If you will forgive me for being elliptical, let me start this discussion with an example taken from a recent Democracy Corps poll that asked white Catholic voters to rate the two parties on a variety of abstract issues:

			    Dems      Reps
Pro middle class		     53        39
Public interest first	     50        37
Cares about people		     49        36
7On your side		     46        43
For families		     44        44
Shares your values		     42        48
Keeping America safe	     30        52
Respects faith		     29        52
Knows what they stand for	     25        58

To clarify, these numbers mean that 53% of the people in the poll thought that "pro middle class," applied more to Democrats than Republicans, while 39% thought the term applied better to Republicans than to Democrats. Apparently, 18% did not feel the term adequately applied to either party.

Now, looking at these numbers I can almost hear the reaction of a thousand pundits at once. In order to approve our electoral prospects, Democrats must improve on national security, respecting faith, and they must develop a backbone! However, I want to argue that is not necessarily the case.

Pundits have a weird problem that emerges whenever they try and explain what Democrats must do to improve their electoral prospects. Whenever they see Democrats losing on Issue X but winning on Issue Y, they become convinced that Democrats would gain more by improving ten points on Issue X than they would by gaining ten points on Issue Y. However, that does not make any sense. Moving from a fourteen point lead on "pro-middle class" to a twenty-four point margin on "pro middle class" would gain us exactly as much ground as moving from a twenty-two point deficit on "keeping America safe," to a twelve point deficit on "keeping America safe." Both would move us ten points, and ten points are the same wherever you draw them from.

If all issues are of equal importance to the electorate, and the electorate has equal swing potential on all issues, than it does not matter what issues Democrats make gains on, as long as they make gains. It is not a reasonable argument to claim that just because Democrats are losing among those who cite "terrorism" as they top issue that Democrats must focus on "terrorism" instead of, say "education." If all things were equal, developing an even greater lead on "education" would be just as useful as closing the gap on "terrorism."

This is almost exactly the same mistake many people make with identifying swing demographics in elections. They often assume that the demographics that are closest to 50-50 in terms of their voting patterns are the swing groups. This isn't true. Swing demographics are those that demonstrate the largest change in their voting patterns from one election to the next. I would wager that it would even be possible to determine the most important swing demographics by creating an index that combines how large an electoral demographic is with how much that demographic swings. All you would have to do is multiply the size of a demographic with the size of its average swing, and viola.

Now here is where I come to my weirdo point. The same can probably be done with issues. The more volatile the opinion of the electorate on a given issue, combined with the importance of that issue to the electorate, the more important the issue becomes in winning elections. In other words, using exit polls it is actually possible develop a metric to determine what issues caused a swing between the two coalitions from one election to the next. Comparing exit polls from 2000 and 2004, I have done just that.

2000 Vote By Most Important Issue
		  %   Bush   Gore  Dem Margin	
World Affairs	 12    54     40      -1.7%
Taxes		 14    80     17      -8.8%
Education		 15    44     52      +1.2%
Economy/Jobs	 18    37     59      +4.0%
Health Care	  8    33     64      +2.5%
Other		 33    43     53      +3.3%
Overall 		100		    +0.5%

The "Dem margin" column indicates the overall advantage Gore was able to earn among the entire electorate as a result of his advantage on a particular issue. For example, take "education." Gore led Bush by 8% among those voters who cited "education" as their top issue. 15% of those who voted cited education as their top issue. Thus, Gore took an overall lead of 1.2% among the entire electorate from the education issue, since 1.2% is 8% of 15%. Let me put this another way: had the 85% of the electorate that did not cite education as their top issue split evenly between Gore and Bush, Gore would have won the election by 1.2%.

2004 Vote By Most Important Issue
		   %   Bush   Kerry  Dem Margin   
World Affairs*	  34	60     40      -6.8%
Taxes		   5	57     43      -0.7%
Education		   4	26     73      +1.9%
Economy/Jobs	  20	18     80     +12.4%
Health Care	   8	23     77      +4.3%
Values / Other	  29	73     26     -13.6%
Overall 		 100		     -2.5%

(* = In 2004, "World Affairs" is a combination of "Iraq" and "terrorism.")

Now our task is easy. To show you what I am about to do, again take the example of the "education" issue. In 2000, the Democratic margin on education was +1.2%. In 2004, the Democratic margin on education was +1.9%. Thus, from 2000 to 2004, Democrats swung the entire electorate by 0.7% in their favor through education alone. Here are the swing totals for all issues:

Issue		 Swing from 2000 to 2004
Other / Values		 -16.9%
Economy / Jobs		  +8.4%
Taxes			  +8.1%
World Affairs		  -5.1%
Health Care		  +1.8%
Education			  +0.7%

The difference between Gore's margin on "other / values" voters and Kerry's margin on "other / values" voters was twice as large as the margin between the two on any other issue. On the plus side, Kerry did much better than Gore among "economy / jobs," voters, and almost entirely eliminated the Bush advantage among "taxes" voters. While the margin of defeat Kerry received among "world affairs" voters was greater than the margin of defeat Gore faced, it was not nearly as large a swing as that found among "other / values" voters, "economy / jobs" voters, and "taxes," voters (what few of those there seem to be left).

What does this mean? It means that there are many ways for Democrats to gain from election to election apart from just improving on the issues where we are getting crushed. For example, rather than just increasing our already gaudy margins among "Education," "Health Care" and /or "Economy / Jobs," voters, we could increase the number of voters in those categories. Also, rather than "taking world affairs more seriously," we could close the gap among "world affairs" voters if the number of voters in that category shrank. Or, rather than gaining ten points among world affairs voters, who seem at least to be a static voting group in party preference if not in size, we could gain further among taxes voters, who swung an astonishing 49 points in party preference from 2000 to 2004.

I don't expect everyone to accept this analysis, but I do hope that this discussion at least keeps some people from thinking that the only way Democrats can improve their electoral chances is by improving our image on issues where we are getting beaten badly. Push our issues to the front, push their issues to the back, find the swing issue voters--there are many other ways.