The present study examines the effect of three different structured methods, traditional, independent and problem-solving, of teaching children arithmetic in the beginning of 7th grade in Sweden, age 13 years. The progress made by these students is presented by measures of their arithmetic ability, calculation and quantitative concept, as well as their self-regulated learning skills in mathematics, characterised as internal and instrumental motivation, self-concept and anxiety. The results will be discussed with reference to a socio-cultural perspective. This study has a split-plot factorial design with time as within-subject and type of intervention as a between-subject factor. The results show that there are no significant differences between teaching methods when assessing arithmetic in total and calculation. Students' progress in quantitative concepts is significantly better if teachers teach traditionally or with a problem-based curriculum. In order to develop aspects of self-regulating learning skills, the results show that pupils would benefit in a traditional or problem-solving curriculum. Problem-solving seems to be more effective than traditional and independent work for students' internal motivation. Traditional work and problem-solving are more effective than independent work for students' self-concept. (Contains 2 tables and 3 figures.)