For nine-week periods during two consecutive years, sixth-grade students at the same school were taught identical mathematics content using two different instructional approaches. Year 1 involved low achievers, whereas Year 2 was mixed ability students. The experimental treatment was a problem-centered approach (PCL) where potentially meaningful tasks were posed to the class and solved in collaborative groups. The groups presented and defended their solution strategies to their peers. A statistically similar control group learned via the traditional explain-practice (E-P) approach. The teacher demonstrated procedures and students individually practiced what had been illustrated. Regardless of perceived ability level, the PCL approach appeared to significantly enhance achievement and improve attitude towards mathematics. Low achievers seemed to gain the most, narrowing the gap between them and their mixed ability peers. PCL appeared useful with any curriculum if the teacher understood and properly implemented the components of the approach. (Contains 7 figures and 5 tables.)