Galt's Gulch

Polls - Size Matters - An Observation about sample size

Posted by freedomforall 8 years, 5 months ago to Politics
11 comments | Share | Flag

The linked site displays a massive amount of polling data for the reader to peruse.
There is a pulldown to the upper left labeled "Battleground States" and it allows you to look at recent polls in each state.
The details show the results and the 3rd column shows the Sample Size for each poll.
Examine them for yourself.
It looks to me that in a very high percentage of the polls, the higher the sample size, the better Trump does, and the worse Hitlery does in the results.
Could this indicate that when real votes are counted, the results are much more difficult to manipulate?
SOURCE URL: http://www.realclearpolitics.com/epolls/latest_polls/

Add Comment

FORMATTING HELP

All Comments Hide marked as read Mark all as read

  • Posted by rbunce 8 years, 5 months ago
    Massive non response rate, manipulation of the flawed raw data to achieve the desired result in the final data... only poll that matters is next Tuesday.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by freedomforall 8 years, 5 months ago
    Please also note that the online surveys taken using Survey Monkey are non-probability surveys. For example:
    NBC News|SurveyMonkey Weekly Election Tracking poll is a non-probability survey.
    "A core characteristic of non-probability sampling techniques is that samples are selected based on the subjective judgement of the researcher, rather than random selection (i.e., probabilistic methods), which is the cornerstone of probability sampling techniques."
    The people to be surveyed are chosen by the biased NBC NEWS, not by an unbiased statistical method.
    The results are biased and should not be reported by any media source unless that bias is made very clear.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by DrZarkov99 8 years, 5 months ago
    Fox News very nonchalantly made the point that most of the polls they rely on for their collective numbers oversample Democrats by as much as ten percentage points. I don't understand why they don't use the LA Times poll, since it does meet their criteria, and doesn't oversample either party. They're looking pretty foolish when their numbers show a dead heat, and the LA Times poll shows Trump with a 6 point lead.

    The Arizona Freedom Alliance did a phone survey of 1,000 people in each state, for a total of 50,000 surveyed. Their sample mix was 33% Democrat, 33% Republican, and 34% Independent. The results were a shock, with 67% supporting Trump, and 19% supporting Clinton. Because the Alliance is a conservative organization, like the Heritage Foundation, the media dismissed their results.
    Reply | Mark as read | Best of... | Permalink  
  • Posted by Herb7734 8 years, 5 months ago
    In the past, polls got accurate and more honest a day or two before the election so the pollsters could brag about their results in the next election. Also, exit polling has proved to be a bit more accurate as well. Up until then, polls mean next to nothing.
    Reply | Mark as read | Best of... | Permalink  
    -
  • Posted by $ Olduglycarl 8 years, 5 months ago
    I'd have to agree with you there. Never considered a poll worthy unless the sample size was yuge!
    Reply | Mark as read | Best of... | Permalink  
      -
    • Posted by edweaver 8 years, 5 months ago
      And even then it is manipulated.
      Reply | Mark as read | Parent | Best of... | Permalink  
      • Posted by lrshultis 8 years, 5 months ago
        Not necessarily manipulated, but rather not a random sample. Also the poll should be done with multiple random samples since a random sample can show bias one way or another. That bias happens all the time with disease clusters, where a high indication of some disease does not mean that the whole population, if surveyed, would have the same result.
        Reply | Mark as read | Parent | Best of... | Permalink  

FORMATTING HELP

  • Comment hidden. Undo