A guide to taster training

 

SensCheck Sensory Analysis Software

Having selected the right people for the job, the quality of training used to develop the competence of your professional sensory assessors is the single most important factor in the success of your professional tasting initiatives.

 

Types of taste tests

The type of flavour training you do should be shaped by what you need your tasters to do once you’ve trained them. Common tasks include:

  • Grading the quality of commercial products and trial samples
  • Assessing the degree to which products conform to pre-defined standards
  • Describing the flavour of your products (qualitatively or quantitatively)
  • Determining the degree of similarity between two or more products or batches of products
  • Assessing product shelf life
  • Assessing the quality of raw materials or ‘in-process’ products
  • Screening packaging materials for the presence of taints

 

Types of sensory assessors

Sensory evaluators in a taste booth

Sometimes the sensory task is straightforward. Two products have to be compared to descide if one is different from the other. In such a case it is possible to use sensory assessors who have little formal training.

They need to be instructed in the details of the test protocol, how to assess the samples and how to fill in the evaluation form.

However, if the assessors are asked to do something more complex, such as evaluating a sample and providing a rating for the intensity of 50 or more pre-defined flavour attributes, a comprehensive training and assessment programme is needed.

The main types of tests you may have to train professional sensory assessors for are: 

    • Difference tests, such as the duo-trio, triangle, A-not-A, and difference from control tests
    • Scaling and rating tests, such as ranking, rating and rank-rating tests
    • Threshold determination tests, such as the method of constant stimuli, method limits, and dilution techniques
    • Descriptive profiling techniques, such as flavour profiling, quantitative descriptive analysis, and Flash profiling

 

Selection of sensory assessors

When novices are presented with two samples, one of which has a dominant flavour, they can usually tell there is something different about them. But, if asked, they are likely to provide an ambiguous description of the difference in flavour between the two samples, and to describe the dfference using terms that are not easily understood by everyone.

In contrast, an experienced sensory assessor can identify and name (using widely understood, unambiguous terms) as many as 100 different characteristics in a given product type. Furthermore, they can give an estimate of the intensity of each flavour note. To the untrained sensory assessor, this looks an impossible task. However, to those who are trained in professional sensory assessment of product quality, it is relatively straightforward.

To attract, train and retain trainee tasters with the best aptitude and enthusiasm for professional tasting, good recruitment and training practices are needed.

Potential trainee sensory assessors shouldn’t be recruited at random. They should be selected from a larger pool of talent. 

What types of people make good sensory assessors? Firstly, if the product contains alcohol they must be of legal drinking age; this varies from place to place and from time to time. They should not have a dislike of the types of products which will be routinely tested. They may be regular consumers of the product, but not necessarily so. They should have an interest in flavour and food – this often means they are not being overly cautious or restricted in their food choices. They should have good mental abilities, and concentration skills. They should be in good health and, in particular, not predisposed to, or suffering from, health problems or major impairments of their senses such as anosmia (smell blindness) and agueusia (taste blindness). They should not be taking medicines that might impair their sensory capabilities. They should not suffer from allergies or sensitivity to common sample components, such as sulphur dioxide or quinine.

Results of selection questionnaire for professional tasters

Questionnaires can be a useful aid in the selection of professional tasters.

They should:

  • Encourage the applicant to become a trainee taster
  • Provide a basis for the selection of sensory assessors based on their level of self-declared interest and motivation
  • Provide a basis for their selection based on some aspect of skill or ability
  • Provide a first level screen to identify those whose health might be placed at risk as a result of their participation in professional tasting activities.

 

Responses to recruitment questions should be quantitative, rather than being in the form of comments or yes/no answers. This allows taste panel leaders to make decisions concerning candidate selection to be based on objective rather than subjective information.

Having identified suitable candidates through use of a questionnaire, an interview can be used to choose among the shortlitsed candidate tasters. Interviews can be used to assess ‘soft’ personal factors, including an ability to work as part of a group, attitude, and physical and intellectual ability. On the downside, it can be difficult to maintain objectivity in this type of selection.

 

Screening of sensory assessors

Screening helps trainers identify those with inherent defects in their sensory acuity (including anosmia and agueusia). It also allows the likely response of candidate tasters to training to be predicted, identifying those individuals most likely to perform best after training. For this latter reason, screening should not focus on current competence. For such tests to have the predictive power expected of them, the way in which the training is carried out during the screening phase must be comparable to how it will be done during the training phase. 

Training of sensory assessors

Taste panel evaluation of beverages

Approaches used to train professional tasters, especially those responsible for evaluating consumer products manufactured on an industrial scale, have recently undergone something of a revolution. This has been made possible by developments in our understanding of the determinants of success in sensory assessments and how the process of learning takes place for those involved. Unstructured, free-form ‘learning’ activities have largely been replaced by outcomes-based modular training programmes.

Historically, factories making all sorts of consumer products relied on the views of a single expert taster in assessing the quality of their products. This expert was usually a senior technologist, or in smaller companies, the owner. As a form of external check, a consultant was often retained: their role was to provide an independent expert view from time to time. The role of the expert is somewhat diminished in modern production operations. The first reason for this is physiological. All of us have one or more flavours that we are unable to detect at concentrations that other people can detect easily. This genetically determined ‘smell blindness’ is called anosmia. For most odours we have a 3% chance of being blind to any individual chemical. When tasting products as a group, this is not too important. If one of us is blind to a flavour such as diacetyl, each of the others still has a 97% chance of finding it. Of course whether they are able to do so depends on whether we have selected talented individuals and trained them for the job. Anosmia is most important in the case of single expert assessors, and in this regard it is essential that the expert is aware of their own strengths and weaknesses with respect to detection of flavours.

The second reason is psychological. Everyone is susceptible to bias in one form or another. Our judgements are affected by outside influences. These include the words and actions of others, off-putting sounds, extraneous odours etc. Most sensory test methodologies aim to minimize bias by ‘dilution of the effect. It is difficult however to eliminate bias in the case of a single assessor. But with a group of tasters, the effect of any single bias can be reduced, since each individual bias will affect only one or two assessors.

Both the present and future of professional quality assurance (QA) and quality control (QC) tasting lies in the use of flavour panels – groups of experts – rather than single experts. Using groups we can derive more information from assessments, and increase the precision of our measurements. Panels typically have 8-10 members, drawn from a pool of 12-15 trained assessors.

 

Differences in sensitivity to different flavours among sensory assessors

Evaluation of whisky

When it comes to perception of odours, tastes and mouthfeel characteristics, each of us lives in a world of our own. There is no such thing as the 'average assessor'.

Rather than fight these differences in sensitivity among individuals it is best to embrace them. Professional tasting is 'crowdsourcing' in action. It is the average response of a group of professional tasters that we're interested in.

The individual taster attracts our attention only insofar as we need to be sure that they are behaving as they usually do, and performing to their full potential.

There is a second use to which such differences in sensitivity can be put. When we have assessors who are anosmic (blind) to specific flavour compounds we can use their response to a particular sample to help confirm or deny a hypothesis concerning the cause of an individual flavour note.

For example, if the panel has scored a 'diacetyl-like', 'buttery' note in a product, yet two people who are known to be blind to diacetyl (2,3-butanedione) have detected and scored this character, it clearly cannot be caused by diacetyl.

A different chemical must have givien rise to this flavour note.

 

Flavour standards and their role in development of trainee tasters

Given the flavour complexity of most foods and beverages it's not too surprising that until relatively recently, most quality control programmes in manufacturing sites focussed on grading of product quality rather than describing the details of product flavour. However, over the past 20 years this has changed, and the rate of change has accelerated greatly in the last five years.

AROXA beer flavour information cards

Historically, lexicons of flavour attributes - such as the 'Aroma and Flavour Lexicon for Sensory Evaluation' published by the American Society for Testing and Materials in 1996 - have referenced other foods, beverages, plants, animals and objects when trying to explain the nuances of particular flavour notes. For example:

  • Acetic - "Aromatic characteristic of white vinegar" - reference: white vinegar, 5% acidity
  • Acrid - "A burnt, harsh aromatic, often associated with burnt wood or smoke" - reference: charred bacon, burnt pine or pine needles
  • Decay / putrid - "Flavor associated with aroma of decaying plants or animals" - reference: Liederkranz cheese at its ripe stage
  • Feathers / wet poulty - "An aroma reminiscent of wet poultry feathers" - reference: processed poultry products (lunchmeat, franks)
  • Malty - "Sweet slightly fermented or sour grain note associated with freshly kilned malt" - reference: malty extract, Grapenuts
  • Violet - "A floral aroma note associated with violets" - reference: petals from 10 crushed violets in 25 ml of red wine

 

While such descriptions and references have been an enormously important stepping stone to take us to where we are today, there are clearly some difficulties with this sort of approach. While charred bacon, burnt pine needles, ripe Liederkranz cheese, Grapenuts and violet petals might be readily available to some, in many parts of the world it would be simply impossible to get hold of them. It is also clear that there is significant potential for lot to lot variability - the qualty of crushed violets will vary year to year, and within a year, in addition to differering depending on the variety grown and the place they were grown.

Catty flavour icon

The same document provides examples of a better approach to attribute training. For example:

  • Almond, nutty - "Aromatic of roasted almonds that is not cherry-like" - reference: benzaldehyde
  • Banana - "Aromatic characteristic of ripe bananas" - reference: isoamyl acetate
  • Bell pepper - "Aromatic associated with green bell peppers" - reference: 2-isobutyl-3-methoxy pyrazine
  • Cabbage water - "Aromatic associated with boiled, cooked cabbage" - reference: dimethyl disulphide
  • Catty - "Aromatics associated with blackcurrant leaves, ribes, tomato plants, oxidized beer, cat urine, tom-cat spray" - reference: p-methane-8-thiol-3-one

 

It can be seen that the use of chemical references changes things substantially. Now it is possible to define what we should be smelling or tasting to perceive the flavour attribute. Notwithstanding the significant improvement on the use of foods, beverages, plants, animals and objects to illustrate flavour notes the use of chemicals brings, there are three issues to appreciate:

Chemical structure of p-methane-8-thiol-3-one

  • 'Chemically pure' chemicals obtained from chemical suppliers, and even flavour houses, are generally not 'sensory pure'. They may have a chemical purity in excess of 99%, but if that contaminant present at 1% is 100-times more flavour-active than the parent chemical it will have a significant influence on the overall flavour note generated on addition to the sample.
  • Pure chemicals in their 'neat' form require specialist facilities and staff to handle them. Generally they are dangerous, representing a hazard to health, facilities or the environment. Handling chemicals such as hydrogen sulphide (H2S - the compound responsible for the boiled egg sulphury flavour in many products) is not to be taken likely. The consequences of getting things wrong can be fatal!
  • When handling pure chemicals it is virtually impossible to contain their odour, no matter how efficient the extraction hood used is. Trace quantities will enter the air of the laboratory and inevitably find their way to the taste room thus compromising the assessments made by tasters.

 

The solution to these issues was arrived at in 1990 when Cara Technology founder, Dr Bill Simpson, at the time working for the Brewing Research Foundation in England, developed a method for stabilizing extremely small quantities of ultra-pure flavour chemicals. Rather than having to handle concentrated dangerous chemicals taste panel leaders could now add specific flavour notes to their products in precise quantities by adding unit-dosed food-grade powders.

This approach has been used to train and assess the performance of many thousands of beer tasters over a period of close to 20 years. More recently, the approach has been adopted in other industries. AROXA flavour standards include almost 80 flavour notes, with many more currently in development.

 

Guidlines for selection and training of sensory assessors

There are many useful guides available to assist panel leaders in the training of their tasters. These include some useful ISO standards:

  • ISO 8586 - 1 - Sensory analysis - General guidance for the selection, training and monitoring of assessors - part 1 - selected assessors
  • ISO 8586 - 1 - Sensory analysis - General guidance for the selection, training and monitoring of assessors - part 2 - experts
  • ISO 13299 - Sensory Analysis - Methodology - General guidance for establishing a sensory profile

 

There are also many helpful publications, including:

  • Guidelines for the selection and training of assessors for descriptive analysis - Guideline 37 (CCFRA, 2002)

 

While such reference works are an an invaluable aid to the process of training assessors there remains a lot for the panel leader to work out for themselves. To simplify this process, and make taster training easier for those who may not have decades of training and experience we developed software to help. Our SensCheck software provides templates forms, workflows and reprts to help you selecte, screen and train your assessors. Find out more here.