The concern to inform citizens and policy makers over the world about students’ skills of problem solving is explicit in the PISA programmes that have assessed if secondary students leaving the school system know how to use abilities that are useful in the job market to solve problems that are usually quite different from the traditional exercises used in the classroom. Students were assessed on how well they could handle real life situations that require science skills (OECD 2004, 2007). According to PISA results, students from certain European countries show low average performance on problem solving. One of the reasons suggested for explaining the outcomes in those countries is the parent’s educational level, below to the one found in other European countries (Suter, 2007). Another concomitant explanation may be the nature of problems designed by PISA when compared with those solved by students in regular classes. Unfortunately, still today, many teachers ask students to solve problems that are far away from those that they are going to encounter in real life and that are presented on PISA studies, denying students’ opportunity to develop their problem solving skills. In fact, several authors recommend the use of problems embedded in real life situations, of an open kind and with a low degree of structure, thus giving students opportunities to frame their own problems (Caillot, 2006; Wood, 2006) The present study was carried out in order to find out if the performance on problem solving of secondary science students is affected simultaneously or separately by the nature of the problems and by parents’ educational level. In order to pursue these objectives, three null hypotheses were stated: Student’s performance on science problem solving … (1) … is not affected by their parent’s educational level. (2) ... is not affected by the nature of the problems to be solved. (3) ...is not affected simultaneously by the parents’ educational level and the nature of the problems

A factorial 2x2 design was adopted as a methodology. The participants of the study from three public secondary schools were randomly assigned to each group. As a treatment, two groups performed a set of low-structured science problems of physics and chemistry embedded in real life situations, developed and assessed in a previous pilot study. The other two groups were taught following a traditional way. Each couple of groups was constituted by an approximately equal number of students whose parent’s had, respectively, high and low educational level. Students took, as a pretest and a posttest, two non-structured problems. The data were the students’ performance on science problem solving assessed by the written answers while solving problems during the pretest and the posttest and the parents’ educational level of each student.

The findings suggest that the parents’ educational level must be taken into account when designing science problem solving activities. Depending on parents’ educational level, the nature and degree of structure of the problems should vary. The implications for teacher’s practice of both factors, separately or together, on the enhancement of science problem solving in the classroom will be discussed.