ABSTRACT
Thus, factor analysis, (or multiple regression analysis, to be described later) is generally used to summarize the variations in a representational field for a given population. This technique closely examines the links between profiles of individual responses in this population but not the level and dispersion of these responses. To make up for it, other techniques such as MDS or HCA (see Part One) aim to account for relations between responses through mean distances between stimuli. Generally, these stimuli have been rated by a relatively homogeneous population of subjects (questions arising from the heterogeneity of the study population will be treated in Part Three). Finally, INDSCAL (individual differences multidimensional scaling), a special technique halfway between the two above-mentioned approaches, will be explained later in Part Two. In summary, the main difference between these techniques lies in the type of information analyzed. It is therefore very important for the user to know the rules of application and interpretation of the analytic method chosen. We will begin by considering factor analysis.
