Hello, I've got a doubt about the following point : I have used Indscal in order to process dissimilarity data directly obtained from a pairwise comparison test during which judgments were produced on a seven-category scale labelled from '0' to '6' (very similar to very dissimilar). To my knowledge, one condition to use Indscal is to have at one's disposal dissimilarity data of metric level (interval level or ratio level); the subjects were told to consider the category-scale as metric (e.g. a rating equal to '4' implies stimuli twice more different than a rating equal to '2'), an a posteriori check of this assumption (using the successive interval method) was first carried out. But as far as this assumption is shown acceptable, is it sufficient to be able to process this kind of data with Indscal ?
Can anyone help me to dispel this doubt ? Many thanks. Arnaud. ---------------------------------------------- CLASS-L list. Instructions: http://www.classification-society.org/csna/lists.html#class-l
