Loading…
26
My neighbor's kid showed me a chart that made me rethink polling numbers
Was over at my neighbor's house last weekend and his 16-year-old daughter was working on a school project about polling accuracy. She pulled up a scatter plot of 2020 election polls versus actual results and pointed out how the error margin was way bigger than what news outlets reported. Said most polls only sample 1,000 people out of 160 million voters and act like it's gospel. I asked her why it mattered and she just said "because 3% error means 5 million people could be wrong." Hit different hearing it from a kid. Anyone else ever dig into the actual sample sizes behind those viral polls they keep posting?
3 comments
Log in to join the discussion
Log In3 Comments
olivia_white932mo ago
Oh wow, that's such a sharp observation from her. I was looking at some local election numbers last month and noticed the same thing - the state polls I followed had way bigger swings than the national ones, sometimes 5 or 6 percent off. It really makes you wonder how much trust we put in those little numbers they flash on the news.
8
mason_reed472mo ago
That 3% margin claim really stuck with me too, but did the chart show how many of those polls had even wider error bars in actual local races? Would be interesting to see if the problem is worse with national polls versus state level ones.
4
wilson.olivia2mo ago
Yeah that's a really good point... I've been wondering the same thing myself because local polls seem way more all over the place than the big national ones.
6