SERVEFollow

Stay classy.

Somehow your lack of understanding the difference between applied math and abstract math makes me doubt that. If you are indeed, I assure you, there's a "problem" in your department as there is absolutely no need for any form of statistics for O.D.E, P.D.E., Cal, Linear Algebra or any other Applied math and only the bare concept of probability in Abstract math.

To be fair, I didn't understand the difference until my elective sequence.



P.D.E is not light weight.



And they shouldn't be Math majors. I'll tell you what one of my professors told us, "our job isn't crunching numbers". I took that message to heart. I used to purposely lose points in my Physics classes because I didn't fully reduce my answer. Once I got rid of all of the integrals and down to fractions, I stopped. You can add the fractions your own self.

Crap, might as well mandate accounting courses while you at it. There's a reason why there's a Business Calculus class. I'm not implying that their entire job is "number crunching", I've seen how stats uses Cal in their calculations, but if your goal is to find out statistics, then you're better off in the statistics field. If your goal is working with mathematical theories and concepts and/or equations to solve real world problems, then you should be a Math Major.

Just because they fit together perfectly in a certain area, doesn't mean it should be mandated for math students. The same level of calculations can be done in Economics and social sciences, they shouldn't be mandated either. You're wasting valuable class slots that could be used for something else. Common knowledge as you mentioned "well-roundness", is one thing but mandating an entire class, especially more than one, might be too much. The math version of the prob and queue that I didn't take was also a proofing class. That's how you get your well-roundness in, you cram it all in one or two classes that consists of various subjects.



Wait, so let me get this straight, before I go to sleep.

1. You don't believe that there are any cons with probability sampling that is countered with non-probability sampling.

2. You don't agree that there exists numerous other factors outside of being random that plays a part in the outcome, i.e. with or without replacements, phone vs email vs snail mail, time of phone calls, automated vs human, spoken or written language, home phone vs cell phone, etc.?

3. You don't believe that it's possible to have two completely different results from a 1000 person sample from a divided 300,000,000 population? For example, one survey says 60% for x and 40% for y while the second survey says 60% for y and 40% for x?

4. You don't agree that the more people you sample, the more representative you are of the entire population?

5. You don't agree that the more people you sample from a given population, the less likely number 3 will occur where you have two completely different polls? For example, given a population of 100 with 50 Repubs and 50 Dems, by sampling only 50 it is likely to not represent one party at all or have a favored representation such as 40 and 10. On the other hand, sampling 75, guarantees 25 Dems and 25 Repubs.

If you agree to these statements, then how again am I wrong?

Same thing that makes their chicken so good. I should have guessed.

I'm not sure why I'm even responding as I already provided you with contradicting polls that happened just last month. So I guess one in 4.657 quintillion happens more often than you think, huh? What are the odds of THAT happening? Just last month, what a coincidence.

Anyway, you were arguing that it could go either way, that the poll was representing that the U.S. is divided, now you're arguing that it's nearly impossible to have a varying poll. Hmmm. I guess Fox News and CNN should talk to each other more.

Edit: Just to make sure. You are arguing that it if the nation is roughly divided on an issue, it is statistically low to poll 501 people out of 300,000,000 in a 1000 person sample to favor a particular opinion?




I got it from University of Memphis. I'm sorry I forgot to respond to that question. Ugly knows that, he's just letting his jealously cloud his trolling. People laughing at me for more education and making more money really doesn't affect me as you might think. ;)

Troll harder.

Edited, Apr 23rd 2011 2:15pm by Almalieque

It was 42% to 35%. I don't know the error margin, but I hope it wasn't 7%. In any case, that doesn't matter as my point was that even if it were 49 to 51 with a 2% error, that doesn't tell you what the majority believes, so you can't go around claiming the U.S. believes "x". That was my argument against using the poll in the first place.

Edit:


More like "This is a basic math concept that doesn't involve any need for anything learned in a Statistics course".

Way to completely miss the point and join unrelated topics.

Edited, Apr 23rd 2011 3:11pm by Almalieque

What about 3.5%? Note that there's also 23% of the responders that did not give an opinion at all. Indifferent, not sure, won't say, etc.

No. Are you honestly this dense? The chances of getting exactly 501 out of 1000 if the distribution is 50/50 is still a fair 2.5%, and the chance that there'd be 501 or more in favor on a 50/50 split is 48.7%.

Seriously, just admit you don't understand statistics at all if you're making errors this glaring.



Edited, Apr 23rd 2011 1:21pm by bsphil

The point was that it was a totally different conclusion.



When you denied non-probability sampling, you lost all credibility of understanding stats as well as you proclaim.

If you stop projecting stupidity onto me, this would be so much easier. I don't know why people continuously take the dumbest interpretation of every statement. I mean, do I have to explicitly spell out everything for you? I didn't mean EXACTLY 501, I meant at least 501 because that would make the majority.

So I'll state again. You are arguing that if the nation is roughly divided on an issue, it is statistically low to poll AT LEAST 501 people out of 300,000,000 in any random 1000 person sample to favor a particular opinion? This is under the assumption that each opinion has at least 100,000,000 supporters.

Carls Jr. waffle fries

My bad. Then what's the problem?

Ok, this isn't a proof, but an example as a counter..

Let there exist a sack of 10 marbles (Red, Blue) where there aren't 5 of each color. Pulling out one marble at a time, what's the minimum amount of draws to see which color is the majority? What is the minimum amount of draws that you would have to do in order to guarantee to see which color is the majority every time?

One was non-probability based and the other wasn't. That was part of my argument, doing a random poll with a small sample size on a large population on a divided topic will not yield the same results. The latter poll is how most polls by Fox, CNN, etc. are done.