Generation of experimental designs
FIZZ includes a very powerful tool for the generation of sample presentation designs (evaluation orders). Depending on the type of test used in the session or the questionnaire, the system proposes the most suitable designs and generates automatically the design chosen, and the sample blinding codes.
A unique characteristic of FIZZ is that a new design is obtained at each design generation. The algorithms used generate a different design every time, corresponding to your choices.
For the profile type tests, FIZZ builds Williams Latin Squares (which balance the position and carry-over effects), designs with products split over several sessions, Balanced Incomplete Blocks, and combinations of these designs when you plan repetitions of the evaluation.
Example of a Williams Latin Square, including the sample codes and identification, shown in the evaluation order: