ples quite similarly. Overall there is no structure in the way that the con-
sumer panels are positioned around each sample. It can therefore be con-
cluded from this plot that the consumers perceived the differences
between the samples similarly. This led us to conclude that there is a
cross-cultural consensus with respect to the perception of mealiness.
Having established this cross-cultural perception it was necessary to look
at how the consumers described their perceptions of the apples.
The consumers did vary in how they used the descriptors. Interest lay
in exploring these differences in more detail. Were there differences be-
tween consumers within one panel in how they used these descriptors
and how they described the differences? In order to help answer this
question a curved box plot procedure was developed. Figure 9.6 shows
the curved box plots for the British and Danish consumers.
The curved box plot is similar to a normal box plot but here it is the
angular distribution of the consumer descriptors that is represented. The
box itself accounts for 50% of the attributes and the line inside signifies
the direction of the median attribute vector. Lines or whiskers extend up
to 1.5 times the interquartile range, or to the left-most or right-most at-
tribute vector. Attributes outside of the interquartile range are repre-
sented by an asterisk and can be considered outlying. The use of the
descriptors “crisp” and “mealy” by Danish and British consumers is
168 CONSUMER PREFERENCE
FIGURE 9.5GPA consensus plot from repertory grid study of apple mealiness. Show-
ing variations among the different consumer panels: 1fresh, 2midpoint, 3mealy.