Connecting the dots: A candy comparison of the free-linking task to the projective mapping task (Video abstract)

The free-linking task has been recently proposed as a new method for rapid similarity testing, in which participants are asked to connect those pairs of samples which they consider similar. Previous results showed that both methods provided highly comparable consensus graphics in the first two dimensions, but free-linking provided more discrimination in dimensions three and four. In this work we compared, for the first time, the free-linking task to the well-known projective mapping task.

Popular candies (K = 9), see Fig. 1, were analyzed via their images, using Projective Mapping by N=298 subjects aged between 4 and 66, and using Linking by N=189 subjects aged between 6 and 56. Data from Projective Mapping were analyzed using Multiple Factor Analysis (Fig. 2) and SensoGraph (Fig. 3). Data from Linking were analyzed using DISTATIS (Fig. 4) and SensoGraph (Fig. 5). Stability was checked using Jaccard stability and bootstrapping with adjusted RV.

The results show that the higher discrimination capacity of Linking over Projective Mapping is apparent already at the first two dimensions, without the need of checking further dimensions. Thus, the free-linking task may be considered as an alternative not only to the free-sorting task, but also to the projective mapping task.