Thinking about evidence and causation in same-sex marriage arguments

Please note that this post is from 2015. Evaluate with care and in light of later events.

A recurring theme in criticisms of allowing same-sex marriage — or, as Obergefell did, in finding that bans violated the fundamental right to marriage — is some variation of the “slippery slope.”

There is nothing wrong with considering possible consequences or implications of applying the logic of one situation to others. In fact, doing so is absolutely key to rational thought and effective decision making. But failing to logically consider the chain of events or enough of the full complexities leads to poor, unsupported, and fallacious conclusions.

Thus, opponents of any kind of state-mandated vaccination must deal with the likely consequence — based on evidence from states and nations that have lessened vaccine mandates — that fewer people will be vaccinated and that such an outcome will result in a greater incidence of preventable diseases. In other words, this is a “slippery-slope” argument with both causative and evidentiary support.

The existence of such a consequence, however grave, does not determine that best choice. The impact on liberty resulting from increasing state intervention may justify the deaths of some infants, for example (we allow infants to travel in automobiles, despite the death rate); the harmful effects, or high costs, of some vaccines may lead us to avoid mandating their use (we don’t mandate the anthrax vaccine in children). But without both data and explanations (to support causation versus mere correlation, for example) we cannot balance such issues rationally.

Consider the argument that allowing same-sex unions will reduce the number of opposite-sex marriages. Positing this possibility is part of effective analysis — but is there evidence to support it? In legal-evidence terms, is there any foundation for such a conclusion? Justice Kennedy:

The respondents have not shown a foundation for the conclusion that allowing same-sex marriage will cause the harmful outcomes they describe.

Can we even construct a reasonable causative explanation for such an effect or, indeed, any negative effect on anyone? Again, Justice Kennedy:

Indeed, with respect to this asserted basis for excluding same-sex couples from the right to marry, it is appropriate to observe these cases involve only the rights of two consenting adults whose marriages would pose no risk of harm to themselves or third parties.

In other words, critics have both failed to provide evidence that same-sex marriages will cause anyone harm and failed to articulate a causative process by which this kind of harm could occur through the marriage alone.

In the same-sex marriage context, this evidentiary problem of harm (e.g., the American Sociological Association’s statement that “research shows parents’ sexual orientation has no bearing on children’s well-being”) meant that supporting infringement of the fundamental right to marriage was so difficult.

This is also why opponents, and the dissenters Obergefell, relied so heavily on historical, philosophical, religious, and definitional arguments. The kind of harm these arguments posited was not so amendable to sociological measurement and thus is not so easy to dismiss.

How does one counter through specific evidence Justice Scalia’s assertion that Obergefell is a “threat to American democracy”? How does one balance this argument of Justice Kennedy:

[Same-sex marriage] denial works a grave and continuing harm, serving to disrespect and subordinate gays and lesbians.

against this one of Justice Scalia:

We have no basis for striking down a practice that is not expressly prohibited by the Fourteenth Amendment’s text, and that bears the endorsement of a long tradition of open, widespread, and unchallenged use dating back to the Amendment’s ratification.

These kinds of value-intensive debates are still amenable to evidence-based analysis — particularly historical evidence — but are much more challenging to resolve, in my opinion at least.