Originally Posted by Brian Rush

It's actually rather different than that and more complex, due to the influence of corporate money, which pulls the center of Congress to the right of the national center, but only on economic issues, not social issues.

Of course I do not now have the tools for multi-variate analysis. I do not have a powerful tool for graphics in which I could have overlain a standard bell curve. All in all the extremes like Commies and Kluxers are unrepresented in political results (not that we need a cross burned or someone singing the Internationale in the House Chambers) -- but we have a surfeit of people who would have been comfortable in Franco's Phalange in Congress and few who would be CDU in Germany or Christian Democrats in Italy.

Tycoons, executives, and big landowners have been core financial supporters of fascist movements elsewhere... and if they don't particularly want book burnings, torture chambers, witch-hunts, or concentration camps here they clearly want extreme inequality and economic hierarchy with a government that represents their economic interests to the exclusion of the interests of all others. What is missing from an American fascism is thuggish enforcers who vandalize property and attack opponents. When electoral processes fail the Hard Right, as demographic change assures, will the Hard Right turn to violence?

As a result, on social issues the parties represent the people pretty well. The Democrats are at or a bit to the left of national center on issues like gay rights, gay marriage, women's rights, racial equality, church and state. The Republicans are quite a bit to the right of national center on these issues, representing a values-voter constituency of their own that's pretty far to the right. Thus we get the repeal of DADT, for example, right on schedule with the public's gradual shift of opinion on gay rights, and states enacting gay marriage on the same pattern.

The current Hard Right consists of the most materialistic people in America -- the rapacious rich -- and the people most willing to sacrifice all earthly happiness for the assurance of a happy eternity. Those segments rarely meet, and would probably hate each other if they ever did because our economic elites exercise no restraint in personal life. For an analogue, think of the old Democratic coalition between urban 'ethnic' whites heavily involved in unions and rustic Southern segregationists who would have despised each other if they ever met. Gated communities exist as much to keep out poor whites as they exist to keep out poor members of minority groups (aside from cowed domestic 'help'). The economic component of the Hard Right has no problem with gay rights because it can always shield someone 'incorrigibly' gay.

The non-white, non-Christian, non-Anglo, highly-educated, and LGBT members of the middle class are increasingly seen as rivals for political power... and increasingly competent, relevant, and credible. These people identify themselves against the elite and against the Christian Protestant fundamentalists who constitute the Hard Right and are Democrats. The Hard Right would love to eviscerate their economic success.

On economic issues, though, the national center is to the left of where most of the Democrats in Congress vote. Most Americans want a more thoroughgoing health-care reform than Obamacare, want taxes raised on the rich substantially, want an income distribution in this country that resembles Sweden's, want jobs brought home and have no respect for the idea of free trade if it costs jobs, want corporations and rich individuals prevented from moving money out of the country tax-free, and so on.

The Democrats are a center-right party by European standards; the Republicans are by contrast in the area in which questions would be raised of their loyalty to democracy.

On economic issues, the national center isn't much if any to the right of Bernie Sanders. Congress could move WAY left on economic issues and be very popular. It is prevented from doing so by the veto imposed by corporate money.

Ain't it the truth!

Originally Posted by JustPassingThrough

I explained what the two things are in my above posts. And yet somehow, you can "rest assured" that I don't recognize the difference between the two. In fact, I was using the difference between the two to prove a point. They are alternate means of analyzing a set of data. The median, while a blunt instrument, is useful precisely because it discards outliers that skew the mean.

In other words, what you have shown is that you have no clue what I was talking about.

If you think you have outliers, which ones are they? Typically, outliers are more than 3 sigma above or below the mean, though even that may be arbitrarily restictive. At 8, this is a very small sample set, so sigma will be large. Do you want to argue that any of the results fall that far from the mean?

Originally Posted by JustPassingThrough

If the error is .7 in a set of 8 numbers, that means that somewhere in the sum of those numbers there's an extra 5.6 in Obama's favor. That 5.6 number is not essential to what I was saying, but since there were polls of +7 and +8, it helps illustrate where that error is likely coming from. I probably didn't express it clearly enough. The average is not off by 5.6, the sum of the individual polls is off by 5.6.

All it means is the distribution is curved rather than linear. Wonkette is right. That there is an error is a given (review the Central Limit Theorem). The siz of the error is another thing all together.

Well, JPT's theory is not looking so good today. Obama is still bouncing!

http://www.realclearpolitics.com/epo...bama-1171.html

Originally Posted by Brian Rush

I disagree, although that is a mistake, yes. There's a more fundamental error he's making, though, which is making much of a much of the fact that the mean and median of the polls are not exactly equal. This doesn't come from the numbers themselves, but from what the numbers are numbering -- the things involved; qualitative rather than quantitative reasoning. As such, it isn't going to come out in number-crunching. Let's consider what exactly that implies and what it doesn't.

The mean is found by adding all the poll results and dividing by the number of polls. The median is the number or range of numbers where there are as many polls higher (measured, say, by Obama's advantage) as there are lower.

The reason this is not indicative of much is that the median doesn't take into account the size of Obama's advantage in any one poll except to note that it's higher or lower than some other poll. The mean being higher than the median indicates that a relatively small number of polls show Obama with a sizable advantage, and these few polls are pulling the mean up slightly compared to the median.

At most, it indicates that there are a few polls -- or maybe just one poll -- that's an Obama-leaning outlier. If the mean-above-median result is consistent, and if we can identify the outlier polls consistently, a correction might be argued involving cutting the outliers out of our calculations. Of course, the same could be argued w/r/t Rasmussen at this point in the election cycle, insofar as Rasmussen is certainly and deliberately an outlier the other way.

There is of course no justification for doing what JPT did with the numbers other than his own wishful thinking.

Uh...what you just described is exactly what my point was. It shouldn't even require everything I did, you can just look at that list of 8 polls and see that Pew and National Journal are way off from the others and skewing the average. But I wanted to spell it out, and that is precisely the value of comparing the mean and median. I also showed that every time you discard all polls outside the standard deviation, Obama's lead shrinks. It's a simple point, and there are a variety of ways of illustrating it using basic statistics. The polls that skew in favor Obama are farther off from the center than the polls that skew in favor of Romney. It's not complicated, and nothing I said in those posts was wrong. But the insults have to keep flying, because that's all you know.

Nightmarish -

http://www.politicususa.com/wheels-f...ans-48-41.html

The Wheels Fall Off for Romney as Obama Leads with NASCAR Fans, 48%-41%

According to a new Zogby Poll the wheels are close to coming off the Romney campaign. President Obama leads Mitt Romney among NASCAR fans, 48%-41%.

The Zogby poll revealed that President Obama has a seven point lead over Mitt Romney among likely voters, 48%-41%, and even larger nine point lead when all candidates are included, 49%-40%. The real shocker is that the poll found the president doing very well with some traditionally thought of Republican voting blocs.

Obama leads in almost every region of the country, and only trailed Romney by six points in the South, 41%-47%. Among voters who shop at Wal-Mart on a weekly basis, Romney’s lead is within the poll’s 3.4% margin of error, 45%-42%. Obama leads with voters who are or who have a family member in the military, 54%-39%. Most surprisingly, Obama only trails Romney by seven points with those who consider themselves born again Christians, 40%-47%. The president also leads with NASCAR fans, 48%-41%.

Romney’s problems with NASCAR fans could go back to when he mocked them while attending the Daytona 500 this year. The New York Times reported Romney’s Daytona experience in the most classic of ways, “But the crowd initially booed Mr. Romney, who occasionally struck a discordant note, as when he approached a group of fans wearing plastic ponchos. ‘I like those fancy raincoats you bought,’ he said. ‘Really sprung for the big bucks.’ And when asked if he was a fan of the sport, he mentioned that ‘I have some great friends who are Nascar team owners.’”

Romney mocked NASCAR fans, and then he followed it up by mocking the 47%, some of which I am sure also follow NASCAR. Overall, Romney’s poor showing with NASCAR fans reflects his inability to connect with people. NASCAR is a national sport now, so it isn’t just Southern white people watching anymore. Romney’s NASCAR fail was a sign of a much bigger problem. Romney has tried to suck up to NASCAR fans, but it looks like his insincere advances are being rebuffed.

The fact that President Obama is doing so well with born again voters isn’t much of a surprise. Romney has struggled to win over the South from day one of his campaign. It was the South that rejected Romney all through the Republican primary process and kept the candidacies of Newt Gingrich and Rick Santorum alive. Less than two weeks ago Reuters ran a story about Romney’s religion and his wealth. Romney has struggled to convince born again voters to support him, and a sizable percentage of them are supporting President Obama.

President Obama is leading Romney by large margins with Independents (46%-32%), Hispanics (68%-28%), African-Americans (93%-0%), and moderates (60%-26%). Romney needs to win big with men (Romney +5), seniors (Romney +3), and especially what has been traditionally thought of as the usually conservative Wal-Mart shopping (Romney +3), NASCAR watching voters (Obama +7). In this poll, Romney had no area of support large enough to offset Obama’s margins with other blocs of voters.

In short, things are bad for Romney.

How bad?

The born again and NASCAR crowds are flirting with dumping him for Obama.

Originally Posted by JustPassingThrough

In fact it is, although a lot of it is also just groupthink.

Oh. OK. In that case ... though whoever is organizing the conspiracy shows far more organizational ability than any dictator or corporate executive in history. No dissenters, no defectors, nobody claiming they can make more money by going their own way ... miraculous! I wish whoever it is could come and moderate my club meetings!

Originally Posted by JustPassingThrough

Uh...what you just described is exactly what my point was.

Then you did just about the worst job of mis-representing your point I have ever seen.

In the first place, one glance at one polling sample isn't enough. You need to show a pattern of the mean being more in Obama's favor than the median. And then you need to show a consistent pattern of one poll (say) being an Obama-friendly outlier. This has not been done.

But let's pretend for the moment that it had, and that we've decided that the Pew Research poll which has Obama at +8 is a skewed outlier. Do we take the difference between the mean and the median and multiply it by the number of polls as you did? No, of course not. We just refigure the mean without the Pew poll.

If we do that, the new mean favors Obama by 3.1. It makes the result slightly less lopsided, but still clearly favors the incumbent. (It also brings the mean much closer to the median.)

Now, I'm not saying this is really warranted. As I said, we'd need to do some more research and look for consistent patterns over time before drawing that conclusion. But in any case, what you did as a correction wasn't even close to warranted, and you might have seen that instantly when the median itself favored Obama, just not by quite as much. A "correction" that reverses the outcome altogether clearly has something wrong with it.

Originally Posted by Brian Rush

Then you did just about the worst job of mis-representing your point I have ever seen.

Nope, I didn't. Maybe a little convoluted, but not a misrepresentation.

In the first place, one glance at one polling sample isn't enough. You need to show a pattern of the mean being more in Obama's favor than the median. And then you need to show a consistent pattern of one poll (say) being an Obama-friendly outlier. This has not been done.

Of course. Maybe I'll try doing this analysis on a weekly basis. 8 polls is a reasonably good sample though, and it tells us something about the current average.

But let's pretend for the moment that it had, and that we've decided that the Pew Research poll which has Obama at +8 is a skewed outlier.

It is certainly suspicious, which is what I said. Now, I did jump to conclusions about why that is, based on my personal assumptions, but I said I was doing that.

Do we take the difference between the mean and the median and multiply it by the number of polls as you did? No, of course not. We just refigure the mean without the Pew poll.

Doing what I did in that case is more conservative than throwing the poll out altogether. The National Journal poll is just as suspicious, so I would throw them both out. I did a variety of analyses with those different approaches. I stated my personal best guess in the first post, which is Obama +2.5, but that's partly based on intuition.

If we do that, the new mean favors Obama by 3.1. It makes the result slightly less lopsided, but still clearly favors the incumbent. (It also brings the mean much closer to the median.)

Which supports the notion that something stinks about Pew.

Now, I'm not saying this is really warranted. As I said, we'd need to do some more research and look for consistent patterns over time before drawing that conclusion. But in any case, what you did as a correction wasn't even close to warranted, and you might have seen that instantly when the median itself favored Obama, just not by quite as much. A "correction" that reverses the outcome altogether clearly has something wrong with it.

My correction did not reverse the outcome. You seem to still not be reading it correctly. I explained a couple of times, so you can go back and try to figure it out. I was not saying the average should be adjusted by 5.6. I was saying the sum of the polls should be adjusted by 5.6, before dividing by 8.

Originally Posted by The Grey Badger

Oh. OK. In that case ... though whoever is organizing the conspiracy shows far more organizational ability than any dictator or corporate executive in history. No dissenters, no defectors, nobody claiming they can make more money by going their own way ... miraculous! I wish whoever it is could come and moderate my club meetings!

Most of the media is acting as the propaganda arm of the Democratic Party at this point. It's not exactly a new observation, and if you think they're not, you're the one who's delusional. They've been doing it for a long time, they've just sunk to new depths in the service of Obama. No questions about his handling of the Libya situation. None. Instead we get old Mitt Romney statements as headline news. If you think they're above massaging poll data, you're naive. The elite left is a relatively small group of people. They know each other. They went to the same schools. They live in the same neighborhoods. They go to the same parties. Call it a conspiracy, call it a "movement", call it a coup de'tat. It is what it is.

Two new polls with results giving Obama lead in NC!

The beauty of this is that one of the results is from conservative polling firm National Research (Civitas) giving Obama a 49-45 lead over Team Romney. The other is the indy pollster, Purple Strategies, giving it to Obama 48-46. The latter is within the margin of error, but both polls represent huge swings to Obama since the conventions.

If the trend continues we could be putting lean Romney Indiana back into toss up status. From there, if we start moving into collapse territory, we may be seeing Arizona, Georgia and Missouri flipping with profound impacts on Senate and House elections. Fingers crossed.

Originally Posted by JustPassingThrough

The average differs from the median results by 0.7.

Actually you are are making a signficant figure error. The individual poll results are rounded to a whole percentage. Thus the median value will be a whole number too. The same could have been done with the average of these whole numbers, in which case the average and median would be identical. After all, if the raw data is given to only two signficant figures then why would you give the average to three figures?

But since the average is given to three figures there is only a 10% chance of any given average happening to come out as a whole number. So 90% of the time you will see a "difference" that is not really there.

Also you lost me by multipling the 0.7 number by the sample size. What's that supposed to show?

- that says something about Romney as well -

“Mitt Romney is the only person in America who looked at the way this Congress is behaving and said, ‘I want the brains behind THAT operation.’ ” — Tom Perriello

Originally Posted by playwrite

Mr. Tennyson is one of the 78 million Americans born between 1946 and 1964, known as the baby boom generation. In the most recent New York Times/CBS News poll, likely voters in roughly that age group favored Mr. Obama by 9 percentage points. In a Pew Research Center poll of likely voters, Mr. Obama is ahead, 50 to 44 percent, within that age group.

Oh, this again.

Originally Posted by playwrite

Lark McDonald, 51, who owns a small business in the Denver area, says he voted for Mr. McCain last time, and usually votes a straight Republican ticket, but is leaning toward Mr. Obama. He worries that the Republicans are moving too far right, he said, but he is also concerned they will dismantle the Obama health care program and make major changes in Medicare. “I take care of my father’s medical issues,” he said. “He relies on Medicare, and if we go to a voucher system, it will make everything harder to do.”

Yes, no shit. The idea of privatizing Medicare and fobbing it off to insurance companies and privatizing Social Security and fobbing that off to Wall Street is stupid. OK, the dude above is 51. My guess is he and I both thingkRyan's plan of fucking things up for folks under 55 is for the birds. He should find whatever cohort group thinks this stuff is swell and go with that.

Originally Posted by JustPassingThrough

Show me how. You don't get to just tell people how smart you are, you have to prove it. Brian Rush has completely failed. Your turn.

In your calcuation you are treating the median values of 48 and 45 as if they were 48.0 and 45.0.

But this is not true. You don't know what the first decimal of the poll data are. All you know is that Obama's median is between 47.5 and 48.5. The average falls into this range and so it is not measurably different from the median. Similarly, Romney's median falls between 44.5 and 45.5, and his average is also not measureably different from the median. Thus you cannot say there is any difference because the signal to noise ratio is too low to measure it. It's like a small peak on a chromatogram that is above the limit of detection but below the limit of quantitation. There could be difference but it is too small to measure with the method you are using.

Note I'm not even considering statistical variation, but simply precision of the measurement. You are trying to derive something from what is essentially noise.

Even if you had the full set of data and could then obtain the individual poll result to three signficant figures, and so have the required precision, any difference you would get would not be signficant. The pollsters round to whole numbers because they know giving it to one decimal point would be misleading precision.

By giving the average to one decimal point, you increase the apparent precision of the average by a factor of 10. Because of the central limit theorem, the average of multiple measurements does give a more precise measurement of the value.

Quiz question: To obtain the ten-fold increase in precision implied by the extra signficant figure, how many values would have to be averaged together?

a. 10 b. 20 c. 100 d. 5

How does this value compare to the number of values that were averaged? Was the extra precision implied by added decimal justifiable?

Originally Posted by Mikebert

Actually you are are making a signficant figure error. The individual poll results are rounded to a whole percentage. Thus the median value will be a whole number too. The same could have been done with the average of these whole numbers, in which case the average and median would be identical. After all, if the raw data is given to only two signficant figures then why would you give the average to three figures?

But since the average is given to three figures there is only a 10% chance of any given average happening to come out as a whole number. So 90% of the time you will see a "difference" that is not really there.

The median is 3. The average is 3.7. If you rounded the average you would get 4...

Also you lost me by multipling the 0.7 number by the sample size. What's that supposed to show?

It shows the total amount by which the polls in the average are over the median. In this case, 5.6. The question is where those 5.6 points come from. All from Pew? Or do you spread them out evenly? It doesn't really matter.

Originally Posted by Mikebert

In your calcuation you are treating the median values of 48 and 45 as if they were 48.0 and 45.0.

But this is not true. You don't know what the first decimal of the poll data are. All you know is that Obama's median is between 47.5 and 48.5. The average falls into this range and so it is not measurably different from the median. Similarly, Romney's median falls between 44.5 and 45.5, and his average is also not measureably different from the median. Thus you cannot say there is any difference because the signal to noise ratio is too low to measure it. It's like a small peak on a chromatogram that is above the limit of detection but below the limit of quantitation. There could be difference but it is too small to measure with the method you are using.

Note I'm not even considering statistical variation, but simply precision of the measurement. You are trying to derive something from what is essentially noise.

Even if you had the full set of data and could then obtain the individual poll result to three signficant figures, and so have the required precision, any difference you would get would not be signficant. The pollsters round to whole numbers because they know giving it to one decimal point would be misleading precision.

By giving the average to one decimal point, you increase the apparent precision of the average by a factor of 10. Because of the central limit theorem, the average of multiple measurements does give a more precise measurement of the value.

Quiz question: To obtain the ten-fold increase in precision implied by the extra signficant figure, how many values would have to be averaged together?

a. 10 b. 20 c. 100 d. 5

How does this value compare to the number of values that were averaged? Was the extra precision implied by added decimal justifiable?

What you're saying is partially true, in the sense that the numbers lack precision, and the variations being measured are small. But the RCP average is derived from the whole numbers in each of the polls. By comparing the mean to the median, you're analyzing it on its own terms. And what it shows, quite clearly, is that there are few polls skewing the average in Obama's favor, while most of the other polls show him with a much smaller lead. It also shows that those "pro-Obama" polls overwhelm any arguably "pro-Romney" polls.

Originally Posted by Eric the Green

Well, JPT's theory is not looking so good today. Obama is still bouncing!

http://www.realclearpolitics.com/epo...bama-1171.html

Maybe it is not so much a bounce as a Romney collapse. Whatever. If I am President Obama, then I will take it.

Originally Posted by playwrite

Nightmarish -

http://www.politicususa.com/wheels-f...ans-48-41.html

Mitt Romney is exposing himself as a snob, one sure way to alienate blue-collar workers everywhere. Is snobbery more of a Yankee phenomenon than a Southern phenomenon? Maybe that is why the GOP did so badly in the South until the Rockefeller Republicans largely became Democrats.

Back in 2008 I had no idea that Mitt Romney would be so singularly awful a politician. But he is.

Originally Posted by JustPassingThrough

Most of the media is acting as the propaganda arm of the Democratic Party at this point.

What about FoX News Channel? What about the editorial page on the Wall Street Journal? What about Sinclair Broadcasting? What about Clear Channel Radio?

It's not exactly a new observation, and if you think they're not, you're the one who's delusional. They've been doing it for a long time, they've just sunk to new depths in the service of Obama.

Most journalists are realists. The money-grubbing media want to maintain interest in the election and thus ratings and ad revenue... but they want to maintain their credibility. Face it. Barack Obama is arguably the most adept politician that the US has known in a long time. Maybe you think that he is only good with words... but words can convince people to change their behavior. Do you want people to save, buy, and invest, and develop new and more lucrative skills? President Obama may be just the person. What is wrong with a liberal who achieves the desirable ends of conservatives?

No questions about his handling of the Libya situation. None. Instead we get old Mitt Romney statements as headline news. If you think they're above massaging poll data, you're naive.

The Libyan situation tells its own story. The Romney campaign tried to spin it into gross incompetence of you-know-who, and showed its own incompetence. Such was legitimate news in itself.

Sure -- Scott Pelley and Bryan Williams tell pollsters what the results of the day are to be. Heck, even FoX News Channel uses a pollster who shows some results that the management of FoX News wishes were not true but can't deny.

The elite left is a relatively small group of people. They know each other. They went to the same schools. They live in the same neighborhoods. They go to the same parties. Call it a conspiracy, call it a "movement", call it a coup de'tat. It is what it is.

Where I live, the schoolteachers mostly graduated from the same University, took the same courses in college, and have more in common with schoolteachers fifty miles away than with their blue-collar relatives. Do you call that conspiracy?

Can you accept that Mitt Romney is one of the most incompetent nominees for President by either of the two major parties for a very long time?

Originally Posted by pbrower2a

Maybe it is not so much a bounce as a Romney collapse. Whatever. If I am President Obama, then I will take it.

There was a definite convention bounce, the start of a drop after the bounce, then Romney had some bad days. I'm not ready to call it a collapse yet, but Obama seems to have the momentum for the moment. Let's see what the debates do.

Originally Posted by JustPassingThrough

What you're saying is partially true, in the sense that the numbers lack precision,

Polls are estimates of what looks like random behavior. The vote count is the definitive result, and that has precision.

and the variations being measured are small.

Unless you are discussing Florida in 2000, which really mattered or Iowa and Wisconsin in 2004 or Missouri and North Carolina in 2008 which did not matter in the respective Presidential elections.

But the RCP average is derived from the whole numbers in each of the polls. By comparing the mean to the median, you're analyzing it on its own terms.

In view of the polarization of the American electorate along ethnic, regional, religious, rural-urban, and occupational lines, President Obama has long been understood to have no chance of winning Wyoming and no chance of losing Vermont. Can you accept that cultural geography and demographics now dictate what is possible in American presidential elections? The mean and the median will differ.

Barack Obama won 69,499,428 votes, 52.87% of all cast, and won 365 electoral votes or 67.8% of the total, and John McCain won 59,950,323 votes, 45.60% of all votes cast and won 173 electoral votes or 32.2% of the total in 2008.

Franklin Roosevelt won 25,612,916 votes, 53.39% of all cast, and won 432 electoral votes or 81.4% of the total, and Thomas Dewey won 22,017,929 votes, 45.89% of all votes cast, and won 99 electoral votes, 18.6% of the total in 1944.

Surely you would say that there is no statistical significance of the difference between 52.87% and 53.39% of the vote, but the difference between 67.8% and 81.4% of the electoral vote is huge. Maybe the popular vote was less polarized where it was free (the vote in some Southern states in 1944 would not now qualify as the result of free elections), but FDR would have won without the South. Outside of the South, FDR's strongest state was Utah (hard to believe these days), which he won by 20% in a free election, and his weakest state was Kansas, which he lost by a margin of 20% in a free election. Aside from the former Confederate States, FDR won or lost only one state by more than 20%. Barack Obama won ten states and DC by more than 20% and lost six states by more than 20%. FDR would have won without the Southern states.

So what is the difference? That Barack Obama was an unusually-polarizing character? Hardly; such also showed in the 2004 election, with Kerry winning only three states and DC by more than 20% but losing thirteen states by more than 20% in a close election. In 2000, the closest election in the popular vote in American history, Gore won four states and DC by margins of 20% or more and lost eleven states by margins greater than 20%. Maybe it is the difference between a late 3T turning into a 4T and a 4T approaching a definitive and triumphant end.

And what it shows, quite clearly, is that there are few polls skewing the average in Obama's favor, while most of the other polls show him with a much smaller lead. It also shows that those "pro-Obama" polls overwhelm any arguably "pro-Romney" polls.

If you look at the statewide polls, with extensive attention being paid to a few swing states likely to decide the election -- CO, FL, IA, MI, NH, NM, NV, NC, OH, PA, VA, and WI. Nobody has any mystery on how Texas or New York is going to vote. Michigan, New Mexico, Pennsylvania, and Wisconsin have effectively slid out of contention. The polarization of the states makes such easy this time.

The states elect the President; the People do not.

Originally Posted by B Butler

There was a definite convention bounce, the start of a drop after the bounce, then Romney had some bad days. I'm not ready to call it a collapse yet, but Obama seems to have the momentum for the moment. Let's see what the debates do.

Of course. The debates will be the last chances for Mitt Romney to make his case.

Originally Posted by The Rani

Well, not exactly a recommendation. Maybe a suggestion.
Obviously, some people think that their shit doesn't smell.

I believe "Olfaction" to be the most bless-ed of the 5 Senses i/r/t Man/Mankind.

Prince

PS:I suppose it's all in the "Nose of the Beholder"!



Case in point!

Originally Posted by JustPassingThrough

The median is 3. The average is 3.7. If you rounded the average you would get 4...

No. The averages rounded are 45 and 48. 0.7 is the difference of averages expressed to an inappropiate degree olf precision. This different is essentially noise. It contains no information and anything you with it is gigo.

If you done the simple exercise I gave, you would see that it is inappropriate assume the averages given to three places are as valid as the raw data.

Do you have any background is this stuff? It seems to me that you are at sea.









It shows the total amount by which the polls in the average are over the median. In this case, 5.6. The question is where those 5.6 points come from. All from Pew? Or do you spread them out evenly? It doesn't really matter.[/QUOTE]