Analyzing Product and Individual Differences in Sensory Discrimination Testing by Thurstonian and Statistical Models
Total Page:16
File Type:pdf, Size:1020Kb
Downloaded from orbit.dtu.dk on: Oct 06, 2021 Analyzing Product and Individual Differences in Sensory Discrimination Testing by Thurstonian and Statistical models Linander, Christine Borgen Publication date: 2018 Document Version Publisher's PDF, also known as Version of record Link back to DTU Orbit Citation (APA): Linander, C. B. (2018). Analyzing Product and Individual Differences in Sensory Discrimination Testing by Thurstonian and Statistical models. DTU Compute. DTU Compute PHD-2018 Vol. 480 General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. Users may download and print one copy of any publication from the public portal for the purpose of private study or research. You may not further distribute the material or use it for any profit-making activity or commercial gain You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim. Analyzing Product and Individual Differences in Sensory Discrimination Testing by Thurstonian and Statistical models Christine Borgen Linander Kongens Lyngby 2018 PhD-2018-480 Technical University of Denmark Department of Applied Mathematics and Computer Science Richard Petersens Plads, building 324, 2800 Kongens Lyngby, Denmark Phone +45 4525 3031 [email protected] www.compute.dtu.dk PhD-2018-480 Summary Sensory discrimination tests are used to gain information about products by using the human senses to evaluate the samples. More specifically, a sensory discrimination study is conducted when the the aim is to investigate whether products are perceptibly different. Such studies are often considered for food, beverages as well as personal care products. An example is when a company gets a new supplier of an ingredient in one of their products. It is of high impor- tance to investigate how this change of the ingredient affects the product. Even though the chemical composition of the product changes, it does not necessarily mean that people can detect the difference. These days, people become more and more interested in how to improve their health. This is also reflected in the companies’ desire to make their products healthier without changing how the product is perceived by their customers. Therefore, it is important to con- duct sensory discrimination tests when ingredients are changed. This thesis is concerned with the analysis of product and individual differences in sensory dis- crimination testing. Sensory discrimination tests become more and more advanced raising a need for new types of analysis of sensory discrimination data. This thesis contributes with the development of Thurstonian models and how these can be aligned with well-known statistical models. Generalized linear mixed models are used in many applications. However, it is not common to consider such compli- cated models when considering sensory discrimination tests. Actually, sensory discrimination tests are often analyzed by too simplistic methods, ignoring im- portant variables, such as individuals, that affect the results of the analysis. One focus of this project is to propose a way to incorporate such effects in the models when analyzing data from sensory discrimination studies. These mod- ii els, including random effects, are called Thurstonian mixed models. Considering generalized linear mixed models for sensory discrimination studies opens up for many possibilities. It becomes possible to gain information about the individu- als, the so-called assessors, as well as making more proper conclusions regarding the products. Moreover, the estimates of product and individual differences are obtained on the d-prime scale. Often multiple sensory attributes are considered in a discrimination study. These can be analyzed individually by the Thurstonian mixed models we are introduc- ing. This thesis is presenting a multivariate analysis to gain knowledge about the product and individual differences across the sensory attributes. This is achieved by analyzing the product and individual differences, on the d-prime scale, by principal component analysis. Sensory discrimination tests are sometimes conducted to investigate the perfor- mance of sensory panels or to compare different laboratories. In such tests, mul- tiple d-prime values can be obtained. For sensory discrimination tests, which lead to binomially distributed responses, we propose a new test statistic for the comparison of multiple d-prime values. The test statistic we suggest is an improved way of analyzing multiple d-prime values compared to a previous suggested test statistic. Resumé Sensoriske diskriminationstest bliver brugt til at opnå information om produkter ved at bruge de menneskelige sanser til at evaluere prøverne. Mere specifikt bliver et sensorisk diskriminationstest brugt når det ønskes at undersøge om produkter er mærkbart forskellige. Sådanne studier bliver ofte brugt til fødevarer, drikke- varer og produkter til personlig pleje. Et eksempel er når en virksomhed får ny leverandør af en ingrediens i et af ders produkter. Det er vigtigt at undersøge hvordan denne ingrediensudskiftning påvirker produktet. Selvom den kemiske sammensætning af produktet ændres betyder det ikke nødvendigvis at menne- sker kan opdage forskellen. For tiden bliver folk mere og mere interesserede i hvordan de kan forbedre deres helbred. Dette afspejles også i virksomhedernes ønske om at gøre deres produkter sundere uden at ændre hvordan produkterne opfattes af deres forbrugere. Det er derfor vigtigt at lave sensoriske diskrimi- nationstest når ingredienser udskiftes. Denne afhandling beskæftiger sig med analysen af produkt og individ forskelle i sensoriske diskriminationstests. Sensoriske diskriminationstests bliver mere og mere avancerede hvilket øger be- hovet for nye typer af analyser af data fra sensoriske diskriminationstests. Denne afhandling bidrager med udviklingen af Thurstonske modeller og hvordan disse kan kombineres med velkendte statistiske modeller. Generaliserede lineære mixe- de modeller bliver brugt i mange anvendelser. Imidlertid er det ikke almindeligt at betragte sådanne komplicerede modeller når data fra sensoriske diskrimina- tionstests betragtes. Faktisk bliver data fra sensoriske diskriminationstests ofte analyseret med for simple modeller som ignorerer vigtige variable, som indivi- der, hvilket påvirker resultaterne af analysen. Et fokus for dette projekt er at foreslå en måde at indkorporere sådanne effekter i modellen når data fra sen- soriske diskriminationsstudier bliver analyseret. Disse modeller, som medtager iv tilfældige effekter, kaldes Thurstonske mixede modeller. At betragte generali- serede lineære mixede modeller for sensoriske diskriminationsstudier åbner op for mange muligheder. Det bliver muligt at få information om individerne, de såkaldte ’assessors’, såvel som at drage mere passende konklusioner omkring produkterne. Derudover er estimaterne af produktforskelle og individforskelle på ’d-prime’ skalaen. Ofte bliver mange sensoriske egenskaber betragtet i et diskriminationsstudie. Disse kan analyseres enkeltvis ved brug af de Thurstonske mixede modeller vi in- troducerer. Denne afhandling præsenterer en multivariat analyse for at få viden om produktforskelle samt individforskelle på tværs af de sensoriske egenskaber. Dette opnås ved at analysere produktforskelle og individforskelle, på ’d-prime’ skalaen, ved ’principal component analysis’. Sensoriske diskriminationstests bliver indimellem udført for at undersøge præ- stationen af sensoriske paneler eller for at sammenligne forskellige laboratorier. I sådanne tests er det muligt at få mange ’d-prime’ værdier. For sensoriske diskriminationstests, som giver binomialfordelte responsvariable, foreslår vi en ny teststørrelse til at sammenligne adskillelige ’d-prime’ værdier. Teststørrel- sen vi foreslår er en forbedret måde at analysere mange ’d-prime’ værdier på sammenlignet med en tidligere foreslået teststørrelse. Preface This thesis was prepared at Technical University of Denmark, Department of Applied Mathematics and Computer Science, Statistics and Data Analysis sec- tion, in partial fulfillment of the requirements for acquiring the Ph.D. degree in Applied Statistics. The project was funded by the Technical University of Denmark and Unilever U.K. Central Resources Limited. The project was su- pervised by Professor Per Bruun Brockhoff. Occasionally, Rune Haubo Bojesen Christensen has been a co-supervisor. The thesis deals with the analysis of product and individual differences in sen- sory discrimination testing. Sensory discrimination testing is a type of testing used in sensory science, where people are used as the measurement instruments. The main focus is developing methods aligning Thurstonian methods with sta- tistical models. The thesis consists of three research papers and a book chapter. An introductory part gives an overview of the thesis. Background and aspects that were not considered in the papers are considered in the thesis. Lyngby, 01-July-2018 vi Christine Borgen Linander List of contributions This thesis is based on the following research papers and a book chapter [A] Linander, C. B., Christensen, R. H. B., Cleaver, G. and P. B. Brock- hoff (2018) Individual differences