The purpose of this study was to investigate the effects of multiple representation-based instruction on seventh grade students' algebra performance. The study was conducted on four seventh grade classes from two public schools lasting eight weeks. For assessing algebra performance, three instruments called translations among representations skill test, objective based achievement test, and Chelsea diagnostic algebra test were used. The analyses were conducted by using multivariate covariance statistical model. The results pointed out that multiple representation-based instruction had a significant effect on students' algebra performance compared to the conventional teaching. In addition to this, students from experimental groups found this way of teaching fruitful.