Polls? What good are they if you are never included?

M

MrBigglesworth

New member
All I have ever heard in the last few years is Political Poll this or that.

Bush was up in the polls during his convention.

Gore likewise.

My question is this. How come I have never been invited to any of these polls relating to political issues? How can these polls be COMPLETELY accurate without asking everyone of the citizens of the US of A?

This makes me question the validity of ANY political poll as if we asked more and more people their feelings of any particular issue at hand then the final result may be DRASTICALLY different than if only 100 or 1000 or 10,000 people are polled on the issue.

/rant off

Thanx!


edit= Also have any of you guys/gals particpated in a poll of this nature?
------------------
Try to take away my gun...and you will see my 2nd Amendment Right in ACTION!!! -Me

FOR THE CHILDREN!!!!

[This message has been edited by MrBigglesworth (edited August 23, 2000).]
 
Well, if they were accurate in any sense of their nature, then we should go ahead and count the total as the final vote. I see these polls as bullsh*t anyway, since it tends to swing voters. Thats my opinion.. but remember: The media is only going to tell you what they think you should know. :mad:

------------------
God, Guns and Guts made this country a great country!
 
Well, I took statistics in college, and I assure you a poll of 10,000 people CAN accurately represent the views of the nation at large... If it's done right. The problem is that these political polls you hear about in the media AREN'T done right, for a number of reasons.

It's EXPENSIVE to do a poll right; After you've randomly selected the people you're going to poll, you have to do everything in your power to contact every one of them; Not just phone calls all hours of the day till you get ahold of them, but personal visits if they don't have phones. And that's the first place these polls fall down: They use random dialers, so they only poll people who have phones, and answer them. And they conduct the poll during business hours, so they don't get the people who are at work during the day, mostly just housewives and unemployed, with a few home businesses thrown in.

It's time consuming to do a poll right. All those calls back to people who didn't answer, instead of moving on to somebody else.

It takes skilled employees.

And, most important for these media polls, if you do it right, YOU CAN'T DICTATE THE RESULT IN ADVANCE! Most of the polls you read about during campaigns are just propaganda, guys; Biased questions, unrepresentative samples, inadequate numbers of people sampled. Nobody should take them very seriously.

------------------
Sic semper tyrannis!
 
There a form of control. If you don't agree with whatever the pc outcome is then there is something 'wrong' with you.

Then in their 'caring and inclusive' way they'll label you an 'extremist', 'meanspirited' and then say the dreaded words "It's for the children".

Polls mean nothing but what 'they'
want you to think.

Sgt.K
 
I'm glad I glanced down, I was just about to begin a thread on this very topic. After hearing a poll quoted on the radio just now, I'm thinking - how can a poll of 1,000 people possibly reflect the views & opinions of the millions that live in this country, to a margin of error of +/- 3% ??? My college statistics is rusty (actually had to retake it to get through grad school) but I can't see how that's possible. With all the political slants, ethnicities, ages, races, economic classes, social classes, family stautus, careers, etc. in this nation, how can 1,000 people possibly be a large enough sample ?
 
I'd like to know about these various self-defense surveys, are they done correctly?

The DOJ says there are about 80,000-100,000 defensive gun uses yearly, and I think they use census takers to gather their information.

Other surveys get numbers from 700000-2500000 DFG yearly, but some claim this is due to small sample size, and false positives.

Any perspectives?
 
Polls are useless drivel even if they did include me.
From American Heritage Dictionary:
Poll..."A canvassing of a selected sample group of people to analyze public opinion on a particular question"
Survey..."To examine or look at in a comprehensive way; to inspect carefully; to scrutinize."

A modern poll (as Brett noted)is merely a reflection of the opinion, at the moment, of those sampled. Look how many outfits put out multiple daily polls these days. Think much work is put into them? Anyone with a website can "conduct a poll"....

Have a run in with the Pres of your company today?....GW Bush would take some serious negatives. Tomorrow the boss sincerely apologizes in public....reverses yesterdays poll. Get a big tax refund? Gore/Clinton would score favorably today.

Modern polls are merely designed to collect opinions, and then used to influence other opinions.....surveys take into account statistics and collect bonafide data. Most folks opinions are like underwear, changed everyday

------------------
"Quis custodiet ipsos custodes" RKBA!
 
Polls are just fine.

Accurate polls are hard to do.

500 person poll +/-5% error margin
750 person poll +/-4% error margin
1000 person poll +/-3% error margin

One night polls suck. They can be 500 ADULTS who may or may not be registered to vote. Registered voters is fairly accurate and 'likely voters' polls are the most accurate. If you see a poll of 1000+ likely voters, it will be more accurate than any others.

One note on the accuracy. The accuracy comes from a trend. If you do one poll one week and Gore is winning then do another the next week and Bush is winning and it alternates-don't pay attention. If a trend starts to develop over 4-6 polls, then take a look. Zogby did close to 6 polls showing Bush ahead by 3-5% until this last one. That is 6 polls of 1000+ LIKELY VOTERS. It may not be close to the final numbers, but it should be able to predict the winner.

madison
 
Madison46: That's the uncertainty due to sample size, given an ideally distributed sample. If the sample is not representative, all bets are off!

For instance, a common technique used by polsters is to sample when and where convenient, and then "adjust" the results to compensate for the amount that the sample diverges from being representative of the population by race, sex, political affiliation, and so forth. The problem with this is that if you get a sample that's 90% female because you called during the day and got housewives, you can't simply multiply the men by 9, and leave it at that; The sampling error on the men is much higher as a percentage, because the number of men was so small, and then that percentage applies to the "adjusted" figures. So you might have polled 1000 people, but if 90 percent of them were women, and you adjust for that, your error is NOT what you'd expect if you had a sample size of 1000; It's an average between the error on 900 women and 100 men.
Then there's the error due to bad questions. The error due to non-responsive samples. (Can't assume that they have the same opinion, after all!) The error due to samples who lie.

The expected error given with polls is usually an absolute minimum based on sample size alone, without taking any of these other sources of error, hard to quantify but still very real, into account.

------------------
Sic semper tyrannis!
 
You must log in or register to reply here.
Back
Top