A naïve method to treat a series of dichotomous variables as a single number

There are many ways to treat a set of dichotomous variables. Factor analysis is one of them, just to quote one of the most used methods. The goal is always the dimensions reduction. For example, a battery of ten dichotomous items may be reduced to a few variables, according with some numerical benchmark (eigenvalues for example). This is done in order to facilitate the comprehension of the underlying phenomenon and to simplify the use of the analytic tools (contingency tables, regression, structural equation models and so on). In this exercise I try to show a way to analyse statistical relations between the set of dichotomous variables (as a dependent variable) and a set of independent features. The dependent set of dichotomous variables will be treated as a single numeric variable without a loss of precision: the poster will explain how to do. It must be warned that this is only an experimental work, in many cases results are not yet satisfying. Nevertheless they may serve to show a way.