Did you know that Fox News hasn't published a single poll showing a general election matchup between John Edwards and a Republican?

On the other hand, Fox has released 23 matchups featuring Hillary Clinton and a Republican -- and 10 with Barack Obama.

Fox must know what a comprehensive analysis of publicly available polling data shows: John Edwards is the most electable Democrat, leading Republicans by 8 points on average, compared to 5 points for Obama, and 3 points for Hillary.

Over the last few days, I conducted a comprehensive analysis of all the publicly available polls showing general election matchups that I could get my hands on.

I explain the details of my methodology below, but the key point is that the analysis was limited to polls including all three Democrats matched up against the same Republicans. In other words, instead of comparing a Hillary Clinton - Rudy Giuliani matchup taken in March to a John Edwards - Mike Huckabee matchup from December, this analysis only makes apples-to-apples comparisons: Democrats matched up against the same Republicans in the same timeframe by the same polling organization.

As the above chart shows, since July 1, John Edwards consistently receives the highest support and holds the Republican to the lowest support. As a result, John Edwards would enter the 2008 general election with the largest lead of any Democrat.

What makes this especially impressive is that Edwards has actually gone through a national campaign -- he's been on the ballot in all fifty states. Like Hillary, he's already been through the Republican attack machine -- and he's still the strongest Democrat.

Given these numbers, it's not suprising that Fox hasn't been releasing any poll data on John Edwards' electability. On the other hand, they've been dumping a flood of data about Hillary Clinton. 23 of the 33 general election matchups Fox has released this year have included Hillary -- 13 of them featuring Hillary versus Republicans in without any other Democrat tested in the poll.

In those 13 polls, Hillary has an average lead of 7% -- well above her 3% average lead in all other polls. Is it any surprise that Fox would release data making our weakest potential candidate appear the strongest? In the polls where they've also tested Obama, Obama has had a higher average margin than Hillary. Fox's solution to this inconvenient problem? Just pretend that Obama (and Edwards) don't exist and exclude them from most polls!

(Of course when Hillary brags about her electability, she includes these Fox polls in her analysis. Who could blame her? It's probably the only really nice thing Rupert Murdoch has ever done for her, besides contributing to her campaigns.)

_______________________________________________________

Just for fun, I put most of this data in a YouTube video clip. ***WARNING*** It is really wonky! Only for geeks! (Or Ben Harper fans!)

How I conducted this analysis:

First, I developed a database of all polls taken in 2007, drawing from the polls at Real Clear Politics and adding a single Rasmussen poll from mid-October that RCP had neglected to include in its listings. I also included the most recent CNN poll, which RCP had not fully included at the time I pulled their data.

After completing this step, I had 485 poll matchups in my dataset (e.g., Clinton-Romney, Obama-McCain, Edwards-Thompson, etc.).

One of the real challenges in comparing performance in polls is that not all polls test all Democrats. Since any good comparison must be apples-to-apples, I sifted through the polls and identified each one where Clinton, Edwards, and Obama were each matched up against identical Republicans. I expanded this to include Rasmussen Reports polls where each of the three Democrats were matched up against identical Republicans in the same time frame, generally a few days apart.

After doing this, I was left with 228 datapoints -- 76 identical polling questions for each Democrat, matched up against the same exact Republican opponents.

For the charts I presented above, I included all data from July 1 to present -- 90 data points for 30 identical matchups for each Democrat. I'll present the full data below, but the matchups for the entire year don't change the electability story. Indeed, they make Hillary look worse, because she only won 45% of the apples-to-apples polls in the first half of the year.

Why is it so important to do an apples-to-apples analysis?

Beyond the obvious, analyses that don't do apples-to-apples can yield misleading results.

Take the example of Fox News, for example.

This year, Fox has released 33 polling general election polling matchups, 23 featuring Hillary Clinton and 10 featuring Barack Obama.

According to Fox's polling, Hillary Clinton wins 78% of the 23 matchups with an average margin of 6%, compared to 70% for Obama with a 4.7% margin.

However, in the 20 polls that are apples-to-apples, Hillary and Obama both win the same percentage (70%) and Obama actually has a higher victory margin than Hillary (4.7% to 4.1%).

Indeed, while Hillary Clinton has a 7% average margin in the 13 Fox News general election matchups excluding both Obama and Edwards, in all other matchups (198 in all) she has just a 3.1% margin.

Now, returning back to the analysis at hand, let's take a look at all the data, including results from January 1 to present.

As you can see, Edwards has the strongest performance of the Democrats in terms of margin of victory not just for July 1 - present, but for January 1 - present.

Barack Obama also does very well. In fact, he has a slightly higher winning percentage than John Edwards. It's important to understand what that means, however. For example, since July 1, Edwards has had an 80% winning percentage and Obama 83%. What's the difference? A single poll, by Quinnipiac, matching Democrats up against McCain. Edwards trailed that matchup by 1, and Obama led it by 2.

Average margin is actually a more important statistic. It's the same idea as tracking runs scored versus runs allowed in baseball. For years, there was a myth that the teams that win one-run games are the best teams in baseball. It turns out, that idea is total hogwash. A team's total net runs scored is a far more accurate way of predicting a team's future success. If it's the all star break and you're betting on whose going to win the World Series, you want a team capable of blowing its opponents out night after night, even if another team that has gotten lucky in some close games has a slightly higher winning percentage. (I could go on with this comparison, but I won't. If you're interested in learning more, read about Bill James Pythagorean Theorem of baseball.)

What does John Edwards large lead mean for Congressional elections?

If we nominate John Edwards, the presidential election won't be as close as it would be with Obama or Hillary. As a result, the party will be able to devote more of its resources to Congressional elections -- and a huge Democratic majority in Congress is essential to enacting a progressive agenda.

What's the bottom-line?

The bottom line is clear: based on an apples-to-apples comparison of publicly available general election polls, John Edwards is the most electable Democratic candidate. Barack Obama is also a strong candidate, but the difference between him and Edwards is significant: about 3 points in marginal performance since July 1 and 1.5 points for the entire year. (Moreover, Edwards has already been through a national campaign and the GOP attack machine, while Obama hasn't.)

Both John Edwards and Barack Obama are clearly stronger than Hillary Clinton, however, who trails both Obama and Edwards. She averages just a 3% margin since July 1, 2 points below Obama and 5 points below Edwards. For the year, she averages just a 1% margin. Because she is a weaker candidate than Edwards or Obama, the Democratic Party will be forced to spend more of its resources on her election, which will have a negative impact on our ability to win a huge Congressional mandate.

If electability is your number one priority, John Edwards should be your candidate.