Quality and consumer acceptability 345 properties over time. The test sensitivity can allow expensive capital expenditure requirements to be made with a high level of confidence When more than two products are compared, ranking tests comparison tests are used. Typically, a ranking test is used when no control is available or required and assessors are asked to rank products in order of intensity for a specific sensory characteristic. A multiple comparison test is used when a control sample is used as an anchor point and assessors are asked to evaluate the intensity of the difference if any Other tests might be required from time to time to evaluate the sensitivity of specific compounds and/or chemicals. These tests are important in taint evaluation or the detection of materials difficult to assess. Threshold tests are used to detect at which concentration level, a compound can be detected Dilution technique determines the smallest amount of compound that can be detected in a product. Gillette et al. (1984)use this technique to assess the level of heat in red peppers 12.3.2 Descriptive tests These tests are used to identify sensory characteristics of a chilled product and quantify them. Panellists are selected on their ability to describe and discriminate between samples. They are presented with variants of the product and asked to describe them. After a period dedicated to confirm and define agreed terms and scale, the panel is then presented with the samples, one at a time and asked to give a score. Stone et al. (1974) has described in detail one of these descriptive techniques calledquantitative descriptive analysis. Results are analysed by means of univariate and multivariate analyses. Analysis of variance(O"Mahony 1986) is used to measure any difference between samples for each attribute Principal component analysis is a technique used to reduce the amount of dimension or sensory terms into a manageable format, usually two or three dimensions. Procustes analysis is mainly used for the assessment of individuals performance and efficiency(Arnold and Williams 1986). Graphs are often used to summarise the results into condensed and meaningful information. Spider graphs(see Fig. 12.1), give a general overview of sensory differences by th overall shape of each product, providing an individual product fingerprint. Other sensory techniques have emerged in recent years in the evaluation of food quality by sensory evaluation. In particular, the time-intensity technique takes into consideration the temporal dimension of tasting behaviour( Cliff et al. 1993). This technique is particularly interesting for the assessment of spicy oroducts or for primary taste evaluation, such as sweet sensation in artificially sweetened drinks and is used by Matysiak and Noble(1991) Figure 12.1 shows an overview of how four samples of chilled smoked salmon differ in sensory terms from one another, using a spider graph. For each attribute, the samples are more intense away from the centre. In particular salmon d had a more orangy colour and sample F had the least moist surface Sample I was the least salty and smoky producproperties over time. The test sensitivity can allow expensive capital expenditure requirements to be made with a high level of confidence. When more than two products are compared, ranking tests or multiple comparison tests are used. Typically, a ranking test is used when no control is available or required and assessors are asked to rank products in order of intensity for a specific sensory characteristic. A multiple comparison test is used when a control sample is used as an anchor point and assessors are asked to evaluate the intensity of the difference if any. Other tests might be required from time to time to evaluate the sensitivity of specific compounds and/or chemicals. These tests are important in taint evaluation or the detection of materials difficult to assess. Threshold tests are used to detect at which concentration level, a compound can be detected. Dilution technique determines the smallest amount of compound that can be detected in a product. Gillette et al. (1984) use this technique to assess the level of heat in red peppers. 12.3.2 Descriptive tests These tests are used to identify sensory characteristics of a chilled product and to quantify them. Panellists are selected on their ability to describe and discriminate between samples. They are presented with variants of the product and asked to describe them. After a period dedicated to confirm and define agreed terms and scale, the panel is then presented with the samples, one at a time and asked to give a score. Stone et al. (1974) has described in detail one of these descriptive techniques called ‘quantitative descriptive analysis’. Results are analysed by means of univariate and multivariate analyses. Analysis of variance (O’Mahony 1986) is used to measure any difference between samples for each attribute. Principal component analysis is a technique used to reduce the amount of dimension or sensory terms into a manageable format, usually two or three dimensions. Procustes analysis is mainly used for the assessment of individuals’ performance and efficiency (Arnold and Williams 1986). Graphs are often used to summarise the results into condensed and meaningful information. Spider graphs (see Fig. 12.1), give a general overview of sensory differences by the overall shape of each product, providing an individual product fingerprint. Other sensory techniques have emerged in recent years in the evaluation of food quality by sensory evaluation. In particular, the time-intensity technique takes into consideration the temporal dimension of tasting behaviour (Cliff et al., 1993). This technique is particularly interesting for the assessment of spicy products or for primary taste evaluation, such as sweet sensation in artificially sweetened drinks and is used by Matysiak and Noble (1991). Figure 12.1 shows an overview of how four samples of chilled smoked salmon differ in sensory terms from one another, using a spider graph. For each attribute, the samples are more intense away from the centre. In particular, salmon D had a more orangy colour and sample F had the least moist surface. Sample I was the least salty and smoky product. Quality and consumer acceptability 345