Remember time. One way to think about this dilemma is that evidence is a result of observing reality. Evidence are about the present and the past. Politics on the other hand is a way to change reality, which will change the future observations to give different evidence as a result. Politics is about making changes now to change the future. Evidence and politics are connected but separated by time.

Indeed. But questions of how to evaluate results are always answered by way of some kind of preference, desire, or ideology, right? Different ethical principles can be in conflict; there's no objective measurement of the good.

I can answer that one: look at other countries. Look at what they do and compare the results. Some countries use taxes to provide health care, some fund it partially by taxes, some don't use taxes at all. Compare the results: which system gets the best health care for the least amount of money?

That is indeed a good way to gain insight into consequences, but still you will have to make ambiguous decisions when formulating the measure of good healthcare—e.g., do you consider equality for rich and poor?—and when judging tradeoffs like those involved in budgeting tax revenue, etc.

I agree that this is the main point: What are you trying to achieve? If your aim is to increase average happiness, what do you consider to be a good measure of happiness? And what kind of average are you using? Most "ideology" relates to this kind of arbitrary preference, I think.

But there are also other problems with evidence: How well do you trust the authors and institutions that produce or report the evidence? And if the evidence is probabilistic, as most evidence is, outside mathematics, what are your prior probabilities (in a Bayesian sense)?

In the case of international comparisons of health care, and other situations in which there is no repeatable experiment, there's also the problem of whether A caused B or whether A and B are both consequences of something else, perhaps something that the authors had considered and dismissed with some non-numerical justification, or perhaps something that wasn't even considered.

I personally usually don't have strong preferences in politics. Most things make sense (or no-sense) in some light.

That's why I have trouble choosing a party. Most issues are a close decision for me, so they don't clearly fall into one party's political view.