Conservatives have been out on the Internet for the past week trumpeting an Oregon study that found Medicaid expansion was associated with only small improvements in physical health, plus large financial gains for the poor, plus substantial increases in mental health. They really like this study, and think it rebuts all other evidence that's ever been gathered on the health benefits of health insurance, up to and including allowing them to make sweeping claims about mortality that were not actually researched in the Oregon study.
The natural interpretation of this is that conservatives are rabidly opposed to government programs that help poor people, but don't want to bite the bullet that refusing the help poor people leads to human suffering.
But perhaps it's all really a fussy concern for science. Conveniently, conservatives right now have an excellent opportunity to prove me wrong about this. The whole reason the Oregon study arose was that Oregon had enough money to give some additional people Medicaid benefits but not everyone. So they chose to run a lottery. A lottery is great from the viewpoint of experiment design, but the sample size here was tragically small. Meanwhile, the Florida legislature just refused to expand Medicaid to anyone even though the state's Republican governor says he supports expansion. And Florida isn't alone. Texas won't be expanding Medicaid either. In fact, most of the deeply conservative states of the south and the plains won't be expanding Medicaid. But these states—especially Texas and Florida—have much larger uninsured populations than smallish Oregon ever did. What's more, the amount of money around to finance Medicaid expansion is really big. So if conservatives in Texas and Florida really want to get to the bottom of the whole Medicaid question, they now have a golden opportunity. They ought to apply for a waiver to get federal money to expand Medicaid to cover half the eligible population. Do a lottery, just like in Oregon. And do follow-up studies. That would give us a much larger sample size and much stronger evidence.