Researchers have argued that the limitation of Free Choice profiling is the requirement for assessors to use scales which can lead to inconsistencies in their scoring. In addition, presentation of samples in Free Choice is sequential and monadic whereas in Flash all samples are presented side by side for ranking. Both of these factors can make Flash more sensitive to sample differences and show Flash to be ‘better’ at recovering the same sample space as QDA type profiling.
Our research found that whilst this was the case, the difference from Flash was small and assessors preferred the Free Choice task. Ranking samples sets of six or more gets quite tiring and would be almost impossible for strongly flavoured samples. For skin creams you soon run out of space!
Data analysis for both Flash and Free Choice is complex for the non-statistician although there are good software tools available such as XLStat. Both suffer the problem of interpretation of the resultant sample space due to idiosyncrasies in use of language.
The Napping task is well received although we found that trained assessors found adjusting to the task quite difficult. Untrained respondents can default to grouping the samples primarily on the basis of appearance but asking assessors to make separate Nappes for appearance, aroma, flavour and texture can help to get round this. Napping is ideal for screening product sets prior to either consumer research or other forms of profiling. The lack of descriptive information makes it application very limited where the need to understand the rationale behind the differentiation of the product groups is required.
Think about your objective particularly how much description you need. Napping is particularly suitable for evaluating the effects of brand and other non-sensory factors. Think too about your samples…Napping and Flash require all samples to be presented together so may not be suitable for hot or frozen foods. Also about your assessors…all of these techniques work well with trained and untrained assessors but you may get more meaningful descriptors from a group who have some familiarity with the product set. Having said that your objective may be see how your consumers view the world and these techniques work well. You can try different techniques to elicit vocabulary in Free Choice and Flash…Repertory Grid is one. For all techniques, remember that consumers especially may use non-sensory factors to discriminate the products.
Essentially it’s lack of detail. The outcome of all of the Rapid Methods is a map or sensory space. There is no comparison of specific attributes; you won’t be able to draw spider profiles of histograms comparing samples for the intensity of specific attributes. The methods don’t work therefore for investigational studies to determine in detail the impact of ingredients and processes on specific sensory characteristics although they will tell you about gross changes in sample groupings as a result of these changes. They are also not suitable for Sensory Claims support. The panel are not trained in the use of their language or criteria so the data are very ‘in the moment’ and context (or sample set related). It is difficult to remove samples from the analysis and probably more importantly, almost impossible to add samples later. This makes the methods unsuitable for storage trials or Quality Assurance purposes.
The Rapid Methods are however, fantastic for selection of samples for QDA trials or for consumer testing. They help us see how consumers discriminate samples and the types of descriptors and criteria they use to do so.