This paper questions the traditional chalk and talk approach to teach science. The author argues that experiments should be arranged for investigating relationship between variables rather than "to prove" a big idea.