AS 2542.2.5:2014 Sensory analysis
This example is similar to Example 1 except that there are two different alternative sweeteners, ‘(not A) 1 ’ and ‘(not A) 2 ’. After evaluating reference sample ‘A’ (sucrose), the 20 assessors are each presented with a test series containing five samples of ‘A’, five samples of ‘(not A) 1 ’ and four samples of ‘(not A) 2 ’.
The results from a test such as this may be used to provide several comparisons between the three sweeteners as follows:
(a) To determine whether the three samples ‘A’, ‘(not A) 1 ’ and ‘(not A) 2 ’ are recognized as being significantly different. In this case the results are interpreted in a similar way to Example 1 (see Paragraph B1), except that the total is extended to all six boxes and there are 2 degrees of freedom (df). Yates correction is not required. Calculated 2 χ is 4.65, which is significantly less than the critical value of 5.99 for a 5 percent risk in a bilateral hypothesis. This indicates that the ‘three sweeteners ‘A’, ‘(not A) 1 ’ and ‘(not A) 2 ’ are not perceived as being different.
(b) To determine whether ‘A’ is recognized as being significantly different from the others as a group (‘(not A) 1 ’ + ‘(not A) 2 ’). In this case the results are interpreted as shown in Example 1 (see Paragraph B1) and Example 2 (see Paragraph B2), with the responses for ‘(not A) 1 ’ and ‘(not A) 2 ’ being combined as ‘not A’.
(c) To determine whether the two samples ‘(not A) 1 ’ and ‘(not A) 2 ’ are recognized as being significantly different from one another. In this case the results are interpreted as in Example 1 (see Paragraph B1) with column ‘A’ deleted.
(d) In the same way as Item (c), ‘A’ may be compared with ‘(not A) 1 ’ (column ‘(not A) 2 ’ deleted), or ‘A’ with ‘(not A) 2 ’ (column ‘(not A) 1 ’ deleted).
NOTE: There is no point in carrying out the calculations associated with Items (b), (c) and (d) if, as in the case of the particular set of results given in this Example, the calculations carried out for Item (a) show that there is no significant difference between the three samples under test.