General overview 2012: the state of play
Welcome to all the new readers. Traffic is booming. First, a brief update. Several weeks ago I showed that convention bounces (and other fluct...
Honoring labor…
…by not laboring.
Preliminary vibes in the Meta-margin and EV estimator do not point toward any bounce for Romney. Recently, those measures pinpointed the Ryan VP bounce at an effective 3 percentage points at its peak. Today there was a small tick toward Obama, suggesting the possibility of a negative bounce. We’ll see more tomorrow.
Don’t forget ActBlue, left sidebar. Or the NRSC. Happy Labor Day.
Topics:
Even the fairness-driven Nate Silver agrees.
https://twitter.com/fivethirtyeight/status/242691461743779840
Eric Cantor tweets today: “Today, we celebrate those who have taken a risk, worked hard, built a business and earned their own success.” Not a word for the masses.
It is in the same league as Gina Rinehart’s comment: “There is no monopoly on becoming a millionaire. If you’re jealous of those with more money, don’t just sit there and complain. Do something to make more money yourself – spend less time drinking, or smoking and socializing and more time working.” She is the world’s richest woman. How did she do that? She inherited the Australian mining empire from her parents.
Tapen
@Tapen Sinha: In Norway we have an apt saying, which roughly translated goes like this:
“Money is a lot like manure. If you spread it around, it can do a lot of good — but if you pile it all up in one place, it smells like sh*t”
@Olov Grinde
Det er morsomt.
Du er en sosialist! 😉
Tapen
@Olav: I used to hear that from my mother (in English). It must have come from her Olsen mother.
@Dr Wang: I have heard more than one pundit on both sides of the aisle say something that could be construed in some degree as “The more you know Mitt, the less you like him.” So, a negative bounce is possible. It will be interesting to see today’s number. Tomorrow will be affected by the convention.
@Dr. Wang: This is partially in response to today and partially in response to the question of (intentional or accidental) bias.
My understanding of statistics comes from reliability and from signal processing. Thus, I interpret the margin of error as the noise in the signal. Also, since it has a square root of N function, it is an analog of shot noise or Poisson distribution. As such, the error could be , say 3%, positive one day and 3% negative the next resulting in an apparent 6% change in the result while there is actually no change, but rather there was a change in the sample.
Since we know that the raw data is manipulated, I would expect the manipulated data could be examined in the frequency domain to verify the bias because of the change of the distribution from the expected value from raw data. If nothing else, the renormalized frequency distributions of the different polls should be very similar, if no bias is entered by the ‘filtering’ process.
This would be analogous to checking bookkeeping for the last number distributions to compare those to the expected values.