Twitter and email info

Wednesday, October 12, 2016

A few observations on polling.

Once political season rolls around, their is no way to avoid the non-stop polling.  We are currently less than a month away from the most important election in our political cycle, one month from knowing if we will have a President Hillary Clinton or a President Donald Trump.  We also have other elections taking place simultaneously, which will decide the makeup of the next Congress.  One group of organizations that have grown fat in recent years because of politicking are the polling companies.  Not long ago, most Americans would be lucky to recognize three separate pollsters, but lately, the media has decided that the American people just can't get enough of polls.  I can count at least twenty companies right off the top of my head.  Granted, pollsters collect and analyze data about far more than just political opinion; but in today's hotly contested electoral environment, the money is in the current campaigns that are coming to a close in early November.

Campaigns utilize polling information to determine candidate weaknesses and strengths.  They are also useful to help campaigns decide where and when to spend advertising resources.  Why would a candidate continue to buy airtime in a state in which he/she trails by twenty percentage points, when other states are much more competitive.  For instance, Donald Trump isn't spending money in California and Hillary Clinton has no ads running in Kansas.  But polls are also used to manipulate perceptions.  You see, polling organizations are inexorably tied to the big media corporations.  Many broadcasting companies, including Fox News, CBS, ABC, NBC and CNN all have their own polling capabilities.  Sometimes they work in tandem with newspapers, for example Fox News/Wall Street Journal Poll, and the CBS/NY Times Poll.  It should come as no surpirse that polls have become a major weapon in national politics.  Lets take a look at how it works:

Its possible to manipulate people with the framing of the polling questions, but in this instance, lets keep it simple.  The question is, if the Presidential Election were held today, and you had to choose between Republican Donald Trump and Democrat Hillary Clinton, for whom would you vote?  As the poll is constructed, the sample must be created.  Polls are scientific, so people can't just be called randomly.  you must first create a sample.  How many Republicans are you going to include, and how many Democrats? How many Independents?  What percentage of males and females?  If you chose to structure your poll by picking a specific percetnage of African-Americans, or Asian-Americans, you can certainly impact the end result.  How you structure your polling sample is called the methodology.

When it comes to polls, it always benefits to read the small words at the end of the survey results. This is where they usually hide their methodology.  For example, you may discover that the sample included 39% Democrat,  30% Republican, 5% Independent, and the rest "no affiliation".  Common sense tells you that the Democratic candidate has a big advantage because more Democrats participated.  Believe it or not, I took this particular sampling example from a major poll published this week and broadcast on every news channel.  Surprise, surprise, Hillary Clinton has a big lead.  The pollsters will justify their methods by arguing that their sampling is in line with the "latest" information regarding what percentage of the electorate is Republican or Democrat.  If the country is 39% Democrat and 30% Republican, then it only makes sense that hou would use the same percentages in your sampling.  The problem is, the percentage fluctuates dramatically, not only nationally, but locally.  And if you are polling for the state of Ohio, wouldn't make sense to try and determine the percentages of Republicans and Democrats in Ohio, instead of defaulting to the perceived national figures?

To sum it up, don't think too much about poll numbers.  Back in 2012, we were told that Romney was going to win a close election.  Romney lost, and it wasn't really close.  Not only are polls subject to manipulation, they can be innaccurate for a host of other reasons.  Sometimes people get annoyed when their dinner is interruptd by a phone call, and they lie.  Other times, people don't understand the process.  Our recent political history, going back to the famous "Dewey vs. Truman" election in 1948, is replete with examples of polls that produced results different from the actual election.  This particular Presidential election is a hard one to handicap.  Although everything appears to be falling Clinton's way (including the polls), there continues to be that feeling in the back of my mind, that we are in for a big surprise.  Trump voters are angry and motivated- they will flock to the voting booths.  The Clinton voter continues to be less-than inspired.  Also, I think its safe to assume that we haven't heard the last about Wikileaks emails, imvestigations, and the extra-curricular activity of former President Bill Clinton.  So don't take the daily deluge of polling information too seriously.  Unless, of course, the figures point to a Trump victory.  😉

No comments:

Post a Comment