The starting point for shaping a behavior, according to the authors, should be: Group of answer choices At baseline level A little below baseline level to assure success A little above the average for baseline to communicate reinforcement is contingent on improvement Two steps above the first mistake