The effects of prior neutral messages on resistance to evaluative communications