XLSTAT: Sample size calculator for sensory discrimination tests

 

Data for Example

In this tutorial, we want to perform a triangle discrimination test to determine if a difference is perceived between two cookies. For reasons of time and cost, we want to limit the number of respondents while ensuring that we are in good statistical conditions. Thus, we will look for the number of respondents necessary to obtain a power of the test equal to 0.9.

Setting up

Once XLSTAT is activated, select the Sensory menu/Discrimination tests/Power-Sensory discrimination tests.

The dialog box appears.

In the General tab, we enter the analysis parameters: 0.05 for the first type error, 0.9 for the power, and d-prime=0. Click Ok. 

Interpreting the Results

In order to obtain results with a power of 0.9, it will be necessary to summon a certain number of respondents, depending on the minimum proportion of correct answers. For our test, we expect to obtain at least 50% of correct answers, which according to the table and graph below indicates that 53 respondents are needed.

 
For more details on sensory discrimination tests, please refer to the Designs for sensory discrimination tests as well as the Sensory discrimination tests.