Have you noticed how flexibly media types interpret electoral victory margins--almost always in the way that favors the right-wing loons?

Three thoughts:

• That old "Joe-mentum"

I suppose it's remiss of those of us in the anti-Lieberman camp to fail to honor His Holiness's stunning electoral triumph in yesterday's Connecticut Senate primary. Okay, here goes:

This represents the senator's most stunning electoral triumph since his famous virtual-three-way-tie-for-third-place (actually, fifth-place) finish in the 2004 New Hampshire Democratic presidential primary. Yes indeed, boys and girls, he's got that famous Joe-mentum going for him once again.

• Meanwhile in Roveworld . . .

When I blither on, as I often do, about the genius of Karl Rove, disgruntled liberal friends who want to believe that he's not that all-fired a genius frequently point out the thin margins by which his candidates typically win their races. (By contrast, I tend to focus on that phrase "his candidates typically win their races.") Somehow it hasn't prevented the Rove-bots from taking over all three branches of the federal government plus all those statehouses and legislatures.

And somehow the thinness of those victory margins never seems to matter. The extreme case was the margin by which George W. Bush, er, "defeated" Al Gore in 2000.

• It's all relative, or something

You know how media "analysts" like to point out that a poll result that shows a 5-percent margin with a 5-percent margin of error actually represents "a statistical dead heat"? It doesn't really, of course, because that takes into account only one possibility: that the error swings fully in one direction--technically possible, of course, but not statistically probable.

Well, in this case, that last Quinnipiac poll showed Ned Lamont with a 6-percent lead over the Holiest of Joes, which unquestionably represented a significant tightening over Quinnipiac's earlier 13-percent margin. It certainly wouldn't be surprising that the race tightened a bit more in the time between the last poll and the actual election, especially when you consider how much money the Lieberman camp dumped into the race, not to mention how masterfully Holy Joe plays the role of Just Joe the Honest Progressive for unwitting media shills, a role that he unfortunately plays only for gullible radio and TV audiences.

Now let's review. The poll showed 6 percent, which by election day narrowed a bit to 4 percent. And the verdict is . . . The poll sucks, and Joe-mentum wins, Joe-mentum wins. Or maybe, just possibly, the poll was pretty accurate, and in addition Lamont suffered some further slippage but still won by a margin pretty consistent with what the poll predicted?

Nah, it's probably a statistical dead heat.