Investigating Tourists’ Destination Choices – An Application of

Network Analysis

Marion Karl 1* and Christine Reintinger1

Received: 13/06/2016 Accepted: 19/07/2016

1 Department of Geography, LMU Munich. Postal address: Luisenstrasse 37, 80333 Munich, Germany.
* Corresponding author: tel: 0049 (0)89 2180 4039. e-mail: marion.karl@lmu.de.
Abstract

A better understanding of the complex destination choice process is highly relevant, both for academia
and practice. Tourism research tends to focus either on actually executed or hypothetical destination
choices. However, a discrepancy exists between these two types of destination choices which has
hardly been investigated. Moreover, past research often studies tourists and their attitudes, needs or
perceptions of destinations but not how destinations’ attributes affect destination choices. To
approach these two research gaps, this study concentrates not only on actual but also on hypothetical
destination choices to better understand differences in the evaluation of alternative destinations. This
study furthermore examines the role of the destination itself to discover the influence of destination
characteristics on destination choices. Therefore, network analysis and set theory are combined in a
new research approach which allows to analyse destination choices with varying closeness to reality
whilst preserving destination information. The analysis is based on a quantitative survey of German
tourists’ travel decision-making behaviour. The results reveal changes in destination choices from
multidimensional hypothetical choices to unidimensional actual and past choices. Furthermore, only
few destinations have a consistent position whilst most destinations are either more relevant for
hypothetical or actual destination choices.

© 2017 Varna University of Management. All rights reserved

Keywords: destination choice, travel decision-making, network analysis, set theory, Germany

Citation: Karl, M., C. Reintinger (2017) Investigating Tourists’ Destination Choices – An Application of
Network Analysis. European Journal of Tourism Research 15 pp. 112-130

Introduction high importance for tourists compared to other

Travel decisions are multilayered decisions with elements of the travel decision (Fesenmaier and
interdependent elements (e.g. destination, type Jeng 2000; Oppewal, Huybers, and Crouch
of accommodation) that “evolve in a decision 2015). Choosing a destination is a complex
process over time” (Dellaert, Ettema, and Lindh process where only the result is visible in the
1998, 313). A type of travel decision that form of tourist flows from a source market to a
receives special attention in tourism research is destination. Until a destination is finally chosen
the choice of a destination, probably due to the by a tourist it has to successfully complete

112 RESEARCH PAPER

 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

several stages of the destination choice (DC) (actual DCs) but also on dreams and
process. The reasons it is preferred to other imaginations of travelling occurring before the
alternative destinations during these steps are decision to travel has been initiated
manifold, interrelated and depend on external as (hypothetical DCs). The purpose of the paper is
well as internal factors (Um and Crompton to better understand differences in the
1990). evaluation of alternative destinations concerning
actual and hypothetical DCs. This study
Tourism research tends to focus either on furthermore aims to analyse the relevance and
actually executed or hypothetical DCs which role of the destination itself at different DC
covers up an existing discrepancy between sections in order to emphasise the importance
these two types of DC. A research gap exists in of destination characteristics in the DC process.
the understanding of DC as the discrepancy
between actual and hypothetical DCs has rarely A reason why destinations have not been paid
been investigated. For example, reasons why more attention in research on DC, in particular
tourists decide to travel to some destinations in quantitative studies, may be the
while others remain hypothetical “dream” methodological challenge to capture the
destinations are not clearly identified. Actual multiplicity of destinations that are considered
behaviour and imagination of travelling cannot during the DC process. This study therefore
be seen as equal (Decrop 2010; Karl, applies network analysis as it allows to identify
Reintinger, and Schmude 2015). This implies structures in the evaluation of alternative
that actual and hypothetical DC do not proceed destinations whilst preserving destination
under the same premises and that there are information. Based on a quantitative survey on
differences in the way tourists evaluate DC behaviour, alternative destinations of
alternative destinations for actually planned or different DC sections and their relations as
executed and hypothetical future holidays. competitors are analysed using network
However, knowledge about the discrepancy and analysis. This allows to capture the role of
how to overcome this discrepancy can be highly destinations on the basis of their position in the
relevant for the tourism industry. network. The results demonstrate that only few
destinations have a consistent position in all DC
DC is furthermore a negotiation process sections whilst most destinations are either
between tourists’ needs and what destinations more relevant in the hypothetical or actual DC
offer (Ankomah, Crompton, and Baker 1996). sections. Exemplary destinations are chosen to
While many past studies focus on tourists and illustrate how destination characteristics
their attitudes, needs or perceptions of determine the role of a destination in the DC
destinations, information on the destinations sections. The exemplary destinations
such as the geographic location is rarely furthermore allow to further look into
captured. There are exceptions in the form of interlinkages and relations between actual and
case studies which give important insights into a hypothetical DCs. The research approach and
specific destination (e.g. Botha, Crompton, and results of this study offer a wide range of
Kim 1999) but not into the DC process in a more application opportunities for practitioners which
comprehensive way. The lack of consideration are discussed in the conclusion of this article.
of the destination itself pushes one of the most
important aspects of DC into the background: Literature review and research questions
The destination and its distinct characteristics Various theories, approaches and methods
are important determinants of DC (Karl et al. have been applied to investigate DC from output
2015). oriented microeconomic models (e.g. Seddighi
and Theocharous 2002) to behaviouristic
To approach these two research gaps (i.e. models focusing on tourists’ behaviours rather
discrepancy between actual and hypothetical than the actual final choices (e.g. Choi, Lehto,
DCs; consideration of destination Morrison, and Jang 2012). Two approaches
characteristics), this study investigates DC dominate in tourism research to analyse DCs: 1)
sections that vary in their closeness to reality. focus on actually executed choices for example
The focus is not only on the outcome of DC by the observation of DCs (e.g. Botha et al.
113
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

1999) or output-oriented models to predict DCs This means that the type of destination changes,
(e.g. Papatheodorou 2001; Seddighi and resulting in a discrepancy between initial travel
Theocharous 2002); 2) focus on hypothetical dreams and actual travel behaviour (Karl et al.
choices with methodologies such as choice 2015). The discrepancy could be attributed to
experiments (e.g. Ferns and Walls 2012; changes in the decision criteria that determine
Sarman, Scagnolari, and Maggi 2016; the different stages of the DC process. While
Sharifpour, Walters, and Ritchie 2014). facilitators dominate at the beginning of the
However, a discrepancy often exists between process, inhibitors or travel constraints play a
actual and hypothetical DC caused by more important role towards the end (Um and
constraints that prevent tourists from Crompton 1992). Although past research agrees
implementing their hypothetical dreams (Decrop that the DC process changes and that different
2010; Karl et al. 2015). types of destinations are more or less relevant
depending on the stage of the DC process, the
This study is therefore embedded in set theory differences in the comparison and evaluation of
as encompassing concept since set theory alternative destinations between actual and
embraces the idea of different kinds of choices hypothetical DCs are not yet clear.
with varying closeness to reality. Another
advantage of set theory is that it allows to Another challenge in the investigation of DC,
investigate all destinations that are considered which has rarely been faced in past studies, in
during the DC process instead of only focusing particular in quantitative surveys, is the
on the final decision. This helps to identify integration of the actual destinations. However,
reasons for the rejection of destinations before it is the combination of tourist and destination
the final choice and provides important input, attributes that forms the DC (e.g. Bekk, Spörle,
both for the theoretical understanding of DC and and Kruse 2016; Karl et al. 2015): A destination
the practical knowledge for tourism will only be chosen as a final destination in case
management and marketing. that the destination’s characteristics and the
tourist’s preferences match. Bekk, Spörle and
Set theory is an approach which has first been Kruse (2016) introduce tourist-destination
introduced in consumer behaviour research to personality similarity, a concept taken from
investigate choice opportunities and has been person-environment fit theory, which is based on
adapted to tourism research to explain DC tourists’ individual perceptions and destination
processes in quantitative as well as qualitative attributes. Their results suggest that the level of
studies (e.g. Crompton 1992; Decrop 2010; similarity between these perceptions influences
Jang, Lee, Lee, and Hong 2007; Prentice 2006; how satisfied tourists are with the destination
Um and Crompton 1990). According to set and how likely they would recommend the
theory, DC can be described as a multistage, destination to others. Another example are
sequential, and funnel-like process with multiple facilitators or inhibitors that lead to the selection
alternatives (Crompton 1992; Sirakaya and or rejection of a destination. These factors are
Woodside 2005; Um and Crompton 1990). The linked to the destination as well as to the tourist
complexity of the DC is simplified by distributing (e.g. risk-avoiding personality – preference of
all destinations among hierarchically structured safe destination) and can only be fully
sets (Crompton 1992). The final decision is then understood through the integration of
made between a few destinations from a smaller destination details in the research methodology
subset (Crompton 1992). More detailed (Karl 2016). There are few studies that consider
information on set theory can be found in a destination attributes to a limited extent: Work
former study by Decrop (2010). Past set theory on the investigation of inhibitors or travel
studies underline the complexity of a DC constraints that influence the narrowing-down
process which changes from the first initial idea process of alternative destinations tend to
to the final choice of one destination (e.g. concentrate on the choice of a particular
Decrop 2010; Um and Crompton 1992). A destination but not on alternative destinations
longitudinal study from Decrop (2010) implies (e.g. Botha et al. 1999). Input-output DC models
that travel decision-making and DC processes that include attributes of the entirety of final
are becoming more realistic towards the end. destinations (e.g. Marcussen 2011), neglect
114
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

alternative destinations that did not pass the different destinations is mentioned in
whole DC process. Set theory provides a quantitative surveys. This impedes statistical
possibility to analyse destinations and their analyses (e.g. combinations of destinations;
relevance at different stages of the DC process. links between destinations) and demands for
However, most past studies on DC using set alternative approaches. In order to avoid the
theory concentrate on the choice sets itself. In concealment of destination information through
particular, the size and type of choice sets are clustering or typing methods, this study applies
investigated rather than the actual destinations network analysis as a tool to investigate DC.
and their characteristics (e.g. Crompton and This allows the analysis of hypothetical future,
Ankomah 1993; Woodside and Lysonski 1989; actually planned and executed past choices
Um and Crompton 1990). Consequently, whilst preserving information on the
specific features of destinations, which are in destinations. This approach helps to better
some cases crucial determinants of DC, are not understand how alternative destinations are
taken into account and tourists’ DC set evaluated during the DC. Network analysis does
structures are only partly explained. not only permit the inclusion of destination
information but also information on linkages
This paper aims to address two research gaps between competing destinations to identify
that have been identified in the literature review: clusters of competing or compatible destinations
1) recognition of a discrepancy between actual in different sections of the DC.
and hypothetical DC; 2) consideration of the
destination in DC research. Therefore, DC is not “Network analysis, derived from graph
investigated in its entirety as a process, instead theory, attempts to describe the structure of
the focus is on certain sections of the DC that relations (displayed by links) between given
are distinguished by their closeness to reality entities (displayed by nodes), and applies
and represent actual and hypothetical DCs. The quantitative techniques to produce relevant
study moreover includes specific information on indicators and results for studying the
the alternative destinations to better explain characteristics of a whole network and the
differences of actual and hypothetical DCs. position of individuals [actors] in the
network structure” (Shih 2006, 1031).
This study proceeds on the assumption that
differences exist between hypothetical and Network science was introduced in social
realistic DCs which are reflected in the science to investigate relationships between
evaluation structure and relevance of alternative stakeholders of various kinds. In tourism
destinations in each DC section. Destinations research, network analysis is applied for
that are perceived as more realistic should be example to investigate stakeholder relationships
more relevant in the realistic actual choices in destinations (e.g. Baggio, Scott, and Cooper
while destinations associated with stronger 2010; Pforr 2006; Scott, Cooper, and Baggio
travel constraints should be less relevant at the 2008), interrelations between destinations (e.g.
same time. Based on this assumption, two Baggio 2007; Shih 2006) or tourist attractions
research questions lead this study: within a destination (e.g. Stienmetz and
Fesenmaier 2015). Further information on
 Research question 1: What are the network science with a specific focus on the
differences in the comparison of alternative application in tourism research can be found in
destinations at different DC sections with Baggio et al. (2010). In contrast to former
varying closeness to reality? research, our study is not based on relationships
 Research question 2: Which relevance or of stakeholders in a particular destination or
role do destinations have in different DC competing destinations and attraction points but
sections with varying closeness to reality? on specific sections of DC and relations between
alternative destinations in these sections.
Methods
Network Analysis Data Collection
DC sets are comprised of a multitude of This study is part of a survey on the structure
destinations. Consequently, a high amount of and determinants of tourists’ DC processes.
115
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

Data were collected in Munich, Germany 2013 activated”. In our study the future consideration
using a standardised questionnaire in personal set has been operationalised as follows: “Please
interviews (n = 835). Trained interviewers name up to six other destinations which you
approached the respondents in 18 public spaces have not visited yet but would like to visit in the
where people tend to spend time instead of just future.”
passing. To avoid bias due to personal
preferences, every second or third passers-by The relevant set and past DC depict a more
was approached depending on the passenger realistic picture of the DC towards the end of the
frequency. To ensure that only potential DC process without destinations which cannot
travellers were integrated in the survey, be visited (easily) due to constraints. The
participants were asked a screening question at relevant set has a crucial position in the DC
the beginning of the survey (i.e. “Are you process. This set comprises all alternative
planning to conduct a main holiday (at least four destinations which tourists are considering for
overnight stays) within the next twelve the next planned holiday and the final
months?”). Only potential tourists at the age of destination is taken from this set (Crompton
14 years or older were accepted since, in most 1992). In this study, the relevant set represents
cases, younger tourists influence but are not the realistic section of DC that is happening at
actively involved in DC (Decrop 2006). present. The relevant set is based on
Ankomah’s et al. (1996) as well as Crompton’s
The questionnaire was completed in personal (1992) late consideration set with a limited time
interviews in an average time of 15 minutes. between DC and commencement of a journey
Only the sections that focus on DC and (in our study: twelve months). “Which
sociodemographic variables are included in this destinations are you considering for your next
article. In the questionnaire set theory is applied main holiday (i.e. at least four overnight stays)?”
to analyse DC behaviour at more hypothetical Past DCs represent actual choices in contrast to
and realistic stages of the DC. Theoretical planned choices (relevant set) or hypothetical
models of DC using set theory (e.g. Crompton choices (future consideration set): “Where have
1992), describe rather complex, differentiated you spent your main holidays in the last three
and detailed DC set structures. However, in years?“. It signifies a realistic section of DC in
quantitative surveys it is difficult to cover such the past with actually executed holiday trips
complex structures. Therefore, the focus of this where respondents were able to overcome
study is on three sections of DC representing travel constraints.
hypothetical future, realistic planned and
executed behaviour: the future consideration Respondents were not restricted to a specific
set, the relevant set and past DC. geographical scale concerning the destinations
in the questionnaire. An explanation for this
Destinations in the future consideration set are approach can be taken from Fyall’s (2013)
alternative destinations for a future holiday statement on destinations:
which the respondent has not visited yet but “Destinations come in all shapes and
wishes to visit in the future. The future sizes and can be found in a variety of
consideration set illustrates the range of geographical settings such as in urban,
destinations that a respondent would be rural and coastal environments.
interested in and is drawn to visit in the future. It Destinations can be countries or a
represents the hypothetical section of DC. The collection of countries, a distinct state,
restriction to destinations which have not been country or province, or in fact represent a
visited before excludes own experience or local city, town or resort, national park,
images based on own experiences, both strong area of outstanding national beauty or
influencing factors of DC and image formation coastline” (Fyall 2013, 118).
(Litvin, Goldsmith, and Pan 2008). It is similar to
Crompton’s (1992, 422) initial set which consists After elimination of invalid questionnaires, a final
of “all the locations that might be considered as sample size of 835 questionnaires is used for the
potential destinations for a vacation before any network analysis. Table 1 displays the socio-
decision process about a trip has been demographic profile of the respondents. The age
116
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

Table 1. Profile of respondents (n = 835).

n Percent
Gender
female 418 51.2
male 399 48.8
Age
14 – 19 years 64 7.7
20 – 29 years 242 29.0
30 – 39 years 123 14.7
40 – 49 years 100 12.0
50 – 59 years 132 15.8
60 – 69 years 92 11.0
> 69 years 72 8.6
Highest level of education achieved
Apprenticeship 28 3.4
Junior high school 57 6.9
Secondary school 123 14.9
High school 267 32.4
University or college 320 38.8
Other 30 3.6
Occupation
Retired 109 15.5
House wife/husband 18 2.6
Student 223 31.8
Workman 24 3.4
Employee 243 34.6
Civil servant 60 8.5
Unemployed 7 1.0
Other 18 2.6
Household income per month
< 750 € 131 15.7
750 – 1,499 € 79 9.5
1,500 – 1,999 € 71 8.5
2,000 – 2,499 € 53 6.3
2,500 – 2,999 € 73 8.7
3,000 – 3,499 € 61 7.3
3,500 – 3,999 € 42 5.0
4,000 – 4,499 € 41 4.9
4,500 – 4,999 € 45 5.4
5,000 – 7,499 € 57 6.8
> 7,499 € 49 5.9
n/s 133 15.9
Household size
1 227 27.6
2 306 37.2
3 131 15.9
4 108 13.1
>4 51 6.2

group 20 to 29 years is particularly dominant interpretation of the results. Moreover, this study
since the survey took place in a city with a high is an explorative study that aims at developing
amount of students. However, specific and testing a new methodology to investigate
characteristics of the location of data collection DC. The goal was not to depict a representative
(e.g. high proportion of young professionals) are picture of German tourists’ DC. Nevertheless, a
considered throughout the data analysis and comparison with a representative survey of
117
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

German tourists travel behaviour (FUR 2016) together by respondents in the same DC
shows a high consensus regarding the realistic section. The network analysis includes weighted
DC sections. as well as dichotomised data (i.e. distinction only
between ties being absent or present;
Data Analysis Hanneman and Riddle 2005) to investigate both
The wide range of destinations named in the the strength and variety of links between
survey created a dataset with information that destinations in order to reveal dominance as
was too differentiated for the quantitative well as diversity of destinations in the DC
network analysis. An aggregation of all sections.
destinations on the national level was necessary
to detect patterns and structures of alternative Social network analysis provides a multitude of
evaluations. Information on the original tools, metrics and algorithms. Since this study
geographical scale was used to better interpret does not examine social relationships but DCs
the results of the network analyses. and competing alternative destinations, only
certain metrics are applicable here. Several
Different software tools are used to analyse metrics were considered that allow examining
network data: R version 3.0.2 (R Core Team the way that nodes are connected, the distances
2013) for the transformation of original survey between nodes and the kinds of structures that
into network data; UCInet version 6.507 characterise the networks. However, some of
(Borgatti, Everett, and Freeman 2002) for the these metrics were not applicable with the
analysis of the structures and properties in the available data set or did not provide further
networks; NetDraw version 2.138 (Borgatti insight into DCs than the chosen metrics. In the
2002) to visualise network data; and SPSS end, network density, degree centrality, network
version 21.0 (IBM 2012) for statistical analyses centralisation and cutpoints were chosen to
of the original survey data. investigate DC network structures in this study.
The chosen metrics are unambiguous
The transformation of the original survey data on measurements and indicators for different
destinations into network data results in an aspects that are relevant for the analysis of DCs.
undirected and symmetric network. All Table 2 provides an overview of the chosen
destinations named in one DC section form a network metrics, measurement levels and
specific network. Nodes which are seen as the indicative meanings in this study.
“elements of a system” (Fortunato, Latora, and
Marchiori 2004, 1) represent destinations Network Density Δ
mentioned in a particular section. Links in the Network density “describes the general level of
networks show that two destinations are named linkage” among the nodes in a network (Scott
by the same respondent in the same section. 2013, 69). In this study, network density
The weight of a link corresponds to the number indicates the degree of connectedness between
of times where two destinations are named destinations. It is measured using both

Table 2. Dimensions of network analysis.

Network metrics Measurement of Indicator of/for
Network Density Δ structure of evaluation degree of connectedness between destinations
dichotomised cross-linking between destinations
weighted strength of connections between
destinations
Degree Centrality CD role of destination degree of compatibility and uniqueness of a
destination
Network Centralisation CA structure of evaluation degree of hierarchical order of destinations
Cutpoint structure of evaluation degree of stability and connectivity between
role of destination destinations
degree of bonding force of a destination

118
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

dichotomised and weighted data. In a binary direct adjacent neighbours to calculate the
network, dichotomised network density (i.e. ratio centrality of a node, degree centrality focuses
of links present to all possible links; see formula only on directly adjacent neighbours
1, Wasserman and Faust 2009) implies whether (Wasserman and Faust 2009). However, non-
a network consists of a high number of direct links are difficult to interpret in a context
marginally connected (i.e. low dichotomised that is not based on human or corporate
network density) or strongly interwoven (i.e. high networks. Therefore, degree centrality (i.e.
dichotomised network density) destinations. In a amount of direct links to adjacent neighbouring
valued network, weighted network density (i.e. actors; Wasserman and Faust 2009) is applied
sum of links divided by all possible links; see in this study to identify the role of individual
formula 2, Wasserman and Faust 2009) destinations within the network. Since degree
incorporates the strength of a link or how often centrality indirectly depends on the size of a
destinations are mentioned together in a DC network, it is calculated in our study in a
section. The inclusion of the weighted network normalised way which is independent of the size
density factors in destinations that are strongly of the network (see formula 4, Freeman 1978-
connected and build a cluster of highly relevant 1979; Wasserman and Faust 2009). Degree
central destinations with specific roles in a DC centralities are calculated with dichotomised
section. In particular, the comparison of data as, in this study, the focus is on the variety
dichotomised and weighted network densities of destinations that are connected by a particular
allows to draw conclusions about the diversity of destination to understand how compatible a
destinations at a certain DC section. For destination is with other destinations. Degree
example, a high weighted network density, centrality is, moreover, able to demonstrate
compared to the dichotomised network density, whether a destination is the main destination for
is evidence for strong ties between few tourists (i.e. low degree centrality) or mostly part
destinations. of a group of competing destinations (i.e. high
degree centrality).
Binary network with dichotomised data

𝐿 2𝐿 𝐶𝐷 (𝑛𝑖 ) = d(𝑛𝑖 ) = ∑ 𝑥𝑖𝑗 = ∑ 𝑥𝑗𝑖 (3)

∆𝐷 = =
⁄
𝑔(𝑔 − 1) 2 𝑔(𝑔 − 1) (1) 𝑗 𝑗

′
d(𝑛𝑖 )
Valued network with weighted data 𝐶 𝐷 (𝑛𝑖 ) = (4)
𝑔−1

2 ∑ 𝑣𝑘 𝑑(𝑛𝑖 ) 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑑𝑖𝑟𝑒𝑐𝑡𝑙𝑦 𝑎𝑑𝑗𝑎𝑐𝑒𝑛𝑡 𝑛𝑜𝑑𝑒𝑠

∆𝑊 =
𝑔(𝑔 − 1) (2) xij value of the link between i and j
g number of nodes in the network
𝑣𝑘 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑙𝑖𝑛𝑒 𝑙𝑘 0 ≤ 𝐶′𝐷 ≤ 1
𝐿 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑙𝑖𝑛𝑘𝑠
𝑙𝑘 𝑙𝑖𝑛𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑛𝑜𝑑𝑒 𝑖 𝑎𝑛𝑑 𝑗 Network Centralisation CA
𝑔 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑛𝑜𝑑𝑒𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑛𝑒𝑡𝑤𝑜𝑟𝑘 Network centralisation describes the structure of
a network based on differences in the level of
Degree Centrality CD degree centralities (see formula 5, Wasserman
Algorithms to measure centrality in a network and Faust 2009). It is able to show to which
generally aim to identify the most important, extent “cohesion is organised around particular
central or prominent actors within a network and focal points“ (Scott 2013, 90). While the analysis
are “a measure for an actor’s level of of degree centralities helps to reveal the role and
involvement or activity in the network“ (Prell relevance of certain destinations during the DC,
2012, 97). Various centrality measures are network centralisation allows to understand the
generally applied in network analysis such as balance or imbalance in a network. In this study,
closeness centrality, betweenness centrality or network centralisation is used to identify the
degree centrality. While closeness and hierarchical order of destinations within one DC
betweenness centrality also incorporate non- section. A high level of network centralisation

119
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

implies that few destinations with high degree alternative for tourists whose preferences do not
centralities dominate the DC section while a low overlap otherwise.
level suggests a balanced structure without
strongly dominating destinations. Results
The analysis of the three DC sections (i.e. future
∑𝑔𝑖=1[𝐶𝐴 (𝑛∗ ) − 𝐶𝐴 (𝑛𝑖 )] consideration set, relevant set, past DC) focuses
𝐶𝐴 = 𝑔 (5) firstly on the general structure of each DC
max ∑𝑖=1[𝐶𝐴 (𝑛∗ ) − 𝐶𝐴 (𝑛𝑖 )]
network and secondly on destinations with
particular roles in the network. The first part
𝐶𝐴 (𝑛∗ ) 𝑚𝑎𝑥𝑖𝑚𝑎𝑙 𝑑𝑒𝑔𝑟𝑒𝑒 𝑐𝑒𝑛𝑡𝑟𝑎𝑙𝑖𝑡𝑦 𝑜𝑓 𝑎 𝑛𝑜𝑑𝑒
concentrates on the. entirety of destinations and
𝑔 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑛𝑜𝑑𝑒𝑠 𝑖𝑛 𝑡ℎ𝑒 𝑛𝑒𝑡𝑤𝑜𝑟𝑘
links between destinations in one DC section to
0 ≤ 𝐶𝐴 ≤ 1
investigate the structure of alternative
destination evaluations, relevance of
Cutpoints
destinations and differences between the
A measurement for the analysis of the internal
investigated sections of DC. The second part
structure and connectivity of networks are
focuses on particular destinations and their
cutpoints that bind the network. The removal of
embedding in a network of competing
a cutpoint leads to a drop out of further nodes
destinations to examine and illustrate the DC
which are only attached to the network by this
network structures of certain types of
particular node. The network would
destinations. Table 3 displays the results of the
consequently be divided into several sub-
DC network analysis.
networks (Scott 2013). Hannemann and Riddle
The DC networks, shown in figure 1 to 3,
(2005) state that “cutpoints may be particularly
illustrate the results and discussion sections of
important actors – who may act as brokers
this article. All network figures are created in
among otherwise disconnected groups”. In our
NetDraw using spring embedding technique, a
study, cutpoints are used to reveal destinations
“simple heuristic for laying out arbitrary
which bind other destinations or groups of
networks” (Scott et al. 2008, 174). Nodes are
destinations together and connect them to the
positioned in a way that minimises distances
network. The assumption is that tourists who
between linked nodes. Hence, the most
prefer or travel to different kinds of destinations
frequently linked destinations are situated at the
can agree on this cutpoint destination. This
smallest distance. The size of the nodes
means that a cutpoint destination is an
represents a destination’s level of degree

Table 3. Results of the DC network analysis.

Future consideration set Relevant set Past DC
N 2,097 1,362 2,273
No. of nodes 139 97 109
No. of links 2,584 743 1,892
∆𝑤 / ∆𝐷 0.27 / 0.12 0.16 / 0.08 0.32 / 0.11
𝐶 ′𝐷 USA 0.62 Italy 0.46 Italy 0.57
Australia 0.50 USA 0.45 Spain 0.46
Thailand 0.46 Spain 0.37 USA 0.44
Canada 0.46 France 0.28 Germany 0.41
New Zealand 0.44 Austria 0.25 Austria 0.41
𝐶𝐴 0.50 0.39 0.47
Cutpoints Argentina, Asia, Caribbean, Austria, Baltic States, Egypt, Germany, Turkey,
Mongolia, Lithuania, Bulgaria Egypt, Greece, USA
Russia, Sweden Italy, Madagascar, New
Zealand, South Africa,
South America, Southeast
Asia, Spain, Portugal,
United Arab Emirates,
United Kingdom, USA

120
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

centrality while the strength of links between research question. In all DC sections, weighted
nodes refers to the number of times that and dichotomised network densities are
destinations are named simultaneously. between 0.16 and 0.32 or 0.08 and 0.12,
respectively. These low values can partly be
Network densities and network centralisation ascribed to the size of the networks since rather
provide information on the structure of DC large networks are generally associated with
sections according to the linkage of alternative lower network densities (Scott 2013). Both in
destinations and are chosen to answer the first regard to weighted and dichotomised network

Figure 1. Network of the future consideration set with degree centralities and cutpoints.

Figure 2. Network of the relevant set with degree centralities and cutpoints.

121
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

Figure 3. Network of the past destination choice with degree centralities and cutpoints.

densities, the relevant network has the lowest inhabitants; UNWTO 2013). Destinations with
values while the future consideration and past the highest degree centralities in the future
DC networks have slightly higher values (see consideration network are long-haul non-
table 3). The level of network centralisation lies European destinations, which do not necessarily
between 0.39 and 0.50 in the three networks. occupy important positions in the other sections
The future consideration and past DC networks and do not reflect actual travel patterns of
show the highest network centralisation and German tourists (UNWTO 2013). This is in
consequently the highest variability of degree contrast to the relevant or past DC network
centralities. Here, few destinations with high where destinations with high degree centralities
degree centralities dominate the centre of the are predominantly European destinations at a
network structure and a large amount of short or medium distance to Germany with high
destinations with low degree centralities tourism intensities and strong tourist flows from
constitute the peripheral areas of the network. In Germany (UNWTO 2013). Two kinds of
the relevant set, a weaker cluster of dominating cutpoints have been detected in the DC
destinations and less peripheral destinations networks: cutpoints with high degree centralities
result in a comparatively low level of network that connect destinations with high and low
centralisation. degree centralities which vary regarding touristic
indicators, geographic locations or development
The metrics degree centrality and cutpoint are statuses (UNDP 2014); cutpoints with low
used to investigate destinations and their degree centralities that connect destinations
relevance or roles within DC sections (see with low degree centralities which are rather
research question 2). It has to be noted that the uncommon for the German outbound market
number of responses influences degree (UNWTO 2013).
centralities: Frequently mentioned destinations
have more chances to be linked to other Discussion
destinations, resulting in a higher degree The results demonstrate that the DC sections
centrality than destinations which are mentioned differ concerning the structures (i.e. network
only by a few respondents. In all networks, density, network centralisation) and relevance of
destinations with the lowest degree centralities alternative destinations (i.e. degree centrality,
are destinations with rather low tourism cutpoint) in the DC sections (see research
intensities (i.e. tourist arrivals per 1,000 question 1 and 2).

122
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

The future consideration network is homogeneous since it mainly consists of rather

characterised by a relatively high weighted similar destinations.
network density despite a large amount of
peripheral destinations and a high level of The high level of network centralisation in the
network centralisation. This indicates that a past DC network is a proxy for a network with a
cluster of highly interwoven destinations exists strong centre of highly interwoven dominating
despite a broad variety of destinations. Hence, a destinations. This suggests that German tourists
limited number of destinations is mentioned choose similarly when it comes to the final DC.
simultaneously by many respondents. The Studies on first-time and repeat visitors, such as
future consideration network is, moreover, a Chi’s (2012) investigation of destination loyalty,
network with few cutpoints which is a proxy for a confirm the stable destination preference.
stable network. These results imply that However, the higher network densities
hypothetical future travel plans of German (dichotomised and weighted) and lower number
tourists are rather multidimensional but of cutpoints, compared to the relevant set, are a
nevertheless concentrated on some specific sign of a strong integration of peripheral
destinations. The multidimensionality can be destinations into the network. The strong linkage
ascribed to the absence of travel constraints at of peripheral destinations to the cluster of central
this hypothetical stage of DC. This is in line with destinations indicates that the same
Um and Crompton (1992) who state that respondents travelled to uncommon as well as
inhibitors of travelling such as travel constraints conventional destinations. In the last decades,
strongly influence DC towards the end of the tourists have evolved into hybrid or multi-faceted
process when it comes to actual DC. The tourists who do not travel to the same kind of
beginning of the process, which can be destination or even the same destination in each
compared to the hypothetical DC in this study, holiday (Dellaert et al. 1998; Lohmann and
however, is dominated by positive attributes and Aderhold 2009). A study by Boztug, Babakhani,
facilitators of travelling. Consequently, tourists Laesser, and Dolnicar (2015, 190) shows that
include destinations in the future consideration nowadays hybrid tourists “are the norm, rather
set without taking factors into account that would than the exception” in regard to motivation and
deter them from travelling to destinations which travel expenses. It is therefore likely that tourists
substantially extends the range of different also act in a hybrid manner in terms of DC.
destinations.
A comparison of the results for each DC section
In contrast to the hypothetical future considera- demonstrates that the structure and relevance of
tion set, the relevant network is shaped by low alternative destinations varies which reflects the
network densities (dichotomised and weighted), changes in the DC process. The structure of the
a high proportion of destinations with low degree future consideration network is characterised by
centralities and a large number of cutpoints. This a broad variety of potential destinations for
means that the network itself is rather fragile tourists’ future holidays. This variety is narrowed
with loosely connected destinations and would down in the relevant set and past DCs when
dissolve into several subnetworks in the tourists have to make actual decisions
absence of the cutpoint destinations. This concerning their holidays. This refers to findings
implies a unidimensionality or uniformity in of previous studies on DC which found that DC
German tourists’ actual travel plans for a follows a funnel-like process where travel
particular holiday: Only a limited amount of constraints successively reduce alternative
destinations is considered for one particular trip. destinations (e.g. Crompton and Ankomah
Moreover, the comparatively low but still 1993; Woodside and Lysonski 1989).
significant level of network centralisation Consequently, alternative destinations become
indicates that this network has a strong centre more homogeneous at the end of the DC
with few interconnected dominating process. Karl et al. (2015) for example show that
destinations. This result is confirmed by a former while most tourists consider a greater variety of
study of Karl et al. (2015) where the level of destinations at an early stage of the DC process,
homogeneity is calculated for several DC sets. their relevant sets for a specific trip consist of a
Their study shows that the relevant set is very very limited amount of destinations of a similar
123
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

type. A closer investigation of the destinations regions). The same applies for past DCs with
that form clusters in the centre of each DC destinations on the national or regional level.
section confirms that not only the evaluation
structure of alternative destinations changes but The network analysis shows that the DC
also the relevance of particular destinations (see sections differ in the variety of destinations and
research question 2) and the strongest that different destinations play a key role in
competitors: Central destinations in the actual and hypothetical DC sections. A
hypothetical DC section are mainly long-haul comparison of destinations’ degree centralities
non-European destinations while central in the three DC networks furthermore reveals
destinations in the realistic DC sections are that most destinations do not have a consistent
mainly popular destinations for German tourists. relevance or position in all DC sections. In fact,
Reasons for the changes are for example most destinations vary in their relevance and
temporal situational or permanent structural positions in the DC sections and loose or gain in
travel constraints (Decrop 2010) which lead to a importance from hypothetical to actual DCs.
domination of destinations in the relevant set Three types of destinations in regard to their
and past DC that can be realised more easily. position in the DC sections are identified: core
These destinations are mainly located within destination (i.e. high relevance in all DC
Europe and probably associated with less sections); surrogate destination (i.e. higher
transport costs which might reduce financial relevance in actual than hypothetical DC
constraints. In the hypothetical future sections); and intentional destination (i.e. lower
consideration set, however, respondents do not relevance in actual than hypothetical DC
actively consider alternatives. Therefore, sections). The following exemplary destinations
constraints do not yet apply and do not influence represent the three types of destinations. These
this section of the DC. This is in line with past exemplary destinations allow to illustrate and
research on DC such as Decrop (2010, 110- explain what kind of characteristics a destination
111) who states that the DC process becomes should incorporate in order to classify as core,
more realistic: “Vacationers may move from a surrogate or intentional destination.
preference/ideal value level (dreamed but not
necessarily available destinations) to an Core destinations are destinations with high
expectation level (realistically available degree centralities in all DC networks. The high
destinations) and finally to a tolerance level degree centralities signify a large amount of
(surrogate destinations that represent an competing destinations and a high combinability
acceptable minimum) as far as plans evolve”. with various other destinations. Therefore, core
destinations could be seen as destinations on
Changes between the DC sections moreover which travel companions with different
manifest in the geographical scale of the named destinations in their minds can easily agree.
destinations (i.e. regional, national, Consequently, core destinations can help to
subcontinental or continental). Subcontinents facilitate DCs since tourists normally decide
and continents are in particular mentioned in the together with their travel partners in mutual
future consideration set where the imagination decisions (Jang et al. 2007; van Raaij and
of an actual destination is (still) sometimes Francken 1984). USA can be defined as a core
rather vague. Tourists probably do not have in- destination since it has high degree centralities
depth knowledge about a destination (yet). A in all DC networks, being the first, second and
reason may be that they do not actively consider third most central destination in the future
alternative destinations, for example through consideration, relevant and past DC network
information search, until the end of the DC (C’D = 0.62/0.45/0.44). The role of this
process (Crompton 1992). In the relevant set, destination in the three DC sections indicates
where respondents are closer to the actual that USA as a core destination can be combined
choice for a holiday, imaginations of the with various kinds of destinations. The
destination are clearer and the destinations diversified touristic potential of the USA allows
named in the survey are narrowed down to tourists to realise various types of holidays from
smaller geographical areas (i.e. countries or city tourism in New York to sun-and-sea tourism
in Florida. With its different climate zones, it is
124
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

moreover a year-round destination. A holiday in The last type of destinations are intentional
the USA (and other core destinations) is an destinations with low degree centralities in the
alternative which is acceptable for a broad DC sections which represent actual travel
variety of tourists with various travel preferences behaviour (i.e. relevant set, past DC) where they
and DC structures. are located at the periphery of the networks and
relatively high degree centralities in the future
Surrogate destinations are destinations with consideration network. The discrepancy
comparatively low degree centralities in the between realistic and hypothetical travel
future consideration network but high degree behaviour is the distinctive feature of these
centralities in the realistic sections of the DC (i.e. destinations. Intentional destinations are rather
relevant set, past DC). According to Decrop uncommon destinations for the German
(2010), surrogate destinations replace ideal outbound market with weak tourist flows from
destinations when it comes to the final DC Germany. Russia’s role as a connector in the
(Decrop 2010). Many tourists plan to travel to future consideration network (i.e. cutpoint with
these destinations on their next holiday or have relatively high degree centrality, C’D = 0.33)
already realised a holiday there but do not together with the peripheral position in the actual
dream about visiting them in the future. The DC networks (i.e. relevant set, C’D = 0.05; past
assumption is that tourists who travel to different DC, C’D = 0.18) classify Russia as an intentional
destinations are able to agree on surrogate destination. An intentional destination, like
destinations as an alternative to their other travel Russia is most central at the hypothetical stage
preferences. The destination Italy represents an and becomes peripheral when it comes to actual
example for a surrogate destination as it has an choices because of a substitution by a surrogate
outstanding position (i.e. very high degree destination, probably due to travel barriers.
centralities, cutpoint) in the relevant and past DC Travel barriers concerning Russia are
networks (C’D = 0.46/0.57) but a weaker position presumably visa requirements which complicate
in the future consideration network (C’D = 0.39). travel arrangements for trips from Germany to
The high degree centralities in the actual DC Russia. Studies on visa regulations agree that
sections imply that Italy is a destination that is there is a significant influence of changes in visa
an adequate alternative for many tourists when regulations on tourist arrivals or economic
it comes to the actual DC. Due to the diverse effects of tourism (e.g. Beenstock, Felsenstein,
touristic potential of Italy, tourists with varying and Rubin 2015; Neumayer 2010). Another
travel motives are able to fulfil their needs at the travel constraint may be the language barrier
same destination from sun-and-sea to culture or which might hamper travelling in and to Russia:
nature oriented holidays. Moreover, with about Tapachai and Waryszak (2000), who analyse
ten million arrivals from Germany every year and destination image in DC, demonstrate that the
a market share of about 20 percent (UNWTO absence of language barriers between the host
2013), Germany is an important source market and source country is perceived as a beneficial
for Italy. The high number of German tourists factor in DC.
implies that many German tourists are familiar
with Italy due to prior visit and/or information Conclusion
from friends or relatives who have visited the Choosing a holiday destination is a complex
destination. Both, own experience and word-of- process where only the result is visible in the
mouth information are rated as strong form of tourist flows from one country to another.
persuasive factors in DC (Litvin et al. 2008). The Until a destination is finally chosen it has to
short distance to Germany, the free movement successfully complete several stages of the DC
of persons within the Schengen area and the process where it is compared to its alternatives.
usage of the same currency further facilitate The evaluation structures and the alternative
travelling between the two countries. All these destinations themselves are not the same in all
aspects point towards the fact that Italy can stages of the DC process, in particular
easily replace other destinations which were comparing realistic and hypothetical sections.
initially considered as alternatives but may be Our study contributes to the understanding of
not be realisable for various reasons. DC, both for tourism research and management.
It sheds light on the competing destinations and
125
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

their relation in different DC sections and relevant in either hypothetical or actual DC

identifies types of destinations with specific, sections.
varying relevance in these sections.
In contrast to many past studies on DC using set
Theoretical Contribution theory, our study investigates DC from the
This empirical study combines network analysis, perspective of the destination rather than of the
a methodological instrument for the investigation tourist. This approach helps to understand how
of “relationships among [...] entities, and [...] the alternative destinations operate in different DC
patterns and implications of these relationships” sections. The concretisation of set theory
(Wasserman and Faust 2009, 3), and set theory, through the integration of actual destinations in
a theoretical model with hierarchically ordered the theoretical construct of DC sets furthermore
subsets of alternative destinations, to investi- deepens the knowledge of DC behaviour. For
gate DCs. Metrics from social network analysis example, reasons for shifts in the DC sets or why
are transferred to examine the position, combi- destinations lose relevance in certain DC
nability or interchangeability of destinations at sections can be examined more thoroughly with
certain DC sections. This approach adds new information on the actual destination and its
insight into the understanding of DC and characteristics. This is particularly important
enhances the DC research methodology. An since the characteristics of destinations have a
advantage of the application of network analysis decisive impact on the decision in favour or
is that all destination information is retained against a destination (Ankomah et al. 1996). A
during the analysis. The analysis and combination of set theory and network analysis
interpretation of the results can consequently be offers new possibilities to study travel
conducted with specific information on the constraints from the tourist and destination
destinations, adding a geographic perspective to perspective for a better understanding of
DC literature. Moreover, set theory is able to changes in the DC process caused by internal
address another gap in past research - the and external factors.
discrepancy between actual and hypothetical
DCs. Hence, some limitations of quantitative Managerial Contribution
and qualitative research approaches can be The combination of network analysis and set
reduced through this study’s approach: theory provides a useful tool for a destination’s
Quantitative outcome-oriented research, such management and marketing for several
as Marcussen (2011), includes a range of reasons. This approach is applicable on various
destinations to predict the final destination. geographical scales, global but also regional or
However, the DC process and different sections local. It enables tourism stakeholders to identify
of DC are not captured. Qualitative process- competitors and possible cooperation partners,
oriented research, such as Decrop (2010), both, on a global level and at close proximity.
focuses on different steps of the DC process but Network analysis, in combination with set
is not able to embrace the variety of destinations theory, allows to determine in which DC section
that are part of the DC process. a destination is represented mostly and with
which alternative destinations it has to compete
Extending former research by Decrop (2010), in order to succeed in the DC process. DMOs
our study shows that the structure of DC are then able to identify competitors at different
networks differs considerably between the sections of the DC process and optimise their
selected DC sections from multidimensional marketing strategies accordingly. For example,
hypothetical future choices to unidimensional the core destination USA is mainly competing
actual and past choices. A significant finding to with Australia and Canada in the hypothetical
emerge from this study is that every DC section DC section and Italy, France and Spain in the
is characterised by destinations which are realistic DC sections. While competitors in the
located at the centre or periphery of the network hypothetical DC are rather similar to USA (e.g.
and consequently more or less relevant long distance, highly developed, politically
alternatives. Few dominating destinations play stable and safe, nature-oriented marketing),
an important role in all sections of the DC while competitors in the realistic DC are more diverse
the majority of destinations are more or less regarding for example the distance to the source
126
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

market, touristic offer or climate conditions. positively influence tourists’ perceptions of a

Explanations for these difference are the hybrid destination.
choice behaviour of tourists and the high
combinability of the destination USA with Limitations and Further Research
various kinds of destinations. This is a challenge Network analysis provides many instruments to
of DMOs which could be faced by more investigate different aspects and areas of
individualised marketing strategies. networks. Since this study does not examine
social networks but networks of destinations in
Another factor that is important in this context is tourists’ minds, only certain metrics are
the interchangeability of destinations. Is a applicable. The focus on only a few instruments
destination a destination with a high amount of in this study can be explained by the aims of this
competitors or a unique destination with few or study: a better understanding of DCs and
no direct competitors and therefore hardly specific roles of destinations in different DC
interchangeable with alternatives? Our study sections with varying closeness to reality; a
emphasises that some sun-and-sea straightforward interpretation of the instruments
destinations in the Mediterranean can easily be to facilitate the application in other geographic
exchanged (i.e. strong links between the regions for research and practice purposes.
destinations) probably because the main However, some network analysis tools could
attractive features (e.g. beach, sea, climate) are offer further and deeper comprehension of DCs,
available in all alternative destinations. For in particular in regard to the structure of DC
example, Italy, a sun-and-sea destination, has networks. A possibility to analyse network
destinations with a similar offer (i.e. Spain, structures is core-periphery analysis. Core-
France, Greece, Turkey and Croatia) as periphery analysis helps to detect structures of
strongest competitors in the relevant set. Other densely connected, cohesive cores and loosely
destinations with a more unique touristic connected peripheries of a network (Borgatti
potential are harder to substitute which is and Everett 1999). In regard to DC, core-
reflected in the network structure (i.e. periphery analysis would be able to reveal
destinations with a high number of mentions but strong competitors in the initial consideration set
low degree centralities such as China). This or relevant set and compatible destinations in
knowledge enables DMOs to target unique the past DCs. For the subsequent and further
characteristics that distinguish a destination investigation of DC structures, modularity
from its competitors to reduce analysis might be useful since it is able to detect
interchangeability. possible divisions of a network and to measure
the strength of division (Newman 2006). Both
This study furthermore demonstrates that some approaches are particularly interesting if
destinations with rather important positions in destinations are categorised according to
the future consideration network are not realised specific aspects such as dominating market
when it comes to actually choosing a destination segments, safety and security aspects or
for the next holiday. Information about distance to the source market. This strategy
competitors and travel constraints that lead to could help to identify how similar strongly or
the rejection of a destination is necessary as it weakly linked destinations are in regard to
supports a destination’s shift from more decisive influencing factors of DC. A deeper
hypothetical sections to actual DCs. More analysis of densely connected destinations in
specifically, DMOs can look at similarities the DC sets enables to detect patterns, for
between competing destinations in regard to example, in regard to the interchangeability.
supply, image or marketing strategies but also This would allow the extraction of clusters with
restraining or decisive factors that distinguish competing, interchangeable destinations from
successful destinations from its competitors in the network for in-depth analysis. Further
more or less realistic DC sections. An analysis research questions could focus on the
of situational and structural constraints can help identification of destination types and similarities
to develop marketing strategies which between destinations in strongly linked clusters
specifically address those constraints to using modularity analysis.

127
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

A limitation of this exploratory study is that the incorporate tourist and destination
results highly depend on the area in which the characteristics in the investigation of DC
investigation takes place. While the general choices. The integration of DC tourist types and
structure of the DC sections and, in particular, DC networks would better capture the
the research methodology is transferable to interdependency between tourist and destina-
other research areas, the results, in particular tion in order to understand the complexity of DC.
the destination-related results, are strongly
related to the geographic location of the study. References
The results should be interpreted in the light of Ankomah, P.K., J.L. Crompton, D. Baker.
the national (e.g. importance of international (1996). Influence of cognitive distance in
destinations due to the strong German outbound vacation choice. Annals of Tourism
travel market; Lohmann and Aderhold 2009) and Research, 23 (1), 138–150.
regional context (e.g. importance of the doi:10.1016/0160-7383(95)00054-2
destination Italy due to the close proximity Baggio, R. (2007). The web graph of a tourism
between Southern Germany and Italy in the system. Physica A: Statistical Mechanics
realistic DC sections). It could be useful to and its Applications, 379 (2), 727–734.
conduct cross-cultural studies in other countries doi:10.1016/j.physa.2007.01.008
to investigate similarities and differences in DC Baggio, R., N. Scott, C. Cooper. (2010). Network
network structures. science. Annals of Tourism Research, 37
(3), 802–827. doi:10.1016/j.annals.2010.
Several research questions concerning DC 02.008
networks, especially in regard to the role of the Beenstock, M., D. Felsenstein, Z. Rubin. (2015).
tourist, remain unanswered at present. Visa waivers, multilateral resistance and
Ankomah et al. (1996) state that factors related international tourism: some evidence from
to the tourist as well as factors related to the Israel. Letters in Spatial and Resource
destination shape the DC process. A former Sciences, 8 (3), 357-371. doi:10.1007/
study by Karl et al. (2015) highlights this s12076-015-0137-3
interrelation showing that the interaction of Bekk, M., M. Spörrle, J. Kruse. (2016). The
personal and destination characteristics results benefits of similarity between tourist and
in diverse DC patterns. Other aspects destination personality. Journal of Travel
concerning the tourist like travel experience, Research, 55(8), 1008–1021. doi:
travel preferences or travel motives should be 10.1177/00472875 15606813
investigated in relation to destination character- Borgatti, S.P. (2002). Netdraw Network
ristics to gain a deeper understanding of DC. An Visualization. Harvard, MA: Analytic
investigation of DC using the methodology of our Technologies.
study can therefore enhance DC tourist Borgatti, S.P., M.G. Everett (1999). Models of
typologies like Plog’s (1974) or Decrop and core/periphery structures. Social Networks,
Zidda’s (2006) that focus primarily on tourist 21 (4), 375–395. doi: 10.1016/S0378-
characteristics as explanatory variables. The 8733(99)00019-2
calculation of clustering coefficients (i.e. extent Borgatti, S.P., M.G. Everett, L.C. Freeman.
of clustering; Hanneman and Riddle 2005) for (2002). Ucinet 6 for Windows: Software for
the DC networks could be a first step to reveal Social Network Analysis. Harvard, MA:
whether tourists primarily choose between Analytic Technologies.
several destinations while other tourists select a Botha, C., J.L. Crompton, S.-S. Kim. (1999).
destination from a different group of Developing a revised competitive position
destinations. A further possibility would be the for Sun/Lost City, South Africa. Journal of
identification of clusters of closely connected Travel Research, 37 (4), 341–352. doi:
destinations, for example by means of 10.1177/004728759903700404
modularity analysis. The final step would be an Boztug, Y., N. Babakhani, C. Laesser, S.
investigation of differences between tourists Dolnicar. (2015). The hybrid tourist. Annals
(e.g. travel behaviour, sociodemography) who of Tourism Research, 54, 190–203. doi:
choose from each cluster of densely connected 10.1016/j.annals.2015.07.006
destinations. This procedure would allow to
128
 Karl, M., C. Reintinger (2017) / European Journal of Tourism Research 15 pp. 112-130

Chi, C.G. (2012). An examination of destination FUR (Forschungsgemeinschaft Urlaub und

loyalty: Differences between first-time and Reisen e.V.) (2016). Erste Ergebnisse der
repeat visitors. Journal of Hospitality & Reiseanalyse. [First results of the German
Tourism Research, 36 (1), 3–24. doi: Travel Analysis]. Hamburg: FUR.
10.1177/1096348010382235 Fyall, A. (2013). Destinations. In Fletcher, J., A.
Choi, S., X.Y. Lehto, A.M. Morrison, S. Jang. Fyall, D. Gilbert, S. Wanhill (ed.) (2013)
(2012). Structure of travel planning Tourism. Principles and Practice. Harlow:
Processes and information use patterns. Pearson Education Limited, 117–143.
Journal of Travel Research, 51 (1), 26–40. Hanneman, R., M. Riddle. (2005). Introduction
doi:10.1177/0047287510394191 to social network methods. Riverside, CA:
Crompton, J.L. (1992). Structure of vacation University of California. URL:
destination choice sets. Annals of Tourism faculty.ucr.edu/~hanneman/nettext/Introdu
Research, 19 (3), 420–434. ction_to_Social_Network_Methods.pdf
doi:10.1016/0160-7383(92)90128-C (Accessed on 10.06.2016).
Crompton, J.L., P.K. Ankomah. (1993). Choice IBM (2012). IBM SPSS Statistics for Windows,
set propositions in destination decisions. Version 21.0. Armonk, NY: IBM Corp.
Annals of Tourism Research, 20 (3), 461– Jang, H., S. Lee, S.-W. Lee, S.-K. Hong. (2007).
476. doi:10.1016/0160-7383(93)90003-L Expanding the individual choice-sets model
Decrop, A. (2006). Vacation decision making. to couples’ honeymoon destination
Wallingford, UK, Cambridge,MA:CABI Pub. selection process. Tourism Management,
Decrop, A. (2010). Destination choice sets: An 28 (5), 1299–1314.
inductive longitudinal approach. Annals of doi:10.1016/j.tourman.2006.11.008
Tourism Research, 37 (1), 93–115. Karl, M. (2016). Risk and uncertainty in travel
doi:10.1016/j.annals.2009.08.002 decision-making – tourist and destination
Decrop, A., P. Zidda. (2006). Typology of perspective. unpublished results.
vacation decision-making modes. Tourism Karl, M., C. Reintinger, J. Schmude. (2015).
Analysis, 11 (3), 189–197. Reject or select: Mapping destination
doi:10.1016/j.tourman.2003.11.011 choice. Annals of Tourism Research, 54,
Dellaert, B.G.C., D.F. Ettema, C. Lindh. (1998). 48–64. doi:10.1016/j.annals.2015.06.003
Multi-faceted tourist travel decisions: a Litvin, S.W., R.E. Goldsmith, B. Pan. (2008).
constraint-based conceptual framework to Electronic word-of-mouth in hospitality and
describe tourists' sequential choices of tourism management. Tourism
travel components. Tourism Management, Management, 29 (3), 458–468.
19 (4), 313–320. doi:10.1016/S0261- doi:10.1016/j.tourman.2007.05.011
5177(98)00037-5 Lohmann, M., P. Aderhold. (2009).
Ferns, B. H., Walls, A. (2012). Enduring travel Urlaubsreisetrends 2020. Die RA-
involvement, destination brand equity, and Trendstudie – Entwicklung der
travelers’ visit intentions: A structural model touristischen Nachfrage der Deutschen
analysis. Journal of Destination Marketing [Leisure travel trends 2020. Travel analysis
& Management, 1 (1–2), 27–35. trend study – Development of the German
doi:10.1016/j.jdmm.2012.07.002 tourism demand]. Kiel: FUR.
Fesenmaier, D.R., J. Jeng. (2000). Assessing Marcussen, C.H. (2011). Understanding
structure in the pleasure trip planning destination choices of German travellers.
process. Tourism Analysis, 5 (1), 13–27. Tourism Analysis, 16 (6), 649–662. doi:
Fortunato, S., V. Latora, M. Marchiori. (2004). 10.3727/108354211X13228713394642
Method to find community structures based Neumayer, E. (2010). Visa restrictions and
on information centrality. Physical Review bilateral travel. The Professional
E, 70 (5). doi:10.1103/PhysRevE.70. 056104 Geographer, 62 (2), 171–181.
Freeman, L.C. (1978-1979). Centrality in social doi:10.1080/00330121003600835
networks: Conceptual clarification. Social Newman, M.E.J. (2006). Modularity and
Networks, 1 (3), 215–239. doi:10.1016/ community structure in networks.
0378-8733(78)90021-7 Proceedings of the National Academy of

129
 Investigating Tourists’ Destination Choices – An Application of Network Analysis.

Sciences of the United States of America, Vacation Marketing, 20 (2), 111–123.

103 (23), 8577–8582. doi:10.1177/1356766713502486
Oppewal, H., T. Huybers, G.I. Crouch. (2015). Shih, H.-Y. (2006). Network characteristics of
Tourist destination and experience choice: drive tourism destinations: An application of
A choice experimental analysis of decision network analysis in tourism. Tourism
sequence effects. Tourism Management, Management, 27 (5), 1029–1039.
48, 467–476. doi:10.1016/j.tourman.2005.08.002
doi:10.1016/j.tourman.2014.12.016 Sirakaya, E., A.G. Woodside. (2005). Building
Papatheodorou, A. (2001). Why people travel to and testing theories of decision making by
different places. Annals of Tourism travellers. Tourism Management, 26 (6),
Research, 28 (1), 164–179. 815–832.doi:10.1016/j.tourman.2004. 05.004
doi:10.1016/S0160-7383(00)00014-1 Stienmetz, J.L., D.R. Fesenmaier. (2015).
Pforr, C. (2006). Tourism policy in the making. Estimating value in Baltimore, Maryland:
Annals of Tourism Research, 33 (1), 87– An attractions network analysis. Tourism
108. doi:10.1016/j.annals.2005.04.004 Management, 50, 238–252.
Plog, S.C. (1974). Why destination areas rise doi:10.1016/j.tourman.2015.01.031
and fall in popularity. Cornell Hotel and Tapachai, N., R. Waryszak. (2000). An
Restaurant Administration Quarterly, 14 examination of the role of beneficial image
(4), 55–58. in tourist destination selection. Journal of
Prell, C. (2012). Social network analysis: Travel Research, 39 (1), 37–44.
History, theory & methodology. Los doi:10.1177/004728750003900105
Angeles, London: SAGE. Um, S., J.L. Crompton. (1990). Attitude
Prentice, R. (2006). Evocation and experiential determinants in tourism destination choice.
seduction: Updating choice-sets modelling. Annals of Tourism Research, 17 (3), 432–
Tourism Management, 27 (6), 1153–1170. 448. doi:10.1016/0160-7383(90)90008-F
doi:10.1016/j.tourman.2005.11.008 Um, S., J.L. Crompton. (1992). The Roles of
R Core Team (2013). R: A language and Perceived Inhibitors and Facilitators in
environment for statistical computing. Pleasure Travel Destination Decisions.
Vienna, Austria: R Foundation for Journal of Travel Research, 30 (3), 18–25.
Statistical Computing. http://www.R- doi:10.1177/004728759203000303
project.org/ (Accessed on 10.06.2016). UNDP (United Nations Development
Sarman, I., S. Scagnolari, R. Maggi (2016). Programme). (2014). Human Development
Acceptance of life-threatening hazards Report 2014: Sustaining Human Progress:
among young tourists: A stated choice Reducing Vulnerabilities and Building
experiment. Journal of Travel Research, 55 Resilience. http://hdr.undp.org/sites/
(8), 979–992. default/files/hdr14-report-en-1.pdf
doi:10.1177/0047287515612595 (Accessed on 24.06.2015).
Scott, J. (2013). Social network analysis. Los UNWTO (World Tourism Organization). (2013).
Angeles, London: SAGE. Yearbook of tourism statistics. Data 2007-
Scott, N., C. Cooper, R. Baggio. (2008). 2011. Madrid: UNWTO.
Destination networks. Annals of Tourism van Raaij, W., D.A. Francken. (1984). Vacation
Research, 35 (1), 169–188. decisions, activities, and satisfactions.
doi:10.1016/j.annals.2007.07.004 Annals of Tourism Research, 11 (1), 101–
Seddighi, H.R., A.L. Theocharous. (2002). A 112. doi:10.1016/0160-7383(84)90098-7
model of tourism destination choice: a Wasserman, S., K. Faust. (2009). Social
theoretical and empirical analysis. Tourism network analysis: Methods and
Management, 23 (5), 475–487. applications. Structural analysis in the
doi:10.1016/S0261-5177(02)00012-2 social sciences: Vol. 8. Cambridge, New
Sharifpour, M., G. Walters, B.W. Ritchie. (2014). York: Cambridge University Press.
Risk perception, prior knowledge, and Woodside, A.G., S. Lysonski. (1989). A general
willingness to travel: Investigating the model of traveler destination choice.
Australian tourist market's risk perceptions Journal of Travel Research, 27 (4), 8–14.
towards the Middle East. Journal of doi:10.1177/004728758902700402
130