Apparently variations in policy make it difficult to understand what works

This is quite possibly the most stupid thing we have seen today. Admittedly, it's early in the day as yet but:

Public policy needs high quality evidence of what works

No, that's not the bit which is odd:

Chief social worker Isabelle Trowler spoke at an event on 29 January, organised by the Cabinet Office, to celebrate the achievements of the 10 What Works centres across the UK. Trowler pointed out that without hard evidence about what works in the often-fraught relationship between social workers and families, this area could be at a real disadvantage when the spending cake gets cut. Trowler explained that there appears to be a different approach to youth social work, for instance, in each of England’s 152 county and unitary councils, making it almost impossible to know which is the most effective approach.

We have a rich data set and that makes it more difficult to work out what works? That's not, to put it mildly, how these things work. The more variations we've got then the easier it is to tease out of the data what is effective. Assuming that anything is, obviously.

What worries is that Ms. Trowler is one of those who actually runs the country and if this is the level of knowledge about data and information then matters are worse than we thought they were.

Just by contrast, 152 different sets of policy over a few decades would produce a better data set than we've got for the entirety of macroeconomics in rich nations. Ouch.