They're not using their computer programs to find an ideological bias, but a bias that assigns a vaguely positive sheen to candidates when they win. When Romney won in Michigan, his Pew positives increased. When he won in six more states, he got his "best coverage" of the year to date.
Mark Jurkowitz explicitly proclaimed on CNN "Winning begets good coverage ... and losing begets bad coverage." So Romney merely winning primaries loads up his "positive coverage" numbers, while Obama had no primaries to win.
Pew also pulled this trick back in October, claiming that Rick Perry was getting much better press than Obama. How does that one smell to you? You can see how much Perry's campaign succeeded with the so-called negative-toward-Obama press.
Meanwhile, current Washington Post media reporter Paul Farhi was also dismissing the notion that the public should perceive the press as favoring Obama's re-election. After a nod to conservative academic Tim Groseclose, Farhi presented a liberal named David D'Alessio to proclaim the notion of a liberal bias favoring Democrats was a big nothing-burger.
D'Alessio's so-called "meta-analysis" of bias studies in presidential elections from 1948 to 2008 found that "left-leaning reporting is balanced by reporting more favorable to conservatives," and so over the decades, "The net effect is zero."
The Post reporter even brought in tired radical left-winger Eric Alterman to make the old claim that the media are so traumatically hounded by conservative media critics that they bend over backwards to be kind to Republican candidates. If you buy all these assertions, you'll also buy that Romney travels from state to state on a magic carpet.
The media's homegrown experts are not displaying an actual interest in measuring bias seriously, as much as they are dying to dismiss it flippantly. They don't want the public to suspect the media are more dedicated to promoting Obama's path to victory than the public's "right to know." But the public can smell it.