Methodological approach to evaluate the effects of age at slaughter and storage temperature and time on sensory profile of lamb meat.

A methodological approach, that requires the use of an experimental design, the definition of a sensory profile and an appropriate statistic validation of sensory data, was developed, in order to evaluate the effects of some process variables, such as age at slaughter and storage time and temperature, on sensory characteristics of lamb meat. In this study factorial design (2(3)) was used. The first step was the definition of descriptive terms for determining the sensory profile of meat. A vocabulary of 10 descriptive terms of odour, flavour and texture was developed. Therefore, after an appropriate and rigorous training, a panel of 8 assessors evaluated the lamb products in triplicate using a 100 mm unstructured scale. The following step was the statistic validation of data. Sensory data for each attribute were submitted to a fixed analysis of variance. Results showed that even though the panel was trained, judgements differed among assessors, as demonstrated by the significant assessor x sample interaction which was found for most attributes. In order to minimise this effect, the sensory data were submitted to a mixed analysis of variance that allows the determination of panel performance and selection of significant attributes for sample discrimination. Eight attributes were selected. Finally, analysis of variance with age, frozen temperature and storage time as main effects, was performed on mean values of selected attributes. Meaty odour and flavour, cohesiveness and stringiness were negatively influenced by age, whereas tenderness and juiciness were positively influenced by age. Lamb odour and flavour were positively influenced by frozen temperature. At increasing storage time, meaty odour and flavour and juiciness increased. The descriptive analysis allowed an objective description of the products in terms of perceived sensory attributes. The fixed and mixed models of analysis of variance produced different results. A mixed model appeared to be suitable for validation of panel performances and selection of significant attributes for meat samples discrimination.