Specifically, we used the "scaling replicated conditional rank-order data" (SRCRD) approach, in which subjects are asked to rank stimulus pairs in terms of similarity.
In the second stage, subjects completed the SRCRD ranking procedure on the twelve stimulus terms.