This study aims to (1) determine whether or not there are some differences in effectiveness of flat-sides space learning using problem-based learning approach and the one using
conventional learning approach viewed from critical thinking skills, learning outcomes, and attitudes towards mathematics and (2) search the comparative effectiveness of every variable.
It is a quasi experimental study with control-group design. The population is the eighth-year students of SMPN 2 Sentolo. To obtain the sample, this research made use of purposive
sampling technique. One class of the school was chosen as a control class and another as an experiment class which was homogenous and was in fulfillment of the assumptions of
normality judged from the pretest result. Test on normality assumptions of pretest and posttest data was multivariately analyzed using test of Mahalanobis distance. While, test on
homogeneity of variance-covariance matrix was analyzed using the homogeneity of Box's M test. And, test on hypotheses was analyzed using two-group MANOVA and Post-Hoc procedure. The result of the two-group MANOVA showed that there were differences of effectiveness between flat-sides space learning using problem-based learning approach and the one using conventional learning approach seen from critical thinking skills, learning outcomes, and attitudes towards mathematics simultaneously. The result of the Post-Hoc procedure indicated that flat-sides space learning using problem-based learning approach was more effective than the one using conventional learning approach in terms of critical thinking skills and learning outcomes, but not in terms of attitudes towards mathematics.

flat-sides space, problem-based learning, critical thinking skills, learning outcomes, attitudes towards mathematics