Reframing climate change: how carbon reduction can also reduce poverty and inequality

January 30, 2017

The WDR 2017 on Governance and Law: Can it drive a transformation in development practice?

January 30, 2017

Links I Liked

January 30, 2017
empty image
empty image

Fantastic news. John Ambler, one of the wisest heads in development, has written his memoirs/life lessons, free to Ambler coverdownload

Simon Maxwell has written one of the most comprehensive reviews to date of How Change Happens, striking a nice balance between liking it and finding loads of gaps.

RCTs (Randomised Controlled Trials) came under scrutiny (again). First a study in the BMJ showed that large effects in small RCTs are rarely followed with larger RCTs to double check, but when they are, 43% fail to find an effect.

Then some CGD researchers took exception to Action Aid and Education International commissioning counter-research to their RCT on Charter Schools in Liberia. Are RCTs neutral and objective (so AA & EI are indulging in a ‘blanket dismissal of evidence’), or inherently flawed and/or biased (so some counterbalancing qualitative research is entirely sensible, and at $30k, a lot cheaper than an RCT)? Discuss.

And two comedic highlights of a pretty awful week.

A bad lip reading of the Presidential inauguration. Priceless

Not only are the Dutch setting up an international safe abortion fund to plug the $600m funding gap caused by reinstatement of the US “global gag rule” but they also produced the funniest spoof of the week. The Netherlands explains itself to the new US pres, in his own style. 14m hits and counting. Respect.

 

 

1 comment

  1. A couple of things:

    “First a study in the BMJ showed that large effects in small RCTs are rarely followed with larger RCTs to double check, but when they are, 43% fail to find an effect.”

    (1) These are medical trials, not clear how it maps onto development ones (could be worse, could be better).
    (2) At least in socioeconomic work, replicability is not solely an RCT problem – it’s a problem with all quantitative (and, let’s be honest, quantitative) work. So the thought that needs to be added to the above study is “compared to what?”

    There is also a scrutiny bias to these kind of results. RCTs get more attention, so we find more flaws in them. Who is out there trying to systematically reproduce qualitative work?

    “Are RCTs neutral and objective (so AA & EI are indulging in a ‘blanket dismissal of evidence’), or inherently flawed and/or biased (so some counterbalancing qualitative research is entirely sensible, and at $30k, a lot cheaper than an RCT)? Discuss.”

    The thing you have omitted here is that Action Aid is commissioning research with a foregone conclusion already in place. There’s no question that this particular piece of qualitative evidence is inherently flawed. It is always a good idea to add in some qualitative evidence to the debate, but that evidence needs to be unbiased in its design and intention.

Leave a comment

Translate »