SERVEFollow

Eh, I kind of liked Wendy's fries better before they were changed.

They changed the salt, I believe they now use sea salt. So much more tasty.

Edited, Apr 22nd 2011 1:17am by rdmcandie

Hmm, not really convinced that sea salt is that significant of an improvement. Next time I get the chance I'll see for myself.

I believe that they also use more salt, which would improve the taste as that is what salt does. They are pretty good, but Wendy's has always had the best fast food fries.

Oh, well.. It's your denial.. I sourced it for you. So if you choose to believe otherwise, that's fine, but I choose to live in reality. I do find it quite hilarious that the person claiming that I don't know anything about samples in actuality didn't know anything about samples. I admit, you had me going at one point. I thought you actually knew what you were talking about in reference to samples, just confused on my point. Twas a nice run...

But if you need the last word, I can give it to you. just to make you feel a little bit better.

Edited, Apr 22nd 2011 2:05pm by Almalieque

You're absolutely correct, but that's under the assumption that your random selections are not heavily favorable in one way over the other. That's why I suggested taking in other factors for consideration such as "with replacements", "without replacements" or taking out the random factor all together.

The same people being polled on their opinion on rape would probably be consistent as it's evident that "most" people feel one way. That isn't the case for SSM, so the slightest bit of misrepresentation can and most likely will change your results. The ONLY way to counter that, under a random sample, is to increase the numbers. That's just plain math.

I'm not going to pretend to fully understand that process, but according to the link below, it isn't that trivial. There are much more factors involved in determining the margin of error. It supports the simple mathematical concept that if you're not using replacements, then the more people you use, the more likely you will have a fairer representation. It's a basic math concept.


http://en.wikipedia.org/wiki/Margin_of_error


Edited, Apr 22nd 2011 4:32pm by Almalieque

I didn't read the whole thing either, but I assumed that was basic knowledge. The whole point of sampling is representing the entire population with a smaller population. The problem is, with random sampling, you're not guaranteed to get that representation. So the more people you sample, the closer you are to the full population, which by default is more representative, but not necessarily *accurate*. For example, 20 people is more representative than 10, but that doesn't mean 20 people accurately represent 300,000,000 people.

You misread my point. I was comparing 20 to 10. 20 is more representative than 10, but not necessarily more accurate. This is to show that a higher number is more representative, but not necessarily more accurate.

You're not guaranteed to get a fair representation on something so divided. It's not the number. Once again, 20 people probably represents the nation's opinion on certain topics with little error. The problem is getting a fair representation on a divided topic with such a small number.

Edited, Apr 22nd 2011 5:37pm by Almalieque

Your continuing ability to misunderstand and misrepresent solid, unchanging mathematical fact continues to amaze me.

That's odd, I was pretty sure that 20 was greater than 10 and was a greater percentage for any number equal to or greater than 1. If you believe otherwise, I would love to see you prove it wrong. I don't even need to use my BS in mathematics to prove that wrong... The ball is now in your court..

No, if you don't fix your bias error and simply increase the sample size you're going to get the same errors...but bigger.

Edited, Apr 22nd 2011 12:36pm by Eske