Wednesday, Dec 11, 2019
 
search SGN
SERVING SEATTLE AND THE PACIFIC NORTHWEST FOR 42 YEARS!

click to visit advertiser's website


Javascript DHTML Drop Down Menu Powered by dhtml-menu-builder.com

Last Weeks Edition
   
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
 
 
 
 
 
 
 
 
 
 




 

 
 

 

 

[Valid RSS]

click to go to advertisers website
to Section One | to Arts & Entertainment
posted Friday, November 4, 2016 - Volume 44 Issue 45
Election 2016: What the polls show, and what they don't
Section One
ALL STORIES
  next story
Election 2016: What the polls show, and what they don't

by Mike Andrew - SGN Staff Writer

The last day to vote is November 8. Want to know who's going to win the election? Don't look at the polls. Or if you do, look very carefully and don't believe the first number you see.

A glance at the national polls shows numbers all over the place. CBS News has Clinton up three points. The LA Times has Trump up five. Rasmussen has Trump ahead by three points, but Reuters has Clinton leading by eight.

What's going on?

First, you should know that national polls are not going to be predictive of the outcome. That's because our elections are not decided by the national popular vote. Remember that in the 2000 election Al Gore won the popular vote but lost the election.

US presidential elections are decided on a state-by-state basis. So what if a candidate is gaining in the national polls? If the gains only reflect more voters in a state where the candidate is already ahead, they don't affect the outcome.

Here's an example. You can win Texas by one point, or you can win Texas by 21 points. Either way, you still win 38 electoral votes.

The extra 20 percentage points don't help you at all. In fact, if you put a lot of your resources into getting an extra 20-point margin in Texas, it might hurt you, because it would mean you took resources away from other states where you might have brought in enough votes to win by one or two points.

It's more predictive to look at state-by-state polling. But even there you have to be careful.

Many polls, especially the state-by-state ones, are based on very small samples, sometimes less than 1,000 voters, so the polls will have built-in statistical errors. A typical margin error for a well-constructed poll of average sample size is about four points.

That means that any single number in the poll can be four points more or four points less than reported. So if the poll reports that Clinton has 50% and Trump has 50%, the 'real' number could be 54% Clinton to 46% Trump, or 46% Clinton to 54% Trump.

In tracking polls, where polling is repeated every day or every week to try to capture swings in public opinion, what looks like an eight-point swing could be nothing more than statistical error. One day the errors puffed up Clinton's numbers, the next day they puffed up Trump's, without any real change in voter preference.

For polls like that to be something you should worry about - or celebrate - the swing would have to be pretty monumental, something on the order of 10 points or more.

The other thing to know about polls is that before pollsters record people's opinions, they have to decide what their sample will look like. In others words, before they ask the questions, they have decide whom they're going to ask.

If you pay attention to the news, you're already familiar with the difference between 'registered voters' and 'likely voters.' A poll that questions only people who will probably vote will obviously be more predictive than one that asks people who won't vote.

But then you get into the problem of determining who is actually likely to vote, and figuring that out can be very tricky. Trump dismissed polling from late September and early October showing him down by double digits because Republicans were underrepresented in the polling samples - or so he claimed. By the same token, more recent polls showing Trump beating Clinton are based on samples with more Republican voters than are found in the general population.

The reason is that pollsters sometimes try to base their predictions of who will vote - and therefore the composition of their samples - on the idea of 'enthusiasm.'

When they ask supporters of the two candidates how enthusiastic they about voting for their candidate, Trump's people say they are more so than Clinton's. Therefore, they conclude, more Republicans will show up to vote than Democrats, and more should be included in the polling sample.

But in the end, habit may trump enthusiasm (pun very much intended). It's a well-known phenomenon that some voters show up to vote at every election, year in and year out - almost by habit. Other voters will vote only occasionally, and still others only when they are particularly moved by a candidate or issue.

The habitual voters tend to be older, wealthier, better educated, and whiter than the off-and-on voters, and usually that tends to favor Republicans. But this election may reverse that trend. Clinton seems to be winning college-educated middle- and upper-middle class whites - especially women - by a very large margin, while Trump supporters tend to be poorer and less educated whites, who historically may or may not turn out to vote.

Another interesting wrinkle is party identification. Obviously Democrats support Clinton, by numbers approaching 95%. But Republicans support Trump only by numbers in the mid-80s. In fact, one poll of early voters in Florida suggests that as many as 28% of registered Republicans are defecting to Clinton.

Most Americans now do not identify with either major party, but how they describe themselves can offer clues as to how they might vote. Self-described 'independents' tend to skew right and will typically break for Republicans. This tendency is already showing up in polling, but it's nothing to be worried about because that's what usually happens.

On the other hand, voters who call themselves 'unaffiliated' tend to skew left. If they are telling pollsters they will vote for Trump, or just not vote, that would be a reason for Clinton's team to be concerned.

There's another way to measure support that doesn't involve polling but may be more predictive of success than even the best poll, and that's fundraising.

Hillary Clinton's campaign raised $11.3 million online between October 28 and October 31, largely as a reaction by her hard-core supporters to new FBI inquiries into her emails. The 'record-breaking' figure is the most Clinton has raised over a 72-hour period since she became the nominee at the Democratic convention in July, according to her campaign.

Tell a friend:

Share on Facebook  Share on Facebook

Post to MySpace!Share on MySpace!

    Share on Delicious

Share on StumbleUpon!

SCOTUS will review Virginia Trans case
------------------------------
Election 2016: What the polls show, and what they don't
------------------------------
NAACP sues North Carolina for voter suppression
------------------------------
Washington Secretary of State scrambles to correct voters guide error
------------------------------
We must end gun violence
------------------------------
JESSE'S JOURNAL: Election Day Blues
------------------------------
Andy Hill, one of three pro-equality Republicans, dies at 54
------------------------------
HRC revokes endorsement of Republican Mark Kirk over racist quip
------------------------------
Managing pain and living well with chronic pain
------------------------------
You will go to hell if you vote Democrat, San Diego Catholic church says
------------------------------
LGBTQ characters hit record high on network TV
------------------------------
'Who's in charge of America?'

Girl in Weiner probe joins criticism of FBI director Comey

------------------------------
New York police officer who made Gay porn suspended as investigators look into sex parties
------------------------------
Wingate civil rights trial reveals victim traumatized; former SPD officer Whitlatch claims devastation
------------------------------
Dems sue Trump and GOP, charge voter intimidation
------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------
BREAKING NEWS
------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

------------------------------

click to visit advertiser's website

click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
click to visit advertiser's website
 
 
 

gay news feeds gay news readers gay rss gay
http://sgn.org/rss.xml | what is RSS? | Add to Google use Google to set up your RSS feed
SGN Calendar For Mobile Phones http://sgn.org/rssCalendarMobile.xml
SGN Calendar http://sgn.org/rssCalendar.xml

Seattle Gay News - SGN
1707 23rd Ave
Seattle, WA 98122

Phone 206-324-4297
Fax 206-322-7188

email: sgn2@sgn.org
website suggestions: web@sgn.org

copyright Seattle Gay News 2016 - DigitalTeamWorks 2016

USA Gay News American News American Gay News USA American Gay News United States American Lesbian News USA American Lesbian News United States USA News
Pacific Northwest News in Seattle News in Washington State News