Can students be taught to think? Educators have debated the question for centuries, and current opinion is split. Some "formal discipline" theorists believe, as did Plato, that teaching the rules of any one subject gives students abstract principles that can be employed in other subjects. Others argue that people learn reasoning skills only within a specific context and that generalized problem-solving is very hard to teach.

To test the "formal discipline" theory, two psychologists administered a logic test to 165 college undergraduates -- once in their first year and again in their fourth. None of the students had taken courses that explicitly taught logic skills. Each student was rated for improvements in three kinds of reasoning: statistical/methodological (ability to find patterns in numerical samples); conditional (understanding such relationships as "if p then q"); and verbal (tracking arguments and analogies in words).

The study, reported in the current Journal of Developmental Psychology, found that all students except social-science majors improved in verbal reasoning. But improvements in the other two categories ranged from 60 to 70 percent and varied with field of study: Majors in natural sciences and humanities scored higher in conditional reasoning; social-science majors fared best in statistical/methodological skills.

Thus, the authors conclude, it appears that "highly general inferential rules can be taught" and "the quality of people's everyday reasoning can be improved."