One important skill to be developed in future physicists is modelling reality using physics. Problem-solving is the activity most used to teach, learn and assess physics skills at the university level. However, most of the problems habitually used in physics subjects seem not to encourage developing this skill. Two empirical exploratory studies were conducted. Educational materials were analysed to contrast the hypothesis asserting that most of the academic problems would not help students to develop the modelling skills. Secondly, with the proposed"Physics Problem Understanding Model", two kinds of experimental problems were elaborated. A small group of experts solved the experimental problems in an interview. The hypothesis stated that there would be evident differences between skills associated with the solving procedures of each type of experimental problem. Results served to contrast the cognitive model proposed for sufficiency and supported the hypotheses even though further studies are needed to obtain external validity