direction, respectively, and the arrows represent the different background
variables. The direction of the arrows represents the correlation
between each variable and the canonical axes, and each other, whereas
the length of the arrows represents the relative a contribution of the
variables to the axes and the perception–background relationship.
Generally, the first canonical axes represent an estimated 71.6% of the
variation in the perception–background of pesticide risk and the first
two represent an estimated 86.9%. Selected region, education, and age
were significantly correlatedwith the responses to the dangers of pesticides
(permutation test, p b 0.01) (Table 7), but occupation and gender
were not (Fig. 2). The opposite directions of the arrows for education
and age in Fig. 2 indicated that these two factors had opposite influences
on the awareness of the dangers of pesticides. The location of gender
and occupation near the origin of the coordinates indicated that the effects
of these two factors on the understanding of the dangers of pesticideswere
similar and could not vary the percentage of the explanation
of understanding on the canonical axis. Consequently, selected region,
education, and age were important factors in promoting the understanding
of the dangers of pesticides to the environment and human
health.