A political candidate feels that she performed particularly well in the most recent debate against her opponent. Her campaign manager polled a random sample of 400 likely voters before the debate and a random sample of 500 likel voters after the debate. The 95% confidence interval for the true difference (post-debate minus pre-debate) in proportions of likely voters who would vote for this candidate was (-0.014, 0.064). What was the difference (pre- debate minus post-debate) in the sample proportions of likely voters who said they will vote for this candidate? ​