But generally (and statistically speaking), they tend to vote democrats

I can see arguments for both sides, but I have never seen any empirical studies on it.  I think that it is just conjecture at this point.

the ratio of young people having cellphone is higher than older people

younger people tend to vote for Obama

It's all in a matter of %, off course. So "Obama favoritisim" is more in the 60%-40% than complete landslide...

Didn't read the other responses so sorry if this is a repeat.

The polls are done by taking US Census information.  So say for New Hampshire there are 900,000 people.  (That's just a random number btw). 

They take the sample size of the population, which involves a statistical formula with a 95% confidence level.  You'd think the more people in a state the larger the sample size, but that's not the case.  It's called the law of diminishing returns and you can look it up if you are interested.

So say a sample size for 900,000 is 700.  They poll 900 people (the good polls anyway) by using a computer to randomly pick telephone numbers of people in the state.

The questions are asked by a professional sounding voice.

Most polls have a formula for determining "likely voters."  They can be different for each poll.  Remember not everyone votes.

So in the end of the 900 they get 800 or 700 but it has to meet the sample size formula for that state.

Here this says it better...http://www.surveyusa.com/client/methodology4.aspx

There is a method to the madness, however you have to remember the last two presidential elections, the polls were wrong.  So if ya ask me, especially after studying ad nausea the last couple months, statistics and polls are a waste of time and money.  The only reason these expensive polls are so popular is because the media wants them.

I think of it as wanting to jump into a lake you've never been in before.  You're pretty sure what will happen, the size of the fish  you'll encounter, the temperature of the water, the feel of the bottom, is it sand or mud squishing between your toes, but just for the added benefit of knowing you send three people down to the water before you to check and give you their opinion.

They come back and one or more will confirm what you thought already, and maybe one or more will tell you something you didn't know, but in the end, until you get in the water and find out for yourself, you'll never know what's real and what's just talk.  And that's all polls essentially are, talk.

Many a slip twixt the cup and the lip. 

But which young people?  That is the rub.  One would nornally assume that the young that dump the land lines are the business types that usually vote conservative.  That is why I am not so sure. It could be as you say (and I can see that), and it could be the opposite.  I think it is needed by the polling firms so they can adjust their numbers accordingly, but so far, I have seen no one try to make a study out of it.

From my research that 18-24 group, and in some states that extends up to 34.  The only state I can find in which Obama isn't doing well with women and young people is Minnesota. 

I was only answering for myself as I tend to vote Republican but have been known to vote Democrat too.

But again, banding people by age and declaring they are of one mind in politics is misleading.  So the question of "which young people" asks the question of what are their other demographics.

Yeah, well its just a variable that polls look at...same as gender, race, etc.  I think they're a waste, but I'm not payin for them. 

Hum...

How old are you? I am disapointed to tell you that.. well, cellphone are very much commonly used by everybody of my generation - hippies and technocrats alike.

You missed the point (and I am over 50 BTW).  Yes, younger people do tend to use them.  But which ones have dumped the land lines?  Poor?  MBAs?  Technocrats?  And what are their political demographics.

I concede the point that it is probably young that have dumped land lines.  But I dont think it is homogenous across the young folks.  And therefore, additional information is needed to determine if their exclusion in the polls have a real impact on the numbers, or a neglible effect.

Most of them all! It's a totally generation thing! That's what I've been trying to say.

At least, in Montreal.

Hum... you should think it is. Cellphones are simply more conveniant, and cheap. Seriously, from my point of view, punks, hippies, technocrats, fratboys, etc.. all of them have, in pretty much equal number, cellphones. Older peoples, however, have different habits.

We can opine about that, but for pollsters to have credibility, they should know what it is.  I think they have some homework to do as the trend is increasing, not going away.

Oh, and there is another group (as I just found out - I just got caller ID).

Those with Landlines and caller ID.  LIke I am going to answer an 800 number!