ND: SRA Research- Cramer +10

[cinyc]

Hot Take: This is actually a Heitkamp internal she released so she'd have another newspaper to hold when she wins on election night.

No. The poll is not an internal. It was done by Strategic Research Associates for the NBC TV Stations in Western North Dakota.

[cinyc]

It’s amazing how much we are going nuts over this poll that doesn’t have a memo, any crosstabs, or even dates! All we know is that this poll was conducted in September, but that could be up to four weeks old!

This happens sometimes for TV station-commissioned polls.

Often times because they drip out the results for different races over days. There are other ND races this cycle. I wouldn't be surprised to see House and Governor results over the next few days, followed by more info on the poll.

[cinyc]

Here's the poll memo:

http://media.graytvinc.com/documents/GRAY_ND_September_Survey.pdf

It says that 650 likely voters were polled between September 17-27, 2018.

[cinyc]

All the crosstabs seem to be off, honestly.

Crosstabs are always off, especially when the target group is <2% of the population, like here.

There aren't many black or Hispanic voters in North Dakota. It's overall population is 90% White as of the 2010 Census. The largest minority group is Native Americans at 5.4%. Hispanics are 2%. Blacks 1.2%. Asians 1%.

[cinyc]

The minority numbers are clearly skewed. There is no way that Black voters are tied, or that Hispanic or "Other" voters are favorable to Cramer.

African-Americans are 1.2% of ND's population. Does it matter if it's a wonky subsample?

It's still a bad sign for the polls methodology.

They polled 650 voters. If African-Americans make up 1.2% of the population and quotas were met (doubtful), that means you'd have a sample size of about 8. Care to guess what the MoE for a sample size of 8 is? +/-34%. In other words, meaningless.