Suppose that we use a randomized experiment to test the effect of a campaign ad on support for the Democratic candidate for president, using a control group that does not see the ad and a treatment group that sees the ad. The estimated effect from this experiment is that the ad increases support for the Democratic candidate for president by 5 points on average. But -- due to random assignment error -- the percentage of Democrats in the treatment group was higher than the percentage of Democrats in the control group. Based on only this random assignment error, we should expect that the treatment effect of the ad is truly ___. Group of answer choices A. lower than 5 points on average B. 5 points on average C. higher than 5 points on average