Representasi matematis penting bagi mahasiswa untuk menyampaikan ide-ide kreatif terhadap permasalahan statistik. Pentingnya representasi matematis, maka tujuan penelitian ini adalah menganalisis perbedaan peningkatan representasi matematis mahasiswa calon guru sekolah dasar yang difasilitasi dengan menggunakan model collaborative problem solving. Metode yang digunakan adalah Quasi Eksperiment dengan desain Non-equivalent Pretest-Posttest Control Group. Sampel penelitian sebanyak 70 mahasiswa calon guru sekolah dasar. Hasil penelitian menunjukan terdapat perbedaan peningkatan kemampuan representasi matematis mahasiswa, dimana mahasiswa yang mendapatkan pembelajaran dengan model CPS memperoleh N-gain 0,46 lebih tinggi dari kelompok ekspositori dengan N-gain 0,30. Berdasarkan kemampuan awal matematis (KAM), baik level tinggi, sedang, dan rendah kelompok CPS mencapai peningkatan kemampuan representasi matematis lebih tinggi dari kelompok ekspositori (EPS). Perbedaan peningkatan mengindikasikan bahwa penggunaan model CPS dalam perkuliahan statistika pendidikan lebih baik dari model pembelajaran EPS. Peningkatan ini dikarenakan berkolaborasi dilaksanakan dalam kelompok, namun tujuan utamanya bukan untuk mencapai kesatuan yang didapat melalui kegiatan kelompok, tetapi mahasiswa calon guru sekolah dasar didorong untuk menemukan berbagai pemikiran kreatif yang disampaikan setiap individu dalam kelompok.Kata Kunci: Representasi Matematis, Kemampuan Awal Matematis (KAM), Collaborative Problem Solving (CPS) Mathematics representation are important for students to convey creative ideas to statistical problems. The purpose of this study was analyzed the difference in improvement of mathematics representation of students who are elementary school student teacher candidate facilitated by using collaborative problem-solving model. By using Quasy Experimental method with non-equivalent pretest–posttest control group design. The study sample consisted of 70 students of elementary school student teacher candidate. The result showed that there was a difference increase in mathematics representation in which student who were gained in collaborative problem-solving model identified to have N-gain 0, 46 higher than those who were ekspository group with the N-gain 0,30. Based on the prior-ability mathematics (PAM), the higher level, medium, and lower level of students in CPS group increased high mathematics representation than the ekspository (EPS) group. This increased result indicates that the use of CPS model in lectures of educational statistics was better than EPS. This increase internalized that the collaboration was in group but the main objective was of getting the unity gained in the group, but the students urged to get creative thinking recived individually in group.Key words; Mathematics representation, Prior-Ability Mathematics (PAM), Collaborative problem solving (CPS)