Rapid sensory profiling methods for wine : workflow optimisation for research and industry applications

Brand, Jeanne (2019-04)

Thesis (PhDAgric)--Stellenbosch University, 2019.

Thesis

ENGLISH ABSTRACT: Descriptive sensory analysis techniques are widely used and trusted methodologies. Due to time and cost constraints, the demand for cost-effective methods for profiling is growing rapidly in food and beverage industries including the wine industry. A number of rapid methods have been tested and validated for various food products. However, further work is needed to identify and address limitations of specific rapid methods, especially reference-based methods, when evaluating complex matrices such as wine. The majority of studies employed novice consumers or trained consumers as judges. The wine industry has an advantage over most food industries with: (1) product experts who can serve as judges and (2) having an extensive lexicon in the form of aroma wheels available that can be used as check-all-that-apply (CATA) questions. The objective of this study was to identify cost-effective, rapid sensory methods that can be used for wine profiling by researchers and the wine industry alike. Furthermore, the study aimed to optimise the identified methods and to propose workflows that include sensory methods and statistical procedures suited for wine sensory analysis applications. Four rapid methods were compared to descriptive analysis (DA). The methods tested were CATA, rate-all-that-apply (RATA), Napping, and sorting. Results obtained for the rapid sensory method and DA were similar. It can therefore be concluded that rapid methods are suitable for the sensory evaluation of wine. Industry professionals can therefore be used as sensory judges, and can use a pre-determined lists of attributes as verbalisation tools. CATA and sorting provided the highest quality profiles with the best discrimination between products. Sorting highlights similarities and differences whereas CATA provides more detailed descriptions. In addition, these two methods were found to be easier than rate-all-that-apply (RATA) and Napping to use. Pivot profile (PP), a reference-based method, was validated against a CATA variant, namely frequency of attribute citation (FC). It was concluded that PP should be used with caution because the choice of pivot on the sensory space could have an influence. This method could, however, be useful when direct comparisons between samples are required, such as benchmarking. In addition to sensory method development, a number of statistical procedures were also proposed to assist with the interpretation of rapid method data. A workflow to calculate drivers of quality and a strategy to calculate confidence ellipses for PP data were developed. This study highlights the importance of selecting a fit-for-purpose method. The objective of the experiment being conducted, along with practical restrictions should be taken into account when deciding which method to use.

AFRIKAANSE OPSOMMING: Beskrywende sensoriese evaluerings tegnieke word algemeen gebruik, en word as betroubare metodes erken. Weens tydsdruk en kostebeperkings is daar ‘n toenemende aanvraag na vinniger sensoriese profileringsmetodes wat gebruik kan word in die voedsel- en drankbedrywe insluitende die wynbedryf. Verskeie vinnige metodes (“rapid methods”) is reeds getoets en gevalideer vir die sensoriese analise van ‘n verskeidenheid voedselprodukte. Verdere navorsing is egter nodig om voordele en nadele van spesifieke metodes te identifiseer en aan te spreek. Dit is veral die geval wanneer verwysings-gebaseerde metodes gebruik word om komplekse matrikse soos wyn te evalueer. Die meeste studies wat reeds hieroor gedoen is, gebruik verbruikers as sensoriese beoordelaars. Die wynindustrie het ‘n voordeel bo die meeste ander voedselindustrieë in die sin dat: (1) produkkenners as sensoriese beoordelaars kan dien en (2) omvattende beskrywings van die sensoriese eienskappe van wyn reeds in die vorm van, onder andere, geurwiele beskikbaar is. Laasgenoemde kan as merk-alles-wat-van-toepassing-is (“check-all-that-apply” of CATA) vraelyste gebruik word. Die doel van hierdie studie was om koste-effektiewe, vinnige sensoriese metodes te identifiseer wat vir wynprofilering deur navorsers en die breë wynbedryf gebruik kan word. Dit het ook ten doel gehad om laasgenoemde metodes te optimaliseer en ʼn getoetste werksvloei voor te stel, wat sensoriese metodes en statistiese prosedures insluit. Vier vinnige metodes is vergelyk met beskrywende sensoriese analise (“descriptive analysis” of DA). Die metodes is merk-alles-van-toepassing (“check-all-that-apply” of CATA), gradeer-alles-van-toepassing (“rate-all-that-apply” of RATA), Napping, ʼn spesifieke variasie van projeksiekartering (“projective mapping” of PM), en sortering. Die vinnige sensoriese metodes en beskrywende analise het soortgelyke resultate opgelewer. Vinnige metodes blyk dus geskik te wees vir die sensoriese analise van wyn wanneer vooraf-opgestelde lyste met sensoriese eienskappe gebruik word as verbaliseringstap en produkkenners uit die industrie as beoordelaars optree. Die hoogste kwaliteit profiele, wat die beste tussen produkte kon onderskei is met behulp van CATA en sortering verkry. Sortering het ooreenkomste en verskille tussen produkte uitgelig, terwyl CATA meer gedetailleerde beskrywings opgelewer het. Volgens die paneellede is beide hierdie metodes makliker om te gebruik as RATA of Napping. Draaipuntprofilering (“pivot profile”, PP), ‘n verwysingsgebasseerde metode, is gevalideer en vergelyk met ‘n variasie van die CATA metode, naamlik frekwensie van eienskap-aanhaling (“frequency of attribute citation”, FC). Weens die invloed van die keuse van die draaipuntmonster op die sensorieseruimte met PP, moet hierdie metode met versigtigheid gebruik word. Hierdie metode kan wel waardevol wees wanneer ‘n direkte vergelyking tussen produkte verlang word, byvoorbeeld wanneer een produk teen ‘n ander een wat as maatstaf dien, vergelyk word.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/106061
This item appears in the following collections: