In previous works we developed and assessed a teaching program, ARDESOS v.1, with which we aimed to improve the fundamental skills of critical thinking. The results obtained were positive, but modest. After analyzing the limitations of the program we introduced certain modifications and assessed the new version. The changes involved designing the activities programmed by means of rubrics and making the students perform them with less direct orientation from the instructor. In sum specificity and initiative proved to be the key variables in the improved program, ARDESOS v.2. Based on the data collected we have seen a significant improvement of the new version over the old one in the following aspects: a) version 2 improved all the fundamental dimensions, mainly in the pre-and post-test measurements, to a significant extent (Student's t test); b) the effect size (Cohen's d) was significantly higher, and finally c) these improvements in the program elicited better performance. Accordingly, an improvement in critical thinking can be achieved via an instruction design that addresses the factors that really induce change. Currently, with these results we have been successful in adding a new improvement to the instruction, which we have re-evaluated.