Selecting a researchable problem