246
A. Terzi´ c and D. Demirovi´ c Bajrami
5 Methodology
The need for identification of heterogeneous visitor segments, based on differences in motivation factors, has been the most reliable approach when striving to
understand different user groups in international travel settings to protected areas
(Weaver and Lawton 2005). Extant research illustrates that values can be used to
predict general travel behaviour, destination choice, leisure activities, preferences
and trip length, as well as mediating other factors on tourists’ behaviour (sociodemographics) (Hedlund et al. 2012). However, Hvenegaard (2002) outlined that
cognitive-normative typologies used to identify ecotourists based on their motivations, attitudes, and values were usually tautological and restricted to the data
collected on specific sites with a relatively small sample size. This study used
a random sample based on standardized methodology applied on a large general
sample, while data were collected from the European Social Survey (2016).
6 Results
Principal component factor analysis was first applied to delineate the underlying
dimensions of basic human values of respondents defined as ‘potential tourists’.
Hence, 21 human values variables were analysed by testing inter-correlations.
Bartlett’s test of sphericity was statistically significant (p < 0.001) and the Measure of
Sampling Adequacy (KMO) was 0.858, indicating that all variables were acceptable
for conducting factor analysis, while the Cronbach’s Alpha calculation was 0.882 (n
= 41,293, df = 21). The reliability alpha calculated for all factors indicated satisfaction of the criterion (above 0.60). Table 1 presents the results of Varimax rotation
with three factors identified with a total of 45.371% of variance explained. Three
factors were labelled, based on the consideration of basic value sets (Schwartz 2001)
in the context of basic tourist segments:
• Factor 1 labelled as ‘Self-enhancement’ explained 16.573% out of total variance
with reliability alpha calculated to 0.793 and incorporating eight variables;
• Factor 2 labelled as ‘Self-transcendence’ explained 14.87% of total variance with
reliability alpha calculated to 0.750, incorporating seven variables;
• Factor 3 labelled as ‘Conservatism’ accounted for 13.92% out of total variance
gathering six variables with reliability alpha calculated to 0.670.
Based on the identified three groups (factors) of basic human values in the previous
step, cluster analysis was conducted. The main aim of this step was to segregate the
ecotourist segments. The K-means non-hierarchical cluster analysis and centroids of
three clusters were used for clustering the respondents, while obtained values were
used for the final cluster solution. The results in Table 2 showed that respondents
were classified into three clusters: Cluster 1 was composed of 35.28% of the total
A. Terzi´ c and D. Demirovi´ c Bajrami
5 Methodology
The need for identification of heterogeneous visitor segments, based on differences in motivation factors, has been the most reliable approach when striving to
understand different user groups in international travel settings to protected areas
(Weaver and Lawton 2005). Extant research illustrates that values can be used to
predict general travel behaviour, destination choice, leisure activities, preferences
and trip length, as well as mediating other factors on tourists’ behaviour (sociodemographics) (Hedlund et al. 2012). However, Hvenegaard (2002) outlined that
cognitive-normative typologies used to identify ecotourists based on their motivations, attitudes, and values were usually tautological and restricted to the data
collected on specific sites with a relatively small sample size. This study used
a random sample based on standardized methodology applied on a large general
sample, while data were collected from the European Social Survey (2016).
6 Results
Principal component factor analysis was first applied to delineate the underlying
dimensions of basic human values of respondents defined as ‘potential tourists’.
Hence, 21 human values variables were analysed by testing inter-correlations.
Bartlett’s test of sphericity was statistically significant (p < 0.001) and the Measure of
Sampling Adequacy (KMO) was 0.858, indicating that all variables were acceptable
for conducting factor analysis, while the Cronbach’s Alpha calculation was 0.882 (n
= 41,293, df = 21). The reliability alpha calculated for all factors indicated satisfaction of the criterion (above 0.60). Table 1 presents the results of Varimax rotation
with three factors identified with a total of 45.371% of variance explained. Three
factors were labelled, based on the consideration of basic value sets (Schwartz 2001)
in the context of basic tourist segments:
• Factor 1 labelled as ‘Self-enhancement’ explained 16.573% out of total variance
with reliability alpha calculated to 0.793 and incorporating eight variables;
• Factor 2 labelled as ‘Self-transcendence’ explained 14.87% of total variance with
reliability alpha calculated to 0.750, incorporating seven variables;
• Factor 3 labelled as ‘Conservatism’ accounted for 13.92% out of total variance
gathering six variables with reliability alpha calculated to 0.670.
Based on the identified three groups (factors) of basic human values in the previous
step, cluster analysis was conducted. The main aim of this step was to segregate the
ecotourist segments. The K-means non-hierarchical cluster analysis and centroids of
three clusters were used for clustering the respondents, while obtained values were
used for the final cluster solution. The results in Table 2 showed that respondents
were classified into three clusters: Cluster 1 was composed of 35.28% of the total