The purpose of this study was to investigate the effects of pair problem solving technique incorporating Polya's problem solving strategy on undergraduate students' performance in conceptual and algorithmic questions in chemistry. The subjects of this study were 89 students enrolled from two first year chemistry classes. The experimental group was a class of 44 students who received pair problem solving technique incorporating Polya's problem solving strategy and the control group was a class of 45 students who received only Polya's problem solving strategy. Students' achievement of conceptual and algorithmic questions in chemistry was measured using as post-tests, Conceptual Chemistry Question Test (CCQT) and Algorithmic Chemistry Question Test (ACQT). Test of Logical Thinking (TOLT) was used as a covariate in this study. The results of Analysis of Covariance (ANCOVA) showed that students in experimental group had significantly better performance on both conceptual and algorithmic questions in chemistry. The results of this study for students' problem solving performance are discussed. (Contains 4 tables.) [Abstract is provided in English and Spanish.]