Sunday 19 December 2010

lol Poll Spinaments

I was just reading an article on CNN headlined "Bad News for Palin" in which a new poll stated that 6/10 Americans would not consider voting for Palin. I'm personally not for or against her as I haven't really paid much attention to her policies. My initial impression in 2008 was that she was a bit of a gimmick, and my 2nd impression is based on the only episode I watched of Palin's Alaska in which she seemed like that loud relative that is often out of line but is someone you can depend on to stick up for you when you're in trouble.

Anyways, I just found the spin on the poll results to be kind of silly. I'm going to round this to make it simple. 6/10 said they would not consider her. 3/10 said they would consider her. 1/10 said they would definitely support her. This is apparently bad news for Palin if you believe the media. I'm no statistics or poll expert but a little bit of rational logic works wonders here.

Assumptions to keep this simple:
The US is 50/50 Republican and Democrat
All Democrats said they would not consider her.
All people that said they would consider her are Republicans.

If 50% of the no consideration results are Democrats, that means the remaining 10% are Republicans that would not consider her. Lets then break down the Republican vote which will double the remaining numbers to total 100% of the (R) vote:

20% will not consider her.
60% will consider her.
20% will definitely support her.

Now maybe the general US population has not been exposed as much to the effect of splitting the vote among multiple candidates as we are here in Canada with 4 major parties and 1 or 2 legitimate fringe parties which lets you form majority governments with less than 50% of the vote, but these poll results are hardly bad news when you consider the number of candidates in the primaries.

Lets say there are 10 candidates that make it to the debates. Her 20% definite support alone is double her fair share. If she were then to receive only her fair share of the 60% considering her, she'd be sitting at 26% of Republican support. This definitely puts her into contention when you split the vote. If she were to receive no extra share of support from the 60% considering her, she would still have more than her share of total support until they were down to 4 people due solely to the 20%. 25% can put you in the top three vs two other 25% support candidates with the remaining 25% split up between a number of long shots.

So already, 2 years before the election, she's already ahead of, if not leading, the pool of Republican candidates. And in the end if she were to win the nomination, it's most likely those 20% of Republicans that said they would not consider her would suck it up to prevent a Democrat from winning, or at least not vote at all, it's not like those 20% are ever going to vote (D).

No one would say, wow Phil Ivey only has 5% of the chips at the final table of the WSOP, yet he's estimated to win 20% of the time, that's terrible news for him. Contrary. That's great news for him. Yes it's true that Phil Ivey and Sarah Palin will both not win the majority of the time. That holds true for all politicians as a whole. The majority will lose most of the time in races when there's one seat and more than 2 people running for it. Simple math.

Kind of reminds me of the financial 'experts' they put on the news saying XYZ stock is probably going to drop today. They base their analysis on the fact that the stock is expected to drop 70% of the time. They then go on to ignore the pertinent fact that if it drops 70% of the time and loses 10% of it's value, but goes up 30% of the time and gains 100% of it's value, it's a better bet to buy it than not due to expected value. Of course the experts are more concerned with appearing right most of the time than actually saying something right most of the time.

I find it odd that these are the people that we let govern mainstream thought and public discourse. Whether through shear ineptitude or outright playmaster spin they somehow always make statistics and polls say whatever they want them to say, even if it's exactly the opposite of reality.

No comments:

Post a Comment