This morning, danah boyd summed up the problem with the open data movement in a simple declarative statement: “Transparency alone is not the great equalizer.”

Delivering a short keynote at the Gov 2.0 Expo here in Washington, D.C., boyd, whose most notable area of research is the sociology of social networks, explained that open data is not enough to create the kind of participatory democracy that is a focus of the Gov 2.0 movement. When data comes out, people engage in new forms of spin, she said, and seek to control the interpretation of the data that is exposed.

“Information is powerful, but interpretation is more powerful,” she said.

If there was one major concept to take away from this expo so far, that would be it. People who are hip to the we-gov (as opposed to e-gov) concept are beginning to see that in order to bring netizens in as partners in governance, they need to be data literate, and need to be empowered with an understanding of what data actually means. Otherwise data — data that is useless to anyone except an intellectual elite — is largely just another tool for public relations, or a way to lower costs.

That latter case is actually a win-win in many cases, such as the cases we heard about repeatedly at Gov 2.0 Expo where cities like Boston and San Francisco have created entire economies around their open data. Through opening up data owned by the agencies that run public transit, the process of getting applications to help people use public transit — applications that calculate and present wait times or help plan trips, for example — becomes faster and more efficient. It costs government nothing. Small business makes money. Everyone walks away with a big grin.

Cammie Croft of the U.S. Dept. of Energy and Judith Freeman of the New Organizing Institute approached the former case, though, in a talk they gave yesterday, only from the other side: Instead of just opening data, they said, a cause or government trying to interact with their constituents and build a movement should go a step further and also explain to them what the data means. I don't think they would put it this way, but going that step further, using data as an engagement tool also has the potential to become a means of control in precisely the way boyd was warning against. An organizer would perhaps instead be teaching constituencies how to read, understand, and cross-reference any data — or, better yet, provide access to people who will do teach those things.

There will be “a power struggle” over what data means, boyd said.

I have already pointed out another problem that data often has: it can be useless because it was collected for another purpose and does not really say what you think it says, or difficult to use because it requires skill, beyond simple data literacy, to clean. There are right now only a few people who have skill enough to do that. I ran into Clay Johnson, the director of Sunlight Labs, yesterday, and this came up; he suggested that if those kinds of skills don't proliferate, then at the very least, folks should introduce geeks of his caliber (not, I think, how he would phrase it) into their networks.

An example of both this problem and this solution found me at this expo. Joseph Fung, director of IT and collaboration strategies at the Ontario-based Centre for International Governance and Innovation, told me about Justin Kozuch, who is creating a database to classify the digital media industry in Canada. Kozuch won a competition where the winner gets prize money for a project to improve Canada's digital media industry.

The law firm organization that hands out the award gave it to Kozuch because, Fung explained, the classification system Canada uses for business does not have very good classifications for digital industries.

“A lot of their revenues, a lot of their work, a lot of their employment information get classified not optimally,” said Fung, whose think tank is backed in part by an endowment established by Jim Balsillie and Mike Lazaridis, the co-chief executive officers of Research In Motion. (That's RIM, the makers of the BlackBerry.)

How are you going to make good analysis about digital media companies when you can't compile aggregate data on how big the companies are, on average, or how much they make? It's difficult and introduces more potential for inaccuracy.

“The information presented to policymakers can't be trusted because the data is dirty,” Fung told me.