It is the purpose of this paper to report a practical and a efficient method for the observation of food preference. The results were as follows :
The investigated foods were classified under 6 divisions by the method of cluster analysis.
It was recognized that the tendency of food prefernce made a change by differences between the sexes or discrepancy of age.
The results of multiple regression analysis for 12 kinds of foods selected from 6 divisions, the coefficient of determination towards the food preference degree showed 91.4～86.7% in range.
The estimation of linear discriminant function for food preference degree with 12 kinds of foods as predictor variables made it possible to classify the samples into 3 groups with misclassification ratios 4～16% in range. Especially, the misclassification between higher and lower level groups of food preference degree was not found at all.