[T]he accompanying analysis says “Barack Obama’s appeal to younger voters and John McCain’s support among older voters may have created a situation where the outcome will turn on the preferences of middle-aged voters — particularly those in their 40s.” You see analysis of this sort all the time, but it’s all based on a mistake — there’s not a demographic electoral college where “winning” particular sub-samples of the population is the key to victory and therefore it’s important to focus attention on the most evenly divided demographic groups. If John McCain persuades an Obama-supporting 25 year-old to switch to his camp, that has just as big an impact as one 45 year-old one 65 year-old or one 85 year-old.
Now, as Matt notes, there are voting categories that merit targeting more than others, because they have more persuadable voters than others. And due to the vagaries of geography, some voting categories are very important in determining who wins key states in the actual Electoral College. Finally, a vote that will otherwise certainly be cast for your major-party opponent is more valuable to capture than one which may otherwise not be cast at all, or will be cast for a minor-party candidate.
But the general rule is: a vote’s a vote.