Thought about scientific consensus recently? We have a new DP @I4Replication that probes into the famous replication debate between Acemoglu, Johnson & Robinson (AJR) and Albouy - and how experts assess this debate. We find that they disagree. 1/8 econstor.eu/bitstream/1041…
We did a survey among experts. How do we define "expert"? We did not mass-email or post the survey on social media. Instead, we emailed it exclusively to people who cite one of the debate papers or methodologically similar papers. It seems we successfully recruited experts: 2/8
Respondents are also fairly familiar with the debate papers, that is, AJR’s original paper from 2001, as well as Albouy’s comment and AJR’s reply published simultaneously in 2012. 3/8
Our short survey provided a concise and neutral summary of the debate papers. Following this summary, we asked respondents how convincing they found the respective arguments. Some preferred AJR's arguments, others preferred Albouy's. 4/8
A specific gem for AJR aficionados, if you ask me, is the huge difference in how experts judge AJR's *general* theoretical claim (institutions matter for growth) vs their *specific* theoretical claim (settler mortality matters for institutions, which matter for growth). 5/8
Back to the debate. The final and main question of our survey was simply whether respondents, overall, agree with AJR or with Albouy. There is a slight tendency towards Albouy, but a significant proportion of experts is in the pro-AJR camp. We interpret this as no consensus. 6/8
So, what does this mean? Nothing bad in any case if your account of science is a very deliberative one. Yet the more you believe in an authoritative role of science – which most economists probably do – the more troubling the lack of consensus is. 7/8
Our paper goes straight to economics' self-identification as a Popperian discipline. It is hard to maintain that ideal if (non-confirmatory) robustness replications always end up in unresolved debates like this one – which we believe they mostly do. 8/8
• • •
Missing some Tweet in this thread? You can try to
force a refresh
