Failure to accept retractions: A contribution to the continued influence effect