CNN has a new “generic congressional” poll out that states the Democrats will win back the house:

Specifically with the CNN poll, of the 1003 people CNN questioned, Democrats were talked to more than Republicans (by 9 points, 34% dem, 25% rep, 40% neither), so it is no shock that the Dems won the poll by 10%.  The exit polls for 2016 (which CNN took part in) had 33% Republican turnout AND those exit polls were limited to 27 states, which gave more votes to Hillary.  The point is, CNNs “poll” has less Republican turnout than the skewed 2016 exit polls.  

CNN is not the only one with a poll out showing a Dem land slide, there are some other beauties on realclearpolitics

There are 435 congressional districts in the United States.  Congressional districts are supposed to be made up of roughly three quarters of a million people and voter turnout ranges from tens of thousands to hundreds of thousands of votes per major candidate. Even if you spread this CNN poll out across all districts, you could only have two to three responses per district.  That is not representative of anything, it’s called “too small of a sam!ple to have any significant meaning.”

If people in San Franfreako turnout in droves on election day, it makes no difference at all.  This was demonstrated in 2016 with the electoral college, where huge turnout in California did nothing to help Hillary win Michigan, however in congressional races, it is even less consequential, because the voting in one congressional district has no effect on who wins in the other, so San Franfreakans cannot even stop Devin Nunes from being re-elected.

In addition to this, these are nameless candidates.  You are not being asked if you are going to vote for some ultra-left wing socialist, just a “democrat” or a “republican”.

In summary, we have:

• A poll of 1003 people
• that is representative of 435 congressional districts
• asking a generic question
• based on some pollsters opinion of how the political parties will turnout

It doesn’t really mean anything.