Effects of "chained" study on spontaneous relational discovery