Welcome to DU!
The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards.
Join the community:
Create a free account
Support DU (and get rid of ads!):
Become a Star Member
Latest Breaking News
General Discussion
The DU Lounge
All Forums
Issue Forums
Culture Forums
Alliance Forums
Region Forums
Support Forums
Help & Search
What polls tell about the progress of the Sanders campaign and the current state of the primary race
Last edited Sat Dec 5, 2015, 12:42 PM - Edit history (1)
This forum is chockablock with discussions of polls (mostly from Clinton supporters saying "the primary is over" and Sanders supporters saying "it's not over and polling shows Clinton is vulnerable int he general election"
, but much of the discussion misses the point.
Nate Silver knows as much about polling as anyone, and this past week, his 538 blog had lots of excellent discussion of how to read and use polls, and here are some take away points:
* there is a wide and consistent divergence from the live phone-based polling to the automated-phone- and internet-based polling (Trump's polling best illustrates this phenomenon but it is true for Clinton, too);
* this wide divergence in the two schools of polling has not been fully explained yet, but pollsters with the most accurate records think the live phone-based polling is the more-likely-accurate model;
* in addition to focusing on the polls that use traditional sampling methods, the polls which focus on likely voters are probably the more reliable indiactors;
* the election process is fascinating to a self-selected group of people who choose to discuss the election on an internet website (i.e., you), but most voters in Iowa and New Hampshire and elsewhere are only starting to pay attention;
* the overwhelming majority of voters are still undecided (even in Iowa where 65% of the voters decide in the last month before the caucus and New Hampshire where 71% of voters decide in the last month before the primary and especially in the 48 other states where the voters decide even later than in Iowa or New Hampshire) so you should read polls with the understanding that "undecided" is currently the prohibitive and overwhelming leader in all of these polls;
* polling is an important tool for campaigns but even in Iowa and New Hampshire (and even more so for national polling), predicting results based only on polling (especially early polling) is not very accurate.
With all of that said, the polling trends for Sanders look good.
In New Hampshire (where the primary is still more than two months away), the live polls of likely voters are very good for Sanders:
If you increase the smooting (i.e., you draw the graph line so it captures the overall trend line more than the rises and falls based on each individual poll result), the trend is even more encouraging for Sanders:
In Iowa (where the caucus is still more than 8 weeks away), the live phone polls of likely voters are very close and a graph of the trend lines presents more encouraging news for Sanders:
If you decrease the smoothing (to focus on the rise and fall of individual poll results rather than the overall trend lines), the graph shows an extremely tight race:
While the national polls are useless as a prognostication tool, they do provide useful data about whether the candidate's message is getting across and whether the national trend is moving in the right direction. The national live polling of likely voters confirms that Sanders is getting his message across and the campaign is heading in the right direction:
Neither side should make more of this polling than it is, but for those of us supporting Sanders, this is nice confirmation that the campaign is on the right track and progressing as we would hope.
27 replies
= new reply since forum marked as read
= new reply since forum marked as read
