Annals of Tourism Research, Vol. 30, No. 2, pp. 323­341, 2003
 2003 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0160-7383/03/$30.00
www.elsevier.com/locate/atoures
doi:10.1016/S0160-7383(02)00060-9
IMPLICATIONS OF POTENTIAL
GREEN TOURISM DEVELOPMENT
Sung-kwon Hong
Jae-hyun Kim
Konkuk University, South Korea
Seong-il Kim
Seoul National University, South Korea
Abstract: Literature reviews suggest the marketing concept should be applied to green
tourism planning. Based on a survey of potential such tourists, this research identifies the
types of attributes and their corresponding levels that they seek. It also estimates the extent
of their influence upon intention to visit by market segment. This study is based on Yangpy-
eong County in South Korea prior to introducing green tourism. Although this is a case
study, other nations can apply the conjoint choice model to develop a suitable combination
of factors for attracting potential green tourists, and refer to the findings of this research
for guidance in operating their green tourism destinations. Keywords: South Korea, green
tourism, market segmentation, conjoint choice model.  2003 Elsevier Science Ltd. All
rights reserved.
Résumé: Les implications des possibilités de dévelopment du tourisme vert. Les examens
de la littérature sugge`rent que l'on doive appliquer le concept du marketing a` la planification
du tourisme vert. Basée sur une enque^te de tels touristes potentiels, cette recherche identifie
les types d'attributs at les niveaux correspondants que ces touristes cherchent. De plus, elle
estime l'importance de l'influence de ces attributs sur l'intention de visiter selon les différ-
ents segments du marché. Cette étude est basée sur la région de Yangpyeong, en Corée du
Sud, avant l'introduction du tourisme vert. Bien que cet article soit une étude de cas, d'autres
nations peuvent appliquer le mode`le de choix conjoint pour développer une combinaison
appropriée de facteurs pour attirer des touristes verts potentiels et se référer aux conclusions
de cette recherche pour des indications utiles pour la gestion de leurs destinations de tour-
isme vert. Mots-clés: Corée du Sud, tourisme vert, segmentation du marché, mode`le de choix
conjoint.  2003 Elsevier Science Ltd. All rights reserved.
INTRODUCTION
As South Korea's economic structure has transformed into a manu-
facturing-oriented economy since the 70s, it has faced the same rural
problems as many other developed countries: falling income levels,
decreasing and ageing populations, and an increasing necessity for pre-
Sung-kwon Hong is Professor in the Department of Horticultural Science, Konkuk Univer-
sity (Seoul 143-701, South Korea. Email <skhong@kkucc.konkuk.ac.kr>). His research inter-
ests are tourism choice behavior. Seong-il Kim, Professor in the Seoul National University
Department of Forest Resources, conducts research on alternative tourism development. Jae-
hyun Kim, Associate Professor in the Konkuk University Department of Forest Environment
Science studies forest resources management based on public participation.
323
324 GREEN TOURISM
serving rural traditions and environments. Under these circumstances,
the central government began to recognize the importance of green
tourism as one strategy to revitalize rural areas' economies in the 80s.
In 1984, it established 12 tourism farms as a pilot project and since
then, the project has expanded on an annual basis. In addition, it
established resort complexes in such areas as a way of implementing
the Rural Area's Tourism Income Source Development Project in 1989
and supported the Homestay Village Project in 1991.
There were some successful cases among rural tourism destinations
that were developed with the support of the central government. These
ventures contributed to the financial well being of the individual oper-
ators, and to some extent, activated the local economy. However,
except for these few successful outcomes, most of the projects failed
to realize their expected benefits (Korea Forest Service 1999).
Impressed with Japan's green tourism (GT) promotion policies and
their fruitful results, South Korean central administration started to
extend assistance to local governments and rural residents in the late
90s in order to introduce a version of the Japanese GT (a type of rural
tourism) concept into agricultural areas. Similar to South Korea, Japan
has a long tradition of rice-farming. Both countries share many cultural
characteristics and have faced similar rural problems due to a very
small ratio of farm land per household. For these reasons, the South
Korean government highly valued the potential success of GT. It also
expected that this tourism type would magnify the marketing effect by
highlighting the "green image" of rural areas. So far, the government
has adopted top-down policymaking and facility-oriented rural tourism
development. However, GT is a form of community-based tourism in
which rural residents participate voluntarily, aiming mainly at
expanding the income of rural households and the host community as
a whole by actively selling home-grown agricultural products. It pursues
environmentally friendly small-scale development and aims to make
urban citizens realize the values of rural areas and the importance of
their culture by providing them with opportunities to experience non-
urban life and culture. In doing so, it is also designed to help these
destinations recover their pride in rural life (Kim 1999).
The South Korean government has not only made changes in rural
tourism policies; in addition, a major social change may occur. The
government is working to introduce a five-day work-week system. Cur-
rently, citizens work an average of 47.9 hours per week. The govern-
ment is planning to pass this system into law in efforts to enhance the
quality of life by expanding leisure time and to create additional jobs
by shortening working hours. When the system is adopted officially,
leisure and tourism behavior is expected to be one of the first changes
(Min, Koh, Choi and Shin 2001). It is expected that average tourism
patterns will change from single day trips to trips including an over-
night stay and move from simple sightseeing to experience-oriented
trips.
Various tourism efforts should be made to capitalize on the changes
in GT policies and the reduced work week. However, establishment
and operation costs of facilities will be high since present rural tourism
325HONG, KIM AND KIM
destinations are outdated, offering recreational and accommodation
opportunities of the past. This impedes the goals of the new policies
by limiting the number of rural residents able to participate in GT
businesses, and thus hinders the ability of the rural economy to
strengthen and stabilize through the increase in secondary incomes
(Kim, Han and Lee 2001). This suggests that simply introducing tour-
ism in specific areas is not the real issue. The more critical issue is
whether or not introducing GT will increase the number of tourists
enough to ultimately contribute to a rise in incomes for rural house-
holds.
One way to address this issue is to apply marketing techniques. How-
ever, since potential tourists have different needs, they all cannot be
satisfied simultaneously. It is essential to segment them into target mar-
kets and match their needs with the skills and resources of GT oper-
ators. The latter should then offer attribute levels within destinations
that are expected by these potential target markets in order to attract
and increase the probability of repeat visitation. Although the litera-
ture suggests the necessity of applying marketing to GT (Sharpley
2002), few businesses have adopted these techniques (Reichel, Lowen-
gart and Milman 2000). Therefore, the present study examined a GT
destination which was to be developed in Yangpyeong County in order
to generate results which tourism operators can apply to their business
development and destination planning programs, particularly those
focusing on increasing the income of rural households. First, the
potential market was segmented based on constraints. Then, the types
of salient attributes and their corresponding levels affecting GT partici-
pation intention were identified, as well as the extent of their influ-
ence. These were calibrated using a conjoint choice model for each
segment in order to identify the most suitable target markets for the
tourism operators, and to specify which attributes and corresponding
levels the target markets seek.
GREEN TOURISM DEVELOPMENT
Being rooted in Europe, green tourism is often used interchangeably
with rural tourism in general (Yamazaki, Oyama and Ohshima 1997).
This study opts for the former, which has also been used in place of
rural tourism in Japan since the early 90s. With the launch of the World
Trade Organization in 1995, it became necessary for Japan to abandon
a number of policies, including diverse agricultural subsidies and tax
reduction which had supported the price of agricultural products.
Japan, which has an average farm size of 1.55 ha per household
(Ministry of Agriculture 2001), has attempted to adapt to changing
trends by focusing on stabilizing the rural economy through efforts to
improve the secondary income of rural households. One of the main
efforts has been to support GT. This policy was initiated by the central
government, defining green tourism as "staying-type leisure activities
to enjoy the nature and culture of the destination and interaction with
local residents in rural and mountainous areas rich in natural land-
scape" (The Ministry of Agriculture 1992). More specifically, the
326 GREEN TOURISM
government has tried to keep these natural resources and mountain-
ous areas environmentally sound in order to develop and maintain
traditional agriculture and forestry businesses in efforts to revitalize
local economies by promoting exchange with urban citizens and to
alleviate social problems in such areas, including decreased popu-
lations (Miazaki 1997).
Japan's GT is similar to rural tourism in Europe in that it is conduc-
ted within the natural environment, and offers tourists opportunities
to experience local culture and rural lifestyle via long stays. Here, resi-
dents involved in agriculture and forestry play important roles in man-
aging the environment. Unlike Europe, GT in Japan is more concen-
trated on the sustainable environment of rural areas, and is managed
by the residents, although the capital needed for businesses is shared
by the central and local governments as well as rural residents. The
key to revitalizing local economies is interaction with urban citizens,
in which buying and selling, such as the direct sale of agricultural pro-
ducts, is critical. While the sale of rural products is important, more
emphasis should be placed on human interactions. In this respect, pro-
grams providing urbanites opportunities to taste the host community's
unique home-made food and to enjoy the farming experiences of the
destination are important attractions, and a useful tool to promote
exchanges between rural and urban citizens. Japan's GT is a type of
rural tourism that utilizes both the rural culture, which has been culti-
vated under the long tradition of agriculture and forestry, as well as
the nature of rural areas, such as forests and secondary nature in the
form of farm lands as attractions (Kim 1999; Kim et al 2001; Miazaki
1997).
The existing research conducted in rural tourism falls into three
categories: the introduction and development of rural tourism by
countries (Busby and Rendle 2000; Clarke, Denman, Hickman and Slo-
vak 2001; Unwin 1996), investigations into the operational character-
istics of each national rural tourism using descriptive methodologies,
and studies of the sociodemographic characteristics of such tourists
and variables affecting their participation.
Research into operational characteristics has been conducted in
many countries. Oppermann (1996) investigated the status of rural
tourism in southern Germany in terms of the number of beds, room
occupancy rates, seasonal demand, and the contribution of tourism-
related incomes toward total net income based on the type of accom-
modation. Getz and Carlsen (2000) surveyed Australia's family owned
and operated businesses in rural tourism to identify their goals, nature
of the operation, and satisfaction and problems which stemmed from
their businesses. Regarding vacation farms, Weaver and Fennell (1997)
surveyed Canadian operators for their operational characteristics such
as average vacation farm size, average number of beds, number of full-
time employees, reasons for establishing their vacation farm, and busi-
ness barriers and opportunities faced by the operators. Fleischer and
Pizam (1997) analyzed Israel's national census to characterize the
development of bed and breakfast operations. They examined
elements such as numbers of businesses, beds, and employees. In
327HONG, KIM AND KIM
addition, Reichel, Lowengart and Milman (2000) reviewed manage-
ment and marketing strategies to enhance the service by measuring
the gap between the quality of service expected by rural tourists and
the actual quality of service.
A wide range of research into the sociodemographic characteristics
of rural tourists has also been conducted in numerous countries
(Fleischer and Pizam 1997; Garcia-Ramon, Canoves and Valdovinos
1995; Koth and Norman 1989; Pevetz 1982). By compiling the existing
research, rural tourists almost everywhere are generally in their 30s or
40s, highly educated, higher income earners accompanied by their
young children. However, rural tourists of Hokkaido, Japan are differ-
ent from those of Europe in that 75% of them are in their 20s and 30
(Department of Agriculture 1992).
Salient attributes affecting rural and green tourism include accom-
modations, seasonality, experiences, and events. The accommodation
types are diverse, ranging from campgrounds, self-catering facilities,
and bed and breakfast, to fully catered establishments, including hotels
and motels in rural communities. In Germany, family groups prefer
to stay on farms, with an average length of stay of 16 days (Oppermann
1996). Average length of stay in Spain ranges from 4.5 to 11.4 days
depending on the region (Garcia-Ramon et al 1995). To the contrary,
a survey of Japanese preferences for GT shows they favor farm inns
which provide two meals of traditional food (53%), accommodations
providing hotel-class services (21%), cottages (16%) and bed and
breakfast (9%), and average length of stay is four days, three nights
(Ministry of Agriculture 1992). From this, it is assumed that the type of
accommodation preferred and the length of stay vary among countries
because of social and cultural differences.
Seasonality was significant in most of the national surveys. The room
occupancy rate in Germany is 93­94% between July and August, and
only 13­15% between November and January (Oppermann 1996).
Similarly, in Spain the summer months are the most popular (Garcia-
Ramon et al 1995), and summer time users account for 98.7% of the
total users in Hokkaido, Japan (Department of Agriculture 1992). The
characteristics which most clearly differentiate GT from mass tourism
are experiences and events. Japan provides experiences and seasonal
events only rural areas can offer, including picking grapes or mush-
rooms, catching insects, observing the stars, harvesting potatoes, catch-
ing fish or other water-based activities, and planting sweet potatoes,
among others (Korea Forest Service 1999; Yoo, Lee, Peon, Joe and
Yoo 1998).
As to the present study and its location in South Korea, Yangpyeong
is a representative county adjacent to metropolitan cities. Its pro-
portion of the population engaged in farming has consistently
declined from 45.2% in 1990 to 35.7% in 2000. In the face of consist-
ent pressure for development due to its geographical proximity to
Seoul, which has a population of approximately 10 million (Figure 1),
the county has been making continued efforts to become South
Korea's representing GT destination with the initiative of its principal
policy aimed at maintaining well-preserved nature and culture. It pro-
328 GREEN TOURISM
Figure 1. Location of Study Area
motes growing organic produce by minimizing the use of agrochem-
icals and fertilizers in designated areas, and continuously educates far-
mers in organic farming by establishing and supporting training
courses for farmers at the so-called College of Environmental Agricul-
ture. In addition to this, the county regularly holds the Firefly Festival
and Scarecrow Festival with environmental themes. When the first
"Digging Out and Buying Doduk (Cdomopsis lanceolata T.) in the Moun-
tains" event was held in 2000 targeting Seoul residents, approximately
35,000 participated in the event. Residents of the county also volun-
tarily organized the Myungdal-Ri Green Tourism Implementation
Committee, in which the locals develop programs to enhance their
own business competency and visit other successful GT destinations
for benchmarking.
A Conjoint Choice Model Approach
In marketing, a product is defined as "a bundle of want-satisfying
attributes" (Howard and Crompton 1980) and the success or failure
of a product depends on consumers' need satisfaction. Under this
definition, in order to manage GT destinations successfully, operators
329HONG, KIM AND KIM
should provide tourists with more utility than competing alternatives
by offering them the type of attributes they expect. By doing so, the
probability of the destination to being chosen (purchased, visited) is
increased because tourists select the alternative that maximizes their
utility based upon their evaluation of available attributes and their cor-
responding levels. Each tourist is likely to give up one benefit to gain
another. For instance, one is likely to abandon a low (level) price
(attribute) to gain a nature-learning (level) experience (attribute) pro-
vided in green tourism.
Existing research approaches to choice behaviors can be broadly cat-
egorized as either conjoint models (which predict behavior by con-
sidering the preference of respondents for hypothetical alternatives)
or logit models (which analyze real choice data). Both determine pref-
erence or choice by analyzing the utility of attributes related to choice
with an algebraic rule. The conjoint model makes it possible to predict
tourists' preferences for new situations, for instance a destination to
be developed, because it can simulate these new arrangements as hypo-
thetical alternatives. In addition, correlation between attributes can be
removed since this model constructs hypothetical alternatives by a frac-
tional factorial design. However, it has to make ad hoc assumptions of
error distribution necessary to translate its predictions into choices, for
it is a preference, not a choice model. As a consequence, the credibility
of results is undermined. The logit model is the representative of the
revealed preference method and can predict the choice behavior with
high external validity because it uses real choice data for analysis. How-
ever, it controls neither the correlation among attributes used, nor
eliminates the influence of external attributes not used in the study,
nor predicts changes in demand in new situations (Hong 1995; Louvi-
ere and Hensher 1982; Louviere and Timmermans 1990; Louviere and
Woodworth 1983).
In order to minimize the drawbacks of the two models while combin-
ing their advantages, Louviere and Woodworth (1983) suggest a con-
joint choice model which can make direct predictions on choices in
the form of the logit model by calculating the part-worth of attribute
levels obtained from choice-type rather than preference data. It also
eliminates problems arising from using real choice data since it offers
hypothetical alternatives like the conjoint model and resolves ad hoc
assumptions of error distribution used to translate its predictions into
choices by meeting the requirements necessary to calibrate the logit
model.
p(a Ci)=
exp(Va)
j Ci
exp(Vj)
(1)
where Ci=ith choice set, p(a|Ci)=probability of choosing alternative a
in Ci (i=1, 2, ..., I), and Va=preference (utility) of alternative a.
330 GREEN TOURISM
Determination of Variables and Hypothetical Alternatives
There is limited domestic research on salient attributes and their
levels affecting GT participation. Hence, they were determined by
reviewing and compiling previous research findings, such as the type
of accommodation, the type of expenses, events, type of dining, daily
costs, visiting season, and amenities of the place of accommodation
(Department of Agriculture 1992; Fleischer and Pizam 1997; Garcia-
Ramon et al 1995; Kang 1999; Korea Forest Service 1999; Luzar,
Diagne, Gan and Henning 1998; Ministry of Agriculture 1992; Opper-
mann 1996; Yamazaki, Oyama and Ohshima 1997; Yoo et al 1998).
According to the Korea National Tourism Corporation (2000), the
average length of stay for Seoul residents in terms of mass tourism is
two days/one night (41.7%) and three days/two nights (34.4%). Thus,
three days/two nights or more, two days/one night, and day trips were
chosen as the suitable lengths of the stay.
It is well documented that constraints exert direct and indirect
influences on both leisure preference and participation (Crawford,
Jackson and Godbey 1991; Jackson and Henderson 1995). Neverthe-
less, little research on constraints has been conducted which can be
directly applicable to GT. Therefore, the authors utilized a 17-item
Likert-type scale by referring mainly to Crompton and Lamb's (1986)
constraints taxonomy. These taxonomy variables were used as a means
to segment the market and to examine perceptions of constraints for
participation in GT. External physical constraints which cannot be con-
trolled by operators or potential tourists were excluded, and only items
which can be applicable to GT were selectively chosen. Further, some
additional items judged necessary for this research were selected from
domestic constraint-related studies and added to the constraints
chosen (Hong 1998; Um 1994) (Table 1).
The hypothetical alternatives for taking an extended GT trip to
Yangpyeong, either three days/two nights or more (3D/2N-M
alternative) or two days/one night (2D/1N), were constructed using
all attributes explained earlier. Thus, 32 hypothetical alternatives were
generated for both the 3D/2N-M and 2D/1N alternatives, by using
1/96 fractional factorial design. In order to construct the hypothetical
alternative for a day trip (1D alternative), type of experiences, events,
daily costs, and visiting season were used. Thus, 32 hypothetical alterna-
tives were generated using one-quarter fractional factorial design. Each
of the three alternatives, 3D/2N-M, 2D/1N and 1D, are comprised of
32 hypothetical alternatives. The choice sets were generated by ran-
domly selecting one alternative out of each of the 3D/2N-M, 2D/1N
and 1D, and adding the base alternative "would not take a green tour-
ism trip to Yangpyeong County". Thus, 32 choice sets consisting of 4
hypothetical alternatives were made.
Because each choice set is composed of three alternatives con-
structed by changing the levels by attribute and one base alternative,
it is demanding for one respondent to evaluate all the 32 sets. So, to
make the evaluation easier, each respondent was provided with 8
choice sets randomly chosen from the 32 sets for this evaluation. Thus,
331HONG, KIM AND KIM
Table 1. Rotated Factor Patterns of Constraints
Constraints Factor 1 Factor 2 Factor 3 Factor 4
A negative image of tourism .6880 .1093 .1917 .1486
itself
Physically difficult in taking a .6147 .0154 .0057 .1176
tourism trip
No interest in tourism .5913 .0525 .0660 .0165
A lack of awareness of the .5196 .0338 .0522 .1303
benefits which tourism offers
A lack of someone with whom .4783 .1396 .0608 .0962
to take a tourism trip
A lack of compatibility with .3942 .1108 .0512 .0526
other tourists
A lack of time .1189 .5530 .0414 .0691
Inconvenient location .0242 .4667 .0044 .0945
A lack of information .0365 .4653 .1757 .0765
Work commitments --.0804 .4241 .0184 .0714
Unable to resolve to take a .2934 .4083 .0455 .0829
tourism trip
Good quality service personnel .0587 .0683 .5367 .1039
in tourism sites
Proximity to other tourist .0242 .0552 .4913 .2219
attractions
Adequate in diversity of .0251 .0928 .4474 .1512
amenities such as resting
lounges, restaurants and wash
rooms
Good quality of amenities such .2040 .0515 .4452 .1221
as resting lounges, restaurants
and wash rooms
Too crowded .0797 .1384 .1014 .5887
Unable to afford taking a .0836 .2710 .0563 .3533
tourism trip
Variance Explaineda
20.87% 13.74% 9.54% 6.59%
a
Percentage of variance explained by each factor ignoring the other factors.
at least 4 respondents were required to evaluate all choice sets. The
sample size was set as 608, a multiple of 4. A polling agency conducted
sampling and data collection targeting Seoul citizens. Respondents
were selected by cluster sampling with probabilities proportional to
size. Age was limited to 19-59. In the first stage, 58 Tongs (South
Korea's smallest administrative unit, equivalent to a census block) were
selected. In the second stage, 8­12 residents were selected in each of
the Tongs chosen. Then, face-to-face interviews were conducted on a
total of 608 Seoul residents. The data was collected from October 30
to November 16, 2000.
332 GREEN TOURISM
Market Segmentation
One of the purposes of this research is to identify the extent of the
influence of attribute levels upon the intent to participate for each
tourism segment. To this end, both factor and cluster analyses were
used simultaneously. The former is used for condensing the infor-
mation contained in a number of original variables (i.e., constraints)
into a smaller set of new, composite dimensions (i.e., factors) with a
minimum loss of information. The latter is a method of grouping
objects (i.e., respondents) based on predetermined clustering criteria
(i.e., constraint factor score) so that the objects in one cluster (such
as market segment) are very similar to each other. The within-cluster
homogeneity and between-cluster heterogeneity are shown high in the
generated clusters of objects.
To segment the market based on constraints, a maximum likelihood
factor analysis with an oblique rotation was performed to determine
the factor structure of the constraints scale because of high correlation
between factors. Based on the overall Kaiser's MSA of 0.77, it was
acceptable to perform this analysis. Considering both eigenvalues
greater than one and the scree plot, four factors were judged appropri-
ate (Table 1). With the sample size of 608, extracted ones were inter-
preted on the basis of variables with factor loadings of 0.35 or more
(Hair, Anderson, Tatham and Black 1995). The first was titled "general
blocking factor" since it was comprised of variables which make partici-
pation in tourism itself hard. The second was titled "individual block-
ing factor" since it was comprised of problems arising from the lack of
time, information and accessibility of individuals. The third was titled
"destination attraction factor" because it mainly consisted of variables
relating to services and attractions of mass tourism sites. The last was
titled "destination dissatisfaction factor" because it consisted of vari-
ables pertaining to crowding and the high prices of mass tourism sites
in South Korea.
A cluster analysis was carried out by using the four extracted factor
scores as independent variables and the three segmented markets were
identified based on pseudo-F statistics. Their characteristics were inter-
preted by the mean factor scores of each cluster (Table 2). Market
segment 1 (SEG 1) accounted for 33.6% of the total respondents, and
was predominantly influenced by factors 2, 3, and 4. Respondents in
this segment were labeled as an "anti-mass tourism group" who tend
to be short of time and have strong negative attitudes toward mass
tourism. Segment 2 (SEG 2) was influenced by all factors evenly, and
was labeled as a "no interest in tourism group". These respondents did
not have a strong desire for tourism and appeared to have little time
for it. They are expected to prefer mass tourism destinations when
they are given an opportunity to take a trip. Segment 3 (SEG 3) was
comprised of respondents with both the desire and time for tourism
and appear to not be dissatisfied with mass tourism. Since this group
was favorable toward tourism, they were labeled as the "variety seeking
group", which might regard GT as a new alternative.
The Chi-square test of independence showed that SEG 1 and SEG
333HONG, KIM AND KIM
Table 2. Mean Factor Scores of Segmented Groups
Segment Factor 1 Factor 2 Factor 3 Factor 4 Number of
Respondents
(%)
1. Anti-Mass .1284 .3624 .4607 .5887 204 (33.6%)
Tourism
Group
2. No Interest .8918 .4798 .5893 .3282 187 (30.8%)
in Tourism
Group
3. Variety .6478 .7541 .0747 .2706 217 (35.7%)
Seeking Group
3 had distinct sociodemographic characteristics. SEG 3 had higher pro-
portions of males, white-collar individuals, college students, and
singles, and a lower proportion of business owners and housewives. In
addition, SEG 3 had higher proportions of those in their 20s and lower
proportions of those in their 40s. When comparing the characteristics
of these two segments, SEG 3 appears to be composed of those who
have enough free time and a positive attitude toward exploring new
things, whereas SEG 1 tends to be greatly influenced by work and/or
family commitment constraints.
Calibration of the Conjoint Choice Model by Segment
The extent of influence of attribute levels was calibrated with the
conjoint choice model using the maximum likelihood method. For
this study, the results of SEG 1 and SEG 3 were compared and inter-
preted, while SEG 2 was excluded because it was seen as a "no interest
in tourism group" (Tables 3 and 4). The attribute levels were defined
using dummy coding. The calculated McFadden 2
of SEG 1 and SEG
3 were .1995 and .1966, respectively, which are acceptable explanatory
levels for the general class of pseudo 2
measures used in qualitative
choice analyses (Greene 1995; Maddala 1991).
Constants. Most of the constants were negatively calibrated because
GT is a rather new term for the public, and because of the influence
of financial and psychological risks of visiting a rural destination which
may not match one's expectations. The values of the constants indicate
the characteristics of each segment well. Although SEG 1 was "anti-
mass tourism group", it does not imply that the alternative they seek
is GT. Since this group is influenced strongly by "individual blocking
factors" (such as a lack of time), it is expected that this group will
choose destinations very cautiously when they have an opportunity to
take a tour. Therefore, the constant for 3D/2N-M alternative was cali-
brated as the largest negative (Table 3). One of the major reasons for
this result is that South Korea still has a six-day work-week system. Since
334 GREEN TOURISM
Table 3. Model Estimation Results: Segment 1(Anti-Mass Tourism Group)a
Attributes Levels 3D/2N-Mb
2D/1Nc
1Dd
Parameter P[|Z| z] Parameter P[|Z| z] Parameter P[|Z| z]
Estimate Estimate Estimate
Type of Log cabin .5701 .0065 .5674 .0036
Accommodation CONDOe
.3476
.0952 .2208 .2817
Farm inn .3268
.1042 .2800 .1806
Motel .0000 .0000
Type of Foragingf
.4257 .0228 .4246 .0275 .5140 .0266
Experience Nature .1305 .5129 .1944 .3490 .1030 .6516
learningg
Recreational -.3844
.0796 .2669 .1937 -.0476 .8415
typeh
Farmingi
.0000 .0000 .0000
Events Provided .3887 .0086 .2159 .1380 .2224 .1762
Not provided .0000 .0000 .0000
Dining Buy one mealj
.5136 .0043 .0255 .8841
Buy two meals .4831 .0166 .0976 .5973
or more
Prepare all .0000 .0000
meals yourself
Daily Costk
$25/day 2.7827 .0000 2.4473 .0000 2.7671 .0000
$50/day 2.1115 .0000 1.5286 .0000 1.7631 .0000
$75/day 1.3259 .0000 .8901 .0001 .7296 .0158
$100/day .0000 .0000 .0000
Visiting Season Spring -.0554 .7953 .4424 .0337 .1534 .5027
Summer .0337 .8668 .2175 .3083 .0178 .9412
Fall .2240 .2491 .1418 .5076 .2217 .3529
Winter .0000 .0000 .0000
Amenities at the Without .0455 .7518 -.0489 .7438
Place of telephone, TV
Accommodation With .0000 .0000
telephone, TV
Constant -1.9152 .0000 -1.2188 .0006 -1.5214 .0000
a
The figures shown in boldface and  indicate statistically significant confidence levels of 95% and 90%, respectively.
b
"Three days/two nights or more" alternative.
c
"Two days/one night" alternative.
d
"One-day return trip" alternative.
e
Timesharing condominium apartments.
f
Possible to experience harvesting mushrooms and wild vegetables such as Doduk.
g
Possible to observe wild birds, wild flowers and stars with a guide.
h
Possible to enjoy many water-based activities or sleighing.
i
Possible to experience farming such as taking care of paddy fields.
j
Buy meal at place of accommodation.
k
Based on 4 persons.
SEG 3 is a "variety seeking group" with positive attitudes toward chal-
lenging new ventures, the constants for the three alternatives were cali-
brated higher than those of SEG 1. This segment was supportive of
GT for the 2D/1N alternative so much so that there was no statistical
difference with the base alternative (Table 4). The reason why the con-
stant for 1D alternative was calibrated as more negative than that for
335HONG, KIM AND KIM
Table 4. Model Estimation Results: Segment 3 (Variety Seeking Group)a
Attributes Levelsb
3D/2N-Mc
2D/1Nd
1De
Parameter P[|Z| z] Parameter P[|Z| z] Parameter P[|Z| z]
Estimate Estimate Estimate
Type of Log cabin .5904 .0030 .4674 .0105
Accommodation CONDO .8043 .0000 .3016
.1039
Farm inn .2389 .2185 .2953 .1185
Motel .0000 .0000
Type of Foraging .2148 .2413 .1573 .3805 .2329 .3030
Experience Nature learning .4439 .0203 .0148 .9378 .1460 .5226
Recreational type .3759 .0471 .1966 .3044 .2265 .3389
Farming .0000 .0000 .0000
Events Provided .1917 .1554 .0348 .7915 .4793 .0028
Not provided .0000 .0000 .0000
Dining Buy one meal .2686
.0985 .0431 .7832
Buy two meals or .1413 .4557 .1606 .3585
more
Prepare all meals .0000 .0000
yourself
Daily Costs $25/day 1.8954 .0000 2.1150 .0000 2.3633 .0000
$50/day 1.2624 .0000 1.4468 .0000 1.2049 .0000
$75/day .6066 .0032 .8345 .0000 .7539 .0081
$100/day .0000 .0000 .0000
Visiting Season Spring .2690 .1513 -.1966 .3029 .3234 .1560
Summer .0296 .8756 .1807 .3560 .0441 .8523
Fall .0857 .6493 -.4220 .0299 .2356 .3173
Winter .0000 .0000 .0000
Amenities at the Without -.0434 .7443 -.0564 .6861
Place of telephone, TV
Accommodation With telephone, .0000 .0000
TV
Constant -1.0801 0.0002 -.3153 0.2900 -1.4657 .0000
a
The figures shown in boldface and  indicate statistically significant confidence levels of 95% and 90% , respect-
ively.
b
"Level descriptions as per the Table 3.
c
"Three days/two nights or more" alternative.
d
"Two days/one night" alternative.
e
"One-day return trip" alternative.
2D/1N alternative in both segments is that respondents felt a day trip
was not enough time for GT and 2D/1N is the longest length they feel
they could spare for tourism in a week.
Type of Accommodation. SEG 1, which seems to have lost the most
interest in mass tourism, prefer different types of accommodations
from the popular types, such as motels for a 3D/2N-M alternative. They
prefer log cabins, timesharing condominium apartments (CONDO),
and farm inns, respectively. In particular, it is notable that the prefer-
336 GREEN TOURISM
ence for farm inns and motels was statistically different in SEG 1 while
no statistical difference was shown in SEG 3. Perhaps it is because the
latter was composed of a larger proportion of young respondents.
Thus, they preferred CONDO and log cabins to motels since they value
convenience and privacy. However, there was no significant difference
between motels and farm inns which are considered as the most desir-
able accommodation type from the perspective of tourism operators.
The 2D/1N alternative appears to be less sensitive to the type of
accommodation than the 3D/2N-M option, lowering the preference
gap for type of accommodation. However, people in both groups pre-
ferred log cabins the most. This is perhaps because this type of accom-
modation is seen as acceptable for a single overnight stay since it offers
a unique experience.
Experiences, Events and Meals. The attribute best reflecting the
characteristics of GT is the type of experiences. Since the 3D/2N-M
alternative offers sufficient time for tours, this was statistically signifi-
cant at two levels for SEG 1 and SEG 3. For both the 2D/1N and the
1D alternatives, however, there was no statistical significance in the
level for experience in SEG 3, which is probably due to time limi-
tations. Additionally, only one type of experience, "foraging experi-
ences," was significant for SEG 1.
SEG 3, which is composed of relatively young respondents, expressed
the strongest preference toward "nature learning experiences", fol-
lowed by "recreational type experiences". SEG 1, which consists of a
higher proportion of housewives and those in their forties, had no
statistical difference in preferences between "nature learning experi-
ences" and "farming experiences". Perhaps this is because people in
SEG 1 tend to regard having experiences in nature as somewhat bur-
densome, and view "nature learning experiences" as better suited for
younger generations. However, "foraging experiences" were statistically
significant only in SEG 1 revealing that these individuals value pol-
lutant-free produce and care about the health of family members. This
interpretation is backed by the fact that "foraging experiences" was
preferred to "farming experience" for both the 2D/1N and the 1D
alternatives. The same context holds good for SEG 1 preferring "farm-
ing experiences" to "recreational type experiences."
In the case of events, only the 3D/2N-M alternative was significant
for SEG 1, and the 1D alternative was significant for SEG 3. This sug-
gests that SEG 1 wants tourism offering a wide range of experiences
that they can enjoy with enough time, as opposed to mass tourism for
which schedules are tight. SEG 3 seeks a variety of experiences and is
willing to enjoy events for one day.
The 2D/1N alternative turned out to be less sensitive to the type of
dining, since it is a relatively short trip. However, for the 3D/2N-M
alternative, SEG 1 preferred buying meals as opposed to preparing
them. Perhaps this is because they tend to enjoy tourism with enough
time free from preparing meals, which is the most tedious household
chore, given the demographic characteristics of this group. Buying one
meal was the only level which showed significant preference for SEG
337HONG, KIM AND KIM
3. Since this group is most likely to take a tourism trip with friends, it
seems easy for them to prepare meals themselves outdoors, and it may
be difficult to buy two meals a day since they prefer "nature learning
experiences" which are usually conducted some distance from accom-
modations.
Other Salient Attributes. Daily costs was the largest coefficient in the
two segments, as expected. This reflects that costs are the most influ-
ential attribute regardless of market segment. Unexpected results were
found regarding visiting season and provision of amenities at the place
of accommodation. SEG 3 responded that winter is more appropriate
for GT than fall in the case of the 2D/1N alternative. It appears that
this group expects GT to offer more unique experiences in winter.
Both groups showed no significant difference in preference between
presence and absence of amenities such as telephones and TV. It can
be interpreted that they do not care about these amenities.
CONCLUSION
Since most existing research focuses on outlining rural tourism's
status by country, or describing the characteristics of the operators and
users, research has failed to provide concrete implementation
methods. For GT to become successful, the market should be seg-
mented, then attributes must be provided at a level that satisfies these
potential tourists' wants. Once the attributes and corresponding levels
are identified by segment, the extent of their influence on intention
to take a trip should be tested prior to implementation in effort to
prevent economic loss from failed ventures. By using both factor and
cluster analyses, this research segmented the market by constraints
which are representative of antecedent variables of leisure preference
and participation targeting potential green tourists in South Korea.
Then, the conjoint choice model was used to estimate to what extent
the selected salient attributes and their levels affect the intention to
take a trip.
This research adopted the conjoint choice model for estimating the
coefficients of attribute levels for two reasons: one, when business start-
ups or existing operators intended to change existing attractions or
existing types of services, demand in this hypothetical situation would
have to be estimated by using intention as a dependent variable; and,
two, it is well known that intention is a direct predictor of behaviors
(Fishbein and Ajzen 1975). Even though demand in the hypothetical
situation can be estimated by both the conjoint choice model and the
conjoint model, ad hoc assumptions of error distribution should be
carried out for the conjoint model. Due to this drawback of the con-
joint model, this research used the conjoint choice model.
On average, respondents in the two segments preferred "would not
take a green tourism trip to Yangpyeong County" over the choice of
"would take a trip." Even though rural tourism has been in operation
in South Korea for more than 20 years, it has not been successful. In
1999, only 1% of all tourists chose farm land and orchards as desti-
338 GREEN TOURISM
nations (Korea National Tourism Corporation 2000). Given this, it is
reasonable that more respondents indicated they would not take a GT
trip. However, the number of respondents who chose the 2D/1N alter-
native was larger than those who chose the 3D/2N-M or 1D alterna-
tives. This suggests it is highly possible for GT to succeed in South
Korea. Perhaps this result was because the longest holiday people can
afford to take, except for vacations and summer leave, is a 2D/1N trip,
due to the present six-day work-week system. If a five-day work-week
system is adopted, GT will stand a better chance of contributing to the
revival of the rural economies in South Korea.
Respondents preferred log cabins and CONDO to motels. Farm
inns, which are the most desirable accommodation type from the per-
spective of operators, were significant but its preference degree was
low in SEG 1. Respondents appear to be already well aware of the
convenience of log cabins and viewed them as complimenting GT,
since they are a typical type of accommodation in recreational forests.
Due to their convenience, CONDO were preferred by SEG 3, which
consisted of a high proportion of college students and singles. Given
the fact that one of the major revenue sources of GT is the provision
of accommodations, this research suggests that it is desirable to
develop log cabin accommodations, which operators are able to pro-
vide and tourists prefer.
Types of dining were significant only for the 3D/2N-M alternative.
Particularly, SEG 1, the "anti-mass tourism group", considered buying
meals important. This can be interpreted as an anticipation of the
opportunity to taste the unique food of the area, as opposed to mass
tourism, which offers food choices they can find anywhere. Since SEG
3 preferred "nature learning experiences" the most, it may be difficult
for them to buy two meals considering their tourism schedule.
Although bed and breakfast are not in operation in South Korea, the
development of bed and dinner should be considered, given the
characteristics of SEG 3.
The typical attributes that differentiate GT from mass tourism are
the types of experiences and events. SEG 1 preferred "foraging experi-
ences", while SEG 3 preferred "nature learning experiences" and "rec-
reational type experiences." This reflects the fact that each segment
has different needs. Thus, programs which satisfy the needs of target
markets should be provided in order to attract new tourists and
encourage repeat visitation. This research used 4 levels of experiences
and 2 levels of events; however, more levels could have been used in
order to estimate the influence of more diverse types of experiences
and specific types of events on intention to visit. Since it is too
demanding for respondents to evaluate a large number of attributes
and/or attribute levels, follow-up studies could use a pairwise conjoint
approach in order to address this problem (Wang, Oppewal and Tim-
mermans 2000).
Tourism operators of all types have to deal with seasonality in order
to enhance profitability. Although there is no existing research on the
effect of seasonality on GT in South Korea, the rate of accommodating
tourists in mass tourism is 14.4% in July and 27.8% in August, as com-
339HONG, KIM AND KIM
pared to only 6.1% in January and 6.0% in February (Korea National
Tourism Corporation 2000). This is because the most favored holiday
season is summer, which overlaps with students' summer vacation.
Contrary to this, there was no significant preference of summer over
winter for taking a GT trip. This research indicated that for the 2N/1D
alternative, SEG 3 preferred winter over fall for taking a GT trip. It
can be inferred that respondents expect new seasonal experiences
from this type of tourism. More in-depth research needs to be conduc-
ted in this area.
This study used constraints as market segment variables. In addition
to these variables, sensation-seeking should be considered as an alter-
native market segment variable to constraints in the follow-up studies,
since it is highly possible for novelty-seeking groups like SEG 3 to par-
ticipate in GT. Sensation seekers are defined as "novelty-seeking people
who are searching for new, diverse or complex ventures or experiences
and are ready to take risks if they can have the experiences they expect"
(Breen and Zuckerman 1999). Research has been conducted in the
leisure field to test the relationship between the personality construct
of "sensation-seeking" and attitudes to several characteristics of
national parks (Galloway and Lopez 1999).
Although many findings from this case study may not be directly
applicable to other countries with different sociocultural character-
istics, they are relevant to GT research in that they may be used as a
basic model for planning and developing a rural destination adjacent
to a large city. In addition, this research is meaningful for adopting
marketing techniques for GT and suggests a methodology to identify
specific attributes and their corresponding levels for successful GT
businesses. Although existing studies emphasized the importance of
marketing techniques, most of them simply demonstrated the sociode-
mographic characteristics of tourists. This research follows the con-
secutive steps of market segmentation, target market selection, and
identification of the preferences of respondents for various hypotheti-
cal attribute level combinations. In particular, even though it is obvious
that GT operators are most interested in what types of attribute levels
to adopt to make their business a success, no research studies were
found to suggest a specific methodology to answer this question regard-
ing GT. Recently, the conjoint choice model has been widely used for
several areas investigating choice behaviors, due to its proven reliability
and validity. This study demonstrates the model is also applicable to
GT research. A
Acknowledgements--This paper was supported by Konkuk University.
REFERENCES
Breen, R., and M. Zuckerman
1999 "Chasing" in Gambling Behavior: Personality and Cognitive Determi-
nants. Personality and Individual Differences 26:1097­1111.
Busby, G., and S. Rendle
2000 The Transition from Tourism on Farms to Farm Tourism. Tourism Man-
agement 21:635­642.
340 GREEN TOURISM
Clarke, J., R. Denman, G. Hickman, and J. Slovak
2001 Rural Tourism in Roznava Okres: A Slovak Case Study. Tourism Manage-
ment 22:193­202.
Crawford, D., E. Jackson, and G. Godbey
1991 A Hierarchical Model of Leisure Constraints. Leisure Sciences 13:309­
320.
Crompton, J., and C. Lamb
1986 Marketing Government and Social Service. New York: Wiley.
Department of Agriculture
1992 The Direction of Needs of Urban Residents for Farm Inns in the Pursuit
of Making Rural Areas Better Places. Sapporo City: Hokkaido Government.
Fishbein, M., and I. Ajzen
1975 Belief, Attitude, Intention and Behavior: An Introduction to Theory and
Research. Reading: Addision-Wesley.
Fleischer, A., and A. Pizam
1997 Rural Tourism in Israel. Tourism Management 18:367­372.
Galloway, G., and K. Lopez
1999 Sensation Seeking and Attitudes to Aspects of National Parks: A Prelimi-
nary Empirical Investigation. Tourism Management 20:665­671.
Garcia-Ramon, M., G. Canoves, and N. Valdovinos
1995 Farm Tourism, Gender and the Environment in Spain. Annals of Tour-
ism Research 22:267­282.
Getz, D., and J. Carlsen
2000 Characteristics and Goals of Family and Owner-Operated Businesses in
the Rural Tourism and Hospitality Sectors. Tourism Management 21:547­560.
Greene, W.
1995 LIMDEP User's Manual and Reference Guide, Version 7.0. Bellport:
Econometric Software.
Hair, J., R. Anderson, R. Tatham, and W. Black
1995 Multivariate Data Analysis. Englewood Cliffs: Prentice Hall.
Hong, S.
1995 Application of Hybrid Conjoint Analysis to Improve Competitive Power
of Theme Parks in Seoul and Its Suburbs. Journal of Korean Institute of Land-
scape Architecture 23:1­16.
1998 An Exploratory Study on the Introduction of Loyalty to Segmentation of
Theme Park Users. Journal of Korean Institute of Landscape Architecture
26:1­11.
Howard, D., and J. Crompton
1980 Financing. Managing and Marketing Recreation and Park Resources.
Dubuque: Brown.
Jackson, E., and K. Henderson
1995 Gender Based Analysis of Leisure Constraints. Leisure Sciences 17:31­51.
Kang, M.
1999 Ecotourists' Travel Motivations and Attitudes in Korea: Scales Develop-
ment and Comparative Analysis on Tourists Group. PhD dissertation in forest
resources, Seoul National University.
Kim, B.
1999 A Study on Green Tourism in Japan. Journal of Korean Institute of Forest
Recreation 3:1­13.
Kim, B., S. Han, and J. Lee
2001 Comparative study on green tourism between Korea and Japan. Korean
Journal of Tourism Research 16:83­103.
Korea Forest Service
1999 Green Tourism and Activation of Mountain Villages. Seoul: Korea For-
est Service.
Korea National Tourism Corporation
2000 1999 National Survey on Traveling. Seoul: Korea National Tourism Cor-
poration.
Koth, B., and W. Norman
1989 The Minnesota Bed and Breakfast Market: A Guest Profile. St. Paul: The
University of Minnesota Tourism Center.
341HONG, KIM AND KIM
Louviere, J., and D. Hensher
1982 Design and Analysis of Simulated Choice or Allocation Experiments in
Travel Choice Modeling. Transportation Research Record 890:11­17.
Louviere, J., and H. Timmermans
1990 Stated Preference and Choice Models Applied to Recreation Research:
A Review. Leisure Sciences 12:9­32.
Louviere, J., and G. Woodworth
1983 Design and Analysis of Simulated Consumer Choice or Allocation Experi-
ments: An Approach Based on Aggregate Data. Journal of Marketing Research
20:350­367.
Luzar, E., A. Diagne, C. Gan, and B. Henning
1998 Profiling the Nature-Based Tourist: A Multinomial Logit Approach. Jour-
nal of Travel Research 37:48­55.
Maddala, G.
1991 Limited-Dependent and Qualitative Variables in Econometrics. New
York: Cambridge University Press.
Miazaki, T.
1997 Green Tourism and Japan's Rural Areas. Association of Agriculture and
Forestry Statistics of Japan.
Min, S., H. Koh, S. Choi, and H. Shin
2001 Pre-Requisites for Implementing the Five-Working-Day-Week System.
CEO Information 311:1­28.
Ministry of Agriculture
1992 Survey of Green Tourism on Urban Residents. Tokyo: The Ministry of
Agriculture Forestry and Fisheries of Japan.
2001 Agricultural and Forestry Statistical Yearbook. Tokyo: The Ministry of
Agriculture Forestry and Fisheries of Japan.
Oppermann, M.
1996 Rural Tourism in Southern Germany. Annals of Tourism Research
23:86­102.
Pevetz, W.
1982 Agriculture and Tourism in Austria. Paper Presented at Symposium of
Agriculture and Tourism Report, Mariehamn.
Reichel, A., O. Lowengart, and A. Milman
2000 Rural Tourism in Israel: Service Quality and Orientation. Tourism Man-
agement 21:451­459.
Sharpley, R.
2002 Rural Tourism and the Challenge of Tourism Diversification: The Case
of Cyprus. Tourism Management 23:233­244.
Um, S.
1994 Development of a Multi-Item Scale for Measuring Theme Park Service
Quality. Journal of Korean Institute of Landscape Architecture 22:25­38.
Unwin, T.
1996 Tourist Development in Estonia: Images, Substantiality and Integrated
Rural Development. Tourism Management 17:265­276.
Wang, D., H. Oppewal, and H. Timmermans
2000 Pairwise Conjoint Analysis of Activity Engagement Choice. Environment
and Planning A 32:805­816.
Weaver, B., and D. Fennell
1997 The Vacation Farm Sector in Saskatchewan: A Profile of Operations.
Tourism Management 18:357­365.
Yamazaki, M., Y. Oyama, and J. Ohshima
1997 Green Tourism. Seoul: Ilshinsa.
Yoo, S., Y. Lee, P. Peon, Y. Joe, and S. Yoo
1998 Research on the Current Status of Japanese Type Green Tourism. Paper
Presented at the Conference of Korean Academy of Tourism Agriculture,
Seoul.
Submitted 2 October 2001. Resubmitted 14 May 2002. Accepted 9 July 2002. Refereed
anonymously. Coordinating Editor: Melville Saayman