It seems that recently people are looking everywhere for trends to determine the direction American politics is taking. Of course the two most common arguments are the liberal one, which points to Bush's low approval rating and argues for a liberal movement in politics, and the conservative, which points to the equally low Congressional numbers and argues for a conservative backlash.

Not to add a third argument, but I think both are right, and wrong. The media has clearly been trying to talk down the Bush administration for some time, and by reporting every economic downturn as the next Great Depression, starting even before 9/11 with talk of a "looming recession" only weeks after Bush took office, as well as reporting only negative news from Iraq, and most recently trying to talk up first the oil "crisis" then the subprime "crisis", the press has succeeded. But they succeeded too well. They did not create a distrust of Bush, but a distrust of existing government as whole. The press may have been most vocal in bashing Bush, and the backlash may have fallen mostly against the Republicans this time, as voters tend to associate the presidential party with the government as a whole, but I think 2010 will surprise some. If the current displeasure with governemnt continues, the congress may show some dramatic changes.

But that is not my point here. Instead I want to show how statistical argument, even those which appear plausible in isolation, can often lead to absolutely incorrect conclusions.

My example comes from a comment posted a while ago, arguing that the nation as a whole is moving toward accepting gay marriage:

 In 2000, Californians voted in a constitutional ban on gay marriage by 61% (called Prop 22 then).

Now, 8 years later, the new ban on same-sex marriage has passed with a meager 52% approval.

For anyone bothering to follow the trajectory of this social movement, HUGE GAINS were made in the past 8 years. It is only a matter of time now (2 more years? 4 years?) until California voters vote in same-sex marriage (unless the CA Supreme Court gets to it first and declares Prop 8 unconstitutional like the one before it, using the exact same arguments).

The argument sounds, on its face, like a valid conclusion from the numbers. Then again, so do many invalid arguments. By ignoring context completely it is easy to make the numbers alone say pretty much anything. However, once context is considered, a lot of numbers say something quite different than the numbers themselves would suggest.

First, let us look at two factors that argue that these numbers are not a valid predictor for California, then look at three other numbers that suggest that even were they valid for California, they hardly hold as a pattern for the nation as a whole.

The argument ignores two factors which need to be considered to put this change in context. First, we need to ask why this same question is up for a second vote at all. Normally a constitutional change needs only be approved once, so why a second vote? Simple, the courts struck down the amendment. That being the case, it is likely the numbers changed precisely because of this court action. The first time many voters came out precisely because they wanted to support the amendment. Many social conservatives voted only because of this question. Now, knowing that this amendment may be struck down again, just as the first one was, and the general election itself having no real social conservative candidate, a number of supporters probably stayed home, making the margin closer.

And, having mentioned the election, let us look at that as our second factor. This election, pitting a popular liberal candidate against a relatively unpopular moderate Republican, was unlikely to bring out the social conservative in great numbers. It did bring out two other demographics, black voters and college students and other liberal "new voters". The black vote, not surprisingly, went in favor of banning gay marriage, so that did not cause the change, but the combination of low turnout among social conservatives, as well as the high turn out of very liberal new voters could clearly push the numbers closer. IN fact, given the dynamics of this election, that the proposition passed at all attests to rather strong, broad based support.

But let us suppose the numbers are right, and California is inching closer to true gay marriage. Does that prove the rest of the nation is moving in that direction?

Well, no.

First, California is a very liberal state. It has pockets of conservatism, but they are dwindling over time. A number of influences have pushed California ever more leftward. First, immigration, creating a massive Hispanic voting block, has given the Democrats majorities in areas they traditional did not control, allowing them to gerrymander districts to provide even more solid majorities. Second, through a combination of confiscatory taxes, elevated property prices through smart growth plans, miserable energy policies, and anti-business legislation, not to mention the insane federal ninth circuit which consistently supports such policies against any challenges, they have driven both individuals and corporations out of the state, while insane welfare policies in place such as Berkley have drawn in new state supported charity cases, a reliable liberal voting block. All of which has served to push California, already to the left of the nation in the late 80's, ever farther to left. So there is simply no way one can fairly take California as a bellwether of the nation as a whole. It would be like basing your projections on Massachusetts or Texas. California is no more characteristic of the nation than those two other states.

And the numbers support that. Assuming that the numbers mean what they say about California, then California is bucking the trend of marriage amendments, as ever more states nationwide are passing marriage amendments, even some states traditionally considered relatively liberal. If the trend in California is toward repealing such amendments (which I don't believe) then California is out of step with the nation as a whole.

And finally, the last bit of evidence. Not only is the nation moving toward a more conservative position on marriage, but, in general, where voter initiatives have been enacted, conservative measures in general have been quite popular and successful. Whether removing racial preferences, enacting term limits, blocking tax increases, or anything else, the conservative side has won time and again in ballot initiatives. Granted, from time to time the liberal side has won, but most often because of a massive PR campaign, or, as I suggested above, after the court has struck down one or more initiatives, leaving voters thinking such initiatives are futile. However, the overall trend has clearly been toward the conservative end, which would indicate a move toward legalizing gay marriage is clearly out of step with the rest of the US.

All of which simply shows that numbers alone, without context or analysis, can obviously mislead. As I should know well, having been mislead by the bias of the presidential polls into thinking McCain would make a better showing than he did. In my case, I ignored the fact that the focus on economic matters, coupled with McCain's vacillation on the bailout, would ruin whatever credibility he had in areas such as defense, and in the case of the post above, the poster ignored both factors internal to California, and factors which set California apart from the rest of the US.

Or, to make it simple, never accept statistical trends without asking whether or not they can be explained some other way.

POSTSCRIPT

I wrote before on a solution to the whole gay marriage debate. It is a solution I doubt will ever be enacted, being anathema to both liberals and social conservatives, as well as removing some subjects from government control, which is rarely popular with any politician. But here are my two essays for those who are interested:

Solving the Gay Marriage Debate
Updating an Old Post

It may not satisfy many groups, but I have to say, were it instituted, after a century or so, I am sure people would look back at our current system in disbelief, the same way we question a Papacy that ruled a temporal empire, or the use of state authority to enforce laws against heresy and witchcraft. What seems self evident and common sense at any point in time can be absurd in a few generations. We need to always keep that in mind.

On the other hand, I have myself argued that we should avoid making dramatic changes, so perhaps we are better off simply making small incremental changes rather than the broad reforms I suggested. Some of my arguments for a sl;ow pace of reform can be found here:

In Praise of Slow Changes
Predictability
The Problem With Evolving Standards
Why Judicial Activism Hurts
Interpretation and Activism

So, the question boils down to this, in my mind, assuming we could get support for my solution, which is a big "if", would we be more harmed by the drastic change than we would benefit from the reform itself. And that is a question I am not presently able to answer.