AllExperts > Conservatives 
Search      
Conservatives
Volunteer
Answers to thousands of questions
 Home · More Conservatives Questions · Answer Library  · Encyclopedia ·
More Conservatives Answers
Question Library

Ask a question about Conservatives
Volunteer
Experts of the Month
Expert Login

Awards

About Us
Tell friends
Link to Us
Disclaimer

 
 
 
 
About Dennis
Expertise
Any questions concerning politics, govenment, political science.

Experience
Former elected official, and educator in the field

 
   

You are here:  Experts > Arts/Humanities > Political Science > Conservatives > Polls.

Conservatives - Polls.


Expert: Dennis - 9/3/2004

Question
Which poll was considered the most accurate for predicting the presidential election between Gore and Bush for the 2000 election?
  Which one for the 2004 election?

             Thanks  

Answer
The Zogby poll was considered the most accurate in 2000 because he caught the last weekend shift after the DUI information was released by the Democrats on the Thursday before the election. Everybody else stopped polling on Friday.

This year the nod is being given to the Rasmussen poll because he is doing a daily "rolling" poll of voters "most likely to vote" in the Presidential election.

A daily poll is really expensive if you do a different 1,000 voters each day. I haven't looked at his methodolgy, but if he is actually interviewing 800-1,000 DIFFERENT voters each day, he will have the most accurate, up-to-date info on the election.

However, saying that, I think he is probably doing a small sampling each day, and probably after an entire week goes by, he ends up with an interview population of 1,000 "new" subjects.

You have to remember though, you want "those most likely to vote" and an even breakdown of Democrats, Republicans, Independents, rural, city, young old, etc. To do that each day (and have them home and willing to answer the phone, THEN answer your questions) you have an almost Herculean task.

I'm willing to bet that he has his core demographics, and then samples a "random" selection of registered voters who say they're definitely going to vote in 2004. Check his margin of error numbers, and you'll get an idea of what kind of sampling he's using. If he has a 4-6% margin of error, then he isn't doing EXACTLY what he says he's doing. If its 3% and below, then he is going the extra mile, and his numbers are probably pretty accurate.

One caveat though - national numbers don't mean squat. You will want to look at the numbers from the individual "battleground states." Those 17 states will indicate which way the election is going.

Add to this Answer    Ask a Question



  Rate this Answer
   Was this answer helpful?
Not at allDefinitely              
   12345  

     
About Us | Advertise on This Site | User Agreement | Privacy Policy | Help
Copyright  © 2008 About, Inc. About and About.com are registered trademarks of About, Inc. The About logo is a trademark of About, Inc. All rights reserved.