When both long and short versions exist, the developer should report precision evidence for scores on each version, preferably from separate administrations to independent samples. Predictions carried over from the full test can rest on assumptions that fail, and short forms often use nonrandom item subsets. Direct, independent evaluation of both versions is therefore preferred where feasible.
The graph holds this control, the 0 it maps to, and the evidence behind each claim, over MCP and REST.