The purpose of the present study is to determine the effects of two multi-modal representations, the use of text and graph for learning, on problem solving, academic achievement and retention when used in a process-based instruction (PBI). The study was designed as quasi-experimental study complying with pretest-posttest control group design. The study group consists of (N=30+34=64) students from the department of classroom teacher education in the Education Faculty of a university from west of Turkey in 2015-2016 academic year. The data in the study were collected through problem solving inventory, texts written and graphs drawn by the students and academic achievement test. For the data analysis, independent-sample t-test, Kruskal Wallis H-Test and descriptive analysis techniques were used. According to the findings obtained in the present study, it can be argued that while there is no significant difference between the academic achievements and problem solving skills of the students carrying out their learning activities according to drawing-modal representation and those of the students carrying out their learning activities according to writing-modal representation, a significant difference in terms of their retention was observed.