Lies, Damned Lies and Statistics - As you saw in the comments to both of John's gay marriage pieces this week, the inclusion of statistical data about gay couples really intrigues me. For one, the data seems almost completely out of context and provides little information on the methodology used to obtain the figures.

For example, in yesterday's AP story about gay marriage the following statistics appeared:

Massachusetts has one of the highest concentrations of gay households in the country with at 1.3 percent of the total number of coupled households, according to the 2000 census. In California, 1.4 percent of the coupled households are occupied by same-sex partners. Vermont and New York also registered at 1.3 percent, while in Washington, D.C., the rate is 5.1 percent.

These stats come from a report by the Urban Institute funded the Human Rights Campaign Foundation. This study analyzed the data from the 2000 Census, looking specifically at same-sex unmarried partners who live together. The data does not include gays who are single and only reflects the couples who chose to select the "unmarried partner" option on the census form. The way in which the percentages were calculated was to take the number of same-sex unmarried partners and divide by the total number of same-sex and opposite-sex unmarried partners plus married opposite-sex partners.

Then today's Washington Post article on gay marriage referenced entirely different numbers:

If it were a state, the District of Columbia, where 1.5 percent of all households are formed by same-sex couples, would lead the nation in this category. Massachusetts, with a 0.7 percent rate, is tied for third with Washington state, behind Vermont and California, each with 0.8 percent.

The Post says that its information comes from the Urban Institute study. However, when I read the text of the study, and pulled up information from the Census, the Post's numbers simply don't match up.

In trying to hunt for additional background, I found an old story from The Advocate that contained another misleading factoid:

In most states, gay male couples and lesbian couples are represented in almost equal numbers--the remarkable exception being Washington, D.C., where gay male couples make up 73% of the reported same-sex couples.

Yes, when you compare Washington, DC to states, the percentage seems "remarkable." But, the statewide data isn't all that relevant for this particular comparison. Instead, compare the percentage of gay male couples vs. lesbian couples in *counties* and you'd see that San Francisco County, California also has 73% male couples; New York County, New York has 72% male couples; and Suffolk County, Massachusetts has 62% male couples. Conversely, you also need to look at counties that seem to be more popular with lesbian couples, including Dane County (Madison), Wisconsin (38% male couples), Santa Fe County, New Mexico (43% male couples) and Chittenden County (Burlington), Vermont (36% male couples).

Time and again in this survey, the treatment of DC as a state makes it a statistical anomaly that reporters seem to find significant. The journalists must wonder if there is an undiscovered wave of gay patriotism that is causing us to flock to the nation's capital. But when you compare DC to similar counties (like San Francisco, New York and Suffolk) then things start to fall more in line. Washington also piques my interest because the District, Alexandria City, Arlington County and Baltimore Counties all appear on the list Top 25 Counties for unmarried gay/lesbian partners even though the Washington-Baltimore region does not appear on the list of Top 25 Metropolitan Statistical Areas. I'm sure it probably has something to do with Simpson's Paradox, and with the heavily hetero outlying suburbs dragging down the average, but it's just another reminder of how easily data can be manipulated to prove a particular point.

So, yes, California and New York and Massachusetts and Vermont have more coupled households who are gay than other states do. And that's why these states are the primary battlegrounds for gay marriage rights. But that's also precisely why we need to remember that the fight will continue in Santa Fe and Bloomington and Austin. Therefore, we must also strategize how to effect change in these places, where gay couples have a strong presence, but may not have statewide political clout to drive legislation.