What he said...

I've tried several times to make the point that there is no such thing as "value free" interventions and policy decisions. This chapter makes the point better than I can:

Policy research is commonly presented as a value-free endeavour but, logically speaking, practical conclusions cannot be derived solely from facts. In research methodology, to attempt to do so is termed the `Naturalistic Fallacy.' Consequently, the term `evidence-based policy'--if interpreted literally--constitutes a contradiction in itself (oxymoron). This chapter will also deal with several other fallacies that are encountered in empirical addiction research, will reveal some popular concepts to be inconsistent and illogical advocacy tools, and will argue that the specific role of a researcher is completely incompatible with the role of an advocate for certain ideas and/or interest groups.

...

Statements can be dichotomised into descriptive statements (factual judgements about what is) and prescriptive statements (value judgements about what ought to be done). The latter are referred to as ethical judgements. It is well accepted in philosophy (Hume's Law) that ethical judgements cannot be derived logically from empirical facts alone. Any syllogism to arrive at ethical conclusions requires at least one ethical premise as well. The flawed idea of basing ethical conclusions purely on empirical facts was called the `Naturalistic Fallacy' by Moore, and has become a well-known term in research methodology. On the basis of facts, only factual conclusions can be derived. However, this does not mean that ethical issues cannot be a topic for empirical research--the opposite is true. It is essential to identify implicit value judgements and to make them explicit. Only if all implicit value judgements in our scientific reasoning are made explicit is it possible to analyse whether the ethical premises are consistent with each other,* with basic ethical principles and with factual evidence. I have previously termed the process of identifying and analysing implicit values in research `ethical evaluation', but this term is ambiguous since it could mean either `to conduct evaluations according to ethical standards' or `to evaluate the ethical content in research.' More precise is the term `evaluation of implicit ethics' or the almost synonymous `evaluation of implicit values.'