I guess you don’t see how contemptible this is. You clearly don’t “get it” about the basics of this argument and yiu want to cheerily pretend it’s a little personal tiff. It’s not. You’re wrong, I’m right and it’s really not a matter of opinion. I may have slightly overemphasized the MOE but you seem clueless about it.
I still don’t know what I’ve said that you disagree with. Why not quote a post and point out the specific words that I’ve said that you believe are factually incorrect? ISTM you’re still arguing with a straw man.
Holy shit who cares.
Be patient… We’re this close to mutual understanding and togetherness! Just a few more posts is all we’ll need!
Gotta agree. Arguing with andy is like wrestling elmo flavored jello. Sorry I got dragged in.
Best to just declare victory. Sure, you could actually try to explain, specifically, what you think I’m wrong about, but declaring victory is much, much easier.
I don’t need to declare victory. Reality does the job for me.
It’s because of this:
In post 2410, CarnalK makes an assertion that the higher MOE means it is not an especially accurate poll. In post 2415, you say that we don’t know if it’s more or less accurate. **That is the disagreement.
**
Then, in post 2419, CarnalK says that the larger MOE means the poll is less trustworthy. In post 2420, you agree.
Post 2415 and 2420 by you are contradictory.
“Less trustworthy” and “less accurate” are not the same things. For all we know, that Monmouth poll is right on the nose, presenting reality perfectly at the moment it was taken. It might be perfectly accurate, for all we know. But it does have a larger MoE, which means a greater possibility of lower precision than a poll with a smaller MoE. And, of course, as Richard Parker mentioned, there are many other factors that go into the accuracy of polls. All of this is in my understanding, of course.
A subtle but still very significant difference, IMO. I see no contradiction.
No. That is not what MOE is. It is a measure of the likelihood of getting a result by chance. Accuracy, by contrast, encompasses much more.
Elmo flavored jello. Kids love it.
Congratulations! I’d love to learn what I’m wrong about, but I guess I’m not bright enough, and of course I can’t expect someone as brilliant as you to deign to expend the effort to try and explain it to dullards like me, right? No one can expect that kind of charity.
Best wishes to you once again!
CarnalK: You also seem to think that MOE means any result in that range is equally likely. That is also incorrect. The most likely reality is the one reported by the sample. In only a small percentage of cases would the underlying reality be at the MOE margin (by chance as opposed to some other driver of inaccuracy unrelated to MOE, of which, again, there are many).
So what, specifically, do you think I’m wrong about?
Remember when you said this poll was nice for people like you who don’t want Biden to win? You were wrong about that because this poll is pretty useless. Next question?
Actually, it’s still “nice” for me. That was a statement of feeling and nothing more. Did you really believe that I was using “nice” as some sort of factual indicator or descriptor? If so, then I can happily correct this misconception.
No, that’s not true. Please decipher “For results based on the Democratic voter sample, one can say with 95% confidence that the error attributable to sampling has a maximum margin of plus or minus 5.7 percentage points (unadjusted for sample design).” in a way that matches your interpretation. AFAIS, you’re just wrong.
Not sure if there is actual disagreement or just loose language. But just to summarize:
Whether the numbers reported in a poll reflect the underlying reality depends on a whole bunch of factors. These include: statistical chance, the demographic weights applied by the pollster, sampling bias resulting from the methodological choices and resources of the pollster, whether the pollster fudges (or non-reports) because of herding effects, how recently it was conducted, and others.
The margin of error is a handy measure of only one of those many factors–statistical chance. It accounts for the possibility that you got more Sanders voters by dumb luck rather than the proportion of the population. Of course, dumb luck works in both directions. So on average it has no effect. It follows that if you calculate, for example, the margin of error based on 75% instead of 95% you get a much smaller window.
Now, if you have a poll with a big margin of error, that might make it less powerfully predictive than another pol with a smaller one, but only if the other factors don’t outweigh the difference. If you have a grade A pollster with a poll that was conducted recently with an MOE of 6% and you have a Grade C pollster with a poll that was conducted weeks earlier with an MOE of 3%, the former is going to be more accurate than the latter much of the time. If you were making a bet based on only one data point, you’d probably pick the one with the bigger MOE.
You are mistaken about how MOE works if you think all the results in the range are equally likely or the reported result is not the most likely. I don’t know what more to tell you. You can go read the Wiki page.
Here’s the wiki page: Margin of error - Wikipedia
I think the graph at the top would be particularly helpful for this discussion.