Agriculture Reference
In-Depth Information
used to evaluate the relative importance of attributes and the extent of agreement among
consumers with respect to two or more rankings. In identifying consumer preferences,
researchers ranked the importance of the attributes following [37].
Cluster analysis was used to separate consumers into groups using the variables: age, educa‐
tion and employment status. Cluster analysis is a technique used for combining observations
or objects (answer, person, opinion, etc.) into groups or clusters. The aim of cluster analysis is
to classify observations into relatively homogeneous groups called clusters such that each
cluster is as homogeneous as possible with respect to the clustering variables [35-36]. The
Kolmogorov-Smirnov normality test was used to check whether the clustering variables
showed normal distribution, and then the Kruskal Wallis test was used to compare different
groups of clusters.
Multidimensional Scaling (MDS) was used to obtain a perceptual mapping of consumers'
preferences for fresh fruit and vegetable attributes. Given a matrix of perceived similarities
between attributes of fresh fruit and vegetables, MDS plots the attributes on a map such that
those attributes that are perceived to be very similar to each other are placed near each other
on the map, and those attributes that are perceived to be very different from each other are
placed far away from each other on the map.
7.2. Results and discussion
In this study, consumers were grouped into three clusters. The mean of the variables used in
the analysis is presented by clusters in Table 6. There were statistically significant differences
among clusters on the variables; age, education and employment of consumers in the sample.
The mean age (37.19) is the lowest in Cluster 1 and the highest (77.54) in Cluster 3. Education
level is the highest (5.63) in Cluster 1, whereas Cluster 3 has the lowest level (3.87). Employment
status changes from employed in Cluster 1 (2.06) to unemployed in Cluster 3 (2.89). Therefore,
cluster 1 is labeled “Young professional”, while the cluster 2 and cluster 3 are labeled “older-
employed” and “oldest-unemployed”, respectively.
Clusters
Kruskal Wallis Test
Variables
1
2
3
Chi-Square
Asymp. Sig
Age
37.19
58.15
77.54
339.960
0.000
Education
+
5.63
4.57
3.87
24.101
0.000
Employment
++
2.06
2.33
2.89
92.656
0.000
+
1: Less than high school, 11: Professional/doctorate degree;+
+
1: Part time, 2: Full time, 3: Unemployed
Table 6.
Cluster analysis by age, education and employment
