It doesn't look to me like John McCain was "unhinged" or "irate" or losing his "cool" in his recent videotaped airplane confrontation with the NYT's Elisabeth Bumiller. He was simply employing the debating tactic he often uses when confronted with a question he can't answer safely--which is to bully and intimidate and interrupt the questioner, using up all the available conversational space until the "questioning" moves on. (To get a word in edgwise, whoever is confronting him would have to be ready to engage in an undignified shouting match, which most are unwilling to do.) McCain used the same technique in the Republican debates when confronted with questions he didn't want to answer on immigration.
Because this is intentional, strategic behavior it isn't a sign McCain is unstable or uncontrolled or overemotional or irrational. But it's a sign that, no less than Obama, he may have been underprepared for the fall campaign by his charmed life as a national press favorite. McCain's bullying evasion is the second campaign tic--the first is his habit of reflexive, righteous blunderbuss denials**--that he's apparently been able to get away with over the years. Neither is likely to hold up over a multi-month presidential race. And the bullying, unlike the righteous denial, doesn't even temporarily make McCain look good.
**--Indeed, Bumiller was asking McCain about one of his earlier reflexive, sweeping denials that later turned out to be inaccurate. ...
3:35 A.M.-- The Trouble with Dials: Despite lots of dismissive punditry--It's a cliche! Badly executed! And look at Obama's swift response!--Hillary's "3 AM" ad appears to have worked. Intriguingly, the ad also worked despite performing poorly in the MediaCurves.com sample of 554 Democrats hooked up to reaction meters (on which they registered their agreement or disagreement).
Which seems to demonstrate a problem I've always had with Frank Luntz-style "dial" groups: The meters measure the voter's visceral reaction to whatever the candidate is saying. If the voter hates abortion, and Candidate A attacks abortion, the meter goes up. If the voter is pro-choice, the meter goes down. What the meter doesn't capture is actual rumination--even fleeting doubts or flashes of confidence. The reaction loop's too short for that. So if something Candidate B says, in the course of defending a right to abortion, actually makes a pro-life voter think twice about the issue, that will happen later, after the meter has moved on (and probably after the meters are locked up and everyone's gone home). Indeed, the voter's immediate reaction to a candidate who prompts reconsideration of a long-held position may be more negative than usual, reflecting the voter's annoyance at being challenged and forced to think. ....
In short, the meters are good at measuring effective pandering, not at measuring effective persuasion. And sometimes candidates do persuade! ... In the case of the "3 A.M." ad, the MediaCurves "undecided" voters were viscerally turned off when they learned it was an ad for Hillary. Their dial-graphs plummet downwards. But a lot of "undecideds" seem to have been affected, non-viscerally, in a different way later. .... 4:37 P.M. link
Obama tried pandering to Latinos in California. It didn't work. He lost. He tried pandering to Latinos in Texas. It didn't work. He lost. He tried pandering on NAFTA in Ohio. It didn't work. He lost. Maybe he should try, you know, not pandering! That would fit better with his claim to practice "a new kind of politics and a new kind of leadership," no? ... Update: Going negative, as recommended by Dick Morris, does not fit well with that claim--it seems like the sort of campaign mistake that might actually cost Obama a large chunk of support when he's mathematically almost home. Halperin's brief on this subject is persuasive. ... There must be other ways for Obama to "start acting like he has a pair." Like by dramatically not pandering! ... 2:57 P.M.