Tourism Management 48 (2015) 54e63

Contents lists available at ScienceDirect

Tourism Management
journal homepage: www.elsevier.com/locate/tourman

Are the days of tourist information centers gone? Effects

of the ubiquitous information environment
Seong Ok Lyu 1, Jinsoo Hwang*
Division of Tourism, Dongseo University, 47 Jurye-ro, Sasang-gu, Busan 617-716, South Korea

h i g h l i g h t s

Tourism authorities reconsider the current TIC operation due to mobile devices.

The ubiquitous information technology affects tourist demand for TIC visits.

Tourists make a sequential decision about their TIC use behaviors.

The diffusion of Internet decreases TIC demand quantity by 27.6 percent.

The use of SNS increases the number of visits to TICs by 28.4 percent.

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 21 February 2014
Accepted 3 November 2014
Available online

The advent of on-line and cloud access through mobile devices has led to a reconsideration of the
traditional role of tourist information centers. Derived from a sample of Korean domestic tourists, this
paper suggests means by which tourist information centers can respond to new demands for travel
information. Study results indicate a negative relationship between general Internet usage and demand
for tourist information center visitation. Using different econometric models, this study denotes that
heavy social media users tend to show more visits to tourist information centers. This nding may be due
to the characteristics of social media users and equally implies that the conventional engines of tourism
promotion should engage more in social media.
2014 Elsevier Ltd. All rights reserved.

Keywords:
Tourist information centers (TICs)
Ubiquitous information environment
Mobile devices
Information search behaviors
Social network services (SNS)

1. Introduction
The revolution of information communication technology has
fundamentally transformed people's everyday lives. A number of
prior studies (e.g., Buhalis & Law, 2008; Ho, Lin, & Chen, 2012) have
provided empirical evidence supporting that tourist behaviors in
searching for relevant travel information are not free from the
irreversible changes. With the extensive diffusion of mobile devices, such as smartphones and wireless tablet computers, tourists
are now able to access travel information anywhere and at any time
(Okazaki & Hirose, 2009). Frommer (2013) indeed released a survey
report stating that almost nine out of ten smartphone users in the
United States reported searching for travel information, using a
variety of applications, for an average of 93 min a month. The

* Corresponding author. Tel.: 82 51 510 2679; fax: 82 51 512 1853.

E-mail addresses: seonglyu@gmail.com (S.O. Lyu), jinsoohwang@hanmail.net
(J. Hwang).
1
Tel.: 82 51 510 2613; fax: 82 51 512 1853.
http://dx.doi.org/10.1016/j.tourman.2014.11.001
0261-5177/ 2014 Elsevier Ltd. All rights reserved.

tourism industry in Korea may not be immune to the effect of

increasing importance of mobile devices in that the country ranked
rst, worldwide, in the smartphone penetration rate in 2012 (Korea
Herald, 2013). The Korea Communication Committee (2013) reported that the number of smartphone users in the country reached
more than 35 million in December, 2012, indicating that about
seventy percent of Koreans were able to access online travel information with their mobile devices.
The rapid evolution toward the ubiquitous information era is
likely to have a substantial effect on several travel information
providers, particularly tourist information centers, known as the
most conventional external source (Beatty & Smith, 1987). According to Wong and McKercher (2011), the typical mission of
tourist information centers (hereafter called TICs) is the direct
provision of travel information about tourist destinations, which
visitors nd helpful for making their optimal decisions. These
tourism facilities also play a key role in determining tourists' rst
impressions of particular destinations, and ultimately improving
the quality of their experiences (Perdue, 1995). Accordingly, many

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

tourism organizations have made efforts to diversify their strategies for successful TIC operations (Dimanche & Taylor, 2006).
The advent of the ubiquitous information environment raises a
signicant question about those traditional functions of TICs. The
increasing amount of travel information available with the use of
the cutting-edge mobile devices suggests that tourism organizations have to reconsider the current operation of TICs (Wallace,
Walton, & Crabtree, 2009). While there is a consensus that the
effortless access to online information sources requires different
roles of TICs, only a few studies (e.g., Connell & Reynolds, 1999;
Wallace et al., 2009) have examined how the development of information communication technology affects the importance of the
tourism facilities. In particular, there is limited information on the
intricate relationships between tourist demand for TIC visits and
the multifaceted nature of the ubiquitous information
environment.
The major goal of this study is to provide constructive TIC
management implications, which can help the conventional engines of tourism promotion in adjusting to the dramatically
changing information environment. This study also aims to
examine how and what aspects of the rapid technology development, characterized by the wide dissemination of mobile devices,
affect tourist demand for TIC visits. With a sample of Korean domestic tourists, this study intends to enhance our understanding of
how tourists make their decisions to visit a variety of TICs through a
comparison of several econometric models.
2. Literature review
2.1. Tourist information centers (TICs)
Tourism authorities have regarded TICs as the most important
element in tourism promotion, encouraging tourists to stay for a
longer period of time and spend more money within their territories (Fesenmaier, Vogt, & Stewart, 1993; Tierney, 1993).
Ballantyne, Hughes, and Ritchie (2009) indicated that TICs
contribute signicantly to the socio-economic and recreational
well-being of particular communities by facilitating tourists to
experience local attractions and events, and to nancially support
local retails, accommodations, and restaurants. In this regard,
tourism organizations in many destinations have exerted a variety
of efforts to develop the tourism facilities and implement successful
TIC operations (Stewart, Lue, Fesenmaier, & Anderson, 1993).
A report released by the Korea Association of Travel Agencies
(KATA) (2012) showed that there are more than 300 TICs available
across Korea. Those facilities are known to employ about 1200 staff
members, each with the ability to communicate in at least one
foreign language, and to provide useful conveniences such as
computers, printers, fax machines, water fountains, and public
toilets. While most TICs in the country were established prior to the
2002 FIFA World Cup to offer travel-related information to international tourists, more than nine out of ten visitors were reported
to be domestic tourists in 2008 (Korea Tourism Organization (KTO),
2009). Different from TICs in the United Statesdcommonly known
as welcome centers or visitor centersdwhich are chiey
located in the border areas of each state, Korean TICs are operated
near tourist attractions in local areas (e.g., national parks, historic
and cultural sites), and major transportation facilities (e.g., bus
terminals, train stations, airports). Those tourism facilities are
administrated by several public organizations, such as the tourism
departments of local governments, KTO, and KATA, as well as by
different public companies, including the Korea Expressway Corporation and the Korea Railroad.
A large percentage of prior studies on TICs were presented in
major tourism journals for a decade, until the mid-1990s. This may

55

be because several state tourism departments in the United States

made efforts to launch and improve their facilities during that time
period. Accordingly, those studies on the chief engines of tourism
promotion focused on limited topics: behavioral differences between TIC visitor and non-visitor groups (e.g., Howard & Gitelson,
1989; Stewart et al., 1993), tourist motivations for visiting the facilities (e.g., Fesenmaier, 1994; Gitelson & Perdue, 1987), and TIC
site preferences (e.g., Pennington-Gray & Vogt, 2003; Perdue,
1995). While a couple of studies (e.g., Fesenmaier & Vogt, 1993;
Tierney, 1993) attempted to utilize an economic approach,
concentrating on the impacts of TIC development and services on
tourist expenditures, there is no comprehensive research available
on tourist demand for TIC visits.
2.2. Tourist information search behaviors & mobile devices
Tourist information search is often referred to as a dynamic
process wherein individuals use various amounts and types of information sources in response to internal and external contingencies to facilitate travel planning (Fodness & Murray, 1997, p.
506). While searching for relevant travel information, tourists
attempt to reduce uncertainty and seek to enhance the quality of
their trips (Lehto, O'Leary, & Morrison, 2004). In order to obtain
appropriate travel information during the travel planning stage,
and to make adequate onsite decisions about attractions, accommodations, activities, and restaurants, tourists utilize two different
ways of searching for information: internal and external (Kim,
Lehto, & Morrison, 2007). Internal search represents the application of travel information derived from long-term memory to solve
their problems (Okazaki & Hirose, 2009). Past experiences of using
travel products and visiting particular destinations comprise an
important basis for internal search behaviors (Engel, Blackwell, &
Miniard, 1995). When this search cannot provide useful information, tourists utilize external sources such as friends and relatives,
brochures, magazines, and travel agents, in addition to TICs, to acquire information and expand their knowledge (Wicks & Schuett,
1991).
According to Fodness and Murray (1997) classifying travel information sources by using two criteria of commercial and
communication characteristics, TICs fall into the quadrant of noncommercial and personal communication sources. Based on this
taxonomy system, many Internet websites accessible with mobile
devices can be classied as impersonal sources regardless of commercial nature. The use of mobile devices is known to provide more
advanced computing capabilities and easier accessibility to information sources than traditional equipment (Charlesworth, 2009).
Thus, prior studies (e.g., Wang, Park, & Fesenmaier, 2012; Xiang &
Gretzel, 2010) suggested that the unique exibility in both time
and space of mobile devices allows tourists to successfully solve
their problems, share experiences, and compile memories.
Concomitant with the progress in information communication
technology, there is a wealth of research examining how tourists
adapt to the ubiquitous information environment. In particular, the
technology acceptance model (TAM), originally proposed by Davis
(1989), has been the most popular conceptual framework to help
understand the effects of the technology development on tourist
information search behaviors. Basically, the TAM is known to be
advantageous for viewing the complicated relationships among
personal belief, attitudes, and behavioral intentions (Hsu & Lu,
2004). Among several elements in the model, the two main concepts of perceived usefulness and perceived ease of use serve as the
fundamental factors to predict users' levels of technology acceptance (Davis, Bagozzi, & Warshaw, 1989). The concept of perceived
usefulness is dened as the extent to which the use of a technology
improves an individual's performance, while perceived ease of use

56

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

is referred to as the extent to which a person believes that using the

system is free of effort (Venkatesh & Davis, 2000). Using this
framework, a plethora of research (e.g., Agarwal & Prasad, 1999;
Hsu & Lu, 2004) has provided supportive evidence that these two
components play a central role in determining tourists' attitudes
toward the technology advancement.
2.3. TIC operation & ubiquitous information environment
The advent of the ubiquitous information era has encouraged
many people to believe that TICs, as traditional tourism promotional facilities, become relics of the old days. This belief is attributed to the transformation in tourist information search behaviors,
which is stimulated by the wide dissemination of mobile devices
(Wang et al., 2012). In response to the progress in the travel information environment, tourism organizations are eager to develop
a variety of applications that mobile device users can gain onsite
information. For example, the Korea Tourism Organization (KTO)
recently began distributing its free smartphone applications, Visit
Korea (KTO, 2011). The applications, available in both Korean and
English, involve various types of travel information such as popular
travel courses, seasonal menus of local restaurants, transportation
schedules, bank and ATM locations, and even useful Korean phrases
for international tourists. At the same time, several local governments in the country have recently attempted to reduce the
number of TICs, and to restructure the operation systems (KATA,
2012).
While there is a general agreement about the changing roles of
TICs in the ubiquitous information environment, limited research
effort has been made to explore how the development of information communication technology affects the importance of TICs.
The work of Connell and Reynolds (1999) worth noting in that it
discussed the signicant inuences of recent technological progress on TIC operations. Utilizing in-depth interviews with TIC
workers, their study offered different TIC management recommendations, which could help the traditional tourism facilities to
adapt to the rapid developments in information communication
technology. While emphasizing the importance of customeroriented services, they proposed alternative tasks for TICs
through making full use of existing technologies (p. 507).
A couple of prior studies focused on useful management strategies using several advanced technologies to enhance TIC visitor
satisfaction. For Kentucky TIC users, Wallace et al. (2009) recommended several management strategies: the implementation of
web-based social media services to spread real-time trafc alerts,
the use of static kiosks or kiosk-type displays to offer reliable information, and the utilization of subscription-based systems to
disseminate customized information. McCann (1999) also noted
that travel information kiosks using multimedia technologies
allowed TIC visitors to quickly and easily acquire adequate information. Despite the meaningful contribution to practical TIC operations, these studies failed to provide opportunities to more
comprehensively understand which aspects of technology development threaten the survival of TICs, and how tourists formulate
TIC visit demand in this ever-changing information environment.
3. Research hypotheses
Based on several prior studies on consumer decision-making
process (e.g., Duan, Manning, Morris, & Newhouse, 1983; Jang,
Ham, & Hong, 2007), this study posits that tourists make a
sequential decision about their TIC use behaviors: whether or not to
visit a TIC / how many times to visit the facilities. Accordingly, we
utilized two different dependent variables representing tourists'
previous experiences of TIC visits (VISIT), and their total number of

uses of TICs (USES). The initial dependent variable of VISIT entails

binary outcomes regarding whether or not to visit a TIC during last
12 months. Once they reported prior experiences of visiting a TIC,
respondents were asked to indicate their levels of USES: how many
times they visited the facilities during the same period of time.
A set of independent variables were comprised of several items
encompassing the behavioral characteristics of travel information
search, personal attitudes toward mobile devices and social
network services, and several socio-demographic features. These
variables were incorporated to examine their effects on the two
dependent variables. Table 1 indicates the descriptive statistics and
proposed research hypotheses for each variable used in this study.
The variable NDT demonstrates the total number of domestic
trips taken during the past 12 months. With 3191 visitors to
Michigan welcome centers, Pennington-Gray and Vogt (2003)
found that the probability and frequency of visits to TICs increase
as the number of domestic trips increases. Accordingly, it seems
reasonable to hypothesize a positive coefcient sign for the NDT
variable (H1). The independent variables INT and SNS were individually employed to investigate how the usage of the Internet and
social network sites affects respondents' demand for TIC visits.
Since online-based information sources and social media serve as
important channels for tourists to search for travel information and
share their travel experiences (Huang, Basu, & Hsu, 2010), many
people using these tools are less likely to visit TICs. Therefore, this
study expected to nd a negative relationship between these
behavioral characteristics and TIC use demand (H2-1; H2-3). While

Table 1
Detailed information about each variable.
Variables

Description

Mean (S.D.)

Hypotheses
(expected signs)

VISIT

TIC visits during last 12

months 1; non-visit 0
Total number of visits to TICs
during last 12 months
Number of domestic travels
during last 12 months
Use of Internet as the main
information source 1
Daily time spent using mobile
devices: 1 h or more 1
Use of SNS to upload personal
travel experiences and
pictures 1
Perceived usefulness of mobile
devices 1
Perceived ease of use of mobile
devices 1
Use of mobile devices to search
for travel information 1
Experiencing uncertainty of
travel information from mobile
devices 1
Satisfaction with travel
information from mobile
devices 1
Age bracket: 20s 1
Age bracket: 30s 1
Age bracket: 40s 1
Age bracket: 50 and over 1
Male 1
Monthly income: below
US$3000 1
Monthly income: US$3000 to
US$6000 1
Monthly income: US$6000 and
more 1
Unmarried 1

0.42

DV

USES
NDT
INT
DTM
SNS

PUM
PEM
UMI
EUI

SAT

AGE1
AGE2
AGE3
AGE4*
GND
INC1*
INC2
INC3
MAR

Note: *Reference category.

0.85 (1.27)
4.77 (3.74)

H1 ()

0.93

H2-1 ()

0.84

H2-2 ()

0.39

H2-3 ()

0.82

H3-1 ()

0.74

H3-2 ()

0.85

H4-1 ()

0.73

H4-2 ()

0.60

H4-3 ()

38.34 (10.67)

H5-1 ()

0.52
0.25

H5-2 (/)
H5-3 (/)

0.54
0.21
0.61

H5-4 (/)

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

there is limited information on the associations between the

amount of use of mobile devices and the demand for TIC visits, it
may be logical to assume that the probability and frequency of visits
to TICs decrease as the amount of daily time spent using mobile
devices increases. This is because individuals with higher use levels
of mobile devices are more likely to indicate positive attitudes toward travel information obtained from the advanced devices over
the use of the traditional information sources (Wang et al., 2012). In
this regard, the DTM variable indicating daily time spent using
mobile devices was hypothesized to show a negative association
with our dependent variables (H2-2).
This study included the two different explanatory variables PUM
and PEM, representing the core elements of the technology
acceptance model (TAM): perceived usefulness and perceived ease
of use concerning diverse mobile devices (Davis, 1989). Table 2
exhibits the results of a principal component factor analysis with
varimax rotation for these two concepts. Respondents were asked
to show their levels of agreement on seven different measurement
items, using a ve-point Likert format with values ranging from 1
(strongly disagree) to 5 (strongly agree). After calculating the
arithmetic mean scores of the two concepts, we collapsed each
score into the dichotomous dummy responses with 1 (mean scores
above 3) and 0 (mean scores 3 and lower). Consistent with prior
studies (e.g., Agarwal & Prasad, 1999; Hsu & Lu, 2004), indicating
that individuals with perceptions of usefulness and ease of use of a
particular technology show positive attitudes toward the technology, and probably engender less frequent uses of old technologies,
this study hypothesized a negative relationship between these two
variables and TIC use demand (H3-1; H3-2).
We also made use of multiple independent variables associated
with respondents' experiences using mobile devices to obtain
travel information during a typical domestic trip. The variables UMI
and SAT involved dichotomous outcomes about whether respondents actually utilized their mobile devices to gain travel information and were satised with the information from diverse
mobile tools, respectively. According to Connell and Reynolds
(1999), tourists with higher levels of dependence upon mobile
devices to search for travel information, and satisfaction with the
quality of travel information obtained from their devices are less
likely to visit a TIC. These facts led this study to expect consistent
negative coefcient signs for the UMI and SAT variables (H4-1; H43). Based on Wallace et al. (2009), denoting that tourists who
encountered uncertainty about travel information retrieved from
mobile devices were more likely to show positive attitudes toward
TIC visits, this study hypothesized a positive coefcient sign for the
EUI variable (H4-2). Several socio-demographic variablesdage
(AGE), gender (GND), monthly household income (INC), and marital
status (MAR)dwere also incorporated to identify their inuences
on TIC use demand (H5-1; H5-2; H5-3; H5-4). In particular, we

57

hypothesized that younger tourists are likely to be less willing to

visit a TIC because of their stronger familiarity with mobile devices
than with the traditional personal tourism facilities (Fesenmaier &
Vogt, 1993). The research framework for this study is depicted in
Fig. 1.
4. Methods
4.1. Data collection
An online survey was conducted to collect data for this study.
We made use of a sample of 2033 domestic tourists with the
application of a computer-generated proportionate sampling procedure. Respondents were randomly selected from an email list of
approximately 1 million research participants who were administered by a Korean online survey company. During the three weeks
from October to November of 2012, this study collected 621 responses, which was an effective response rate of 30.5%. After
excluding 22 cases with incomplete information and 85 respondents who reported non-use of mobile devices in their
everyday lives, 514 responses were included in the nal data set.
4.2. Models
The vast majority of tourism demand data using offsite surveys
tend to be plagued with the presence of zero observations (Loomis
& Walsh, 1997). The excess zeros result in serious estimation biases
and inconsistencies when employing traditional linear regression
models such as ordinary least squares (OLS) to examine tourist
behavioral patterns (Cai, Hong, & Morrison, 1995). This is because
the plethora of zero observations often leads to a non-normal
distribution in errors around the expected average of dependent
variable (Maddala, 1999).
To address the combination of zero observations and positive
values in many tourism demand data, the Tobit model is often
applied to examine the distinctive inuences of diverse determinants on tourist behaviors of interest. The Tobit model is
commonly given by y*i X0i b i, where Xi denotes the vector of an
individual i's characteristics, b is the vector of coefcients in terms
of independent variables, and i represents the unobserved error
term. The observed variable (yi) can be expressed as follows:
yi 0 if y*i  0; yi y*i if y*i > 0, where y*i shows the unobserved latent variable, and 0 is the lower censoring point. The
probability
of
a
censored
observation
is
Pry*  0 PrX0i b  0 Ff0  X0i b=sg, where F$ is the
standard normal cumulative distribution function. The expected
value of y for the non-censored observations can be shown as
Eyi jXi ; yi > 0 X0i b sffX0i b  0=sg=Ff0  X0i b=sg,
where
f$ is the standard normal density (Cameron & Trivedi, 1998).

Table 2
Results of principal component factor analysis for perceived usefulness and ease of use.
Dimensions and items
Perceived usefulness
Information provided by mobile devices was precise
Information provided by mobile devices was trustable to make travel decisions
Information provided by mobile devices was useful to make travel decisions
Information provided by mobile devices was efcient to make travel decisions
Perceived ease of use
Learning to operate mobile devices was easy
It was easy to become skillful at using mobile devices
It was easy to communicate with others by using mobile devices
Note: Total explained variance 68.973%.
KMO measure of sampling adequacy 0.746.
Bartlett's test of sphericity 1660.741 (p < .001).

Mean

S.D.

3.49
3.46
3.85
3.89

0.63
0.65
0.71
0.70

3.79
3.74
3.76

0.74
0.78
0.78

Variance
explained

Standardized factor
loadings

47.16%

Cronbach's
alpha
0.75

0.831
0.808
0.649
0.658
21.81%

0.89
0.895
0.907
0.867

58

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

Fig. 1. Research framework for this study.

Despite its extensive application in tourist expenditure studies,

the Tobit model is so restrictive that it fails to comprehensively
illustrate an individual's sequential decision-making process:
whether or not to participate / how many times to engage. This is
because the model assumes that the decision to consume a
particular product has the same direction as the decision about the
amount of the product to consume (Amemiya, 1984). While
McDonald and Moftt (1980) proposed a Tobit decomposition
method to separately assess the two different decisions, the
censored data model is still inexible with the failure of normality
and homoskedasticity assumptions (Cameron & Trivedi, 1998). As a
result, several econometric models, such as two-part (Duan et al.,
1983), double-hurdle (Cragg, 1971), and Heckman sample selection models (Heckman, 1979), have been developed to address individuals' sequential decision-making mechanisms. Unlike the
Heckman model, which assumes a close interdependence between
the two decisions, both two-part and double-hurdle models
postulate that the decisions are made in an independent manner,
each of which is also determined by a different array of explanatory
variables.
According to Cragg (1971), individuals must cross two separate
hurdles to report their positive frequency of consumption. The rst
hurdle is associated with whether or not to purchase; the second
one involves how much to consume. In the rst stage of participation decision, a logit (or probit) model estimation is generally
employed for the binary choices between purchase and nonpurchase. Once the rst hurdle has been cleared, another independent procedure is needed to estimate the parameters that
demonstrate respondents' levels of consumption. For the second
phase, the two-part model makes use of OLS or lognormal model
estimations with a truncated sample of participants (Duan et al.,
1983). Unlike the two-part model, a set of count-data models,
including truncated Poisson and truncated negative binomial
regression models, are used in the double-hurdle model estimations (Mullahy, 1986). The probability of an observation with a logit
model in the rst stage and a truncated regression model in
the second stage, for example, can be expressed by
Pryi 0jWi LW0i g; Pryi jjXi ; Wi ; yi > 0 1  LW0i gexp

li lji =j!1  expli ; j 1; 2; , where li denotes expX0i b and

g represents b=s (Greene, 2012).
5. Results
5.1. Sample description
Of the 514 domestic tourists who reported using at least one
mobile device during last 12 months, more than half (52.1%) were
males, with a mean age of 38.3 years. Respondents traveled
domestically an average of 4.8 times during the same period, and
their median monthly household pre-tax income fell in the range of
US$4000 to US$5000. Slightly more than four out of ten (42.4%)
indicated that they visited a TIC for their domestic trips, and reported an average of 2.0 visits. Respondents were also segmented
into two groups based on their previous experiences of visiting a
TIC, in order to identify group differences between visitors and nonvisitors. The two tourist groups disclosed a signicant difference in
the number of domestic trips (t 3.11, p < .05), indicating that TIC
visitors took more domestic trips than non-visitors. This may be an
expected result in that frequent tourists are more likely to visit
several kinds of TICs. Nevertheless, there was no sociodemographic difference between the two segments. Table 3 demonstrates the results of several difference tests.
5.2. OLS & Tobit model estimations
The results of the OLS and Tobit model estimations are presented in Table 4. Those models assume that the two decisions
regarding whether to visit and how many times to visit TICs are
made simultaneously. Unlike the Tobit model, characterizing the
censoring technique, the OLS model was estimated with the full
sample, including zero observations, which implies inconsistent,
biased estimates of parameters. The statistics of variance ination
factors (VIF), ranging from 1.05 to 3.03, displayed the absence of
serious collinearity problems between each independent variable.
Several explanatory factors, representing tourist behaviors to
search for information and attitudes toward mobile devices, as well

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

Table 3
Results of difference tests between visitors and non-visitors.
Variables

NDT
AGE
GND
MAR

Visitors (n 218)

Non-visitors (n 296)

Mean (S.D.)

Mean (S.D.)

5.36 (3.70)
38.72 (11.07)
0.51
0.60

4.33 (3.72)
38.07 (10.37)
0.52
0.62

Difference

t 3.109*
t 0.682
c2 0.183
c2 0.253

Note: NDT number of domestic travels during last 12 months; AGE age;
GND gender; MAR marital status.
*Signicant at the .05 level.

as previous experiences using mobile devices were statistically

signicant, while all of the socio-demographic variables were
insignicant, with the exception of the AGE variable. All signs of
each signicant coefcient were also identical with our hypotheses,
except for the variable of SNS, associated with respondents' experiences using social media. Moreover, there was no difference in
each signicant coefcient signal across the two models.
The estimation results demonstrated that the number of domestic trips (i.e., NDT) was positively associated with respondents'
frequency of TIC visits (H1). As discussed above, this showed that
tourists were more likely to visit TICs if they took domestic trips
more frequently. The signicant negative coefcient signs of the
INT variable provided empirical evidence that tourists who used
the Internet as their main travel information sources were less
likely to visit various kinds of TICs (H2-1). Nevertheless, it may be
interesting that the variable of SNS, representing the use

59

experiences of social network services, indicated signicant positive signs in both models (H2-3). These unexpected signals denoted
that tourists committed to various social media visited TICs more
often. This result will be addressed more extensively below; it may
result from the unique nature of SNS users.
The negative impacts of recent technology developments on TIC
visits were further conrmed through the coefcient signs of the
two main concepts of the TAM: perceived usefulness (i.e., PUM) and
perceived ease of use (i.e., PEM) for mobile devices (H3-1; H3-2).
Our results suggested that tourists with perceptions of usefulness
and ease of use for mobile devices were less likely to visit TICs. The
positive coefcient signs of the EUI variable showed that respondents who experienced uncertainty about travel information
acquired from their mobile devices tended to visit TICs in a more
habitual manner (H4-2). Several variables associated with the age
bracket disclosed signicant negative signs, suggesting that respondents' frequency of TIC use decreased as their age decreased,
compared to the reference category of those 50 and above (H5-1).

5.3. Two-part & double-hurdle model estimations

Unlike the OLS and Tobit models, the two-part and doublehurdle models are based on the important assumption that consumers generally make a sequential decision to purchase a commodity. In the rst stage for binary outcomes, we made use of the
logit model with the full sample to examine distinctive inuences

Table 5
Results of two-part and double-hurdle models.
Table 4
Results of OLS model with full sample and Tobit model.
OLS (y  0)

NDT
INT
DTM
SNS
PUM
PEM
UMI
EUI
SAT
AGE1
AGE2
AGE3
GND
INC2
INC3
MAR
constant

Coef

S.E.

Coef

S.E.

0.077***
0.930***
0.276**
0.642***
0.651***
0.370***
0.161
0.479***
0.127
0.362*
0.239
0.245*
0.065
0.128
0.158
0.196
1.315***

0.013
0.192
0.139
0.103
0.141
0.128
0.158
0.115
0.103
0.189
0.151
0.139
0.097
0.120
0.150
0.139
0.298

0.125***
1.352***
0.463
1.152***
1.094***
0.384
0.121
1.311***
0.143
0.743*
0.633*
0.699**
0.016
0.202
0.248
0.365
0.190
2.00

0.027
0.393
0.308
0.220
0.298
0.274
0.353
0.272
0.223
0.404
0.329
0.305
0.209
0.262
0.326
0.299
0.641

s
F
Adj R2
LR c2
McFadden r
Log likelihood
N

Tobit

13.93***
0.29

757.52
514

135.33***
0.10
629.30
514

Note: NDT number of domestic travels; INT use of Internet as the main information source; DTM daily time spent using mobile devices; SNS use of SNS to
upload travel experiences; PUM perceived usefulness of mobile devices;
PEM perceived ease of use of mobile devices; UMI use of mobile devices to
search travel information; EUI experiencing uncertainty of travel information
from mobile devices; SAT satisfaction with travel information from mobile devices; AGE age; GND gender; INC household income level; MAR marital
status.
***Signicant at the .01 level.
**Signicant at the .05 level.
*Signicant at the .1 level.

NDT
INT
DTM
SNS
PUM
PEM
UMI
EUI
SAT
AGE1
AGE2
AGE3
GND
INC2
INC3
MAR
constant
HosmereLemeshow c2
F
Adj R2
LR c2
McFadden r
Log likelihood
N

VISIT

USES
Two-part

Double-hurdle

Logit

OLS (y > 0)

Truncated poisson

Coef

S.E.

Coef

S.E.

Coef

S.E.

0.065**
0.749*
0.294
0.859***
0.634**
0.093
0.143
1.133***
0.007
0.694**
0.631**
0.705**
0.086
0.228
0.293
0.273
0.121
4.59

0.027
0.397
0.284
0.207
0.288
0.260
0.332
0.251
0.210
0.383
0.308
0.285
0.197
0.245
0.304
0.281
0.610

0.112***
0.934***
0.286
0.396***
0.766***
0.339**
0.510**
0.255
0.182
0.031
0.075
0.065
0.205
0.017
0.122
0.104
1.897***

0.018
0.229
0.202
0.136
0.184
0.171
0.225
0.192
0.138
0.249
0.209
0.193
0.130
0.173
0.215
0.188
0.405

0.066***
0.526***
0.229
0.353**
0.473***
0.192
0.332
0.197
0.102
0.023
0.033
0.010
0.195
0.061
0.026
0.036
0.118

0.014
0.177
0.246
0.152
0.162
0.163
0.250
0.219
0.144
0.245
0.208
0.185
0.128
0.180
0.209
0.175
0.414

10.50***
0.41

312.51
514

535.58
218

115.14***
0.19
249.47
218

Note: NDT number of domestic travels; INT use of Internet as the main information source; DTM daily time spent using mobile devices; SNS use of SNS to
upload travel experiences; PUM perceived usefulness of mobile devices;
PEM perceived ease of use of mobile devices; UMI use of mobile devices to
search travel information; EUI experiencing uncertainty of travel information
from mobile devices; SAT satisfaction with travel information from mobile devices; AGE age; GND gender; INC household income level; MAR marital
status.
***Signicant at the .01 level.
**Signicant at the .05 level.
*Signicant at the .1 level.

60

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

of each predictor on the probability of visit to a TIC. The results of

the logit model estimation are presented in Table 5.
To evaluate the goodness of t by comparing the sample frequency of the dependent variables and the tted probability
within subgroups of observations, we employed the HosmereLemeshow test. As demonstrated in Table 5, the HosmereLemeshow test statistic showed a good t to our data. All
signicant coefcients of the logit model indicated the same signs
as those of the Tobit model, while the magnitude seemed to be
differentiated. Briey, the variable of NDT revealed a signicant
positive sign, suggesting the more domestic trips respondents
took, the more likely they were to visit a TIC (H1). Paralleling the
results of the OLS and Tobit model, the SNS variable revealed a
positive coefcient signal, indicating that respondents had a
higher probability of visiting a TIC provided that they previously
used social media (H2-3). The predictor PUM disclosed a signicant negative effect on the probability of visiting a TIC, implying
that tourists with a perception of usefulness of mobile devices
were less likely to visit a tourism promotion facility (H3-1).
With the truncated sample of TIC visitors, this study investigated tourists' levels of TIC uses by utilizing two different regression techniques. Table 5 also shows the results of the truncated OLS
and truncated count-data model estimations. The former procedure stands for Duan et al.s (1983) two-part model, while the latter
is the double-hurdle model proposed by Cragg (1971). The F statistic in the two-part model was signicant, suggesting the truncated OLS model suitably tted the data. Unlike the estimation
results of the rst-stage logit model, none of the sociodemographic variables revealed any statistical signicance. The
PEM variable, representing perceived ease of use for mobile devices, showed a negative coefcient sign. This indicated that tourists with a perception of mobile devices being easy to use were less
likely to frequently visit TICs (H3-2).
For the double-hurdle model, we estimated a set of parameters
with the application of the truncated Poisson regression model. The
negative binomial model is preferred over the Poisson model
because of the latter's strict assumption that the variance and mean
of the dependent variable are identical (Cameron & Trivedi, 1998).
As shown in Table 1, the data for this study displayed an overdispersed distribution of the dependent variable, commonly caused
by excessive zero observations. Despite the existence of the problematic distribution, the splitting effect of the double-hurdle model
is known to attenuate the strict assumption of the Poisson model
(Greene, 2012). As a result, this study estimated parameters using
the Poisson model.
The truncated Poisson model was believed to t our data well
based on the signicant likelihood ratio statistic (c2df 16 115.14,
p < .001). Compared to the estimation results of the truncated OLS
model, fewer independent variables revealed a statistical signicance. Aside from the variable of SNS, associated with experiences
of using diverse social media, all of the signicant coefcient signs
were consistent with prior expectations. The negative coefcient
sign of the INT variable showed that respondents who accessed the
Internet to search for travel information were unwilling to visit TICs
frequently (H2-1).
5.4. Model comparisons
Among different competing models discussed above, we
attempted to identify a model that best tted our data. The selection of the best-performing model is heavily dependent upon the
complicated trade-offs among several criteria such as t indices,
parsimony, and interpretational easiness (Cameron & Trivedi,
2010). As illustrated in Table 6, this study made use of three
different criteria that are popular for comparing multiple models:

Table 6
Information criteria for model comparisons.
Model

Log likelihood

Degree of freedom

AIC

BIC

Tobit
Two-part
Double-hurdle

629.30
848.28
561.98

18
34
35

1294.59
1768.56
1193.95

1370.95
1921.28
1342.43

Note: AIC Akaike's information criterion; BICBayesian information criterion.

log likelihood, Akaike's information criterion (AIC), and Bayesian

information criterion (BIC). This study excluded the OLS model
from the comparison procedure because the model using the full
sample with the excess zeros is known to deliver biased estimation
results.
To make a better choice between the Tobit model and the two
different sequential models, we rst conducted the likelihood
ratio (LR) tests. The results of the LR tests indicated that respondents' decisions to visit TICs were made with a sequential
process. The double-hurdle approach signicantly improved the
model t, compared to the Tobit and two-part models. The superiority of the double-hurdle model was also conrmed through
comparisons of different information criteria of AIC and BIC.
Because smaller values of the two criteria are commonly
preferred, the double-hurdle model was chosen as the nal model
for this study.

5.5. Marginal effects

It is impossible to directly interpret the magnitude of each
coefcient in count-data models. Thus, marginal effects are
calculated to assess the effects on the conditional mean of the
dependent variable (Cameron & Trivedi, 1998). Table 7 demonstrates the marginal effects of each independent variable in the
double-hurdle model. The marginal effects in the rst decision
model concerning whether or not to visit a TIC showed that
tourists in the 40s age-range were 16.3 percent less likely to visit
a TIC, compared to the base category (i.e., tourists aged 50 and
over), while holding all of the other variables constant. The
invariant negative signs in a set of AGE variables indicated the

Table 7
Marginal effects of double-hurdle model.

NDT
INT
DTM
SNS
PUM
PEM
UMI
EUI
SAT
AGE1
AGE2
AGE3
GND
INC2
INC3
MAR

VISIT

USES

0.016
0.185
0.070
0.208
0.156
0.023
0.035
0.253
0.002
0.161
0.147
0.163
0.021
0.055
0.072
0.066

0.086
0.852
0.274
0.457
0.701
0.259
0.382
0.239
0.131
0.030
0.042
0.013
0.252
0.079
0.033
0.047

Note: NDT number of domestic travels; INT use of Internet as the main information source; DTM daily time spent using mobile devices; SNS use of
SNS to upload travel experiences; PUM perceived usefulness of mobile devices;
PEM perceived ease of use of mobile devices; UMI use of mobile devices to
search travel information; EUI experiencing uncertainty of travel information
from mobile devices; SAT satisfaction with travel information from mobile
devices; AGE age; GND gender; INC household income level;
MAR marital status.

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

higher probability of older tourists to use a tourism facility. A

perception of usefulness for mobile devices decreased the likelihood of a visit to a TIC by 15.6 percent, while the use of diverse
social network channels increased the probability of visiting a
facility by 20.8 percent. Furthermore, respondents who experienced uncertainty about travel information gleaned from mobile
devices were 25.3 percent more likely to enter a TIC.
The marginal effects in the second decision, how many times to
visit TICs, represent the changes in the quantity of observing a
certain outcome when an independent variable changes by one
unit. The results indicated that taking one more trip increased respondents' frequency of TIC use by about 0.086 times. Respondents
who used the Internet to search for travel information showed
fewer visits to TICs by approximately 0.852 times. This nding
demonstrated that the extensive diffusion of the Internet decreased
TIC use demand by 27.6 percent when the average number (3.089
times) of visits to the facilities for those who did not use the
Internet to acquire travel information was taken into account (i.e.,
(0.852/3.089)*100 27.6%). The marginal effect of the SNS variable
meant that various kinds of social media increased the number of
visits to TICs by 28.4 percent, given the mean value (1.608 times) of
the dependent variable for respondents who did not use diverse
social network services (i.e., (0.457/1.608)*100 28.4%).
6. Discussion
The primary objective of this study was to examine the impacts
of recent rapid changes in the information communication environment, characterized by the wide dissemination of mobile devices, on tourist demand for TIC visits. We also aimed to provide
useful TIC management implications to help the traditional elements of tourism promotion adapt to the ever-changing technology landscape, and to look for new tasks that will improve
tourist experiences. With a sample of domestic tourists in Korea,
we estimated several models to demonstrate how tourists make
decisions about TIC visits in this ubiquitous information
environment.
Overall, study results suggest that the development of information communication technology has a negative effect on the
formation of TIC use demand. The marginal effects of the
double-hurdle model showed that the use of the Internet to
search for travel information decreased the demand quantity of
TIC use by 27.6 percent and dropped the probability of visiting a
facility by 18.5 percent. Among the two different major components of the technology acceptance model (TAM), tourists'
perceived usefulness for mobile devices also decreased both the
actual number and the likelihood of visiting TICs. However, the
predictor related to perceived ease of use regarding the electronic devices failed to reveal signicant associations with the
two dependent variables in the double-hurdle model. These
ndings provide supportive evidence that the revolutionary
technology progress has substantially reshaped tourist behaviors
for information search.
It may be interesting that the coefcients of the SNS variable
consistently revealed a positive sign across several models,
which was the opposite of our prior expectations. This nding
indicates that tourists' experiences using a variety of social
network services encourage them to visit TICs more frequently
and increase their probability of visiting a facility. The differentiated effect from other predictors associated with tourist attitudes toward the ubiquitous information environment may be
attributable to the unique characteristics of social media users. It
is well known that the vast majority of social media content
regarding inexpensive goods, product brands, and even serious
political issues are voluntarily generated, circulated, and used by

61

consumers to educate each other (Blackshaw & Nazzaro, 2006).

Because social network sites allow users to effortlessly upload
their travel experiences and photos, as well as to broadcast realtime content using mobile devices, many members of social
media are eager to acquire location-based information about
specic travel destinations and attractions (Xiang & Gretzel,
2010). In this sense, social network service users are likely to
show greater demand for TIC visits to obtain a wealth of travel
information and knowledge from the conventional information
sources.
This nding suggests an important management strategy that
helps deliver timely travel information to current and potential TIC
visitors. In response to the emergence of new media, TIC operators
need to utilize social network sites for successful travel information distribution. Wallace et al. (2009) indeed noted that the
expanded use of existing social media is one of the best ways to
provide cost-efcient travel information to tourists with mobile
devices. Along the same lines, Xiang and Gretzel (2010) stressed
the growing importance of social network marketing that leads to
a better understanding of what types of travel information are
searched by tourists. In other words, the real-time provision of
tailored travel information to different social network channels
can be benecial for increasing TIC visit demand and user satisfaction. Connell and Reynolds (1999) indicated, nevertheless, that
a large percentage of TIC workers reported considerable concerns
about unemployment likely resulting from the recent technology
progress, while showing their unwillingness to accept and learn
new technologies. Accordingly, a variety of efforts are needed for
TIC staff to adjust to the technological dynamics and recognize the
important role of social network sites in distributing adequate
travel information.
Furthermore, the results of the rst-stage logit model estimation revealed that tourists with a feeling of uncertainty
regarding travel information from mobile devices were more
likely to come to a TIC. Although a wealth of travel information is
accessible anywhere and at any time these days, not everyone is
able to nd their desired information (Ho et al., 2012). Many
prior studies (e.g., Okazaki & Hirose, 2009; Wang et al., 2012)
have provided supportive evidence that a large proportion of
tourists expressed signicant dissatisfaction and confusion
about travel information obtained from diverse Internet sources.
In this regard, tourists may be willing to use the classic knowledge brokers to check the quality and reliability of travel information they already obtained by using both online and ofine
sources. While signicant progress has been made in the area of
information communication technology, our nding indicates
that the traditional tourism facilities are still important promotional elements that play a central role in determining the
quality of their clientele's experiences. Moreover, this result
suggests a useful implication that TIC operations can be successfully improved by means of the O2O (Online-to-Ofine)
business model, which has emerged as a new internet-based
marketing phenomenon (Shen & Wang, 2014). According to
this model pursuing a combination of online and ofine business
sectors, consumers are rstly attracted and make payments by
way of online platform, and then visit real stores to gain the
purchased products. With the usage of this O2O approach, many
TICs are likely to provide more realistic and destination-based
travel information to visitors who previously obtained knowledge from a variety of online channels.
Despite several noteworthy ndings, study limitations are
apparent and may point to future research avenues. This study did
not incorporate relevant predictors representing tourist expenditures into the model estimation procedure. Individual expenditures
have often been used as one of the most important right-hand side

62

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

components in tourism and recreation demand functions (Loomis

& Walsh, 1997). Because TICs in Korea are operated by local governments and non-prot organizations, visitors can freely use most
of the equipment and services available in those facilities.
Accordingly, visitor expenditures at TICs are impossible to derive.
This study is restricted from generalizing the results to the entire
population of TIC visitors in that the current study did not include
international tourists in our sample. International tourists, particularly from developing countries, likely show different attitudes
toward the tourism facilities because they are unaccustomed to the
ubiquitous information environment. Future research is required to
provide a better understanding of heterogeneous TIC use demand
among domestic and international tourists. Another special caution
is also needed when generalizing the results due to the absence of a
non-response check.
In summary, there is general agreement that the conventional
engines for tourism promotion are inferior goods in this advanced
information environment, unlikely to allow a dramatic increase in
TIC use demand in the near future. Thus, redressing the critical
drawbacks of mobile devices and searching for new services that
the electronic machines are currently unable to offer may be
practical avenues for TICs to survive in the ubiquitous information
era and ultimately fulll visitor satisfaction.
References
Agarwal, R., & Prasad, J. (1999). Are individual differences germane to the acceptance of new information technologies? Decision Sciences, 30(2), 361e391.
Amemiya, T. (1984). Tobit models: a survey. Journal of Econometrics, 24, 3e61.
Ballantyne, R., Hughes, K., & Ritchie, B. W. (2009). Meeting the needs of tourists: the
role and function of Australian visitor information centers. Journal of Travel &
Tourism Marketing, 26, 778e794.
Beatty, S. E., & Smith, S. M. (1987). External search effort: an investigation across
several product categories. Journal of Consumer Research, 14, 83e95.
Blackshaw, P., & Nazzaro, M. (2006). Consumer-generated media (CGM) 101: Word-ofmouth in the age of the web-fortied consumer. New York: Nielsen BuzzMetrics.
Buhalis, D., & Law, R. (2008). Progress in information technology and tourism
management: 20 years on and 10 years after the Internet: the state of eTourism
research. Tourism Management, 29(4), 609e623.
Cai, L., Hong, G. S., & Morrison, A. (1995). Household expenditure patterns for
tourism products and services. Journal of Travel and Tourism Marketing, 4(4),
115e140.
Cameron, A. C., & Trivedi, P. K. (1998). Regression analysis of count data. Cambridge,
UK: Cambridge University Press.
Cameron, A. C., & Trivedi, P. K. (2010). Microeconometrics using STATA. College Station, TX: STATA Press.
Charlesworth, A. (2009). The ascent of smartphone. Engineering & Technology, 4(3),
32e33.
Connell, J., & Reynolds, P. (1999). The implications of technological developments on
tourist information centres. Tourism Management, 20(4), 501e509.
Cragg, J. (1971). Some statistical models for limited dependent variables with
application to the demand for durable goods. Econometrica, 39(5), 829e844.
Davis, F. D. (1989). Perceived usefulness, perceived ease of use and user acceptance
of information technology. MIS Quarterly, 13(3), 319e340.
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer
technology: a comparison of two theoretical models. Management Science,
35(8), 982e1003.
Dimanche, F., & Taylor, M. M. (2006). An analysis of the differences between state
welcome center users and local visitor center users: a prole of Louisiana
travelers. Journal of Travel Research, 44(3), 348e351.
Duan, N., Manning, W. G., Morris, C. N., & Newhouse, J. P. (1983). A comparison of
alternative models for the demand for medical care. Journal of Business and
Economic Statistics, 1(2), 115e126.
Engel, J. F., Blackwell, R. D., & Miniard, P. W. (1995). Consumer behavior. Orlando, FL:
Dryden Press.
Fesenmaier, D. R. (1994). Traveler use of visitor information centers: implications
for development in Illinois. Journal of Travel Research, 33(1), 44e50.
Fesenmaier, D. R., & Vogt, C. A. (1993). Evaluating the economic impact of travel
information provided at Indiana welcome centers. Journal of Travel Research,
31(3), 33e39.
Fesenmaier, D. R., Vogt, C. A., & Stewart, W. P. (1993). Investigating the inuence of
welcome center information on travel behavior. Journal of Travel Research, 31(3),
47e52.
Fodness, D., & Murray, B. (1997). Tourist information search. Annals of Tourism
Research, 24(3), 503e523.
Frommer, D. (2013). The mobile tourist: how smartphones are changing travel.
Business Insider Intelligence. Retrieved January 5, 2014, from http://www.

businessinsider.com/the-travel-industry-and-the-mobile-boom-2013-8?utm_
sourcefeedburner&utm_mediumfeed&utm_campaignFeed%
3Abusinessinsider%28BusinessInsider%29.
Gitelson, R. J., & Perdue, R. R. (1987). Evaluating the role of state welcome centers in
disseminating travel related information in North Carolina. Journal of Travel
Research, 25(4), 15e19.
Greene, W. H. (2012). Econometric analysis. New York: Pearson.
Heckman, J. J. (1979). Sample selection bias as a specication error. Econometrica, 47,
153e161.
Ho, C., Lin, M., & Chen, H. (2012). Web users' behavioral patterns of tourism
information search: from online to ofine. Tourism Management, 33,
1468e1482.
Howard, D. R., & Gitelson, R. (1989). An analysis of the differences between state
welcome center users and nonusers: a prole of Oregon vacationers. Journal of
Travel Research, 27(4), 38e40.
Hsu, C.-L., & Lu, H.-P. (2004). Why do people play on-line games? An extended TAM
with social inuences and ow experience. Information & Management, 41(7),
853e868.
Huang, J., Basu, C., & Hsu, M. K. (2010). Exploring motivations of travel
knowledge sharing in social network sites: an empirical investigation of
U.S. college students. Journal of Hospitality Marketing & Management, 19(7),
717e734.
Jang, S., Ham, S., & Hong, G.-S. (2007). Food-away-from-home expenditure of senior
households in the United States: a double-hurdle approach. Journal of Hospitality & Tourism Research, 31(2), 147e167.
Kim, D.-Y., Lehto, X. Y., & Morrison, A. M. (2007). Gender differences in online travel
information search: implications for marketing communications on the
internet. Tourism Management, 28, 423e433.
Korea Association of Travel Agencies. (2012). Developing operational strategies for
Korean tourist information centers. Seoul, Korea: Korea Association of Travel
Agencies.
Korea Communication Committee. (2013). Smartphone user behaviors. Seoul, Korea:
Korea Communication Committee.
Korea Herald. (2013). S.Korea tops smartphone penetration rate in 2012. Retrieved
January
12,
2014,
from
http://www.koreaherald.com/view.php?
ud20130625000550.
Korea Tourism Organization. (2009). Report on tourist information centers in Korea.
Seoul, Korea: Korea Tourism Organization.
Korea Tourism Organization. (2011). Tourist information center opens in Myeongdong. Retrieved January 6, 2014, from http://english.visitkorea.or.kr/enu/FU/
FU_EN_15.jsp?cid1882253.
Lehto, X. Y., O'Leary, J. T., & Morrison, A. M. (2004). The effect of prior experience on
vacation behavior. Annals of Tourism Research, 31(4), 801e818.
Loomis, J. B., & Walsh, R. G. (1997). Recreation economic decisions: Comparing benets
and costs. State College, PA: Venture Publishing.
Maddala, G. S. (1999). Limited-dependent and qualitative variables in econometrics.
Cambridge, UK: Cambridge University Press.
McCann, R. (1999). Public-access interactive computers at state welcome centers in
the United States: 1991 and 1995 studies. Journal of Travel Research, 37(3),
249e255.
McDonald, J., & Moftt, R. (1980). The use of Tobit analysis. Review of Econometrics
and Statistics, 62, 318e321.
Mullahy, J. (1986). Specication and testing in some modied count models. Journal
of Econometrics, 33, 341e365.
Okazaki, S., & Hirose, M. (2009). Does gender affect media choice in travel information search? On the use of mobile Internet. Tourism Management, 30(6),
794e804.
Perdue, R. R. (1995). Traveler preference for information center attribute and service. Journal of Travel Research, 33(2), 2e7.
Pennington-Gray, L., & Vogt, C. A. (2003). Examining welcome center visitors' travel
and information behaviors: does location of centers or residency matter?
Journal of Travel Research, 41(3), 272e280.
Shen, C., & Wang, Y. (2014). Online to ofine business model. Retrieved October 14,
2014, from http://www.diva-portal.org/smash/get/diva2:750707/FULLTEXT01.
pdf.
Stewart, W. P., Lue, C., Fesenmaier, D. R., & Anderson, B. S. (1993). A comparison
between welcome center visitors and general highway auto travelers. Journal of
Travel Research, 31(3), 40e46.
Tierney, P. T. (1993). The inuence of state traveler information centers on tourist
length of stay and expenditures. Journal of Travel Research, 31(3), 28e32.
Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology
acceptance model: four longitudinal eld studies. Management Science, 46(2),
186e204.
Wallace, C. Y., Walton, J. R., & Crabtree, J. D. (2009). Best practices for providing
traveler information services to motorists at rest areas and welcome centers.
Kentucky Transportation Center Research Report. Retrieved January 6, 2014,
from http://uknowledge.uky.edu/ktc_researchreports/100/.
Wang, Y., Park, S., & Fesenmaier, D. R. (2012). The role of smartphones in mediating
the touristic experience. Journal of Travel Research, 51(4), 371e387.
Wicks, B., & Schuett, M. (1991). Examining the role of tourism promotion through
the use of brochures. Tourism Management, 11, 301e312.
Wong, C. U. I., & McKercher, B. (2011). Tourist information center staff as knowledge
brokers: the case of Macau. Annals of Tourism Research, 38(2), 481e498.
Xiang, Z., & Gretzel, U. (2010). Role of social media in online travel information
search. Tourism Management, 31(2), 179e188.

S.O. Lyu, J. Hwang / Tourism Management 48 (2015) 54e63

Seong Ok Lyu is an Assistant Professor in the Division of
Tourism at Dongseo University. He earned his doctoral
degree from Michigan State University. His research interests involve tourist decision-making process and destination marketing as well as economic valuation for
tourism and recreation resources.

63
Jinsoo Hwang is an Assistant Professor in Division of
Tourism at Dongseo University, Korea. He received his
Ph.D. degree in Hospitality Management from Kansas State
University. He specializes in hospitality and tourism marketing and research methods.