Impeachment polls, what’s real & what isn’t - Part 2
Bottom Line: Anytime you’re analyzing polls analytically it’s important to see what the political sampling was and adjust for the real political conditions on the ground. In the case of the much-reported Fox News impeachment polling, they sample over-sampled Democrats by 8%. Now it’s not as easy as pointing to the 8% and saying – see it’s 8% too many Democrats. There are more Americans who identify as Democrats than Republicans. That’s the next layer of understanding.
The best way to gain an idea of what current conditions are nationally is to look at Gallup’s monthly political ID polling. Gallup has polled monthly on party ID of Americans since 2000 and as a result provide the best real-time temperature check at any given time of political preferences. Most recently 2% more Americans ID as Democrats than Republicans. In Florida there are 1.8% more Democrats than Republicans. Leave it to Florida to resemble almost exactly the rest of the country with just an ever so-slightly more conservative bend. Anyway, this means that the Fox News poll over-sampled Democrats by six points. So, any use of this poll would have to factor in a 6% oversampling of Democrats, or 6.2% if you want to infer what Floridian’s think. What’s more is that they didn’t breakout the splits on Democrat responses vs. Republican responses to the question to be able to analytically provide you with adjusted accurate results – meaning there’s no credible way to specifically use their polling to provide a specific answer to the question. But there’s enough here to offer a general conclusion.
Based on the Fox News Poll’s oversampling of Democrats, I can empirically say that most Americans don’t support impeachment. And this exercise is an initial exercise into how to analytically use polls during election cycles and how I’ve been able to effectively use them to accurately project outcomes in elections going back to the 2000 cycle, including accurately projecting the wins for George W. Bush in 2000 and Donald Trump in 2016. The problem with polls is taking them at face value. The information within accredited polls is still highly instructive in context.