• Log In
  • Sign Up
    • Last week, I came across an interesting polling statistic from CNN: While the amount of Americans supporting impeachment and removal from office (50%) hasn’t changed over the course of the impeachment hearings, there has been an increase in the amount of women who support impeachment and removal (61%) and a decrease in the amount of men who support impeachment and removal (40%).

      What this  got me to thinking about is polling as a whole, how accurate they are, and how much stock we should put in them. The only poll we can be certain about in politics is the final one. Namely, the elections themselves. Until people actually place their votes, we’ll never know for sure who will win any election. In other words, are polls nothing more than just something to entertain us until the actual election happens? 

      One other thing this poll got me to think about is what role does gender play in how we view the world? The fact that 61% of women support impeachment and only 40% of men is pretty wild. Why would it be the case that so many more women would support impeachment than men? I understand that in other polls, you can’t really see a gender gap, but in this one there most definitely is. 

      Below is an interesting article I came across that helped she some light, at least for me, on the subject. Curious to get all of your thoughts on this topic. 

    • Thanks, @slamdunk406, this is a huge topic and I don't know much about polling but when I see a political-type poll, I always check the raw data to see the source, the sample size and weighting, and the way the question was asked. I have written surveys for organizations and companies and found it is not an easy task on many levels.

      A lot of political polls are crafted with acquiescence bias or else they ask loaded questions. The order of the questions asked can also skew results and some polls use compound questions, meaning you are giving one answer to two questions.

      The sample sizes are really critical too, as they need to be at least 1,000 but not much larger than 2,000 or else the reliability is not any better. And sometimes, the "buzz" pulled out from the original poll is misleading when you consider what people were actually responding to.

      I think you have it right when you say polls are nothing more than something to entertain us with! And I really don't know why there would be a focus on gender here either. As a woman, I cannot understand why anyone needs to divide me out from people in general in terms of my opinion on something. When it comes to counting, I want to be considered equal with, not different from, every living person!

    • Thanks for weighing in, @Jain! As you said, there are a lot of variables to polls that make them tough to trust. The questions asked, the way they're asked, etc. Basically, the element of bias is one that can make things really tricky. Plus, it also matters where the polling is taking place. Each region has its own culture that shapes the way people view things. I would imagine that makes it harder to get accurate polls at a nationwide level than at a local level. But then again maybe not.

      I definitely do think that polling at its core serves two purposes: #1. To entertain us. #2. To give political campaigns some sort of insight into how they're doing. I like to compare it to sports. You can look at statistics to see how teams match up against each other, but no matter how much you break it all down before the game, you really don't know for sure what will happen until the game is actually being played.

    • I am surprised polls are as accurate as they are. When you stop to ask who actually answers pollsters, does anyone pick up their phone anymore to callers they don't know? How can you guage the pulse of a nation from 2,000 people?

      I am no authority on how accurate they are but my purely unscientific sense is they are about as accurate as weather reports and more accurate than sports predictions.

    • I really wonder if polling has been adversely affected by the massive tech changes we’ve seen over the last decade or more. Back in the day, people were fascinated by the glimpses they got of “metadata” that pollsters provided. Nowadays, we tangle with issues regarding metadata all the time.

      Twenty years ago, people were much more impressed with poll results and anxious to compare their own opinions to the poll results. Today, we compare (and broadcast!) our own opinions to “the masses” every day on social media, and insulate ourselves from divergent opinions so much so that those divergent opinions are sometimes seen as “the enemy.”

      Now, when a poll result turns out different than our own metadata experience, we just chalk it up to pollsters calling the wrong people, or manipulating the question, or having some ulterior motive.

      I think polls have very little impact on how we see the world these days. Now, they seem more like an old-fashioned pastime than a scientific sampling.