Sep 08 2008

Baffling Yourself With BS Statistics

I have to laugh at one of the liberal polling experts on the left, someone over at a blog called Fivethirtyeight. Today Nate Silver shows how statistics, when appropriately rigged, can even delude the statistician:

Among the many difficulties that we face in this unique election cycle is figuring out exactly how long one should expect a convention bounce to last. Under ordinary circumstances, bounces are actually fairly persistent, lasting for perhaps several weeks …

Of course, we don’t know what two bounces will look like when laid down on top of one another (with a one-week gap intervening). Our best guess was something like this:

Well, isn’t that special! He drew a chart with a line on it. Where did get the line? Well, it seems he pulled it out of his … but I digress. Today’s whopping 10 point lead among likely voters in the Gallup/USA Today poll pretty much busts this BS, just like a decade of global cooling has busted the climate ‘models’ from the Church of Al Gore/IPCC. The guy predicts at maximum 2 point lead for McCain after his bounce – which just goes to show his chart is now garbage.

The problem with polls and projections on them this year is this is ANYTHING but a normal election. The incumbent is not running, nor is anyone from his administration. The contenders are Senators (who can only win in a situation like this). The Dem contender is the most inexperienced of all four candidates, and we have a woman ready to shatter the glass ceiling in the White House. We are at war and our economy is being crippled by high energy prices that a Democrat Congress allowed to balloon under their watch.

And most importantly – the media and the rest of the Political Industrial Complex in DC and NY are being rightfully blamed for the current partisan mess. Americans are running from the fringes and trying to find any broom or fire house they can to begin cleaning out the DC swamp.

Pollsters don’t have models for this environment, which is why they kept blowing the primary projections. Pollsters and their models resist massive upheavals and change in the statistics (as any good statistician should) as possible outliers or transients. But to claim the current upheavals are all outliers and transients is to grasp onto the reason why, in traditional elections, it is good to not jump on upheavals in the numbers.

This year prudence is simply making sure the upheaval is as real as it seems, not some special wisdom that things will be getting back to normal soon. I have seen change sweep by in front of us and many times the timid are too afraid to say what is happening.

I predicted in the March of 2007 The Surge and Awakening in Iraq would be an unstoppable wave. The reason was clear, to see any wave at all meant a fundamental, wide, overpowering shift in the Muslim Street had to occur to see the changes. For Arab Muslims to take up arms against al-Qaeda meant crossing lines that would never be uncrossed. It meant al-Qaeda had become so violent its one time allies realized they must destroy them to save themselves.

While not as dramatic, the change in the US electorate is no less firm. To drop Obama and support Palin (which is the prime dynamic here) means Obama and his media surrogate’s treatment of women has been so abhorrent that people who would never consider voting GOP have come to the realization the women’s movement was being destroyed by their one time allies, and the only way to save it from – at minimum – another decade of paralysis if not defeat meant changing horses to the GOP.

When Obama starts losing middle America it is over. And the media gave even the most socially guilty thinkers enough reason to forget about racism and go back to gender equality. Men have a deservedly bad rap in many ways (the exceptions being men like Todd Palin). And I would say, given their flight from responsibility in their own families as Bill Cosby has noted, the black man is in the worst shape of all categories of American males. The media, in questioning Palin’s abilities to be mother and have a career, in attacking a young unwed (but soon to wed) mother, have given enough people reasonable cover to move from Obama to Palin. And once they move there is nothing Obama can do in 2 short months to make these people uncross that line.

This liberal poll watcher is praying conventional statistics will save his preferred choice. What he forgot is that his preferred choice had the opportunities to become one with the electorate (especially the women voters) and blew it to the point people are jumping ship. If McCain’s bounce was equal to Obama’s they would have cancelled each other out and there would be near parity. Clearly there is more than that going on. Nice graph, but meaningless.

3 responses so far

3 Responses to “Baffling Yourself With BS Statistics”

  1. markg8 says:

    Even Silver says it’s meaningless. I suggest you actually read the article.

  2. AJStrata says:


    You’re new here. This NASA Engineer with an BS in Biology knows math, statistics and when people are even unconsciously fooling themselves.

    Silver is waking up to what is happening – slowly.

  3. Redteam says:

    I read the full article, I’m not sure he said it’s meaningless. I interpreted it as he ‘hopes’ it’s meaningless.
    I can tell you for sure the commenters on that site ‘hope’ it’s meaningless.