Like everyone says here, utilize those 3D models that DAT Bootcamp (or other softwares) provides. SLOWLY rotate each 3D model and see how the "top" image matches with the "front" image, how the "end" image matches the "front" image, etc. Try to understand how the EDGES of the top, front, and end all match up in the 3D model and then keep practicing on conjuring up the 3D model in your mind by "gluing" those top, front, end images by their corresponding edges. You don't have to be able to visualize ALL the single details of the entire 3D structure in at once. Take only two images at a time (top and front, front and end, etc.) and then eliminate answers right away. From my experience you can eliminate one answer choice right away (several times even two) by just focusing on two sides at a time.