Temporal precision and resolution of TDS and TCATA used with consumers

Abstract

Temporal Dominance of Sensations (TDS) and Temporal Check-All-That-Apply (TCATA) make it possible to capture temporal changes in the perception of sensory attributes within a product. However, temporal measurements bring new sources of uncontrolled variability that can make them imprecise, particularly with consumer panels. This article aims to demonstrate the importance of considering this imprecision when interpreting temporal data. First, the levels of imprecision due to non-repeatability or disagreement were assessed. The precision of TDS and TCATA was shown to be low, mainly due to disagreement caused by high heterogeneity in total durations of perception. Then, a new framework of analysis has been introduced. It allowed testing of differences in the median times of selection or unselection of attributes. A new representation of TDS and TCATA data has been proposed to visualize and objectify uncertainty in periods of perception. The conclusions obtained with the new framework were compared to those obtained with the usual analyses based on the evolution of citation rates over time. It was shown that the interpretation of TDS/TCATA curves and difference curves can be tricky without guidelines regarding the minimum size of differences that can be considered significant and interpretable. Finally, it was shown that the temporal resolution of the methods was about 10% of the total duration of the perception. These findings question the ability of TDS and TCATA to capture subtle temporal differences. Therefore, recommendations were made for interpreting temporal data in a way that robust conclusions can be drawn.