XLSTAT: Time-Intensity Check-All-That-Apply (TCATA) Analysis in Excel

 

Background

Temporal Check-All-That-Apply (TCATA) is a sensory analysis method that extends the traditional CATA test. In a classic CATA test, a panelist checks the attributes that describe a product once, usually at the end of tasting. However, perception changes over time: some flavors appear and then fade, while others persist. TCATA records these changes by asking panelists to continuously indicate which attributes they perceive. For each combination of panelist, product, and attribute, this produces a series of binary values over the entire tasting period.

This approach matters because it reveals the dynamics of sensory perception rather than a single final snapshot. It shows when an attribute becomes significant, when it disappears, and how two products differ over time. 

 

Dataset for Example

The data in this article come from a TCATA experiment in which 50 panelists rated 6 orange juices (products) using 6 attributes over a 30-second period. To keep the example simple, the dataset has been reduced to 3 products and the 0 to 20 second window.

Each row of the file corresponds to a combination of product, panelist, and attribute. Columns D to X contain a binary vector of length 21, covering the instants from 0 to 20 seconds. A value of 1 means the attribute is selected at time t, and 0 means it is not.

These data are described in detail in Ares et al. (2016), Comparison of TCATA and TDS for dynamic sensory characterization of food products, Food Research International, 78, 148–158.

 

Setting Up

1. Select XLSTAT, then Sensory, then CATA Data, then TCATA.

2. The TCATA dialog box appears.

General tab

• Select the Binary format, because the TCATA data (columns D to X) are coded as 0 and 1.

• Select the Variable labels option: the first row of the selected TCATA data is then treated as the time vector.

• Select columns A, B, and C for the Panelists, Products, and Attributes fields, respectively.

• Enable the chart options you need. [List the chart options to select]

Running the analysis

• Click the OK button to launch the computations.

• The results are displayed on a new Excel sheet.

 

Interpreting the Results

Summary table

The first result is the summary table, which lets you verify your data. It confirms that 3 products were evaluated by 50 panelists, using 6 attributes, over a period of 0 to 20 seconds. Check these values before continuing with the analysis.

Bar charts

The bar charts show the citation proportion of each attribute for each product. They give a first idea of how the products are characterized by the different attributes.

Smoothed curves and reference lines

The smoothed curves show how the citation proportion of each attribute changes over time for a given product. The dotted lines are the significant reference lines, which represent the citation proportions of the other products grouped together for the same attribute. When a reference line is significant, the corresponding curve appears with a thicker line.

For example, for product 2, the Orange flavour attribute is significantly more present from second 7 to second 20 compared with the other grouped products. Conversely, the Off flavour attribute is significantly less present from second 14 to second 20, because its curve falls below the reference line.

Significant differences between products

This chart compares products pairwise and highlights significant differences in citation proportions over time. For example, the Sweetness attribute is significantly less cited in product 1 than in product 2 between seconds 7 and 20. In contrast, the Off flavour attribute is significantly more cited in product 1 than in product 2 between seconds 16 and 20.

Product trajectories (PCA)

The final representation is the product trajectory chart from the PCA. It offers a synthetic view of the relationship between products and attributes over time. In this example, product 2, located on the right, is associated with Orange flavour and Sweetness. Product 1 is first characterized by Sweetness and Orange flavour, then mainly by Off flavour. These conclusions are consistent with the product curves presented earlier.

Note

Check that the chart options you use match those available in your version of XLSTAT, since labels may vary between releases.

Related Articles

• XLSTAT: CATA Check-All-That-Apply Analysis in Excel

• XLSTAT: Temporal Dominance of Sensations in Excel