<![CDATA[This research is based on the background that teachers still use the one-way lecture learning model which the teacher dominates, thus making students passive, lack of students' mathematical communication skills, especially in learning mathematics.. This study aims to: (1) determine the differences in the initial abilities of students in the experimental class and control class; (2) find out whether students' mathematical communication skills given the Three Step Interview type cooperative model are better than the conventional model; (3) find out how the influence of the learning model cooperative type Three Step Interview on students' mathematical communication skills. This research is an experimental study with a Nonequivalent Control Group research design. Data collection techniques using a mathematical communication ability test which consists of four question in the form of a description. Te date analysis technique was carried out by normality test, homogeneity test, average difference test, and effect size. Based on the calculation of hypothesis test 1, obtained t_(count) 1.056 and t_(table) 1.685 at a significance level of 5%, which means t_(count) > t_(table) ( 1.056 <1.685) Then H_(0) is accepted. Test hypothesis 2 obtained t_(count) 3.016 and t_(table) 1.685 at a significance level of 5% which means t_(count)> t_(table) (3.016 > 1.685) then H(0) is rejected and H(1) received. The effect size test is 0.9796, so the interpretation criteria are high. So it can be concluded that: (1) there is no difference in the initial abilities of the experimental class and control class students; (2) the mathematical communication ability of students who are given the Three Step Interview type of cooperative learning model is better than the conventional learning model; (3) the effect on the cooperative learning model Type Three Step Interview on students' mathematical communication skills is high]]>
