Lies, Damned Lies, and Polls. Who to Believe?

Lies, damned lies, and statistics was an expression made famous by Mark Twain but it is equally as true about modern polling data. 

I have noticed a trend in recent years where polls are all over the place. I doubt this is a new trend and has likely been happening for decades but the internet gives us  instant access to information few had before. Crime rates are actually lower as well compared to previous decades but thanks to Facebook, Twitter, email, 24 hours news channels, and just the internet in general, you would think this is the most dangerous crime period in history and you would be dead wrong.

In some cases polls taken just a week apart have completely conflicting and even opposite numbers.The polls we choose to believe also tend to be the ones that support our own personal bias. A poll where our candidate is seven points ahead has to be accurate, but one that has our candidate seven points behind was obviously biased and too small a sample and therefore wrong. Humans can’t help being biased. It is simply the way we were built and likely a survival instinct to protect our tribe. We like to belong to groups and separate the “Us” from “Them” where they are not to be trusted.

You gotta love the guy that came up with the phrase” Within the statistical margin of error of +/- 4%”. That phrase gives pollsters a free pass if they really screw up. I will not call out individual polls as either good or bad but I will try and give you some tips on how you can judge which ones are more likely to be accurate and also ask you to take them all with a grain of salt.

1. How many people were interviewed. Be careful not to simply trust a poll that interviewed 1600 people and ignore another poll that only sampled 800. Although it is true the margin error of 4% can go slightly down with larger samples,  it levels off at a certain point.
2. Who was interviewed. Polls that sample a true representation of likely voters will be more accurate. Polls that have a larger sample of only over 65, or only Atlanta metro, or more men, more women, etc. should not be as trusted as one that offers a more accurate representation. The best polls will come as close as possible to the various groups likely to vote and in the same percentages they are likely to vote.
3. Were cellphones and/or live interviews included. This is a fairly new one for the list. Ten years ago, nearly everyone had a landline phone and most also had a cellphone. But more and more people have cancelled their landline phone and now use their cellphones as their only phone and who can blame them given the price.No reason to pay an extra bill when you don’t need to.  Polls that rely solely on landlines should no longer as trusted.
4. Who sponsored the poll and what was their purpose? Pollsters don’t work for free and they aren’t cheap. Did the person or group paying for this poll have any reason to hope for a certain outcome that may have influenced the results?
5. Look at the questions for bias and also word choices. People that may share the exact same views might answer differently depending on how the questions are phrased. Words that sound alike, such as “prophets” and “profits” or “very” and “fairly” can often be confusing on a phone call. Respondents should be aware of what is being asked of them, not “tricked” into giving responses.

There are many other factors to consider but that is at least a good start to understanding why polls can be so different. This past week we have such a case right here in Georgia where two polls had nearly opposite numbers.

11alive News in Atlanta paid for a SurveyUSA poll.

Deal, 48 percent;  Carter, 39 percent; Andrew Hunt, 4 percent.

David Perdue, 48 percent; Michelle Nunn, 41 percent;  Amanda Swafford, 3 percent.

About a week later Channel 2 Action News paid for a survey by Landmark.

Carter, 44 percent; Deal, 40 percent; and Hunt, 4 percent;

Nunn, 47 percent; Perdue, 40 percent;and Swafford, 3 percent.

Margin of error for both is about the same at around 4.2% +/-. So who is right and who is wrong? In reality, poll numbers in August mean very little since they will likely change after the debates and ad campaigns get in full swing. But it is still good to know which poll we can put more trust in as we see more polls closer to the election date.

Landmark employed live interviews via land lines and cell phones which is now considered to be the gold standard.The SurveyUSA poll was a combination of automated calls to land lines and an Internet questionnaire conducted via smart phones but did not include live interviews or calls to mobile phone numbers.

Other Key Differences:

SurveyUSA survey presumed African-American turnout at 26 percent. Landmark survey presumed African-American turnout at 29 percent. Historically 29% is the average number. I would have to give the accuracy to Landmark on this poll given the mitigating factors.

Which poll you choose to believe is up to you. But at least you now have a few tools to more objectively analyze poll results. Be careful of obviously skewed polls from companies with a known left or right leaning history. Georgia is no longer a deeply red state as some may believe. It you wanted to pick a color it would be more accurate to say it is a purplish-red state. 

As an independent with no party loyalty, I base my preference of two simple factors.

1. Can I trust them? Do I like them? Do I think they will keep their word over campaign promises? Are they smart and articulate? Do they come across as honest, trustworthy, and just a decent human being or more of a slime ball? Would I enjoy having a conversation with them over dinner. This is what we  call the gut feeling or likability test. 
2. I then make a checklist of issues. I find issues of importance to me. For Governor,  their foreign policy views are inconsequential since a Governor would not deal with foreign policy to any significant degree. For Senator that would be an important issue. My topics are specific to the race. Then I simply go down my checklist and see which candidates comes closest to my positions. 

I have made no secret of my disdain for Nathan Deal on this website. I think he is quite possibly THE worst governor in the country. As much as I would like to hide my bias against Nathan Deal, I simply can’t. Here are a few articles I have written explaining how I came to my have my views on Shady Deal as I like to refer to him. They explain why I hope he does not get four more years in office to wreck our state even further than he has already done.

If Deal manages to pull out a victory, I do not think it will be because a majority of Georgia’s voters necessarily support him. Rather it would from people voting against Jason Carter. In all fairness the same thing would likely apply if Jason carter won. Probably around 5% to 15% always hold their noses when they vote. This is also known as the “voting against lesser of two evils” principle.

Carter is still a relatively unknown entity and has to prove he is more than simply the grandson of former President Jimmy Carter. Deal’s campaign has already started trying to paint Jimmy Carter as anti-Israel to try and win over older voters who tend to be very pro-Israel in the hopes this will tarnish Jason by association. This may actually backfire long-term as more Americans are much more educated about the Middle East thanks to Iraq, Afghanistan, 9/11, and the Arab Spring than they used to be. Poll numbers have begun to shift on Israel. This is especially notable among younger Americans.  Both parties are still very much pro-Israel and the argument is more about which one is more Pro than about any party being Anti-Israel.

Most Americans understand Hamas is a terrorist organization, but they may not feel that necessarily justifies all the billions in aid to a country with a GDP per capita getting closer to our own anymore either. The images of around 1,000 or more dead Palestinian civilians with a ratio of around 70% innocent to 30% Hamas killed doesn’t help. It is also very true that Hamas is just as responsible since they use civilians as human shields. But Deal and Perdue cannot count on this as a wedge issue either to bring out voters as in past elections.  Deal shouldn’t bring it up at all since he is running for Governor and should focus instead on problems in Georgia and not Gaza.

Perdue and Nunn are both unknowns with no political track record. Like Jason Carter, these two also have to step out of the shadow of the more famous members of their family that share their last name. Former Senator Sam Nunn is Michelle’s father, and former Governor Sonny Perdue is David’s first cousin. Sonny Perdue was not really a very popular Governor (even among Republicans)  so I am not sure if his connection to David will help much beyond simple name recognition. Sam Nunn left the Senate as a very popular Senator and even today is held in high regard among many Republican voters. Voters under 35 years-old will not be as familiar with him though since he retired back in 1995.

I am currently undecided in the Senate race and will wait for all the debates and do some more research before I make up my mind. Michelle Nunn seems to be running pretty far to the right compared to other Democratic Senators in blue states which is understandable in Georgia. Perdue will have a hard time painting her as a liberal when she shares or is even to the right of even some Republican Senators from other states on certain issues.

Will Georgia have a Run-off?

The answer is almost certainly yes. I think at this point in time it is highly unlikely that anyone can win 50% plus 1 vote for Governor or in the race for the Senate. If you don’t like either candidate from the Republican or Democratic party and you really like Andrew Hunt (Libertarian) for Governor then you should vote your conscience. You have the security of knowing your true voice was heard then you can always get a second chance in the run-off once it is narrowed to only two candidates. The same applies in the Senate race with Amanda Swafford (Libertarian).