Revealing the nature of contemporary tourism research: Extracting common subject areas through bibliographic coupling

Revealing the Nature of Contemporary Tourism Research: ExtractingCommon Subject Areas through Bibliographic Coupling

YULAN YUAN1, ULRIKE GRETZEL2 and YUEN-HSIEN TSENG3*1Department of Travel Management, Jinwen University of Science and Technology, New Taipei, Taiwan2UQ Business School, University of Queensland, Queensland, Australia3Information Technology Center, National Taiwan Normal University, Taipei, Taiwan

ABSTRACT

This study aims at identifying recent subject areas and citation patterns of tourism research by applying a bibliometric tool to analyze over2000 publications since 2008 from 10 tourism journals. Twelve subject areas were identified, and interesting citation patterns were revealed.From this analysis, the tourism research was found to be quite dynamic but at the same time very much focused on marketing and managementaspects. This article contributes to ongoing efforts to reveal the nature of tourism research and its knowledge bases by applying the methodologyof bibliographic coupling. Copyright © 2014 John Wiley & Sons, Ltd.

Received 26 May 2013; Revised 03 February 2014; Accepted 18 February 2014

key words bibliographic coupling; tourism knowledge; subject analysis; citation patterns; multistage clustering; content analysis

INTRODUCTION

Academic journals reflect academic endeavor and provide amedium to disseminate and exchange scholarly knowledgeamong professionals. They have become an essential platformfor scholars and practitioners to learn the known and to identifythe unknown. As stated by Van Doren and Heit (1973),‘Academic journals mirror the direction of a discipline’sresearch’ (p. 67). Thus, examining articles published in theseacademic journals can help identify the influential literatureand research frontiers. Moreover, such introspection can revealthe evolution of a field and important structural properties(Xiao & Smith, 2006).

Subject areas resemble the branches of knowledge in amature field of inquiry (Sheldon, 1990). Tourism is generallyconsidered as a multidisciplinary field diffused with knowl-edge from several disciplinary areas (Tribe, 1997). To capturethe essence of tourism as a field of inquiry, researchers haveproposed different tourism-related subject areas. For example,Tribe (1997) distinguished the business of tourism from thenonbusiness aspects of tourism. Jafari and Ritchie (1981)identified three principal dimensions that comprise the studyof tourism: tourists (study of people), the tourism industry(tourism goods and services) and the settings (socio-culturalfabric and physical environment). The study of the tourismsettings can be further split into four subject areas: naturalresources, man-made resources, socio-cultural resources andthe encounter between tourists and residents. The interrelation-ships among these three principal dimensions were identifiedas the fourth dimension of the tourism knowledge domain(Jafari, 1990, 2001). As tourism research develops, the bodyof knowledge grows, and topics split off and new subject areas

emerge (Burdge, 1983). While scholars find it increasinglydifficult to keep up with the growing literature in their fieldsof research, students and young scholars are particularlyoverwhelmed by the volume of literature that they must famil-iarize themselves with to identify foundational literature relatedto the subject area on which they can focus. Additionally,increased neoliberalism and managerialism in higher education(Tsaur et al., 2008) demand concrete descriptions of whattourism study involves, its homogeneity and the works mostprominently cited in the body of related literature. Under-standing the other disciplines that influence the literatureon tourism is also important. Consequently, the application ofbibliometrics/scientometrics in tourism research becomes im-portant (Benckendorff, 2009; Cheng et al., 2011; Hall, 2011).

Bibliometrics is the science of measuring and analyzing afield of science of interest (Moed, 2005). Bibliometricmethods have been used to examine citation patterns thatcould reflect the development of academic efforts frommetalevel examinations of field development to microlevelexaminations of individual scholarship (Hall, 2005). Therange of foci in the bibliometric studies of tourism researchincludes the following types: (i) the research productivityof individual scholars (Sheldon, 1991; Zhao & Ritchie,2007; McKercher, 2008; Park et al., 2011; Way et al.,2012) and institutions (Hall, 2005; Park et al., 2011); (ii)knowledge flow and social networks (Howey et al., 1999;Benckendorff, 2009; Ying & Xiao, 2012); (iii) topics andlong-term development trends (Swain et al., 1998; Xiao &Smith, 2006; Ballantyne et al., 2009; Benckendorff, 2009);(iv) journal rankings (McKercher et al., 2006; Jamal et al.,2008) and journal development (Cheng et al., 2011); and(v) most frequently cited scholars and works (Benckendorff& Zehrer, 2013).

Citation and content analyses are the most commonapproaches in these studies. If performed manually, the analysesare time consuming and labor intensive (Dembkowski &

*Correspondence to: Yuen-Hsien Tseng, Information Technology Center,National Taiwan Normal University, Taipei, Taiwan.E-mail: [email protected]

Copyright © 2014 John Wiley & Sons, Ltd.

International Journal of Tourism Research, Int. J. Tourism Res. (2014)Published online in Wiley Online Library (wileyonlinelibrary.com) DOI: 10.1002/jtr.2004

Hanmer-Lloyd, 1995; Swain et al., 1998; Schmidgall et al.,2007; McKercher, 2008), which suggests that efforts could belimited to only a small number of journals across a narrow timeframe (Hall, 2005; McKercher, 2008). Previous studies ontourism examined intellectual patterns primarily based on twojournals, the Annals of Tourism Research and TourismManage-ment (Benckendorff, 2009; Kim, 1998; Swain et al., 1998; Xiao& Smith, 2006). These analyses draw on an insufficient sampleof the literature (Kennedy, 2007). As Tribe (2008) points out,many studies focus on what is present in those two journalsand thus inevitably miss important aspects of a large and diversefield such as tourism. Xiao and Smith (2006) also stronglyrecommend that analytical data be derived from differenttourism journals to ensure a comprehensive picture.

Given the constraints of manual content and citationanalyses, an array of bibliometric analysis software toovercome these limitations is emerging. This article appliesthe software package called Content Analysis Toolkit forAcademic Research (CATAR) to identify subject areas andcitation patterns from various tourism journals based onbibliographic coupling (BC) and content analysis. Adoptingan epistemological perspective, this paper aims to performthe following: (i) identify subject areas in contemporary tour-ism research using CATAR and compare them to existingtypologies and (ii) assess the ability to identify tourism sub-ject areas and citation patterns using an automated approachthat could be used for persistent and consistent observationof the development of tourism research.

LITERATURE REVIEW

Subject areas in tourism journalsA subject area is a branch of knowledge in a field of study thatis researched from different perspectives. Characteristicallyspeaking, subject areas are dynamic. The tourism literatureindicates that the foci of tourism studies have shiftedchronologically and diverged (Jafari, 1990). In an attempt toorganize the body of tourism knowledge and illustrate itsdevelopment, Jafari (1990, 2001) proposed that scholars adoptfour positions when studying tourism-related issues, namelythe advocacy, cautionary, adaptancy and knowledge-basedpositions. These positions provide a firm framework for inves-tigating tourism knowledge (Swain et al., 1998). Researcherswho adopt the advocacy position consider tourism as an indus-try and examine tourism research mainly from an economicperspective. They view tourism as inherently positive forcommunity culture, nature and economic development. Anoppositional position, called the cautionary platform, emergedin response. This position focuses on the negative influences ofthe tourism industry and recommends caution owing to thehidden costs and undesired consequences of tourism develop-ment. The debates between the advocacy and cautionaryplatforms led to the formation of a more synthesizing positionthat explores alternate forms of tourism development to reducethe associated negative effects and build a sustainable future.This position is termed the adaptancy platform. The fourthposition, knowledge-based platform, is a collective positionbased on all platforms. This position adopts a holistic view

and system-oriented approach to understand desirable changesand unwanted consequences in the age of globalization.

Under this reference framework, five main studies haveattempted to identify specific subject areas of tourism researchthat have emerged from these positions and to document thechanges in these subject areas over the past three decades(Table 1). Swain et al. (1998) applied index headwordanalysis, and Xiao and Smith (2006) adopted content analysisto analyze the headword index of the Annals of TourismResearch. Both identified subjects such as impact, development,the USA, international tourism, tourists, organization andmethodology as being part of the top 10 list.

Based on the articles extracted from the Annals of Tour-ism Research and Tourism Management using ThomsonReuters’ Web of Knowledge (WoK) database, Benckendorff(2009) performed keyword and cocitation analysis of 334 ar-ticles from 1994 to 2007 with a focus on those by NewZealand and Australian authors. The analytical resultssuggested the existing links between concepts dealing withsocial impact, communities and resident attitudes. The resultsalso revealed strong connections between the followingconcepts: authenticity and experience, motivation andsatisfaction, behavior, perception and segmentation. Emerg-ing subject areas identified by Benckendorff includeplanning and development, economics and demand, servicequality and satisfaction, heritage and authenticity, andecotourism and conservation. Benckendorff also found thatcertain subject areas are region specific. For example,New Zealand researchers are especially prolific in theresearch on accidents and adventure tourism.

These three studies provide excellent insights into tourismas a knowledge domain. However, the findings are poten-tially limited by being derived from a very small number ofjournals (Kennedy, 2007). This suggests that it is necessaryto examine the same issue using a larger sample of journalsto gain a comprehensive picture of tourism knowledge.Two articles expanded the range of data to remedy the short-comings of previous studies. The works of Ballantyne et al.(2009) grew the data set by obtaining 2868 full-length arti-cles published between 1994 and 2004 from 12 tourismjournals. A content analysis approach was applied, and 21subject categories were identified. The three topics that havemost published articles were tourist/visitor studies, destina-tions and tourism planning, and marketing. Park et al.(2011) later analyzed the contents of 2834 articles and re-search notes from six commonly cited tourism and hospital-ity journals over a ten-year period (2000–2009). Twentysubject categories were proposed based on expert judgment.However, both studies have other shortcomings: The methodused to delineate subject categories is unclear, as are the con-tributors to each topic area. Additionally, the subject catego-ries presented in the two studies were solely yielded viasubjective judgment without any supportive reasoning.

Two problems occur with respect to the above studies.The first is that multiplication of the dimensions of tourismknowledge by the various positions considerably increasesthe research complexity of investigating tourism researchtrends, including the potential number of subject areas thatmust be identified. The other problem relates to the use of

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

Table

1.Com

parisonof

five

major

studieson

trends

andsubjectareasin

tourism

research

Authors

Swainet

al.(1998)

XiaoandSmith

(2006)

Benckendorff(2009)

Ballantyneet

al.(2009)

Parket

al.(2011)

Subject

area

Top

10headwords

Top

13subjectareas

1057

keyw

ords

andtop16

areas

21subjectareas

20tourism

subjectareas

Conferences

Attractio

nsAttitudes

Culturaltourism

Attractio

nmanagem

ent

Destin

ation

Culture

Australian

Businesstourism

Crisisandsafety

Development

Development

Authenticity

Destin

ations

Destinationmarketingandmanagem

ent

Impact

Impacts

Behavior

Economic

issues

Economic

impact

andeconom

etrics

Industry

Internationaltourism

Culture

Ecotourism

Internationaltourism

Marketin

gEcotourism

Environmental

interpretatio

n

Educatio

n

Methodology

Methodology

Experience

Geographicalissues

Organization

Motivation

Heritage

Hospitality

General

marketin

g

Tourist

Organizationandassociation

Impacts

Hum

anresource

Imageandbranding

USA

Planning

Managem

ent

Marketin

gInform

ationtechnology

Resort

Motivation

Managem

ent

Marketsegm

entatio

n

Tourist

New

Zealand

Special

events

Meetin

gs,incentiv

es,conventio

nsandexhibitio

nsUSA

Perception

Sportandleisure

Tourism

planning

Risk

Sustainable

developm

ent

Politics,policy,

legaland

governmentalissues

Sustainability

Service

managem

ent

Tourism

Tourism

education

andtraining

Special

intereststourism

Tourism

impacts

Supplychain

Tourism

planning

Managem

ent

Tourism

policy

Sustainable

tourism

andecotourism

Tourism

trends

Tourism

research

issues

andmethods

Tourism

developm

entandresidence

perceptio

n

Tourist/Visito

rstudies

Transport

Tourists’

perceptio

nandbehavior

Other

Study

materials

Subject

indexof

ATR

Subject

indexof

ATR

Full-length

articles

ofATRandTM

Full-length

articlesof

ATR,C

IT,IJTA,JRR,

JTR,JTS,

JVM,TA,

TCC,T

E,TG

andTM

Full-length

articlesandresearch

notesof

ATR,JTR,T

M,JH

TR,

IJHM

andCHQ

Tim

eperiod

1973–1

998

Six

five-yearperiods:1973–1978,

1979–1983,1984–1988,1989–1993,

1994–1998,1999–2003

Papersauthored

byAustralian

andNew

Zealand

researchers

from

1994

to2007

1994

–2004

2000–2

009

Focus

Indexcitatio

nandheadword

activ

ity,categorizing

index

headwords

toidentifypathsof

tourism

know

ledge

Showingtopictrends

Influences

andinfluencersin

tourism

research

inAustralia

andNew

Zealand

Showingtopictrends

Approach

Indexheadwordnnalysis

Content

analysis

Keywordandcocitationanalysis

Content

analysis

Content

andcitatio

nanalysis

Note:ATR,A

nnalsof

Tourism

Research;

CHQ,C

ornellHospitalityQuarterly;CIT,C

urrent

Issues

inTourism

;IJHM,Internatio

nalJournalof

HospitalityManagem

ent;JH

TR,Journal

ofHospitality&

Tourism

Research;

JTG,T

ourism

Geographies;JTR,Journalof

TravelR

esearch;

TS,Journalof

Tourism

Studies;JVM,Journalof

VacationMarketin

g;JR

R,Journalof

RecreationResearch;

IJTA,Internatio

nalJournalof

Tourism

Analysis;TA,

Tourism

Analysis;TCC,Tourism

Culture

andCom

munication;

TE,T

ourism

Economics;TM,T

ourism

Managem

ent.

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

content analysis. Content analysis is guided by a theoreticalframework that classifies data. The analytical results dependheavily on the unique perspectives of the conductingresearchers. Taking such an approach can introduce authorand method bias into the research results. Different authorsmight devise various subject areas and assign articles tothem. Table 1 shows the diversity in subject areas thatemerges from this approach.

In terms of method bias, keywords and headwords are notcontext free. The same words come with context-specificmeanings, and new meanings may arise during differentperiods (Leydesdorff, 1997). When they serve as the unit ofanalysis, their meanings change with the context. Thus,keywords and headwords may not accurately reflect thefocus of a given set of articles. Moreover, polysemy andsynonyms further deteriorate this situation. Recent studies haveattempted to resolve this problem by (i) including titles and ab-stracts as additional data sources and (ii) using co-occurrenceanalysis of keywords to examine the relationships among sub-ject areas (Benckendorff, 2009). However, the change in themeaning of keywords also stimulates a change in the relatednessamong articles. That is, the use of co-occurrence analysis basedon keywords introduces excessive noise and decreases the effec-tiveness of this method in identifying the relatedness of articlesand subject areas over an extended period of time comparedwith citation-based methods (Leydesdorff & Zaal, 1988).

Approaches for analyzing research documentsand journalsAcademic scholars and researchers exchange and disseminateacademic knowledge through academic journals (Weiner,2001; Hall, 2005, 2011). Analyzing these ‘communicationchannels’ lets researchers answer the five ‘w’ questionsproposed by Lasswell (1964), namely ‘who says what inwhich channel to whom with what effect?’ The analyticmethods include qualitative and quantitative approaches.Qualitative approaches investigate the topics delivered by agiven ‘channel’, as well as when, where and by whom theyare addressed. Studies adopting this approach include thework of Xiao and Smith (2006).

Quantitative approaches use various bibliometric/scientometric methods to describe patterns associated withbibliographic data in a given field (Holden et al., 2005), toidentify relations between citing and cited documents(Garfield, 1994). As Joseph (2003) noted, citations are the‘currency of modern science’ (p. 283), and thus, the assumedinfluence of a work increases with the number of citations itreceives. Citation data can also reveal different types ofconnectivity, which may indicate the health of a field(Borgman & Furner, 2002). Connections that can beanalyzed are not only limited to who is addressing whomvia citations but also include connectivity among subfields,linguistic or geographic communities and methods. Citationpatterns thus provide important insights regarding conver-gence or divergence in a field (Broadus, 1987). Xiao et al.(2012) asserted that citation practices reflect the norms in ascientific community. Furthermore, citation patterns illustratethe intellectual structure of a field or discipline and highlight

relationships among concepts and ideas (Benckendorff,2009) or authors (Benckendorff & Zehrer, 2013).

Various bibliometric methods are used to illustrate quanti-tative features, including BC analysis, cocitation analysis andcoword analysis (Moed, 2005). Coword analysis is alsoknown as keyword or headword analysis and was used bySwain et al. (1998). Bibliographical coupling has not yetbeen applied to tourism research.

Bibliographical coupling method for extraction of subjectareasThe main bibliometric method used in this study is BC. Thisconcept was introduced by Kessler (1963), who consideredscientific papers not only involving the storage of factualinformation but also having logic and dynamics that maysuggest topical relations among citing articles. The relationbetween two or more papers is revealed in the number ofworks shared in their reference lists. Restated, two articlesare bibliographically coupled if their reference lists shareone or more of the same cited documents. The BC methodthus pairs articles based on the same knowledge sources thattwo articles share. As shown in Figure 1, BC identifiesrelationships between papers A and B because they citepaper 1 (left panel). In contrast, cocitation analysis wouldresult in papers C and D being associated because they arecited by paper 2 (right panel). In other words, two papersmust cite the same source to become bibliographicallycoupled, while any two papers listed in another’s referenceform the cocitation relationship (become cocited).

Quantitatively speaking, the more common referencestwo papers cite, the more closely the two papers relate toeach other and the higher is their BC strength. For example,if paper A cites 10 references and paper B cites 15 referencesand if they share five common references, then the topicalsimilarity between A and B can be calculated using anynormalized similarity metrics, such as the dice coefficient(Salton, 1989), as 2 * 5/(10 + 15) = 10/25 = 0.4. Quantita-tively, A and B are bibliographically coupled with a topicalmeasure of 0.4 in this case. This idea can be applied to twogroups of papers, and thus, a corpus of papers can be recur-sively grouped into topical clusters.

Figure 1. Illustration of bibliographic coupling (left) and cocitation(right) methods. Arrows represent the citation from one citing article(arrow end) to the cited one (arrowhead). Dashed ovals illustrate howthe papers might be grouped by bibliographic coupling (documents Aand B) and cocitation methods (documents C and D) respectively.

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

Previous studies had examined the validity and accuracyof BC and suggested that it is more effective for topic cluster-ing than other citation analysis methods. Vladutz and Cook(1984) conducted an experiment using randomly selecteddocuments from the Science Citation Index (SCI) databaseto examine the validity of BC as an indicator of topicalsimilarity between journal publications. Expert assessmentswere used to verify the results. They concluded that utiliza-tion of BC to analyze very large citation databases is feasibleand that valid results were obtained as to subject relatedness(Vladutz & Cook, 1984). More recently, Boyack andKlavans (2010) compared the accuracy of different citation-based approaches, namely cocitation analysis, BC and directcitation, and found that BC is the better option in variouscases. Moreover, the number of citing papers of any givenpaper may change over time. Thus, a time lag is requiredfor papers to build up cocited pairs. The results of cocitationanalyses thus will differ considerably with the time of study.On the other hand, the number of references is fixed andavailable at the time the paper is published. As such, BC doesnot suffer this weakness. To summarize, the above pointsstrongly support the use of BC for our study.

METHOD

Journals for analysisAcademic journals included in the Social SCI (SSCI) arewidely regarded as prominent and influential publicationoutlets and serve as convenience sources for this study.Thirteen out of 37 journals, which were included under thesubject category of hospitality, leisure, sport and tourism inSSCI, were identified by scholars as tourism-orientedjournals (McKercher et al., 2006).

Hospitality, leisure and sport journals were excluded fromour analysis for two reasons. First, tourism, hospitality,leisure and sport are considered to be separate fields, andthe corresponding journals should be treated separately(Jamal et al., 2008). Second, tourism academics are generallyless aware of hospitality journals and vice versa (McKercheret al., 2006). As a result, hospitality journals are not a major-ity source of citations for tourism journals (Kim et al., 2009).

This study adopted an epistemological perspective to explorethe subject areas embedded in tourism journals. In otherwords, it is assumed that the knowledge sources of eachsubject area have a crucial influence on the analytical result.Therefore, the journals whose scope and orientation weredeemed as not pertaining directly to tourism were excluded.The Journal of Hospitality & Tourism Research and theJournal of Hospitality, Leisure, Sport & Tourism Educationwere recognized as hospitality journals (McKercher et al.,2006). Therefore, they were excluded as well, even thoughtheir journal titles include the word ‘Tourism’.

To explore the subject areas, target journals were limitedto only those that have been included in the SSCI since2008, which is the year when a large number of journalswere included in SSCI. This was necessary to make sure thatsome journals that had been indexed longer were notoverrepresented in the sample. At the same time, a five-yearperiod was seen as the minimum amount of data needed toobtain a good representation of the articles published in aparticular journal. Asia Pacific Journal of Tourism Manage-ment and Journal of Tourism and Cultural Change were notincluded in the SSCI until 2009. Thus, these two journalswere excluded from this study. As a result, 10 tourismjournals with their articles published during 2008–2012formed the basis of the analysis. They are regarded byscholars as high quality journals with wide recognition inthe field (McKercher et al., 2006). Table 2 gives an overviewof the 10 journals, listed alphabetically. The selected timeframe is warranted because the main purpose of this paperwas to detect the subject areas contributed through scholarlyefforts in the most recent years, rather than overall tourismresearch trends. Previous research suggests that a five-yearperiod is sufficient to reveal the interrelatedness of articles;and studies based on five-year data have a long tradition inpublished research (Samiee & Chabowski, 2012).

The data used for the analysis were downloaded fromWoK on 24 February 2013. The data set consists of 2545articles published in the 10 journals. Only full-length articleswere analyzed because this type of document providescomplete bibliographic records to support BC analysis. Assuch, book reviews, conference reports, editors’ commentsand research notes were excluded. As mentioned before,

Table 2. Overview of journals included in the study

Journal title (acronym)Year first issue

published

Year firstincludedin SSCI

2011impactfactor

Country ofpublication

Scholar*awareness

Quality*rating

Articlesincludedin study

1. Annals of Tourism Research (ATR) 1973 1982 3.259 USA 85.7 4.6 2772. Current Issues in Tourism (CIT) 1998 2008 0.836 England 44.9 3.6 1833. International Journal of Tourism Research (IJTR) 1989 2008 0.816 England 56.1 3.5 2264. Journal of Travel Research (JTR) 1972 2008 1.579 USA 71.0 4.2 2295. Journal of Travel & Tourism Marketing (JTTM) 1992 2008 0.835 USA 61.1 3.6 2626. Journal of Sustainable Tourism (JST) 1993 2008 1.929 England 59.9 3.8 2337. Scandinavian Journal of Hospitalityand Tourism (SJHT)

2001 2007 0.630 Norway 13.1 3.1 114

8. Tourism Economics (TE) 1995 2008 0.579 England 43.9 3.8 3229. Tourism Geographies (TG) 1999 2008 0.633 England 44.9 3.7 12610. Tourism Management (TM) 1980 1994 2.597 England 79.3 4.3 573

*Scholar awareness and quality rating are based on the work of McKercher, Law and Lam (2006).

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

the year 2008 was used as the cutoff year because 7 out of the10 journals were not indexed by SSCI until that year.

Analytical toolCobo (2011) compared nine bibliometric analysis tools andfound that there are overlapping functions but that no singletool can do all the analyses that the other tools can. This studyused CATAR developed by Tseng (2010a) to conduct thebibliometric analysis because it supports multistage clustering(MSC; for hierarchical classification building) that is not foundin other tools. CATAR is a freeware that combines the strengthsof content analysis and bibliometric methods and is tailored toanalyze those publication records from WoK. The analysis thatCATAR provides can be used to summarize the background ofa research field, obtain an overview of prominent researchtopics, obtain a breakdown analysis of various actors (authors,institutes and countries), track trends and knowledge develop-ment, suggest hypotheses for further exploration and offerevidence-based, data-driven, bottom-up quantitative informa-tion to help describe a scientific field (Tseng, 2010a).

Bibliographic coupling is an embedded analytical functionof CATAR and is built for grouping papers. Publications thathave a similar research focus are clustered into the same sub-ject area based on their shared references. CATAR is also ableto extract keywords from the titles and abstracts for contentanalysis to assist the interpretation of subject areas. Further-more, it yields a quantitative summary from the bibliographicdata for each subject area. Detailed explanations of themethods and algorithms used in CATAR can be found inthe works of Chang et al. (2010) and Tseng and Tsay (2013).

For this paper to be self-contained, the following brieflydescribes the six analysis steps used byCATAR to help readersunderstand the analyzed results presented in later sections:

Step 1: Text segmentation. Each record from WoKcontains about 40 fields, including the title, authors, authors’addresses, publication years, etc. This step identifies at leastseven of them for use in our analysis. Below lists their ab-breviations, meanings and examples:AU: authors, e.g. Dolnicar, S; Grabler, K; Grun, B andKulnig, A.TI: publication title, e.g. Key drivers of airline loyalty.SO: journal title, e.g. Tourism Management.AB: publication’s abstract.C1:authors’ countries.CR: normalized citations, e.g. DOLNICAR S, 2004, JTRAVEL RES, Vol. 42, p. 244.PY: year of publication, e.g. 2011.

Step 2: Similarity computation. After identifying each field,similarities between each pair of articles are calculatedbased on the common citations normalized by the individ-ual citations each article has. This kind of similarity formsthe basis of BC, where it is believed that the more of thesame references two articles cite, the more likely the twoarticles are about the same topic.

Step 3: MSC. Next, a clustering algorithm called completelinkage clustering described in Salton (1989) is used to

detect the underlying knowledge structure of the corpus.The algorithm regards each article as a singleton clusterat first. It then groups the most similar pair of clusters (withsimilarity larger than a threshold) into a larger cluster. Thesame grouping rule applies again to the remaining clustersand newly created ones, where the similarity between anytwo clusters is defined as the minimum similarity betweenany pairs of articles each residing in the opposite cluster.This process repeats until no clusters can be merged. In thisway, each of the articles is assigned to a cluster automatically.For thousands of articles to be analyzed like the case in thisstudy, one-time clustering results may not yield ideal clustersizes and manageable cluster numbers for manual analysis.An MSC strategy was therefore adopted. The MSC involvesthe following process in each stage: eliminate the outliers(clusters having low similarities with others), treat eachremaining cluster as a virtual article (i.e. the references ofthe cluster are the union of all the references of the articlesin the cluster) and cluster the virtual articles in the sameway as described above. The MSC repeats the sameprocess stage by stage. In this way, articles are groupedinto concepts, which are further clustered into topics,which, in turn, can be grouped into categories or subjectareas. Although this is not always the case, it representsan expected ideal knowledge structure for the corpus.Through this MSC, logical self-organization of thecorpus can be derived automatically.

Step 4: Cluster labeling. Although articles are now orga-nized into clusters, analysts need to read through the titlesor even abstracts in individual clusters to know theircontent. To help analysts spot the topic for each cluster with-out much effort, a text mining approach (Tseng, 2010b) isused to generate cluster descriptors automatically. By sortingthe correlations between each pair of term and cluster in de-creasing order, the top five terms for each cluster are selectedas the cluster’s descriptors to help analysts identify and labelthe topic of the cluster. For this paper, two analysts withtourism domain knowledge performed the cluster labeling.

Step 5: Facet analysis. Once the topics have beendetected, it becomes easy to cross tabulate the topics withother facet information. This kind of cross analysis oftenleads to more information than single facet analysis canprovide. For example, it is possible to know the topic dis-tribution of all authors (and, thus, the major domains oftheir expertise), instead of knowing only their productiv-ity. The time series of yearly distribution of the articlesfor each topic can also be listed, and topic trends cantherefore be quantified by various techniques, such asthe slope of the linear regression line that best fits the timeseries (Tseng et al., 2009). Such analyses are the major fo-cus of our study and represent the main differences of ourapproach from the approaches of others. We shall seemore cross analysis examples in the next section.

Step 6: Visualization. To represent the detected knowl-edge structures, two techniques are used: one is the previ-ously introduced MSC and the other is multidimensional

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

scaling (MDS; Kruskal, 1997). Based on the precalculatedsimilarities between each topic, the MSC methodorganizes the topics in a hierarchical way. This creates astructure that is readily available as a folder tree or topictree representation, where detailed information about eachcluster (such as cluster descriptors) can be easily shown.Similarly, from the pairwise similarities, the MDS tech-nique computes the coordinates of each topic in the speci-fied dimensions of the Euclidean space, which are usually 2or 3 for ease of visual interpretation.With these coordinates,a topic map is created, where comparative informationamong the clusters (such as relative cluster size, interclustercloseness and spatial topic distribution) can be easilyvisualized. These two representations allow different waysto explore the corpus. The results from each clustering stagecorrespond to the folders in the topic tree, which reflects thetopic hierarchy imposed on the clustering results. Throughthe topic map, relations among the detected topics are visu-alized in a two-dimensional space, which helps understandthe intercluster relationship by distance and orientation.

RESULTS

Subject areasThe steps described above yielded 12 categories of tourismresearch topics and 41 subtopics. As shown in Table 3, thesetopics are labeled as follows: (i) tourism marketing, develop-ment and management (TMDM); (ii) sustainable tourism (ST);(iii) tourism economics (TOE); (iv) knowledge transfer andstrategic management (KTSM); (v) tourist profile and con-sumption pattern (TPCP); (vi) politics, policy and governmen-tal Issues (PPI); (vii) responsible tourism (RT); (viii) touristmovement patterns and trade (TMPT); (ix) urban tourism(UT); (x) crisis and disaster management (CDM); (xi) meeting,incentives, conventions and exhibitions (MICE); and (xii)economic valuation and business performance (EVBP). Asdescribed in the Method section, these 12 clusters emergedautomatically from several subclusters. For example, TMDMincludes 10 subtopics. Seventy-eight documents are groupedinto its first subtopic, with a focus on stakeholder collaborationand network. Additionally, the subtopics of destination image,travel satisfaction and destination loyalty have attracted theattention of most scholars in contributing to this subject area.

As seen in Table 3, the topic of TMDM can be consideredas the mainstream research topic in tourism research. Therewere 1260 articles classified into this topic, representing55.9% of the 2254 research articles from 2008 to 2012. Thistopic was followed by TOE, accounting for 11.8% of totalpublications. ST ranked third, and KTSM was ranked fourth.Finally, TMPT, UT and CDM received less researchattention. It should be noted that the large share of TMDMis due to its large number of subtopics, which share morecommon references than the other 11 subject areas.

The 12 topics cover 2254 articles of the original 2545.That is, 291 articles (11.4%) were removed as outliers duringthe MSC. As described by Chang et al. (2010), outliers arearticles ‘dealing with independent and probably less-noticedissues (p. 321)’. These articles form numerous small clusters,

a phenomenon that resembles the long tail effect reflected instatistics on online book sales. It could also be that thesesmall clusters represent articles from other disciplines withcitation patterns that do not match citation conventions intourism research and therefore do not map onto the salienttopic domains. Articles included in these small clusters areinteresting since they can be seen as papers that introducenew knowledge and innovation to tourism by drawing onnew literature.

Characteristics of subject areasFigure 2 maps the spatial relations based on BC similarityamong these topics, and each circle in the figure denotes atopical cluster, while the size of the circle reflects the numberof articles classified into the topic. The spatial relationshipamong the centers of the circles indicates the topical similarity(similar citation patterns) among them. For example, TMDM,ST, KTSM and TPCP are located in close proximity on thelower left-hand side of the map, which indicates similarity intheir citation patterns. Additionally, CDM and MICE areclosely related to each other. Figure 2 further shows thatTMPT, UT and EVBP are relatively independent topic groupslocated on the right of the map. Although TMDM is the main-stream research topic, it is TOE that emerges at the center ofthe topic map and thus relates, in the broadest sense, to allthe topics identified in our analysis. Alternatively, using aninterpretation focused on citations, TOE can be said to sharebibliographic sources with the other topics.

Figure 2 clearly shows the popularity of each topic andthe focus of scholars based on the sizes of the specific subjectareas. Studies related to TMDM are undoubtedly the topicthat has attracted the most attention. Studies on this topicaim to provide quality service and increase tourist satisfac-tion from both the supply and demand aspects. The subjectarea consists of 10 subtopics, including (i) stakeholdercollaboration and networking; (ii) community planning andST development; (iii) travel experience and sense of place;(iv) authenticity, heritage and identity; (v) destination image,tourist satisfaction and loyalty; (vi) destination choice anddecision-making; (vii) alternative tourism; (viii) risk and safetymanagement; (ix) effects of information and communicationtechnology (ICT); and (x) place representation. Specifically,destination image, tourist satisfaction and loyalty were themost published subtopics, followed by travel experienceand sense of place. This finding demonstrates that tourismresearch is mainly focused on travel experience and business-related issues. The subtopic of community and ST develop-ment has also attracted considerable scholarly attention.

The second largest topic, TOE, comprises four subtopicsand includes studies focused on demand forecast models,growth management, productivity and competitiveness atthe regional, national and international levels. The subtopicsalso suggest that research efforts have focused on the eco-nomic effects of tourism, particularly international tourism.

The topic of ST refers to studies that adopt an environmentallysensitive perspective to examining tourism effects. Subtopicsinclude the examination of corporate social responsibility, identi-fication of the impacts of tourism development, disclosure ofpower struggles and conflicts of interest among various parties,

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

the calculation and reduction of carbon emissions and theinfluence of climate change on tourism seasons andmanagementof protected areas. Visual examination reveals a close relation-ship between ST and TMDM in the topic map. Such findingsecho the point made by Swain et al. (1998) that the interconnec-tions among subject fields define the meaning of tourism.

Being closely related to TMDM, studies clustered inTPCP examine the consumption patterns and travel stylesof tourists with different profiles. KTSM comprises foursubtopics, including eCommerce, innovation implementa-tion and knowledge management, interdisciplinary researchand curriculum review. The subtopics of KTSM apparentlyshared common references related to strategic management,learning and knowledge creation.

Studies on CDM concern the impacts of disaster and crisison hosts and to the perceptions of guests and how to prevent,react and recover from disaster and crisis. UT is primarilyconcerned with culture, meaning and place making in urbancontexts. EVBP revolves around environmental economics.Studies on this topic focus on the estimation and calculationof value and benefits for public goods, environmental ameni-ties and natural resources and attempt to determine the truevalue of an ecosystem.

Compared with previous studies, topics like education,sport and leisure, research methods and transport did not emergeas major themes. Many topics identified by previous workswere grouped into the same subject areas. For example, attrac-tion management, destination marketing and management,

Table 3. Topic categories extracted

Abbreviation Topics Documents Subject areas Subtopics (number of articles grouped in the subtopic)

TMDM Topic 1 1260 Tourism marketing,development andmanagement

•Stakeholder collaboration and networking (78)10 subtopics •Community planning and sustainable

tourism development (184)•Travel experience and sense of place (228)•Authenticity, heritage and identity (77)•Destination image, tourist satisfaction and loyalty (389)•Destination choice and decision making (180)•Alternative tourism (50)•Risk and safety management (47)•Effects of ICT (11)•Place representation (16)

ST Topic 2 184 Sustainable tourism •Sustainability, ecotourism and environment (74)Four subtopics •Climate change, carbon emission and travel scenario (60)

•Corporate social responsibility and green tourism (36)•Tourism in protected areas (14)

TOE Topic 3 266 Tourism economics •Supply and growth management (19)•Economics in various tourism sectors (32)Four subtopics•Destination competitiveness modeland industry structure (66)•International demand (149)

KTSM Topic 4 132 Knowledge transfer andstrategic management

•eCommerce (51)•Innovation implementation and knowledgemanagement (34)

Four subtopics

•Curriculum review (14)•Interdisciplinary research (33)

TPCP Topic 5 91 Tourist profiles andconsumption patterns

•Culture and demographics (37)•Travel, passenger, holiday experience (42)Three subtopics•Travel style, lifestyle and distance segmented (12)

PPI Topic 6 67 Politics, policy andgovernmental issues

•Policy, policy implementation, governance (32)Three subtopics •Culture, ethnic, role of interest groups (22)

•Effect of tourism organizations (13)RT Topic 7 56 Responsible tourism •Value, attitude, and behavior in natural areas (17)

Three subtopics •Tourism management in natural areas (27)•Willingness to pay (12)

TMPT Topic 8 21 Tourist movementpatterns and trade

•Econometric, evidence using aggregateand country level (13)Two subtopics•Reception, dilemma, destination development (8)

UT Topic 9 27 Urban tourism •Cultural heritage and city image (18)•Planning and management of urban precincts (9)Two subtopics

CDM Topic 10 39 Crisis and disastermanagement

•Attitude formation, Black Saturday, bushfire, Katrina (9)Two subtopics •Crisis, disaster, scenario plan, tourism disaster (30)

MICE Topic 11 70 Meeting, incentives,conventions and exhibitions

•Estimation of economic impact (38)•Culture, festival, game and event (32)Two subtopics

EVBP Topic 12 41 Economic valuation andbusiness performance

•Firm performance and trade (14)•Travel cost, contingent, value (27)Two subtopics

n= 2254 articles.

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

image and branding, tourism planning and special interest tour-ism identified in the work of Park et al. (2011) were groupedinto TMDM. Meanwhile, previous studies did not detect theemergence of certain subtopics, such as climate change, carbonemissions, corporate social responsibility, eCommerce andknowledge management. These topics reveal growing schol-arly concerns regarding the degradation of the ecological envi-ronment. This result also suggests the occurrence of a‘twigging effect’. Twigging is the process whereby subjectareas split into subtopics and may result from new scholarswith different backgrounds entering a field, as well as shiftsin funding to certain areas.

Yearly distribution for each topicTable 4 lists the growth rate of each topic during the past fiveyears, in terms of the slope of the linear regression line thatbest fits the yearly published record numbers of articles.Unsurprisingly, TMDM exhibits the largest growth, whichsuggests that this topic continues to attract scholarly atten-tion. Studies motivated by environmental concerns andsustainability emerged as a focus of researchers, as reflectedin the growth rate of 3.0% (ranked No. 3) for the topic of ST.

TOE has growth of 5.5% (ranked No. 2), revealing concernsregarding the economic influence of tourism continuouslyattracting the attention of tourism scholars. Finally, PPI andMICE have not grown during the past five years.

Topic distribution for each journalTable 5 lists the topic distribution of the 10 journals and thusreveals the individual characteristics of the journals. Thetable suggests that some journals only published articles oncertain topics, while others had a broader focus. TourismManagement, Tourism Economics and Journal of TravelResearch are the only three journals that attract publicationsacross all the topics examined. Most papers published inTourism Management are focused on marketing and manage-ment. A large proportion of articles published by Tourism Eco-nomics dealt with TOE. Tourism Economics also publishednumerous articles on MICE and EVBP. Unsurprisingly, withthe aim to increase the understanding of sustainable tourism,the Journal of Sustainable Tourism contains a significantamount of articles on the topic of ST. The Journal of Sustain-able Tourism also published the largest number of articles onRT and PPI. Current Issues of Tourism would provide a good

Figure 2. Topic map rendered by multidimensional scaling from multistage clustering of over 2000 articles in 10 tourism journals(2008–2012). This figure is available in color online at wileyonlinelibrary.com/journal/jtr

Table 4. Number of articles by topic and by year

YearTopic 1 Topic 2 Topic 3 Topic 4 Topic 5 Topic 6 Topic 7 Topic 8 Topic 9 Topic 10 Topic 11 Topic 12

TMDM ST TOE KTSM TPCP PPI RT TMPT UT CDM MICE EVBP

2008 192 33 38 24 17 17 10 3 4 5 18 72009 190 21 49 23 10 10 11 1 4 1 18 92010 235 47 56 29 16 15 10 4 4 12 10 82011 290 49 66 24 24 14 13 7 8 11 9 72012 353 34 57 32 24 11 12 6 7 10 15 10Growth rate 42.2 3.0 5.5 1.7 2.8 0.0 0.6 1.2 1.0 2.0 0.0 0.4

n= 2254 articles.

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

source for finding articles related to crisis and disaster manage-ment. Additionally, tourism scholars perceived the Journal ofTravel Research as a marketing and management-orientedjournal (Goeldner, 2011), and this is supported by the figuresin Table 5. Annals of Tourism Research primarily publishespapers in TMDM and has very limited publications on othertopics. Papers of TOEmostly appeared in Tourism Economics,Tourism Management, the Journal of Travel Research andInternational Journal of Tourism Research, as identified bySong and Li (2008). Finally, the topics of KTSM and TPCPmostly appeared in Tourism Management and the Journal ofTravel and Tourism Marketing. These findings suggest thatsome tourism journals, such as Journal of Sustainable Tourism,are highly specialized, while others, such as Tourism Manage-ment, have broad coverage.

Geographic distribution of topicsTable 6 lists the top 10 contributing countries for each topic.The geographical origin of articles was determined based onthe data provided by the paper author(s). Countries contribut-ing less than three articles were excluded. A total of 69 coun-tries contributed to tourism research published in the 10journals during 2008 to 2012. Regarding the top 10 coun-tries, English-speaking countries, including the USA, theUK and Australia, accounted for the bulk of tourism journalarticles (45.1%). The USA ranked first in research productionfor seven topics, namely TMDM, TOE, TPCT, PPI, RT,TMPT and EVBP. The UK ranked the first for three topics,namely KTSM, UT and MICE. Researchers from Australialed in terms of research on ST and CDM. Besides theabove-mentioned English-speaking countries, non-English-speaking countries such as China, Demark, France, Italy,Germany, Malaysia, Norway, Portugal, South Korea, Spain,Sweden and Taiwan also ranked in the 10 most productivecountries. China was ranked among the 10 most productivecountries for nine topics and invested particular effort inthe topics of TMDM and TOE. Spain focused on TMDM,TOE and ST, while researchers from Taiwan focused onTMDM, TOE, KTSM and TPCT. Benckendorff (2009)examined papers by Australian and New Zealand researcherspublished in Annals of Tourism Research and suggested that

researchers from different regions may have different foci interms of area of study. Although the results of this studyshowed that the focus of most research topics was not regionspecific, it is notable that CDM was the only geographicallyspecific topic. In this topic, either the research objects, suchas volcano eruption and tsunami, were region-specificphenomena or the authors were residents that lived near thelocation of the crisis or disaster. Taking natural disasters,such as forest fires, earthquakes or hurricanes, as examples,studies by US-based scholars revealed the effects of Hurri-cane Katrina; studies by Taiwanese scholars examined thecrisis management related to earthquakes and typhoons,while Canadian scholars investigated recovery strategies af-ter forest fires.

Most cited documents by subject areaJournal publications are considered as a major source ofknowledge, and references in journal articles indicate theorigin of knowledge. The cited references represent the poolof archival knowledge from which the authors learn the stateof existing knowledge and generate new research ideas.Thus, the most cited references for each topic could be usedto indicate which articles in a given topic should be read.Tables 7a and 7b list the most cited documents in each ofthe 12 identified subject areas during the period from 2008to 2012. Take Topic 2 (ST) in Table 7a as an example, re-search focusing on different aspects examined the environ-mental impacts of tourism, delineated the nature of ST andidentified supply and demand side concerns related to itsmanagement and development of sustainable tourism. Stud-ies on this topic aimed to maintain the value and integrityof ecosystems and to reduce the negative consequences oftravel activities. Scholars who attempted to understand thecharacteristics and behavior of ecotourists thus obtainedknowledge from the work of Weaver and Lawton (2002).Scholars who were interested in assessing the global environ-mental consequences of tourism cited Gössling (2002). Addi-tionally, Weaver (2005) presented a comprehensive modelthat adopts a holistic perspective to integrate ecosystems withassociated human influences. To summarize, anyone interestedin understanding the concepts of ecotourism, the effect of

Table 5. Percentage of articles by topic and by journal

TopicsTopic 1 Topic 2 Topic 3 Topic 4 Topic 5 Topic 6 Topic 7 Topic 8 Topic 9 Topic 10 Topic 11 Topic 12

TotaldocumentsTMDM ST TOE KTSM TPCP PPI RT TMPT UT CDM MICE EVBP

ATR 14.8 7.1 2.6 8.3 14.3 17.9 3.6 19.0 7.4 10.3 5.7 0.0 259CIT 6.1 13.0 2.3 6.8 4.4 13.4 10.7 14.3 18.5 25.6 2.9 0.0 155IJTR 9.6 8.2 9.0 11.4 8.8 6.0 0.0 19.0 11.1 2.6 10.0 4.9 204JTR 11.3 4.9 9.8 8.3 12.1 1.5 8.9 14.3 7.4 7.7 7.1 7.3 222JTTM 13.4 1.6 2.3 18.9 20.9 4.5 5.4 0.0 14.8 10.3 5.7 0.0 240JST 8.0 33.7 0.8 0.8 4.4 20.9 28.6 0.0 7.4 12.8 1.4 4.9 210SJHT 4.0 4.9 1.1 3.0 3.3 3.0 19.6 0.0 0.0 10.3 10.0 2.4 95TE 4.8 2.7 46.6 5.3 4.4 3.0 3.6 23.8 3.7 2.6 37.1 53.7 260TG 5.6 3.8 1.5 4.5 2.2 10.4 3.6 4.8 7.4 2.6 0.0 2.4 103TM 22.2 20.1 24.1 32.6 25.3 19.4 16.1 4.8 22.2 15.4 20.0 24.4 506Total % 55.9 8.2 11.8 5.9 4.0 3.0 2.5 0.9 1.2 1.7 3.1 1.8 100

n= 2254 articles.Note 1: Topics and journals are presented in abbreviations, which are listed in Tables 2 and 3.Note 2: There are 2545� 2254 = 291 articles not included in this table.Note 3: Percentages in each of the 12 topics (in columns) sum up to 100%.

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

Table

6.Topicsby

countryof

author

Topic

1Topic

2Topic

3Topic

4Topic

5Topic

6Topic

7Topic

8Topic

9Topic

10Topic

11Topic

12

TMDM

ST

TOE

KTSM

TPCP

PPI

RT

TMPT

UT

CDM

MICE

EVBP

1260

documents

184

documents

266

documents

132

documents

91documents

67documents

56documents

21documents

27documents

39documents

70documents

41documents

USA

332

Australia

41USA

54UK

26USA

25USA

17USA

20USA

7UK

6Australia

10UK

18USA

24Australia

184

USA

35China

42USA

25Australia

13UK

15Australia

14UK

4USA

5New

Zealand

6USA

12S.K

orea

9

UK

173

UK

30Spain

40China

23UK

11Australia

15Sweden

10China

3Australia

4UK

5Australia

11Taiwan

6China

126

Canada

18UK

26Australia

18China

11Canada

7Finland

4S.K

orea

3USA

5Spain

6Australia

5Spain

90New

Zealand

18Taiwan

26Taiwan

8Taiwan

10China

6Canada

4Taiwan

4Sweden

6

S.K

orea

79Norway

11Australia

26Portugal

7S.K

orea

7New

Zealand

4UK

3China

4Italy

4

Taiwan

69Germany

10Italy

23New

Zealand

7Turkey

6Turkey

3Malaysia

3Canada

3China

4

New

Zealand

57Spain

10Portugal

10S.K

orea

6Canada

5S.K

orea

3Canada

56China

9France

9Spain

6Denmark

4S.A

frica

3Norway

44Sweden

8Malaysia

9Austria

4New

Zealand

4

Note:

Ineach

topic,countriescontributin

gless

than

threearticlesarenotlisted.

The

geographic

originsof

articlesweregivenby

authorsandwererecorded

inWoK

data

with

outanymodificatio

n.n=2254

articles.

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

climate change on tourism and tourists’ knowledge and aware-ness of tourism impacts can locate and read these articles to un-derstand the central aspects of this research cluster.

DISCUSSION

Technology has now reached a point where computer-assistedbibliometric and content analyses can provide evidence-basedsnapshots of the current status and future prospects of thetourism field by analyzing large volumes of data. However,for the sake of future research, it is important to identify theadvantages and limitations of this technique.

Epistemological characteristics of clustersThe subject areas identified by the given data set present thecurrent research efforts of the tourism field from anepistemological perspective. However, such research only re-flects the primary interests of current researchers whose workis published in the SSCI-listed tourism journals. The identifiedsubject areas are based on shared citations, which means that itexcludes works by researchers who did not follow tourismcitation conventions or who built their research based on differ-ent literature. Nevertheless, the analytical results of this studydo provide an important overview of the nature of a substantialpart of the contemporary tourism research. Tables 7a and 7bhelp verify these subject areas. This is important for thosewho advocate for certain areas (e.g. within a curriculum) orfor tourism students who wish to understand the field.

This study presents evidence supporting the idea that theresearch topics within the tourism field grow in diversedirections. The results reveal 12 main subject areas and 41

subtopics. The numerous subtopics and the emergence ofcertain topics signal the growth of the tourism knowledge base.As the tourism field matures, much more of the ‘twiggingeffect’ of subtopics is expected. Nevertheless, 11% ofdocuments were excluded from the result because their citedknowledge is less known to the tourism community. Oneway to interpret this is that many issues have attracted the at-tention of scholars but have failed to accumulate critical mass.

Unlike previous studies (e.g. Ballantyne et al., 2009; Parket al., 2011) that use expert judgment to categorize literature,the BC approach adopted by this study yielded relatively moreobjective subject areas. Articles that cite the same literaturewere automatically clustered into the same subject area. Thisapproach is able to reveal hidden relationships amongsubtopics in terms of knowledge sources. Given that 56% ofpapers fall into the broad category of TMDM, good reasonsexist to believe that the subtopics within this subject area sharea relatively strong cognitive resemblance. Epistemologicallyspeaking, this finding indicates that the efforts of mainstreamcontemporary tourism research remain business related andsuggests the existence of a dominant groupwithin the scholarlytourism community.

The results roughly support the subject areas identified byXiao and Smith (2006) and Swain et al. (1998), who focusedsolely on the Annals of Tourism Research. Subject areasidentified in these papers were primarily grouped into ourfirst subject area: TMDM. This study confirmed thecontinued focus on this broad topic area. Particularly,TMDM was the most frequently studied subject area acrossall journals during the assessment period. Clearly, topicssuch as TMDM, TOE and TPCP focus on the study of peoplewith the intention to provide quality tourism services and

Table 7a. The top five most cited references for each topic from 2008 to 2012

Topics Sources of citation Times cited Topics Sources of citation Times cited

Topic 1: tourism marketing development and management Topic 4: knowledge transfer and strategic managementFornell (1981) JMR, Vol. 18, p. 39 132 Buhalis (1998) TM, Vol. 19, p. 409 18

Anderson (1988) PB, Vol. 103, p. 411 93 Buhalis (2008) TM, Vol. 29, p. 609 16Crompton (1979) ATR, Vol. 6, p. 408 81 Buhalis (2002) TM, Vol. 23 p. 207 15

Butler (1980) CG, Vol. 24, p. 5 80 Novelli (2006) TM, Vol. 27, p. 1141 14Hair (1998) MDA 77 McKercher (2003) TM, Vol. 27, p. 1235 12

Tribe (1997) ATR, Vol. 24, p. 638 12

Topic 2: sustainable tourism Topic 5: tourist profiles and consumption patternsWeaver (2007) TM, Vol. 28, p. 1168 17 Burnett (2001) JTR, Vol. 40, p. 4 15Gössling (2002) GEC, Vol. 12, p. 283 16 Shaw (2004) TM, Vol. 25, p. 397 15Becken (2007) JST, Vol. 15, p. 351 16 Daruwalla (2005) ATR, Vol. 32, p. 549 13Weaver (2002) JTR, Vol. 40, p. 270 14 McKercher (2003) TM, Vol. 24, p. 465 13Weaver (2005) ATR, Vol. 32, p. 439 14 Smith (1987) ATR, Vol. 14, p. 376 12

Topic 3: tourism economics Topic 6: politics, policy and governmental issuesSong (2008) TM, Vol. 29, p. 203 39 Hall (1994) TPPPP 12

Li (2005) JTR Vol. 44, p. 82 38 Hall (1995) TPP 9Oh (2005) TM, Vol. 26, p. 39 37 Tyler (2001) CIT, Vol. 4, p. 210 9

Song (2000) TDMF 35 Dredge (2006) JST, Vol. 14, p. 562 8Dritsakis (2004) TE, Vol. 10, p. 305 34 Bramwell (2007) ATR, Vol. 34, p. 766 8

Pforr (2006) ATR, Vol. 33, p. 87 8Dredge (2006) TM, Vol. 27, p. 269 8

Note: ATR, Annals of Tourism Research; CG, Canadian Geographer; CIT, Current Issues in Tourism; GEC, Global Environmental Change; JMR, Journal ofMarketing Research; JST, Journal of Sustainable Tourism; JTR, Journal of Tourism Research; PB, Psychological Bulletin; TM, Tourism Management.Books include MDA, Multivariate Data Analysis; TDMF, Tourism Demand Modeling and Forecasting: Modern Econometric Approaches; TPP, Tourism andPublic Policy; TPPPP, Tourism and Politics: Policy, Power and Place.

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

increase the profitability of tourism. Such economic-orientedand management-oriented research remains the primaryfocus. However, Annals of Tourism Research, Journal ofTravel Research and Tourism Management were the threeprimary data sources. The analysis shows that the journalshave specific emphases, and thus, subject categorizationbased mostly on these three sources may underestimate theimportance of certain subject areas. Using data from limitedsources is also a major criticism of past research, and thisconcern was further substantiated by the current results.

While Jafari’s (2001) platform concept suggested aparadigm shift in tourism research, the results of this studyindicate that those platforms coexist simultaneously. Addi-tionally, the platform framework appears to require updating.Macbeth (2005) proposed a fifth platform called the ethicsplatform. He stated that tourism should no longer be

perceived as an economic phenomenon and that future studiesshould address pleasure-seeking tourist behavior, the moralityof tourism development and the moral code of the tourismindustry. Closer examination of the interrelatedness of the sub-ject areas investigated in this work confirmed rising concernsregarding the sustainable use of natural resources and the envi-ronmental impact of tourism, which was also identified as animportant topic area in previous research (Swain et al., 1998;Xiao & Smith, 2006; Benckendorff, 2009).

The use of computer-assisted analysis toolsThe subject areas were identified by automatic clusteringmethods based on BC similarity. Although it may yield anoptimal clustering mathematically, the BC similarity doesnot precisely reflect human-perceived similarity. Conse-quently, the clusters derived automatically may not be

Table 7b. The top five most cited references for each topic from 2008 to 2012 (continued)

Topics Sources of citation Times cited Topics Sources of citation Times cited

Topic 7: responsible tourism Topic 10: crisis and disaster managementNewsome (2002) NAT 9 Faulkner (2001) TM, V22, P135 15

Eagles (2002) TNPP 8 Ritchie (2004) TM, V25, P669 9Eagles (2002) STPA 8 Cioccio (2007) TM, V28, P1 7

Manning (1991) SOR 7 Blake (2003) ATR, V30, P813 7Ajzen (1991) OBHD, V50, P17 6 Mansfeld (1999) JTR, V38, P30 6

Cottrell (2003) EB, V35, P347 6 Faulkner (2001) TM, V22, P331 6Ajzen (1980) UAPSB 6 Ritchie (2009) CDM 6

Tratalos (2001) BC, V102, P67 6

Topic 8: travel movement patterns and trade Topic 11: meeting, incentives, conventions and exhibitionsSonmez (1998) ATR, V25, p112 4 Dwyer (2004) TM, V25, P307 16

Kulendran (2000) AE, V32, P1001 4 Fletcher (1989) ATR, V16, P514 15Drakos (2003) JCR, V47, P621 4 Zhou (1997) ATR, V24, P76 14Sonmez (1998) ATR, V25, P416, 3 Hall (1992) HTE 11Pizam (1996) TRIS 3 Adams (1995) AE, V27, P985 9Enders (1991) TERM, V14, P49 3 Dwyer (2000) CIT, V3, P325 9

Judd Dennis (1999) TC 3 Crompton (2001) JTR, V40, P79 9Harrison (1992) TLDC 3 Jones (2001) IJTR, V3, P241 9

Higgins-Desbiolles (2006) TM, V27, P1192 3Enders (1992) KYKLOS, V45, P531 3

Morakabati (2007) TTRP 3Neumayer (2004) CR, V48, P259, 3

Khan (2005) JTR, V44, P171 3

Topic 9: urban tourism Topic 12: economic valuation and business performanceMcKercher (2002) IJTR, V4, P29 6 Haab (2002) VENR 7

Jansenverbeke (1986) ATR, V13, P79 5 Mitchell (1989) USV 6Law (1993) UTA 4 Markowitz (1952) JF, V7, P77 5Law (2002) UTV 4 Creel (1990) AJAE, V72, P434 5Page (1995) UT 4 Loomis (1997) REDC 5

Ashworth (1990) THC 4 Bishop (1979) AJAE, 61, P926 5Silberberg (1995) TM, V16, P361 4 Lang (1994) JPE, V102, P1248 5McKercher (2002) CTP 4Richards (1996) ATR, V23, P261 4Richards (2006) TM, V27, P1209 4

Note: Journals include AE, Applied Economics; AJAE, American Journal of Agricultural Economics; ATR, Annals of Tourism Research; BC, Biological Con-servation; CR, Conflict Resolution; EB, Environment and Behavior; EM, Event Management; IJCH, International Journal of Contemporary Hospitality Man-agement; IJTR, International Journal of Tourism Research; JTR, Journal of Tourism Research; OBHD, Organizational Behavior and Human DecisionProcesses; TERM, Terrorism; TM, Tourism Management; JF, Journal of Finance; JPE, Journal of Political Economy.Books include CDM, Crisis and Disaster Management; CTP, Cultural Tourism: The Partnership between Tourism and Cultural Heritage Management; REDC,Recreation Economic Decision: Comparing Benefits & Costs; SOR, Studies in Outdoor Recreation; STPA, Sustainable Tourism in Protected Areas: Guidelinesfor Planning and Management; TC, Tourist City; THC, The Tourist-historic City: Retrospect and Prospect of Managing the Heritage City; HTE, Hallmark Tour-ist Events: Impacts, Management and Planning; TLDC, Tourism and the Less Developed Countries; TNPP, Tourism in National Parks and Protected Areas;TTRP, Tourism, Travel Risk and Travel Risk Perceptions; VENR, Valuing Environmental and Natural Resources: The Econometrics of Non-Market Valuation;USV, Using Surveys to Value Public Goods: The Contingent Valuation Method; UAPSB, Understanding Attitudes and Predicting Social Behavior; UT, UrbanTourism; UTA, Urban Tourism: Attracting Visitors to Large Cities; UTV, Urban Tourism: The Visitor Economy and the Growth of Large Cities.

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

perfect. In most cases, many clusters would be logical andacceptable, while a few would appear to be flawed. In com-parison, content analysis relying heavily on expert judgmentcould lead to more accurate results but tends to yield incon-sistent subject classification, which makes it difficult forscholars to compare the results analyzed by different researchgroups. Thus, the main advantage of our approach is that theresults will be the same provided the same data set is used.

This study showed that such an automated tool provides anefficient approach to perform content and citation analysis toidentify the subject areas within tourism research. One of themain features of CATAR is the visual representation of quali-tative data, which is useful in delineating and interpreting com-plex sets of relationships. The clustering output (the clusterdescriptors and the map) presented in this investigation revealsreasonable results and thus can derive meaningful insights.However, the interpretation of the results still requires consid-erable domain expertise. As automatic as the process mayseem, CATAR only facilitates but does not replace humaninterpretation of what the clusters represent.

Limitations and future researchThe time frame of the data analysis was limited to five years.Although this is sufficient to reveal dominant subject areaswithin that period, in a dynamic field like tourism, a changein the period or the type of journals included in the data setcould yield different subject clusters, depending on howlarge of a change is made. This limitation suggests two futureresearch directions. One is that journals included in other da-tabases, such as Scopus, can be incorporated to expand thesize and duration of the data set. The other is that researchnotes, conference reports, monographs, dissertations, bookchapters, course syllabi or editors’ comments could also beincluded for a comprehensive study, as journal articles arenot the only source of ideas or knowledge for analysis. Suchpractices would permit testing of the stability of the identifiedsubject areas in a more rigorous way.

CONCLUSION

This article presents a new methodology for subject analysis,which was previously conducted manually in tourism re-search, by combining BC, multistage clustering and contentanalysis. Twelve subject areas from 10 tourism journals wereidentified during the five-year period. The use of BC enablesthis study to group articles with a similar epistemologicalbasis and thus provides insights regarding knowledge crea-tion within contemporary tourism research. For example,the emerging topical cluster of RT was not mentioned in pre-vious studies. Additionally, the references identified as beingthe most cited in each subject area could be regarded as must-know articles. Reading these articles might help graduatestudents, layperson scholars and practitioners to grasp thekey concepts and basic knowledge or identify researchopportunities regarding certain subject areas. Our findingshave further significant practical value in identifying journalslikely to accept particular articles, which, in turn, helpsauthors target the most suitable publication outlets for their

work. Future research can build upon this methodology andapply it in a periodical way to watch the evolution in tour-ism research.

ACKNOWLEDGEMENTS

This research is partially supported by the ‘Aim for the TopUniversity Project’ of National Taiwan Normal University(NTNU), sponsored by the Ministry of Education, Taiwan,R.O.C., and the ‘International Research-Intensive Center ofExcellence Program’ of NTNU and National Science Council,Taiwan, R.O.C., under Grant No. NSC 103-2911-I-003-301.

REFERENCES

Ballantyne R, Packer J, Axelsen M. 2009. Trends in tourismresearch. Aannals of Tourism Research 36: 149–152.

Benckendorff P. 2009. Themes and trends in Australian and NewZealand tourism research: a social network analysis of citationsin two leading journals (1994-2007). Journal of Hospitalityand Tourism Management 16: 1–15.

Benckendorff P, Zehrer A. 2013. A Network analysis of tourismresearch. Annals of Tourism Research 43: 121–149.

Borgman CL, Furner J. 2002. Scholarly Communication andBibliometrics. In Annual Review of Information Science andTechnology,Vol. 36, Cronin B (ed). Information Today: Med-ford, NJ; 3–72.

Boyack KW, Klavans R. 2010. Co-citation analysis, bibliographiccoupling, and direct citation: Which citation approach representsthe research front most accurately? Journal of the AmericanSociety for Information Science and Technology 61(12):2389–2404.

Broadus RN. 1987. Toward a Definition of "Bibliometrics".Scientometrics 12(506): 373–379.

Burdge RJ. 1983. Making leisure and recreation research a scholarlytopic: Views of a journal editor, 1972-1982. Leisure Sciences 6(1): 99–126.

Chang Y-H, Chang C-Y, Tseng Y-H. 2010. Trends of ScienceEducation Research: An Automatic Content Analysis. Journalof Science Education and Technology 19(4): 315–331.

Cheng C-K, Li X, Petrick JF, O’Leary JT. 2011. An examination oftourism journal development. TourismManagement 32(1): 53–61.

Cobo MJ, López-Herrera AG, Herrera-Viedma E, Herrera F. 2011.Science mapping software tools: Review, analysis, and coopera-tive study among tools. Journal of the American Society forInformation Science and Technology 62(7): 1382–1402.

Dembkowski S, Hanmer-Lloyd S. 1995. Computer applications – anew road to qualitative data analysis? European Journal ofMarketing 29(11): 50–62.

Garfield E. 1994. Using the impact factor. Current Contents 29(July18): 3–5.

Goeldner CR. 2011. Reflecting on 50 Years of the Journal of TravelResearch. Journal of Travel Research 50(6): 583–586.

Gössling S. 2002. Global environmental consequences of tourism.Global Environmental Change 12: 283–302.

Hall CM. 2005. Systems of surveillance and control: commentaryon ’An analysis of institutional contributors to three majoracademic tourism journals: 1992-2001’. Tourism Management26(5): 653–656.

Hall CM. 2011. Publish and perish? Bibliometric analysis, journalranking and the assessment of research quality in tourism.Tourism Management 32(1): 16–27.

Holden G, Rosenberg G, Barker K. 2005. Tracing thought throughtime and space: A selective review of bibliometrics in socialwork. Social Work in Health Care 41(304): 1–34.

Y. Yuan, U. Gretzel and Y.-H. Tseng

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr

Howey RM, Savage KS, Verbeeten MJ, Van Hoof HB. 1999.Tourism and hospitality research journals: cross-citations amongresearch communities. Tourism Management 20: 133–139.

Jafari J. 1990. Research and scholarship: The basis of tourismeducation. The Journal of Tourism Studies 1(1): 33–41.

Jafari J. 2001. The scientification of tourism. In Hosts and GuestsRevisited: Tourism Issues of the 21st Century, Smith V, BrentM (eds). Cognizant Communication Corporation: New York;28–41.

Jafari J, Ritchie BJR. 1981. Toward a framework for tourismeducation: Problems and prospects. Annals of Tourism Research8(1): 13–34.

Jamal TB, Smith B, Watson E. 2008. Ranking, rating and scoring oftourism journals: Interdisciplinary challenges and innovations.Tourism Management 29: 66–78.

Joseph KS. 2003. Quality of impact factors of general medicaljournals. BMJ 326(7383): 283.

Kennedy MM. 2007. Defining a Literature. Educational Researcher36(3): 139–147.

Kessler MM. 1963. Bibliographic coupling between scientificpapers. American Documentation 14(1): 10–25.

Kim S. 1998. Content Analysis: Annals of Tourism Research andJournal of Travel Research. (Unpublished MS thesis), Universityof Wisconsin-Stout.

Kim Y, Savage KS, Howey RM, Van Hoof HB. 2009. Academicfoundations for hospitality and tourism research: A reexaminationof citations. Tourism Management 30(November): 752–758.

Kruskal JB. 1997. Multidimensional Scaling and Other Methods forDiscovering Structure. In Statistical Methods for Digital Com-puters, Enslein K, Ralston A, Wilf HS (eds). Wiley: New York;296–339.

Lasswell HD. 1964. The structure and function of communicationin society. In The communication of ideas, Bryson L (ed).Cooper Square Publishers: New York; 37–51.

Leydesdorff L. 1997. Why Words and co-words cannot map thedevelopment of the sciences. Journal of the American Societyfor Information Science and Technology 48(5): 418–427.

Leydesdorff L, Zaal R. 1988. Co-words and citations. Relationsbetween document sets and environments. In Informetrics,Egghe L, Rousseau R (eds). Elsevier: Amsterdam; 87(88):105–119.

Macbeth J. 2005. Towards an ethics plaform for tourism. Annals ofTourism Research 32(4): 962–984.

McKercher B. 2008. A citation analysis of tourism scholars.Tourism Management 29(6): 1226–1232.

McKercher B, Law R, Lam T. 2006. Rating tourism and hospitalityjournals. Tourism Management 27(6): 1235–1252.

Moed HF. 2005. Citation analysis in research evaluation. Springer:Netherlands.

Park K, Philips WJ, Canter DD, Abbott J. 2011. Hospitality andTourism Research Rankings by Author, University, andCountry Using Six Major Journals: The First Decade of theNew Millennium. Journal of Hospitality & Tourism Research35(3): 381–416.

Salton G. 1989. Automatic Text Processing: The Transformation,Analysis, and Retrieval of Information by Computer. Addison-Wesley: MA.

Samiee S, Chabowski BR. 2012. Knowledge structure in interna-tional marketing: a multi-method bibliometric analysis. Journalof the Academy of Marketing Science 40: 364–386.

Schmidgall R, Woods RH, Hardigree C. 2007. Hospitality’s mostinfluential scholars: Fifteen years of citation analyses. Journalof Hospitality and Tourism Education 19(2): 32–4332.

Sheldon PJ. 1990. Journals in tourism and hospitality: The percep-tions of publishing faculty. The Journal of Tourism Studies1(1): 42–48.

Sheldon PJ. 1991. An authorship analysis of tourism research.Annals of Tourism Research 18: 473–484.

Song H, Li G. 2008. Tourism demand modelling and forecasting—Areview of recent research. Tourism Management 29(2): 203–220.

Swain MB, Brent M, Long VH. 1998. Annals and tourism evolving:Indexing 25 years of Publication. Annals of Tourism Research25(3001): 991–1014.

Tribe J. 1997. The indiscipline of tourism. Aannals of TourismResearch 24(3): 638–657.

Tribe J. 2008. Tourism: A Critical Business. Journal of TravelResearch 46(3): 245–255.

Tsaur S-H, Wang K-C, Chang T-Y, Wang W-C. 2008. An Analysisof Research Performance for Tourism, Leisure and HospitalityRelated Academic Institutions in Taiwan. Outdoor ResearchResearch 21(3): 81–99.

Tseng Y-H. 2010a. Content Analysis Toolkit for AcademicResearch (CATAR) Retrieved April 2, 2011, from http://web.ntnu.edu.tw/~samtseng/CATAR/

Tseng Y-H. 2010b. Generic Title Labeling for Clustered Docu-ments. Expert Systems With Applications 37(3): 2247–2254.

Tseng Y-H, Tsay M-Y. 2013. Journal clustering of Library and In-formation Science for subfield delineation using the bibliometricanalysis toolkit: CATAR. Scientometrics 95(2): 503–528.

Tseng Y-H, Lin Y-I, Lee Y-Y, Hung W-C, Lee C-H. 2009. A Com-parison of Methods for Detecting Hot Topics. Scientometrics81(1): 73–90.

Van Doren CS, Heit MJ. 1973. Where it’s at: A content analysis andappraisal of the Journal of Leisure Research. Journal of LeisureResearch 5: 67–73.

Vladutz G, Cook J. 1984. Bibliographic coupling and subject related-ness. Paper presented at the Challenges to an Information Society,Proceedings of the 47th ASIS Annual Meeting, White Plains.

Way KA, Harrington RJ, Ottenbacher MC. 2012. Hospitality Au-thor and University Productivity in the 21st Century. Journalof Culinary Science & Technology 10: 239–258.

Weaver DB. 2005. Comprehensive and minimalist dimensions ofecotourism. Annals of Tourism Research 32(2): 439–455.

Weaver DB, Lawton LJ. 2002. Overnight Ecotourist Market Seg-mentation in the Gold Coast Hinterland of Australia. Journalof Travel Research 40(February): 270–280.

Weiner G. 2001. The academic journal: Has it a future? EducationPolicy Analysis Archives, 9(9). Retrieved from Education policyanalysis archives website: http://epaa.asu.edu/ojs/article/view/338

Xiao H, Smith SLJ. 2006. The making of tourism research: Insightsfrom a Social Sciences Journal. Annals of Tourism Research33(2): 490–507.

Xiao H, Xiao Q, Li M. 2012. The Behavior of Citing: A Perspectiveon Science Communication Across Languages. Journal ofChina Tourism Research 8(3): 334–356.

Ying T, Xiao H. 2012. Knowledge Linkage: A Social NetworkAnalysis of Tourism Dissertation Subjects. Journal of Hospitality& Tourism Research 36(4): 450–477.

Zhao W, Ritchie JRB. 2007. An investigation of academic leadership intourism research: 1985-2004. Tourism Management 28(2): 476–490.

Nature of Contemporary Tourism Research

Copyright © 2014 John Wiley & Sons, Ltd. Int. J. Tourism Res. (2014)

DOI: 10.1002/jtr