Dataset for Example

The dataset consists of 5 chocolate samples, with Chocolate0 as the reference product. Each sample was tested 4 times and each time the assessor gave his opinion on whether or not the sample tested was similar to the reference product and on the certainty of his choice. 

The answers are in the format 1, 2, 3 and 4 symbolizing respectively a sample judged to be similar to the reference product with certainty, a sample judged to be similar without certainty, a sample judged to be different without certainty, and a sample judged to be different with certainty.

 

Setting Up

• Open XLSTAT.

• Click on the Sensory / Discrimination tests menu and select the R-Index feature.

• The dialog box appears.

• In the XLSTAT interface, select Raw data for Data type. Select the Raw data.

• Select the Sample identifiers.

• Click on the OK button to launch the computations.

• The results are displayed on a new Excel sheet.

 

Interpreting the Results

The first important result of the R-Index feature is the response matrix table, which summarizes the information in the dataset. In this table, we can see strong differences in the judgment between Chocolate4 and Chocolate0, which corresponds to the reference product: Chocolate0 was judged to be similar to the reference product with certainty on 3 occasions and similar but without certainty on 1 occasion, while Chocolate4 was judged to be similar to the reference product without certainty on 1 occasion, different without certainty on 1 occasion and different with certainty on 2 occasions.

We then analyze the various R-Index calculated from this response matrix. Firstly, we note that the R-Index for Chocolate4 is 0.969, a value very close to the maximum value of 1, which symbolizes a sample that is totally different from the reference product. The p-value associated with this sample confirms that, at the 5% threshold, Chocolate4 is different from the reference product. It is the only sample in the dataset that leads to this conclusion. Of the other samples tested, Chocolate2 is the most similar to the reference product.

In conclusion, using R-Index analysis in this example not only identified a product that was significantly different from the reference product, but also established a ranking of the products most similar to the reference product.