Experimental testing of the plan and collection of responses y
Validation of the robustness of an analytical method by design of experiments

For each experiment in the design, the levels of the various factors are finally converted back into the real unit u _i so that the tests can be carried out on a practical level. For example, in the case of the first experiment in the X matrix, the X ₁ variable associated with temperature must be adjusted to level -1, which corresponds to a setpoint equal to 24°C ( see figure Area of study of the temperature factor of a CLHPX ₁ furnace as part of a robustness study).

Once all the experiments in the design have been carried out in practice, and the corresponding responses y _i have been measured, the respective effect of each factor on the response is calculated using the box formula :