I was wondering when someone was going to ask me about that! I've been noticing this for the past few years. Here's an example:
And the Snopes version. Notice they say this is 'False', but they don't have all the information.
snopes.com: Martha Raddatz/ABC News Interview
If a Republican were in office, it might appear that SNOPES has a right-wing bias. The point is that SNOPES makes mistakes, and I've noticed plenty of them. I'm not just going to take their word for it like Modbert does. If it's important to me, I will check it out and find out the truth for myself.
This is what they had to say about that:
"Regardless of the number of military personnel interviewed, whether this segment reveals some deliberate agenda on the part of ABC to misrepresent the political preferences of U.S. military personnel
is an argumentative and subjective issue.
On the one hand, one side claims that the ABC report wasn't
supposed to be a representative sampling of party preferences; it was supposed to illustrate that American troops are following the presidential campaign closely and evaluating candidates based on their positions on
all the issues (not just the war in Iraq), and some are even favoring Democratic candidates who advocate U.S. withdrawal from Iraq. Hence, the preponderance of interviews showing soldiers who were not (as many might expect) reflexively endorsing the Republican candidate, John McCain.
On the other hand, critics maintain that by showing only one soldier's expressing a preference for the Republican candidate (prefaced by a laconic Martha Raddatz voice-over intoning, "there
were some McCain backers ..."), by separating the portion of the report in which soldiers discussed their candidate preferences from the portion in which they discussed what issues (other than the war) were important to them, and by identifying the report with titles such as "Whom Are Our Troops Endorsing?" and "Surprising Political Endorsements by U.S. Troops," ABC News presented the piece as being a survey of American troops' presidential preferences without offering a true representative sampling of those preferences"
Seems pretty even-handed to me.