International Biometric Society LINEAR MIXED EFFECTS MODELLING FOR MULTIFACTORIAL SENSORY AND CONSUMER DATA USING THE R-PACKAGES LMERTEST AND SENSMIXED Isabel de Sousa Amorim a, b, Alexandra Kuznetsova a, Søren Bech c Renato Ribeiro de Lima b , Per Bruun Brockhoff a a Technical University of Denmark, bUniversidade Federal de Lavras, cBang & Olufsen A/S Sensory and consumer data are produced in massive numbers within as well food research and industry as within many non-food areas, e.g. high end audio and TV production industry (Næs, Brockhoff & Tomic, 2010, Bech & Zacharov, 2006). A user friendly open source tool, Panelcheck, for high throughput analyses of sensory quantitative descriptive analysis (QDA) data was recently developed (www.panelcheck.com) including visual tools for simple mixed models for such multi-attribute data. Due to the simplicity in use, the tool is now used extensively globally. The scope of the mixed modelling in Panelcheck, however, is limited in several ways. In 2014 a new tool, Consumercheck will be released extending the scope of the tool to also include consumer preference data together with background data on as well products as on consumers. Among the available methods in Consumercheck is a user friendly approach for rather general linear mixed modelling of such data avoiding the mixed model limitations that characterises Panelcheck. The mixed model tools of Consumercheck is based on the newly developed R-packages lmerTest, Kuznetsova, Brockhoff & Christensen (2013) and SensMixed, Kuznetsova, Brockhoff & Christensen (2013) – again based on the lme4-package. In lmerTest, among other things, automated model selection procedures are available to facilitate more easy access to proper mixed modelling for challenging structured situations, Kuznetsova et al (2014). Also recently, the so-called Mixed Assessor Model (MAM) was proposed by Brockhoff, Schlich & Skovgaard (2014) as an improved mixed model analysis of sensory data more properly taking into account the inherent effects of individual differences in perceptive scale use in such data. The MAM approach, however, was introduced in a rather restricted setting. In this poster we present the novel combination of the MAM approach with a more general mixed modelling approach covering now any relevant sensory and consumer data situation. These extensions will be part of the currently developing R-package SensMixed. The approach will be illustrated on a data set from the company Bang and Olufsen A/S, Struer, Denmark. The purpose of this study was to evaluate 90 combinations of 6 car sound systems, 3 reproduction sound pressure level and 5 different tracks. A trained audio panel composed by 10 panellists evaluated the 90 products for 8 different sensory attributes in 2 replications. References Brockhoff, P.B., Schlich, P. & Skovgaard, I. (2014). Taking individual scaling differences into account by analyzing profile data with the Mixed Assessor Model. Submitted to Food Quality and Preference. Bech. S & Zacharov, N. (2006). Perceptual Audio Evaluation – Theory, Method and Application, Wiley, England, 450 pages. Kuznetsova, A., Christensen, R.H.B., Bavay, C., Brockhoff, P.B. (2014). Automated mixed ANOVA modelling of sensory and consumer data. To be submitted to Food Quality and Preference. Kuznetsova, A., Brockhoff, P.B. & Christensen, R.H.B. (2013). LmerTest: Tests for random and fixed effects for linear mixed effect models (lmer objects of lme4 package). R package version 2.0-0. http://CRAN.R-project.org/package=lmerTest/ Kuznetsova, A., Brockhoff, P.B. & Christensen, R.H.B. (2013). SensMixed: Mixed effects modelling for sensory and consumer data. R package version 1.0-1/r44. http://R-Forge.R-project.org/projects/lmertest/ T. Næs, P.B. Brockhoff and O. Tomic, (2010). Statistics for Sensory and Consumer Science, John Wiley & Sons. International Biometric Conference, Florence, ITALY, 6 – 11 July 2014