AS 2542.2.4:2014 pdf – Sensory analysis Part 2.4: Methodology-Duo-trio test.
B.1 Example 1: Duo-trio test to confirm that a difference exists — Balanced reference technique
Bi.1 Background
A tomato soup manufacturer would like to introduce a new and more costly low-salt formula in the hope of gaining a market advantage. Before submitting It to a consumer test in comparison with the old formula, the company wIshes to confirm that the two products can be distinguished sensorlally. The duo-trio test in the balanced reference mode is chosen because the complex flavour of the product makes It important that the assessors’ decision process is uncomplicated. The head of production Is willing to take oty a small chance of concluding that a difference exists when one does not However, because the old product is still very acceptable, he is willing to accept a greater risk of mIssing a difference that does exist.
B.1.2 Test objective
The objective is to confirm that the new product (B) can be di5tinguished from the current product (A) in order
to justify testing wIth consumers.
B,i.3 Number of assessors
To provide the head of production with substantial protection against falsely conduding that a difference exists, the sensory analyst proposes u 0,01. In order to balance the order of presentation of the samples, the analyst decides to recruit 36 assessors,
B.14 Conducting the test
Samples (54 servings of A end 54 servings of B) are prepared. Of these. 18 A samples and 18 B samples are labelled as references The remaining 36 A samples and 36 ‘B” samples are coded with unique random three-digit numbers. The entire collection of samples is then sorted into nine series, each composing four sets of samples as shown below. The fIrst serving in each set Is the relerence, designated A-REF or B-REF as the case may be:
A-REF AB B-REF AB
A-REF BA B-REF BA
Each of the four triads is presented nme times so as to cover the 36 assessors in a balanced random order.
See Figure B. 1 for the worksheet. An example of the scoresheet used is shown in Figure B.2. Bi.5 Analysis and interpretation of results
A total of 28 assessors correctly identify the sample that is the same as the reference. In Table A. 1 in the row corresponding to 36 assessors and the column corresponding to a • 0,01, the sensory analyst finds that 26 correct responses are required to conclude that a perceptible difference exists at the a – 0,01. Therefore, 28 correct responses are sufficient to conclude that the two products are perceptibly different.
B.2 Example 2: Duo-trio test to confirm that two samples are similar — Constant reference technique
8.2.1 Background
A soil drinks company wishes to make certain that a proposed new package does not alter the flavour of the beverage to a point where consumers can detect a difference. The production manager knows that it is impossible to prove that Iwo products are identical, but he wishes to make certain that only a small proportion 04 the population wil be likely to detect a difference if one exists. On the other hand, he is willing to take a fairly large ctiance of incorrectly concluding that the products are different wtien they are not, because this will only mean continuing with the satisfactory old package, pertlaps modifying the new one and then testing again.
B.2.2 Test objective
The objective is to determine if product filled and stored in the new package is sufficiently similar to product
filled and stored in the current package.
B.2.3 Number of assessors
The sensory analyst proposes to use the duo-trio-test with the current product as the constant reference because this product is well-known to the assessors, who will need no time or effort to lanuliarize themselves with its flavoi The analyst then works with the production manager to decide on the levels of risk that are appropriate for the test. It is decided that the maximum allowable proportion of discnniinators should be Pd 30 %. The manufacturer is only willing to take a ft – 0,05 chance of failing to detect that level of discriminators. The sensory analyst recruits 52 assessors for the test.