• Enter your email address to follow this blog and receive notifications of new posts by email.

    Join 37 other followers

  • November 2007
    S M T W T F S
    « Oct   Dec »
  • Recent Bookmarks:

  • Archives

  • Categories

  • Advertisements

Simplifying Polls

Simplifying Polls

My original method for reading polls is very thorough and can be accomplished by anyone who has a high school education and access to the graph I included in my post on the subject. (you’ll want to be familiar with my earlier post to understand where my shortcuts come from) After long contemplation I realized there is an even easier way to analyze any political poll without need for the graph.

When comparing polls from the same agency, we rarely have to worry about differing SD sizes. Since the SD ratio will almost always be 1, all we have to do is memorize certain values along the x-axis of the graph.

The way we find the x-value is simple, take the difference between the two values expressed in the poll and divide that number by half the margin of error.

Once we find the x-value we need to find out what it means. At or below x-value .5 we can assume the race is a tossup since there’s only a 55% chance the favored candidate is actually in the lead. At x value between .5 and 1.5 we can assume the race is “likely” to go to the leading candidate because there’s a 60% to 75% chance the lead candidate is truly ahead and anything above 1.5 is a decided race (“sure thing”)

Here’s an example:

A Reputable Poll puts candidate “A” at 48% and his opponent, candidate “B” has a 47% show of support with a margin of error of 2%. We take (48-47)/(.5)(2) and this gives us an x-value of 1. We know Candidate “A” has somewhere between 60-75% chance of winning this race.

Let’s apply this to a real race:

American Research Group poll. Nov. 9-12, 2007. N=600 likely Democratic primary and caucus voters nationwide. MoE ± 4.

“If the 2008 Democratic presidential preference primary/caucus were being held today between [see below], for whom would you vote?” Names rotated

Hillary Clinton 46%
All other candidate’s combined 42%

The difference between the two values is 4, half the margin of error is 2. Four divided by two is two and we know that an x-value of 2 signifies a “sure thing” so we can confidently say Hillary is the Democratic horse to bet on.*

This method is faster and can normally be done in your head. All one has to do is memorize the equation and the three significant x-values.

Just a reminder: The race for delegates in the nomination process is regional, not national. A victory for Barack in the early primaries might change the national race and these idiosyncrasies need to be noted whenever you’re doing this sort of analysis.

*A quick look over at Intrade shows the contract on Hillary to be the Dem Nominee is currently trading at $70. A quick peak at the graph on my other post shows the odds Hillary will win are well above 80%. The Hillary contract is priced right for purchase.

%d bloggers like this: