This paper reports on a part of a study which was conducted to determine the effect of van Hiele theory based instruction in the teaching of geometry to Grade 10 learners. The sample consisted of 359 participants from five conveniently selected schools from Mthatha District in the Eastern Cape Province in South Africa. There were 195 learners in the experimental group and 164 learners in the control group. The experimental group was given van Hiele theory based geometry instruction and the control group was given traditional method of geometry instruction. A multiple choice geometry test was administered to the participants before and after five weeks of instruction (pre- and post-test design). The results indicated a statistically significant difference in the mean scores in favour of the experimental group. The significant improvement in the performance of the experimental group having more learners at level 2 than at level 0 and level 1 suggest that the van Hiele-based instruction had a positive effect in raising the learners' levels of thinking compared to traditional instruction.