Replies

So look at VA10, which is held by a GOP incumbent, and the polls have her trailing by 7. If you read the latest poll the first paragraph they admit the sample size is very small, a 5% margin of error, which means it is anywhere from -12 to +3. If you just read what they write it sounds like wishful thinking, “lots of college aged residents” and “Trump not well liked”. I am not saying she will win, and all but one of the polls show her trailing, but just seeing it leaning Dem for a seat held by a GOP incumbent is an oddity since incumbents win more than 90%Z of the time. It will come down to get out the vote and the article says the GOP is spending to defend this seat. Further down there is a McLaughlin poll that shows Comstock ahead by 1%. It is the oddball poll, so we have to dig to find out the methodologies and see if there is an obvious error.

Keep digging and you see, 70% of the respondents were reached on a cell phone. That skews to younger. In fact they were 20% short of their target in reaching people 65 and older. They do however reach more males than female, if it matters, and both candidates are female. 11% were undecided or refused to answer. When you narrow down to “types of people they believe are likely to vote” they have her down by 4%, not 7%.

Then look at IL14. It is a GOP incumbent, Hultgren, male against an African American challenger. RCP has only 1 poll, leans GOP +4. But when you go into the details among “types of people who voted in 2014/2016” they have him +6/+8.

NV3 is a tossup. It is an open seat. Only 1 RCP poll and it shows the Dem, Lee, up by 3. But they only polled 173 people so it’s a margin of error of 7.6! That means the Dem could be down close to 5. I couldn’t find any specific details on that poll. But Trump won that district by +1 and Obama won it by +0.8. Given that Obama got out more vote overall nationally (63%+) vs Trump (61%) I think this district leans more GOP than RCP lets on. If you factor in those national GoTV numbers from the presidential elections and realize there is no Obama type figure draw Dems to the polls and in a district won by Trump, the GOP candidate should win that race by 3%-5%.

The flavor is all in the sausage making.

I’ll dig some more when I have time because it interests me.

I don’t follow this guy I just caught this link on another thread, and it just so happens that he mentions the race in VA10. A coincidence since that’s the race I randomly looked at. Scroll down a bit and read his tweets numbered 1-14. Specifically he says that the DNCC has had to go back into VA10 with a pile of more money to support their challenger - a race they thought they had locked to unseat the (R) incumbent.

https://twitter.com/larryschweikart

He thinks the GOP is looking a lot stronger than anyone wants to admit, not just VA10 but he seems to think NV Senate seat leans GOP and the shocker of all would be the OH Senate seat, where all the polls show the incumbent Brown (D) with a double digit lead. Not sure if he is reliable at all - I can’t recall why I know his name.

Anyway I looked at the most recent poll for the OH Senate race. Here are the internals: https://www.suffolk.edu/documents/SUPRC/10_12_2018_marginals.pdf

Keep in mind that Trump won OH 52% to 44%.. He won by 8 points but the polls had him even. The polls had him +2, and he came in 5 points higher than the polls.. But in reading the OH Senate race poll you wouldn’t know that to be true. They have Brown the (D)incumbent, ahead +18%! Get into the questions like “do you support impeaching Trump” and they have it 54% in favor vs 40% opposed. Do you believe a state that voted +8 for Trump in 2016 suddenly wants to see him impeached by +14 in 2018? No way! That’s crazy. They have Trump at 41% approve 53% disapprove. Way off the national average for a state that voted him +8 vs Clinton. They have “Did Trump keep his promises to OH” at -1%. They also have “support/reject” Kavanaugh -6.5% which is way under the national average. None of that makes sense. So it strongly suggests that they have skewed the poll, probably in under-sampling Independents and oversampling Dems. OH is a strange case where they classify party affiliation by whether a citizen can prove they voted in a primary. So according to the SoS for OH, the registrations numbers should be 17%D - 25%R - 58%I. I can’t be certain but this could be a case where they just got sample groups way off from reality.

Overall, I guess the DNC would fear demoralizing their voters if they let on that they were behind, while the GOP would fear their voters will coast if they think they’re ahead. So imo it’s in everyone’s interest to play it like it’s close. Yet OH Senate race they are playing it like they are way ahead. Who knows, they may be... but that poll I read when you look at the internals doesn’t make sense at all.