The current paper explores the effects of providing people with schema training at the outset of learning (compared to at later stages) on mathematical word problems modeled after problems from the Graduate Record Examination. Additionally, the ratio of worked examples to problem-solving practice was manipulated. Participants were randomly assigned to one of four conditions and were tested on near and far transfer problems. Participants provided schema training at the outset of learning outperformed those provided schema training after problem-solving practice, particularly on far transfer problems. Likewise, participants provided with a higher ratio of worked examples to problem-solving practice demonstrated better performance during testing than those provided with more problem-solving practice and fewer worked examples. These findings extend the worked example effect to far transfer problems and suggest that providing schema training prior to learning solution strategies may be effective in improving mathematical problem solving.