contributions
W
The goal of this paper is to explore different factors that impact multilingual users’ code-
switching behaviors during online searching and the roles of languages that they use. The
2. Moreover, the studies in this work have been performed by applying concepts in
PR
The detailed analysis of multilingual users’ code-switching behaviors includes examining when
and how the users code-switch and the reasons for their behaviors. Prior work has examined
multilingual users’ second language proficiency and cultural impact, domain knowledge, other
languages’ web accessibility, and language translation (Berendt & Kralisch, 2009; Kralisch &
Berendt, 2005; Marlow et al., 2008; Nzomo et al., 2012). This work explains multilingual users’
3
code-switching behaviors by applying second language theories and sociolinguistic code-
switching concepts. This study will be based on previous research in both information
technology and sociolinguistics. Outcomes of the study are expected to include design guidelines
The paper aims to provide implications to support website design for multilingual web users,
especially non-native English users, who seek information on English websites. The implications
are for mixed language search interfaces which help designers further understand what
W
challenges multilingual web users face when they switch languages and their information needs,
habits, search pattern, and search strategies. Thus, better user-centered websites may be created
IE
or refined to satisfy more multilingual web users’ information needs.
EV
1.3 Dissertation Summary
This chapter reviews related research in three research areas of different disciplines: second
sociolinguistics and information technology. It aims to provide a wide range of research that has
been done to build a strong basis for this research. This chapter explains L2 web users’ search
behaviors from the factors that impact their second language acquisition to the factors that affect
their second language retrieval. It also describes the related research about code-switching in
4
This chapter presents three related studies that have been done prior to this research. The first
study examined native English web users’ information seeking behaviors in their L2 language,
Spanish. It explored language and cultural factors that impacted their searching and the
challenges they faced. This study provided prior research experiences for this dissertation from
the L2 users’ information seeking perspective. The second study explored how multilingual web
users switched different languages to interact with information on Facebook. It emphasized the
impact of factors such as language proficiency, cultural capital, context, locality, and audience
on web users’ code-switching behaviors on Facebook. It offered effective research methods and
W
prior experiences to allow further investigation of users’ code-switching behaviors on websites.
This chapter first states six research questions for the Phase 1 and 2 studies. This work explores
when, how, and why non-native English web users search for information by switching between
their native language and their second/foreign language and examines the roles of the languages
that they use. L2 web users’ code-switching search patterns during searching have been
This chapter then describes the designed methods for this dissertation. The participants in this
dissertation are native Chinese web users who search for information using English as their
5
second/foreign language. In Phase I self-report diaries and in-context interviews confirm that L2
web users code-switch and explore the reasons that L2 web users switch languages during their
searching. In-context interviews allow users to search for information and switch languages in a
natural way. In Phase I controlled lab experiments and follow-up interviews further examine the
factors that impact L2 web users’ code-switching behaviors and investigate these web users’
search patterns and strategies when they switch languages during searching. Two pilot studies
have been done for Phase 1 and one pilot study has been done for Phase 2 after the IRB approval
in August 2015. In this work, 10 participants have been recruited and completed the experiments
W
in Phase I when the saturation was reached in the in-context interviews. Fifteen participants have
performed the tasks in Phase II when the saturation was reached in follow-up interviews.
IE
Chapter V Results
EV
This chapter reveals the findings from the Phase I and II studies. The data showed that code-
switching occurred during the participants’ information seeking process. They code-switched
PR
when they had language problems, needed translation, or searched for travel, academic, domain
knowledge information, and daily life information such as shopping and medicine. The two
studies figured out the factors that impacted the L2 web users’ situational and metaphorical code-
Chapter VI Discussion
This chapter presents and interprets the L2 web users’ code-switching search models and
strategies for their searching. It also provides design implications which are revealed from the
6
Chapter VII Conclusion
This chapter emphasizes the contributions of this work. It explains the limitations and the future
W
IE
EV
PR