Several bloggers have brought a set of graphics to my attention on the age distribution of political leanings based on a dating website's information. There are two big pictures they post and I haven't seen anyone else put both together, so that's my marginal contribution.
The first shows that people start off life relatively Libertarian - no one should tell me how to live my life, socially or economically - then prefer a bit more income redistribution once they enter the labor market and look up the ladder of success. As they get older and their paycheck grows, they want more economic freedom the story goes, and society should do a better job respecting our values. Finally, they realize they haven't saved enough for retirement and prefer a bit more redistribution again. Overall, about half of the age distribution leans Republican and the other half leans Democrat.
However, this story is based on cross-section data, NOT NOT NOT panel data. We are not following the same people over time. This is more of a generational trend: the generation before mine tended to be hippies, my generation are yuppies and Xers, and then we have another generation that hasn't fully told its story yet but is leaning Democrat. So it's easy to make too much of this graph, however instructive it is.
Another reason not to make too much of this is self-selection. If there is something different about 50-60 year olds who are visiting dating websites than those who are not, that will skew the results. There's also no discussion of folks who opt-out of the questions altogether, or of people who self-identify as Independent. There's also a signaling aspect: they may adjust their answers based on the type of person they are trying to attract.
Leaving those problems aside, the other interesting diagram shows the age-size and policy-space of the tents cast by the Dems and Reps. Notice that the political space for the Republican tent is much smaller than that of the Democrats. Democrats are in fact including polar opposites: Libertarian youth and authoritarian retirees. This means that Republicans have an easier time voting as a block than Democrats, who are easier to pick off and appear less organized and less well-defined. The analyst discussing this uses nuns who favor income redistribution but are opposed to abortion (bottom right quadrant) as an example.
One of the implications I draw from this (another marginal contribution) is that when the "pure economic-freedom" people want to purge the Republican party of social conservatives and the religious right, little do they realize they would remove the largest block of the party.