The arousal-retention interval interaction revisited: The effects of some procedural changes