ABSTRACT
The training study about conjunctive probabilities reported in this chapter was designed to address two issues. First, it examined whether training procedures that fulfill the conditions specified in chapter 4 can yield greater improvements in performance than those found in previous training studies. Second, it tested whether, and to what extent, a frequentistic pictorial representation is superior to a conventional pictorial representation. The training was implemented as a computer program with a flexible user interface, and required that participants get actively involved in solving the tasks. The basic training consisted of teaching participants how to translate numerical information into pictorial representations and how to derive conjunctive probabilities using that representation. The conventional pictorial representations were Venn diagrams and the frequentistic pictorial representations were frequency grids.
