Big headline: “Majority of America wants Trump Charged”. “Just over half of the respondents in a new poll say former President Trump should be indicted for encouraging his followers on Jan. 6, 2021, to march to the Capitol, where a mob stormed the building but was ultimately unsuccessful in overturning the 2020 presidential election results. The poll showed 52 percent of respondents said that Trump should be charged for his role in the Capitol riot.” Wow, 52%.
Survey researchers, not typically known for their humor, have an old joke that goes, “If you don’t believe in random sampling, the next time you have a blood test, tell the doctor to take it all.” The analogy of a small sample of a patient’s blood drives home the point that if selected properly, all that is needed to get a president’s job approval rating (or a cholesterol count) is a sample. Think about a chef who tastes a sauce. If the sauce is well stirred, the chef can try just a spoonful rather than eating the entire potful to assess its flavor. A wonderful theory.
Let’s look a little deeper. The poll was conducted by the Washington Post in connection with ABC News. Not exactly “non-biased” sources.
Look a little more, they based this conclusion on the responses from 1,004 “respondents”. One thousand and four people. Or .0000030424 percent of the US population. Yeah, that’s a real cross section of the voter base. The pollsters will go into deep long explanations on how this is the perfect “random” sample.
There is just one tiny problem as I see it. In working with the Hawaii Legislature on/against mandatory helmets and other motorcycle laws, I found that a poll will say whatever the person/group paying for it wants it to say. How do I know? I’ve done it.
The response to a poll depends on where you ask the questions, when you ask them, and how you ask them. These three things can be slanted and used to prove whatever you want the poll to prove. Example? Okay.
I go to an “upper income” area church, on a Sunday morning, and I ask the first 100 adults walking out of that church the following question: “Because of the increased injury level when being in an accident and for their own safety, should motorcyclists be required to wear a helmet full time, regardless of their age, training, or experience level in operating a motorcycle?”
I’m willing to bet I’d get an 85% “YES” response. Plus, or minus 3% of course. The headline would read, “Majority In Favor Of Helmet Laws”.
Now I go to the local strip club area, on a Friday night to Saturday morning (early), and I ask 50 people going into the clubs, and 50 people coming out of the clubs the following question: “Should bikers, for their own good, be forced to wear a helmet all the time?”
Want to bet against “NO” being the majority answer? Me either. This headline would read, well let’s face it, there wouldn’t be a headline, it wouldn’t fit the narrative.
See, location, timing, and phrasing.
A poll shows exactly what the buyer wants it to say. It shows nothing more, and often a lot less.