Statistical analysis of sensory data: An introduction to methods and dedicated open source software tools

Tuesday, May 30, 08:00–12:00

What is Sensory Evaluation?

Sensory evaluation is at the core of QoE: It addresses the evaluation of products – in the QoMEX context of multimedia technology – by means of human assessors, applying their senses as part of the measurement instrument and test method. The so-called “subjective tests” essentially are a specific instantiation of sensory evaluation.

According to Wikipedia (accessed March 2017), “sensory evaluation […] applies principles of experimental design and statistical analysis to the use of human senses (sight, smell, taste, touch and hearing) for the purposes of evaluating consumer products. The discipline requires panels of human assessors, on whom the products are tested, and recording the responses made by them. By applying statistical techniques to the results it is possible to make inferences and insights about the products under test. Most large consumer goods companies have departments dedicated to sensory analysis. Sensory analysis can mainly be broken down into three sub-sections: 1) Analytical testing (dealing with objective facts about products), 2) Affective testing (dealing with subjective facts such as preferences), 3) Perception (the biochemical and psychological aspects of sensation)”


A number of open source analysis tools are now readily available for easy user access to the analysis of sensory (and consumer) data. These come partly as R-packages and partly as standalone tools. They cover all sorts of experimental data types as well as univariate and multivariate analysis of data. A brief overview of these is given. The focus of the tutorial is then primarily on the inherent challenge of analysing “repeated measures” sensory data by the proper mixed model versions of Analysis of Variance tools. With other words, how to properly account for the individual variabilities and correlations inherent in human perceptual data. User friendly access to such analyses are now available in tools such as PanelCheck (, Consumercheck ( and e.g. a GUI based Shiny version of the SensMixed R-package. With hands-on examples it will be illustrated how to use such tools for practical data analysis. The tools of SensMixed and ConsumerCheck are both based on the developments in the more generic R-package lmerTest.

About the Tutors

Per B. Brockhoff

DTU, Copenhagen, Denmark

Per Bruun Brockhoff is Professor in statistics at the Statistics and Data Analysis Section at DTU Compute. He was Ms.C. in statistics from Department of Mathematics, University of Aarhus, Denmark in 1991 and Ph.D. in statistics from Dep. of Mathematics and Physics, The Royal Veterinary and Agricultural University, Denmark in 1995. During the period 1995-2004 he was assistant and associate Professor in statistics at the same department until September 2004, where he started as Professor at DTU Informatics. From 2008 to 2013 he was Section Head for the Statistics Section of the former DTU Informatics department.

Dominik Strohmeier

Mozilla, Berlin, Germany

Dominik Strohmeier is a Senior Product Manager for Firefox Platform Metrics at Mozilla. He received his M.Sc and his Ph.D. degrees from Ilmenau University of Technology, Germany. After finishing his Ph.D., he joined the Telekom Innovation Laboratories in Berlin as a Senior Research Scientist where his main focus was set on modeling Quality of Experience for Web Services. At Mozilla, Dominik Strohmeier is responsible for informing strategic decisions for the Platform team by prioritizing performance measures through a QoE-driven approach.


  • Brockhoff, PB (2017). Applied Univariate Statistics. Chapter in Sensory Evaluation of Sound, Edited by Nick Zacharov. Taylor & Francis Books Inc.
  • Brockhoff, PB, Amorim, IDS, Kuznetsova, A, Bech, S & de Lima, RR (2016). Delta-tilde interpretation of standard linear mixed model results. Food Quality and Preference, vol 49, pp. 129-139.
  • Brockhoff, P. B., Schlich, P., & Skovgaard, I. (2015). Taking individual scaling differences into account by analyzing profile data with the Mixed Assessor Model. Food Quality and Preference, 39, 156-166.
  • Kuznetsova, A. (2015). Linear mixed models in sensometrics. PhD Thesis. Technical University of Denmark.
  • Kuznetsova, A., Christensen, R.H.B., Bavay, C. and Brockhoff, P.B. (2015). Automated mixed ANOVA modeling of sensory and consumer data. Food Quality and Preference 40, 31-38.
  • Alexandra Kuznetsova, Per Bruun Brockhoff and Rune Haubo Bojesen Christensen (2014). SensMixed: Mixed effects modelling for sensory and consumer data. R package version 2.0-8.
  • Alexandra Kuznetsova, Per Bruun Brockhoff and Rune Haubo Bojesen Christensen (2015). lmerTest: Tests for random and fixed effects for linear mixed effect models (lmer objects of  lme4 package). R package version 2.0-29.
  • Kuznetsova, A., Brockhoff, P. B., Christensen, R. H. B. (2017). lmerTest package: Tests in Linear Mixed Effects Models. Journal of Statistical Software, in press
  • Næs, P.B. Brockhoff and O. Tomic, (2010). Statistics for Sensory and Consumer Science, John Wiley & Sons.