<![CDATA[The low ability of students' mathematical reasoning causes students difficulty solving mathematical reasoning ability questions on several indicators. The purpose of this study is to find out and describe the mathematical reasoning abilities taught with the Modular Instruction learning strategy better than those taught by conventional learning of Grade VIII students of SMP N 3 Merangin in the 2015/2016 Academic Year. This type of research is quantitative research that is experimental in design "true experimental design". Data collection techniques using tests of mathematical reasoning ability in the form of essay test questions. Data analysis technique is t-test. Based on the research results the average experimental class taught with the Modular Instruction learning strategy was 62.53 and the average control class taught with conventional learning was 57.41. The normality test results of the experimental class 0.178 <0.301 while the control class 0.095 <0.318 then the two classes are normally distributed. From the homogeneity test with the F test produces an Fcount of 1.21 ≤ Ftable 2.25 that both classes have homogeneous variants. Hypothesis test results obtained value greater than then be accepted. So it can be concluded that the mathematical reasoning abilities taught with the Modular Instruction learning strategy are better than those taught with conventional learning of Grade VIII students of SMP N 3 Merangin in the 2015/2016 Academic Year.]]>
