Pew has released an expansive survey of the two coalitions that deserves a closer look. It divides the electorate into nine different groups, eight of which are engaged with politics: Enterprisers, Social Cons, Pro-Gov Cons, Upbeats, Disaffecteds, Bystanders, Con Dens, Dis Dems and Liberals. According to Pew, each group's percentage of the population, percentage of registered voters, two-party turnout, and Bush vs. Kerry breakdown went as follows:

	       Pop %  Reg %  Turnout  Bush  Kerry
Republicans
Enterprisers	 9     11	      93     92     1
Social Cons	11     13	      90     86     4
Pro-Gov Cons	 9     10	      73     61    12

Middle
Upbeats 		11     13	      77     63    14
Disaffecteds	 9     10	      63     42    21
Bystanders	10     0	       3     NA    NA

Democrats
Con Dens		14     15	      79     14    65
Dis Dems		10     10	      84      2    82
Liberals		17     19	      63      2    81

I am not really sure why "Upbeats" aren't just considered Republicans.

Democrats are able to stay close because their groups are larger. In fact, Liberals and Conservative Democrats are the two largest groups, and are equal to the size of the three Republican groups, Enterprisers, Social Conservatives, and Pro-Government Conservatives, combined. Republicans are able to maintain their edge largely through their overwhelming strength among the two "middle" groups that are politically engaged: Upbeats and Disaffecteds. It is worth noting that Liberals are not only the largest group, but also the fastest growing. Since 1999, they have doubled in size.

Turnout did in act slightly favor Republicans. By my calculations from the numbers above, had the turnout between all groups except "Bystanders" been equal to their levels of the population, then Bush probably would have won a very, very narrow popular vote victory, but also probably have lost the electoral college (maybe 284-254, with Iowa, New Mexico and Ohio switching). Certainly, there would have been extensive overtime. While it certainly isn't a cure-all for the crushing and systematic problems with the Democratic Party, we can never ignore the importance of improving turnout among our base.

As the graphic on the right illustrates, when it comes to determining the difference in outlook between the two parties, national security is by far the clearest difference:

Anyone hoping to close the so-called "national security gap" between Democrats and Republicans needs to take into account that the clear difference in ideological and policy outlook between members of the two parties is in the area of national security. It is not entirely clear to me how we can close a gap on the issue that most starkly divides the two parties.

Lots, lot more in the study. Give it a look.