A survey with 247 customers was carried out in four german horticultural sh
ops to identify the relationship between plant signs and consumer preferenc
es. Objective were poinsettia plants. Approximatly 50 to 60 percent of the
variance could be explained by the plant signs included into regression ana
lysis and 70% with the two main components of a principal component analysi
s. For each of the shops where the survey took place different plant signs
were identified as relevant. The results show that it is not possible to ex
plain customer decisions by using a simple additive approach where the fina
l decision is seen as the result of the addition of partial decisions. But
the combination of principal component analysis and regression analysis may
help to show which plants signs are most important for growers to satisfy
customers needs. Also it seems to be neccesary to record the customer ranki
ngs not only as scalar values. Confrontation of PCA-results with the result
s of multiple regression points out that using multiple regression has limi
ted validity in this kind of experiments if there are variables with large
ranges and intercorrelations.