could find for 608 state level senator and governor races between 1998 and 2014 comparing those poll results with actual electoral results find the historical margin of error plus or minus 6 to 7 points. that s an error range of 12 to 14, not typically reported 6 or 7. expand on that, please. polls can be wrong for two reasons. we say, sampling error and non-sampling error. it s because you sample a finite number of people and like drawing bulls from the urn. you won t see the entire population. that s the margin of error. not sampling error is because you re not getting the right people or adjusting to the wrong population. and in this particular case, the 2016 election, i think a lot of trump supporters were not responding to polls. and the turnout wasn t what was
so from that same summation of your work, i, too, want to quote. i m going to put it up on the screen. i thought this was remarkable. we examined this is part of your research, 4,221 late-campaign polls, every public poll we could find, for 608 state level presidential senate and governor s races between 1998 and 2014, comparing those poll results with actual electoral results, we find the historical margin of error is plus or minus six to seven percentage points, yes, that s a error range of 12 to 14 points not the typical reported 6 or 7. explain that fleez. we say sampling error and nonsampling error. the sample is because you re sampling a finite number of people, drawing bulls from the urn. you won t see the entire population. that gives you the margin of error you ll see in the newspaper. non-sampling error is because you re not getting the right
bolduan. i m john berman. big political news this morning, or huge in donald trump language. so, what happens when a candidate who calls himself a winner is no longer winning? for the first time in a long time, it is happening. a brand-new poll from cbs news and the new york times shows ben carson surging ahead of donald trump. carson now has a four-point lead nationwide. yes, that s within the sampling error. yes, it s also the first time in a major national poll that carson has topped trump, and another yes for you, it all happens on the eve of the next republican presidential debate. let s bring in cnn s political director, david for more on this. what s behind the change? if you look across every demographic group in this poll, guys, you will see the trend you see overall nationally, which is carson is besting trump group after group after group. first women voters, among women voters you can see carson beats trump 27% to 19%. if you look where that came from
single digits. can supporters give him a boost with $24 million advertising campaign? i m wolf blitzer and you re in the situation room. 25 hours and counting to the big republican showdown right here on cnn. but on this, the eve of the gop debate, here at reagan library in simi valley, california, there are other numbers we want to take a look at and need to look at right now. new poll shows donald trump is still the leader of the pack but he s now just four points ahead of dr. ben carson. 27% to 23%. with the sampling error, that s effectively neck and neck. will carson keep turning the other cheek, or will he go on the attack? it s all on the line right now for jeb bush, marco rubio, and
necessarily gone online but actually signed up with new insurance shortly. so we ll see, we ll stand by for that. much more on that story coming up. other news we re following, last week as you know, crist christie easily won a second term as the new jersey governor. now a new poll showed him in a dead heat to win the 2016 presidential race. the quinnipiac university poll shows him with 4443% support. hillary clinton look at this, 42% support hillary clinton. of course, the presumptive favorite on the democrat side. we should stress that christie s 1% edge is certainly well within the sampling error. let s bring in chief political analyst gloria borger. what do you make of this 42% for hillary clinton, 43% for christie. what year is this, would he have? this is 2013. we are a little early on this. what these poll numbers really tell us is that these two candidates have an awful lot of