Figure 3.

Correspondence analysis with respect to instrumental variables. The between-beverage analysis applied to the data where the "individual" effect has been removed, shows a structure associated with the effect of red wine on the first discriminating axis. Two factors were successively included in the analysis: "individual" effect (6 modalities) and "beverage" effect (4 modalities). The "beverage" effect was taken positively while the "individual" effect was controlled. The analysis is focused on the 1 h-time point.