After the US elections a year ago, I decided that "liberal media bias" was far more harmful to liberals than conservatives. In fact, if I were a Democrat, I'd be getting a little miffed at the recurring pattern of the past two years: throughout the election campaign, my newspaper produces a poll showing my guy way ahead; finds "typical voters" (choreographers of environmentalist dance companies, etc) anxious to blame Bush for the worst recession since Hoover; runs front-page features on how Clinton's flown in to campaign with my man, exuding the rock-star glamour that so enthuses the base, etc.
And then the morning after election night, I wake up to discover that, in a stunning upset utterly predictable to anyone but the expert media analysts, the Democrat got hammered.
But not to worry. Just as your rattled Democratic supporter is beginning to feel a harsh jab of reality in what Slate's Mickey Kaus calls the "liberal cocoon", the media rush to lull him back to the land of make-believe, assuring us that the Democrat defeat is attributable to strictly local factors and is definitely not part of a trend.
Oddly enough, all these non-trends seem to trend the same way: November 2002 - Democrats lose control of the US Senate; October 2003 - Democrats lose the California gubernatorial race; November 2003 - Democrats lose the Mississippi and Kentucky gubernatorial races.
If it weren't for media bias, Steyn suggests, Democrats might be trying to do something about this.