You really don’t need to survey many people to get statistically significant results, assuming your sample is truly random. For a population of 340 million, you only need to randomly sample ~2500 people to get a 95% confidence interval with a 2% margin of error.
A sample of 9000 people would get you closer to a 99%+ confidence interval.
Does a pre-cleared set of volunteers who willingly gave their demographic information in order to participate in online polls count as a random sample?
Well that’s the thing. Do meaningless things make the national news everyday in October? Yes. And here we are in June. They’re gonna throw more money at us until we forget about 2016. And since that’s not going to happen, it’s all this.
You really don’t need to survey many people to get statistically significant results, assuming your sample is truly random. For a population of 340 million, you only need to randomly sample ~2500 people to get a 95% confidence interval with a 2% margin of error.
A sample of 9000 people would get you closer to a 99%+ confidence interval.
Does a pre-cleared set of volunteers who willingly gave their demographic information in order to participate in online polls count as a random sample?
I don’t know how participants in polls are selected, so I’m not really qualified to make assumptions about it.
The methodolgy section of their site lays it all out. It’s a selling point. And no, it’s not random.
In that case, the data is practically meaningless :D
Well that’s the thing. Do meaningless things make the national news everyday in October? Yes. And here we are in June. They’re gonna throw more money at us until we forget about 2016. And since that’s not going to happen, it’s all this.