Yes/no to both. Depending on how you store your data, you have to come up with a way to calculate the polygon / volume that encompasses the volume of space where the object has been in the time in between the two frames.
First, many game engines run the rendering at a different clock than the physics calculation: the latter is most often fixed, while the first runs as often as possible. So, the time slices have no direct relation to the intervals where the position is recalculated. And objects can change form and rotate...
So, it's not that easy. For sprites, the locations in between frames are often ignored, so you can shoot through an object without hitting it. And the 3D calculations are quite complex, so most of the time one or more rough, squarish, bounding boxes ("hitboxes") are calculated somewhat around the path of the object.
Edit: and to do it right, you should check that at the time the bullet hits the bounding box, the object was at that location and not at the other side.