Nova metodologia para interpretação de dados de análise sensorial de alimentos

Abstract

RESUMO:Os métodos atualmente utilizados para realização de análise sensorial têm deixado muito a desejar por não traduzir satisfatoriamente os resultados. Diante deste problema, analisou-se a possibilidade de introdução de um fator que medisse o índice de concordância dos provadores sobre determinado aspecto analisado. Para validação deste coeficiente, foi avaliada a concordância entre julgadores na análise sensorial para três diferentes amostras de suco de pinha (Annona squamosa L.): uma constituída apenas do suco de pinha sem leite e duas com diferentes concentrações pinha/leite. Utilizou-se a Análise de Componentes Principais (ACP) e também a nova metodologia proposta pelos autores, com base na definição do desvio-padrão, aplicada às frequências absolutas das notas dos julgadores, para cada propriedade sensorial estudada. Com esta metodologia, definiu-se uma relação para o cálculo de uma medida de concordância entre os julgadores, denominada de coeficiente de concordância, a qual pode assumir valores de 0 a 100% e também permite comparações que não são possíveis, utilizando-se da ACP. Por fim, concluiu-se que tal metodologia é mais indicada que a ACP, na avaliação da concordância entre os julgadores na análise sensorial.PALAVRAS-CHAVE: coeficiente de concordância, propriedades sensoriais, degustadores. NEW METHODOLOGY FOR DATA INTERPRETATIONOF FOOD SENSORIAL ANALYSIS ABSTRACT:The methods now used, for accomplishment of sensorial analysis, do not translate the results satisfactorily. Because of this problem, it was analyzed the possibility of introducing a factor that measures the agreement index of the tasters on a particular aspect analyzed. To validate this coefficient, the agreement coefficient among the sensory analysis tasters had been evaluated using three different annona samples; one consisting only of the annona juice (without milk) and the other with different concentrations of juice / milk (mass percentage). This evaluation was accomplished by the Analysis of Main Components (ACP) and by the new methodology also proposed by the authors, with base in the definition of the standard deviation, applied to the absolute frequencies of the tasters' notes, for each studied sensorial property. With this methodology it was defined a relation for calculation of an agreement measure among the tasters, denominated of agreement coefficient, which can assume values from 0 to 100%. Finally, it was concluded that such methodology is more suitable than ACP, in the evaluation of the agreement among the tasters in the sensorial analysis, since that methodology results in a measure of the agreement degree and it allows comparisons that are not possible being used ACP.

Production of a product similar to gelatin from chicken feet collagen

“…The answers were recorded on standardized forms describing the intensities of the sensations that the food caused on the evaluators, both in standard hedonic scales. They evaluated the attributes color, flavor, aroma, texture and the availability in consuming the product, using the affective test with structured hedonic scale of 9 points, ranging from 1-I dislike extremely to 9-I like extremely to the questions color, aroma, flavor and texture and the test of predisposition to consumption with scale of 1-not to consume to 6-consume everyday (ALMEIDA et al, 2008 andANDRAE-NIGHTINGALE et al, 2009;MADRONA et al, 2009;SANTANA et al, 2010;STERN JR et al, 2007;SILVA et al, 2010).…”

Influence of storage of refrigerated milk on yield and sensory characteristics of queso fresco

“…To complement the sensory analysis results, the percentage of the coefficient of agreement among judges with respect to the scores given was analyzed using the Consensor software (Silva et al, 2010a). (Brasil, 2011), which recommends at least 3%.…”

Experimental infection of Salmonella Enteritidis in quails submitted to forced molting by feed fasting

“…Treatment means were compared by the nonparametric Mann-Whitney test (5%). Statistical analyzes were performed using the program ASSISTAT, beta version 7.6 (Silva et al, 2010).…”