The TL;DR is that anchors should be used to determine which corner or edge (or center) of the screen to snap the elements to (there are at most 9 combinations anyone should use), and the offsets should be used to determine simultaneously (1) the position of the element relative to the anchor point, and (2) the size of the element.
The value of AnchorMin and AnchorMax should be the same. For example, "0.5 0" represents the bottom center of the screen. The values of the offsets will depend on what you want to draw, and will determine (1) the bottom left corner and (2) the top right corner the element rectangle.
What people often get wrong is that they just use anchors. The problem with only using anchors is that the elements will have an inconsistent aspect ratio for different screen aspect ratios. For example, if the element is 16:9 on a 16:9 screen, then changing the screen to 4:3 will also change the element to 4:3. While that may be desirable for some rare cases, it's not usually what people want. The reasons it's not what people want are (1) it can cause images to stretch, and (2) it does not line up with vanilla UI, which many plugins try to draw UI relative to. A great example of this is the vanilla hot bar. Each slot in the hotbar is always a square, regardless of the screen aspect ratio. If you want the same effect, you need to use anchors + offsets as I explained at the beginning.
One exception to the rule of "don't use just offsets" is that you can probably get away with it inside a nested element, if the root parent element is using offsets.