Performance of hedonic scales in sensory acceptability of strawberry yogurt

• We proposed to evaluate assumptions of analysis of variance (ANOVA) in different hedonic scales.
• Meeting of the ANOVA assumptions is extremely important to use statistical parametric methods.
• Structured scales studied had less freedom of expression of feelings by consumers.
• There is a difference in the product evaluation when used different scales.
• Bipolar structured-numbered and traditional unstructured scales provided better discrimination of the products.

This study sought to use four different hedonic scales for assessment of consumer behavior in sensory acceptance tests of strawberry yogurt. Two nine-point structured scales were used (unipolar and bipolar), along with two unstructured linear scales measuring nine centimeters (hybrid and traditional). For each of the scales the assumptions of homogeneity of variances and normality of experimental errors were evaluated. Also assessed was the discrimination power of the products on the different scales, using the Duncan’s multiple range tests. For this purpose, four different brands of strawberry yogurt were evaluated with regards to sensory acceptability by 400 consumers, where each group of 100 volunteers used a different type of scale to express their opinions about the products. Normality of the errors was verified using the Kolmogorov–Smirnov test and homogeneity of variances of the errors by Levene’s test. For all scales assessed the normality and homogeneity of the experimental errors were observed (p > 0.05), but the unipolar structured scale presented poorer performance with respect to the normal errors. Contrarily, the unstructured scales underperformed in relation to the homogeneity of variance of the error. According to Duncan’s test for the products within each scale, the bipolar structured and traditional unstructured scales showed higher discrimination of the products. The structured bipolar scale and unstructured traditional scale showed satisfactory performance for assessing sensory acceptability of consumers, since they allow higher discrimination of products and met the assumptions of the analysis of variance.