I always keep in mind when I teach that it’s great if kids make scientific discoveries during the course of a year. I have found that, usually, direct instruction is best, followed by hands-on activities and review. It’s difficult to discover something specific. It’s better for them to develop good questions and then to try and answer those questions on their own using guided instruction.
From John Mighton:
It is important to bear in mind that entire civilizations failed to discover the idea of zero as a place-holder for operations such as addition and division. If the Romans couldn’t develop an effective method of division over the course of eight centuries, it seems unrealistic to expect children to discover a great deal on their own.