Romney Looks Like the Next Pres

Discussion in 'Politics' started by jem, Apr 13, 2012.

  1. Tsing Tao

    Tsing Tao

    (insert token canine consumption joke here)
     
    #211     Apr 19, 2012
  2. Because there are more registered democrats then republicans.Call 1000 registered voters,odds are most of them will be registered democrats

    From 2011


    http://www.usatoday.com/news/politics/story/2011-12-22/voters-political-parties/52171688/1


    Registered Democrats still dominate the political playing field with more than 42 million voters, compared to 30 million Republicans and 24 million independents.
     
    #212     Apr 19, 2012

  3. Same thing,more democrat or lean democrat.If there or more democrats then there will be more democrats in the poll results
     
    #213     Apr 19, 2012
  4. jem

    jem


    Do we want the polls to predict the election... or a random persons favorite candidate.

    If we are talking elections and who is looking like the next president... we would all want polls with proper samples.
     
    #214     Apr 19, 2012
  5. 377OHMS

    377OHMS

    Its going to be dog eat dog. :D
     
    #215     Apr 19, 2012
  6. These are registered voters called at random so they are proper samples imo
     
    #216     Apr 19, 2012
  7. Epic

    Epic

    You've obviously never been in charge of sampling for research. The idea isn't to just randomly call 1,000 people. This will not get you a representative sample and will lead to erroneous results. There are many things that effect your sample that result in misrepresentation.

    For example, if you call people on land lines you get a different result than cell phones. If you call land lines after 5pm you get a different result than land lines at 12noon. If you interview people in person you get a different result than if you call them.

    Given all these variable you have two choices;

    1) Use the entire sample and weight the results to better represent the population.

    2) Collect a larger sample and then use a random selection process to re-balance things.

    One way or another the sample must accurately represent the population. You claim that more people are Dem than Repub. While that was certainly true in '08, it has been closing every year since. As of the past couple months there is only about a 1-2% gap. The number of independents also drops during an election year. There is simply no excuse for over-weighting the sample with Dems by up to 11 points, given the current national demographics. That is just sloppy.
     
    #217     Apr 19, 2012
  8. That was true in Dec 2011,just a few months ago



    http://www.usatoday.com/news/politics/story/2011-12-22/voters-political-parties/52171688/1


    Registered Democrats still dominate the political playing field with more than 42 million voters, compared to 30 million Republicans and 24 million independents.
     
    #218     Apr 19, 2012
  9. Epic

    Epic

    Also, a poll of likely voters would be much more accurate. Simply taking anyone over 18 is also very sloppy given that only about 55% of eligible voters turnout, and they aren't necessarily the ones that can be reached on a home phone during the middle of the day.
     
    #219     Apr 19, 2012
  10. Epic

    Epic

    Not talking about how people are registered. I'm talking about polls of what party people claim. Registration is not an accurate gauge. Some states require registration to a party in order to participate in the primaries. Other states don't. This skews the registration. What matters is what people claim. According to all the polls I could find during the last few months, it is very evenly split. That is especially true when you include Dem leaning or Repub leaning people.

    For someone who follows these polls so closely, I'm surprised that you don't know these details. Personally, I don't want to see a poll comforting me that my preferred candidate is ahead. I want to be able to look at the data and know that I'm getting an accurate picture.
     
    #220     Apr 19, 2012