Analysis of Sensory Data

The determination and analysis of sensory properties plays an important role in the development of new consumer products. Particularly in the food industry sensory analysis has become an indispensable tool in research, development, marketing and quality control. The discipline of sensory analysis covers a wide spectrum of subjects physiology of sensory perception, psychology of human behaviour, flavour chemistry, physics of emulsion break-up and flavour release, testing methodology, consumer research, statistical data analysis. Not all of these aspects are of direct interest for the chemometrician. In this chapter we will cover a few topics in the analysis of sensory data. General introductory books are e.g. Refs. [1-3]. 

There are four main types of data that frequently occur in sensory analysis pair-wise differences, attribute profiling, time-intensity recordings and preference data. We will discuss in what situations such data arise and how they can be analyzed. Especially the analysis of profiling data and the comparison of such data with chemical information calls for a multivariate approach. Here, we can apply some of the techniques treated before, particularly those of Chapters 35 and 36. 

---

Almost everyone is now utilizing the computer for statistical analysis of sensory data. Some laboratories also are using computers to gather the data as well (18). A computerized sensory system would benefit most laboratories by freeing workers from laborious data entry and analysis. Also, it would allow for a more thorough analysis of the data. It should not replace inspection of the raw data by the sensory scientist, but allow this to occur more easily. 

Fig. 23.11 Discriminant analysis of sensory data on thirty three lager beers, (a) Result (b) Interpretation of result. Code A = North American beers, B = British beers, C = Continental European beers. (After Brown and Clapperton [54].)...

King, B. M. and Arents, P. (1991). A statistical test of consensus obtained from generalized procrustes analysis of sensory data. Journal of Sensory Studies, 6, 37—48. 

Cocchi M, Bro R, Durante C, Manzini D, Marchetti A, Saccani F, et al. Analysis of sensory data of Aceto Balsamico Tradizionale di Modena (ABTM) of different ageing by application of PARAFAC models. Food Qual Prefer 2006 17 419-28. 

Procrustes analysis has been generalized in two ways. One extension is that more than two data sets may be considered. In that case the algorithm is iterative. One then must rotate, in turn, each data set to the average of the other data sets. The cycle must be repeated until the fit no longer improves. Procrustes analysis of many data sets has been applied mostly in the field of sensory data analysis [4]. Another extension is the application of individual scaling to the various data sets in order to improve the match. Mathematically, it amounts to multiplying all entries in a data set by the same scalar. Geometrically, it amounts to an expansion (or... 

Sensory analysis. Significant differences in intensity were found for all seven aroma terms by analysis of variance (data not shown). Because of a highly significant judge-by-wine interaction, the berry term was excluded from further data analysis. 

Results of cluster analysis of sensory and physicochemical data on beer After CLAPPERTON [50] ... 

The development of rapid sensory profiling methods may have potential consequences on sensory activities themselves, since it broadens the spectrum of available methods and opens way for measurements that were previously not possible. Besides offering new opportunities in the use of sensory data in R D and research projects, this development may also have an organizational impact on sensory services and their relationships with stakeholders. As a result of this evolution, the practice of sensory descriptive analysis certainly becomes richer but also more complex and challenging. 

In spite of these reservations, it may be anphasized that when appropriately applied, rapid sensory profiling techniques are powerful tools for clever use of sensory analysis. Many examples and testimonies of successful uses of these methods, both in industry and in acadania, are presented throughout this book. This wiU hopefully spark the interest of students in sensory programmes, as well as that of sensory professionals, sensory scientists and, more generally, all users of sensory data. May this book help them in their daily work, provide than with some solutions and contribute to fostering innovation in sensory science. 

There are a finite number of methods described in the sensory literature and a very large number of modifications, many of which are based on a need to try a different method, or belief that a modified method will improve results. In most instances, the modifications are driven by a statistical, not a behavioral approach. Increasing the power of a test is a goal and certainly will have a significant impact on the results, but how much more power can be achieved if one is not using qualified subjects. Confidence in results derives from knowing what subjects were used, the chosen design, and the analysis of the data. 

Hunter, E. A. and Muir, D. D. (1995). A comparison of two multivariate methods for the analysis of sensory profile data. Journal of Sensory Studies, 10, 89-104. 

The first visualization and analysis of TDS data were presented 3 years later at the Pangbom symposium (Pineau et al., 2003) and started to be used by several companies. This method is nowadays well established in the sensory domain and has been successfully applied to many product categories. 

More sophisticated statistical treatments of sensory data have been more commonly applied in studies of multiple factors of Upid oxidation on quality of foods, including Multivariate and Principal Component analyses. These procedures attempt to simplify complex relationships of several factors and sets of data into more understandable levels. Multivariate analysis is based on the fact that one measured property generally depends on more than one factor and the classical statistical univariate methods dealing with just one variable at a time are inadequate to analyse complex data. 

The data from sensory evaluation and texture profile analysis of the jellies made with amidated pectin and sunflower pectin were subjected to Principal component analysis (PC) using the statistical software based on Jacobi method (Univac, 1973). The results of PC analysis are shown in figure 7. The plane of two principal components (F1,F2) explain 89,75 % of the variance contained in the original data. The attributes related with textural evaluation are highly correlated with the first principal component (Had.=0.95, Spr.=0.97, Che.=0.98, Gum.=0.95, Coe=0.98, HS=0.82 and SP=-0.93). As it could be expected, spreadability increases along the negative side of the axis unlike other textural parameters. 

Let us try to relate the (standardized) sensory data in Table 35.1 to the explanatory variables in Table 35.3. Essentially, this is an analysis-of-variance problem. We try to explain the effects of two qualitative factors, viz. Country and Ripeness, on the sensory responses. Each factor has three levels Country = Greece, Italy,... 

As an example we try to model the relation between the sensory data of Table 35.1 and the instmmental measurements of Table 35.4. The PLS analysis results are shown in Table 35.8. The first PLS dimension loads about equally high on... 

G. B. Dijksterhuis, Procrustes analysis in sensory research, Ch. 7 in Multivariate analysis of data in sensory science (T. Naes and E. Risvik, eds), Elsevier, Amsterdam (1996). 

---