Welcome to the general election campaign. You’re now seeing an excellent example of how that’s going to go, in the form of a truly ridiculous twitter war over something that Democratic operative Hilary Rosen said on CNN that Ann Romney was insulted by and that…oh, never mind. This kind of thing really, really, doesn’t matter. Which brings up the need for a quick guide to what does and doesn’t matter to the outcome of campaigns.

Political scientists have had a long argument about whether anything in campaigns matters to general election results. That argument has mostly been won by those who have demonstrated that campaigns can have some effect in November. But still: Whole campaigns, including everything from door-to-door electioneering to millions of dollars of TV ads to all those rallies with the bunting and the speeches and the music — all of that probably moves the needle a few percentage points.

That’s very important, no doubt, in a close race! But it also places everything in context. If debates rarely affect the results at all; if vice-presidents rarely affect the results beyond perhaps a small boost in their home state; if perceptions of the candidates’ personality may not matter at all; then what are the chances that a silly flap about what some talking head said in April will have any effect whatsoever? The first rule: Basically ignore the back-and-forth.

Second rule: Polls. We’re now getting into the point at which these polls start being somewhat meaningful. But individual polls fluctuate all over the place. Not only is there a margin of error for any particular poll, but statistically we should expect one in twenty polls to be outside that margin of error. Swings of five to points are totally normal just from the luck of the draw. For a better sense of what’s happening, use the poll-of-polls, such as the terrific one from Pollster.

Here’s what does matter.

Many campaign events are important mainly because of their effect on future governing. The choice of running mate is not likely to have a significant effect on the outcome. But you might want to remember Harry Truman, Lyndon Johnson, Richard Nixon, George H.W. Bush, and others who probably would never have been president had they not first been nominated on a national ticket. And campaign promises matter in one sense: Research shows presidents do try to keep them once in office.

Here’s what’s important in determining who will win: Polling does become meaningful, but only later this summer. And here’s how campaign events can make a difference: Look more for broad themes than single-day flaps. We have pretty good evidence that perceptions of ideology can affect vote choice, so it’s worth paying attention to anything that might really move the way voters think about that.

Most important of all: External events. It’s the events far from the campaign trail that will influence swing voters the most. Chief among them is economic news, but other major events, such as foreign policy crises and natural disasters, especially if they make a real difference in people’s lives, can change votes. These are the real things to watch.