Why does nobody talk about how polls shift after the first 500 responses?

I ran a local survey in Dayton last month and the first 500 people said one thing, but by 1,200 responses the split flipped 9 points, so how do we trust any headline number without the sample size breakdown?

1 comments

1 Comment

jason_stone599d ago

But is a 9 point flip really that big of a deal in a local survey though? Polls are just a snapshot of who answered at that exact moment, and the first 500 people might just be the ones who care the most or have the most free time. By 1,200 you're getting a more average crowd, so of course the numbers are gonna move some. I run into this all the time with online polls, people click through fast and then it settles down. The real issue isn't the shifting, it's that nobody ever shows you the margin of error or how they picked the sample in the first place. If you're gonna trust a headline number, you gotta look at the full breakdown, not just the first batch.