As a stats junkie I will definitely be playing with this stuff... looks like conservatives favor Switchfoot and Lifehouse by about 2-to-1 margins and liberals favor Rihanna and Jay-Z by about 2-to-1 margins. I'll have to come back to this.

1. Are libertarians excluded, or just grouped with another constituency?

2. Should you be tracking the number of conservatives vs. liberals overall on Facebook, or proportionately versus each other?

--Hal Levy / studentsforkucinich.com

This data is excellent and certainly light years ahead of what was available before from social networking sites. However, "the greatest microtargeting engine ever built" might be a bit optimistic. Almost all, if not all of these market segments can already be found on the market. I don't think it's in Facebook's strategic plans to compete with the credit bureaus and other monster data compilers who build models based on thousands of data points for each and every household in America.

This is a great find, thanks for sharing. A free database of millions of users' self-identified political leanings and interests/affiliations is pretty amazing.

But I'll echo some of the words of caution above. First, this data is most useful for understanding the Facebook-using population, which skews more liberal, educated, affluent, younger, and tech-savvy than the rest of the population as you noted in your previous post. So any data gleaned should be interpreted with those facts in mind, or normalized appropriately.

Second, the simple results aren't all that interesting outside of the context of placing ad buys. Of course more conservatives will like Garth Brooks and football, and more liberals will like the West Wing and be vegetarian. Where the real value will come from is when/if this raw data gets put into a data warehouse and goes through more sophisticated mining techniques. That's where the unexpected, "gee I hadn't thought of that" data will come from. But as mentioned above, if money is going to be invested in such an undertaking, better to use the already existing consumer data providers like Choicepoint, instead of self-identified data from Facebook. Also, Facebook knows the tremendous value of mining their data (they certainly do it already) for their own advertising deals, and would never expose such sophisticated tools to the public.

That's right. I'm not sure Facebook is replacing big data vendors like Acxiom any time soon. But you could argue that this is a nice supplement to that data for the younger demographic, who tend not to consume content online and not subscribe to magazines that would make them a data point for the big commercial firms. Plus, I would imagine this is a better guide to cross-referencing media consumption patterns (favorite TV shows, movies, books) than most commercial vendors.

I also think we need to get away from the idea that self-reported data is bad data. If someone cares enough to self-report an interest in your product, that's the kind of customer you want. I'd think you'd want to profile those people moreso than folks who give the right answer when prompted.

I agree. 18-24 year olds are notoriously difficult to market to directly because they don't yet have the credit, voting or consumer histories that their parents do. Facebook could easily be a better media choice for colleges, distance learning centers and other marketers who must really zero in on this market segment. I would imagine that as Facebook grows any direct marketer will be able to overlay their own cutomer/constituent model over Facebook's users and then they are off to the races. And the value of self-reported data depends on how it is collected. It is usually more reliable than inferred/modeled data and thus more expensive to collect and rent.

I am baffled by just one thing in Patrick Ruffini's commentary.
I must say at the onset that I read his reportage religiously, several times a day, on most of the new stories and blog comments on Bill Richardson, my choice for President, and Ruffini's grasp of the breadth of kinds of articles is almost alway top knotch.

Except for just one small fact: Ruffini still lists ex-Iowa Governor Vilsack as one of the Democratic candidates, even though he has withdrawn many months ago, and even despite the fact that I sent Mr. Ruffini 4 emails to that effect to remind him that Vilsack had quit, but to no avail....

Thus, it seems a bit hypocritical for Mr. Ruffini to blast the outgoing Brownback for his lack of internet attention, when Mr. Ruffini seems to have paid little attention to his own website on this change in candidates, or to his own fan's emails reminding him to correct his own website.

Ad Age has an interesting article today on the value of self-reported data on Facebook. It asks the question, "are people telling the truth about their age, etc. in their profiles?" Based solely on your data, I personally think they are. Or at least they aren't lying so much and so deliberately that it makes a big difference at this point. Tastes in bands and tv shows will definitely change and people will normalize their cultural preferences to "fit in". But the basic demographics in your spreadsheet I think are good. They make sense and sometimes that's all you have go on. The only real way to know is to test. If an ad campaign tailored to conservatives works, then the data is good enough.