ConsumerCheck is an open source data analysis software tailored for analysis of sensory and consumer data. Since some of the implemented methods are generic, such as PCA, PLSR and PCR, other data from other domains may also be analysed with ConsumerCheck. The software comes with a graphical user interface and as such provides non-statisticians and users without programming skills free access to a number of widely used analysis methods within the field of sensory and consumer science. Computational results are presented in plots that are easily generated from the tree-controls within the graphical user interfaces. Since the construction of conjoint analysis models is not always straightforward, ConsumerCheck provides three previously defined model structures of different complexity. ConsumerCheck is an ongoing research project and the objective is to implement further statistical methods over time.

One of the frequent questions by users of the mixed model function lmer of the lme4 package has been: How can I get p values for the F and t tests for objects returned by lmer? The lmerTest package extends the 'lmerMod' class of the lme4 package, by overloading the anova and summary functions by providing p values for tests for fixed effects. We have implemented the Satterthwaite's method for approximating degrees of freedom for the t and F tests. We have also implemented the construction of Type I - III ANOVA tables. Furthermore, one may also obtain the summary as well as the anova table using the Kenward-Roger approximation for denominator degrees of freedom (based on the KRmodcomp function from the pbkrtest package). Some other convenient mixed model analysis tools such as a step method, that performs backward elimination of nonsignificant effects - both random and fixed, calculation of population means and multiple comparison tests together with plot facilities are provided by the package as well.

The beverage industry has long revolved around sugar reduction as a response to heightened calorie and health awareness. More recently dietary fibre has also garnered attention to meet the consumer's demands for low calorie and yet more satiating food and beverages. From a health perspective the sweetener Stevia rebaudiana and the fibre β-glucan seem like very good solutions, as stevia is the only natural non-nutritive sweetener and β-glucans have been related to various health benefits besides increasing satiety. However, both also have distinctive perceptual effects on the sensory characteristics of the products they are added to. To gain knowledge on the sensory characteristics of fruit based beverages sweetened with S. rebaudiana and added β-glucans and lime flavour, and how consumers respond to the products, sensory descriptive analysis and a consumer study were conducted. The sensory characteristics of the fruit drinks were affected by stevia and the addition of β-glucans. However, the addition of lime flavour was able to mask the side effect of the aftertaste caused by S. rebaudiana. Further, by adding lime flavour to the fruit beverages, the side effects of increased fibre concentration “Unfresh odour” and “Metallic odour” could also be counteracted to such an extent that the β-glucans containing fruit beverages were evaluated just as favourably in terms of liking, wanting and sensory satisfying as the fruit beverages not containing β-glucans.

We utilize the close link between Cohen’s d, the effect size in an ANOVA framework, and the Thurstonian (Signal detection) d-prime to suggest better visualizations and interpretations of standard sensory and consumer data mixed model ANOVA results. The basic and straightforward idea is to interpret effects relative to the residual error and to choose the proper effect size measure. For multi-attribute bar plots of F-statistics this amounts, in balanced settings, to a simple transformation of the bar heights to get them transformed into depicting what can be seen as approximately the average pairwise d-primes between products. For extensions of such multi-attribute bar plots into more complex models, similar transformations are suggested and become more important as the transformation depends on the number of observations within factor levels, and hence makes bar heights better comparable for factors with differences in number of levels. For mixed models, where in general the relevant error terms for the fixed effects are not the pure residual error, it is suggested to base the d-prime-like interpretation on the residual error. The methods are illustrated on a multifactorial sensory profile data set and compared to actual d-prime calculations based on Thurstonian regression modeling through the ordinal package. For more challenging cases we offer a generic “plug-in” implementation of a version of the method as part of the R-package SensMixed. We discuss and clarify the bias mechanisms inherently challenging effect size measure estimates in ANOVA settings.

Mixed effects models have become increasingly prominent in sensory and consumer science. Still applying such models may be challenging for a sensory practitioner due the challenges associated with the choosing the random effects, selecting an appropriate model, interpreting the results. In this paper we introduce an approach for automated mixed ANOVA/ANCOVA modeling together with the open source R package lmerTest developed by the authors that can perform automated complex mixed-effects modeling. The package can in an automated way investigate and incorporate the necessary random-effects by sequentially removing non-significant random terms in the mixed model, and similarly test and remove fixed effects. Tables and figures provide an overview of the structure and present post hoc analysis. With this approach, complex error structures can be investigated, identified and incorporated whenever necessary. The package provides type-3 ANOVA output with degrees of freedom corrected F-tests for fixed-effects, which makes the package unique in open source implementations of mixed models. The approach together with the user-friendliness of the package allow to analyze a broad range of mixed effects models in a fast and efficient way. The benefits of the approach and the package are illustrated on four data sets coming from consumer/sensory studies.

In descriptive sensory analysis, large variations may be observed between scores. Individual differences between assessors have been identified as one cause for these variations. Much work has been done on modeling these differences and accounting for them through analysis of variance (ANOVA). When the products studied are prone to biological heterogeneity (e.g. fruits, vegetables, cheeses, etc.), variations in the data may be due to assessor differences and/or product heterogeneity. The present paper proposes an approach for quantifying these two sources of variation. For individual differences, an extended version of the assessor model approach is applied. The data set used in the paper is based on sensory evaluations of three apple samples scored by a panel of 19 assessors using seven descriptors in four replicates. The application of the extended assessor model approach to unbalanced data provides more insight into assessor differences and a better test for product differences. These results demonstrate the importance of choosing the right model and taking all potential sources of variation into account.

The assessment of produce quality is a major aspect of applied postharvest biology. Horticultural researchers working on organoleptic quality of fruit need objective methods for the evaluation of sensory properties. The development of sensory methodologies specifically for apples highlighted the problem of handling variation due to fruit variability and assessor differences. The aim of this study was to investigate the weight of within-batch variability in sensory evaluation of apples and to propose a methodology that accounts for this variability. Prior to sensory analysis, for three apple cultivars, apples were sorted into homogenous acoustic firmness categories within each cultivar. The discrimination ability of the trained panel was observed not only between cultivars but also within each cultivar for crunchiness, firmness, juiciness and acidity. Following these results, a mixed hierarchical model for the analysis of the sensory data was proposed to measure the contribution of fruit variability to the variability of sensory scores. The results showed the efficiency of the model in quantifying within-batch variability. Fruit sampling and presentation methods as well as data handling procedures are suggested for obtaining reliable sensory results in the assessment of apple quality. (C) 2012 Elsevier B.V. All rights reserved.

The SensMixed package offers analysis of sensory and consumer data within a mixed effects model framework. The package provides tools for analysis of simple settings, similar to what PanelCheck( www.panelcheck.com ) is giving, as well as advanced tools such as incorporating MAM (mixed assessor model) Brockhoff, Schlich & Skovgaard (2014), handling unbalanced data and allowing for multi-way product structures. The automated identification of important random effects, that uses the methodology introduced in Kuznetsova, Christensen, Bavay and Brockhoff (2015) is also employed in the SensMixed package. Based on the shiny R package, the SensMixed package includes an application that has a graphical user interface (GUI) for the provided tools. Apart from providing the GUI for the tools, the application includes such functionalities as importing the data in different formats, presenting results in tables and plots as well as saving them. A number of modelling options are provided that allow to easily construct and analyze in a proper manner a broad range of complex mixed effects models. All that makes the package together with the application very valuable for sensory practitioners as requires no skills in R-programming and provides advanced statistical methods for analyzing sensory data. The usefulness of the package and the application will be illustrated on examples coming from the sensory studies.

Mixed effects models have become increasingly prominent in sensory and consumer science. Still applying such models may be challenging for a sensory practitioner due the challenges associated with the choosing the random effects, selecting an appropriate model, interpreting the results. In this paper we introduce an approach for automated mixed ANOVA/ANCOVA modeling together with the open source R package lmerTest developed by the authors that can perform automated complex mixed-effects modeling. The package can in an automated way investigate and incorporate the necessary random-effects by sequentially removing non-significant random terms in the mixed model, and similarly test and remove fixed effects. Tables and figures provide an overview of the structure and present post hoc analysis. With this approach, complex error structures can be investigated, identified and incorporated whenever necessary. The package provides type-3 ANOVA output with degrees of freedom corrected F-tests for fixed-effects, which makes the package unique in open source implementations of mixed models. The approach together with the user-friendliness of the package allow to analyze a broad range of mixed effects models in a fast and efficient way. The benefits of the approach and the package are illustrated on four data sets coming from consumer/sensory studies. 2014 Elsevier Ltd. All rights reserved.

towards sensometrics DTU Orbit (06/12/2018) Visualizing multifactorial and multi-attribute effect sizes in linear mixed models with a view towards sensometrics In Brockhoff et al (2016), the close link between Cohen's d, the effect size in an ANOVA framework, and the so-called Thurstonian (Signal detection) d-prime was used to suggest better visualizations and interpretations of standard sensory and consumer data mixed model ANOVA results. The basic and straightforward idea is to interpret effects relative to the residual error and to choose the proper effect size measure. For multi-attribute bar plots of F-statistics this amounts, in balanced settings, to a simple transformation of the bar heights to get them transformed into depicting what can be seen as approximately the average pairwise d-primes between products. For extensions of such multi-attribute bar plots into more complex models, similar transformations are suggested and become more important as the transformation depends on the number of observations within factor levels, and hence makes bar heights better comparable for factors with differences in number of levels. For mixed models, where in general the relevant error terms for the fixed effects are not the pure residual error, it is suggested to base the d-prime-like interpretation on the residual error. The methods are illustrated on a multifactorial sensory profile data set and compared to actual d-prime calculations based on ordinal regression modelling through the ordinal package. A generic ``plug-in'' implementation of the method is given in the SensMixed package, which again depends on the lmerTest package. We discuss and clarify the bias mechanisms inherently challenging effect size measure estimates in ANOVA settings.