Performance of beer sensory panels: A comparison of experience level, product knowledge, and responsiveness to feedback calibration

Abstract

This study investigated the effect of level of sensory experience and type of product knowledge on panel performance during a descriptive analysis. Additionally, the applicability of the feedback calibration method (FCM) was investigated for use with different types of panels. A well‐established expert panel (professional product specific knowledge) was compared to (a) a newly trained product‐specific panel (interest based product specific knowledge) and (b) a well‐established general panel (general product knowledge). Results showed that the overall product sensory profiles generated by the three panels were similar. However, analysis of the panels' repeatability, discriminability, attribute understanding, and use of scale displayed that the product‐specific panel performed more similar to the professional beer panel and was therefore superior compared to the general panel. The applicability of FCM was most effective for the newly trained panel, compared to the more extensively trained panel. Practical applications Sensory descriptive analysis is one of the basic tools in the field of sensory science. However, recent focus in research has been to investigate faster and more cost‐efficient methods, resulting in a decrease in research on the methodology of classical sensory descriptive analysis. The methodology of descriptive analysis still has many unanswered questions regarding panel selection and training. The results of the current study indicate that product specificity is more important than sensory experience when studying complex foods like beer. More attention should therefore be applied to type of panel, when conducting descriptive analysis of complex products. Furthermore, the results show that the feedback calibration method (FCM) is more advantageous for training of panels with less sensory experience, though beneficial for all types of panels. The general guideline is therefore to include FCM as part of the training procedure for all type of panels.

A Comparison of Two Sensory Panels Trained with Different Feedback Calibration Range Specifications via Sensory Description of Five Beers

“…Nine pilsner-type beers were included in the present study (see Table 1a for product name, sample code, beer type, and ethanol content (ABV)). All samples were used in the training sessions, but only five of the samples were used in the final evaluation DA sessions (Tuborg, Carlsberg, Jacobsen, Tuborg SL, and Heineken), as they were shown to exhibit diverse sensory profiles [9]. The same five samples were also included for the six training descriptive analysis sessions (TDA 1–6), while the remaining four samples were included in the training sessions to highlight and represent specific attributes.…”

“…For both panels, ballot training was used with predefined attributes and references. The attributes, references, and target scores were established in our previous studies [9]. To assist panelists in establishing a framework for each attribute, reference beers were used during training to establish minimum and maximum intensities for each attribute (Table 1b).…”

“…Up to 130 h of training and more are reported [6], depending on the chosen method, the complexity of the product, the number of attributes, and the test sensitivity requirements [1,2,3,4]. Studies have shown that different factors related to the training step can influence the performance of a sensory panel [9,10,11,12,13,14,15]; these include, among others, the panel’s sensory experience, product knowledge, and product involvement. During DA training, a common vocabulary and reference frame are established, together with procedures for evaluation and training in the use of the scale.…”

“…Our previous paper [9] and papers by the developers of the method have investigated the implications of using FCM. Findlay et al [39] investigated the suitability of FCM for sensory analysis training and found it a useful and effective training method.…”

“…They concluded that both panels showed similar levels of proficiency and improvement (no difference in training time) and that FCM therefore is a useful tool to improve panel performance in descriptive analysis. Elgaard et al [9] also found that FCM improved panel performance and concluded that FCM was most efficient for a newly trained panel compared to an extensively trained panel.…”