Experimental designs are used to generalise the theoretical
effects of variables (Highhouse, 2009). As we aim to explore
the key factors of response options after NCR (e.g. brand
strength) on PPIs, we decided to conduct an experiment.
Therefore, we applied a 2 (brand strength: strong/weak) 3
(response option: brand/consumer/no response) 2
(reference option: with/without reference) fractional factorial
between-subject design. In this experiment, not all the groups
could be filled (for example, in the non-response groups, it is
not possible to include a reference).