john, explain that. what is the background here? why would they be different if they're cell phone users? >> well, the big problem, of course, is that a quarter of all u.s. adults have cut the cord, are cell phone only. and that's a big problem for pollsters who use only landline phones. we can talk about our surveys and others who do use cell phones, but most at the state level and district level do not. so this new research suggests that those may be bias a little bit. but the issue that those who have cut the cord in the united states look -- are a distinct group. they tend to be younger, they're more likely to be his panic and more likely to live in urban areas. those people dispropossessionatly vote as democrats and that's the potential for bias there. >> research looks at registered voters & conducted by landlines. republicans are at 44% and