Twitter | Search | |
Arun Agrawal
Long thread... Contributions to inaugural World Development Symposium-Experimental Approaches in Development & Poverty Alleviation are now appearing online. Fastest SI ever? 🍾🍾🍾Check them out here: 1/n
Reply Retweet Like More
Arun Agrawal Dec 16
TY, Tony Bebbington, Cathy Boone H. Mcgowan, C. Watkins, J-P Platteau, ; TY also for inspiring the idea to the Nobelists: Banerjee, Duflo, & especially Kremer who also serves on the WD editorial board. Excellent taste, Michael!
Reply Retweet Like
Arun Agrawal Dec 16
All three Nobelists exemplify not only the commitment to rigorous and path-breaking work, but also to long-term, place-based engagement and support for their colleagues in the field and of course for improving lives of those living in poverty and deprivation 3/n
Reply Retweet Like
Arun Agrawal Dec 16
The authorial composition of the Symposium reflects our editors' commitment to diversity – representing differences in gender, setting (higher+lower income), profession (scholars and practitioners), seniority, & reputation. I am proud of this diversity too; 4/n
Reply Retweet Like
Arun Agrawal Dec 16
RCTs have changed development research; like many influential ideas they also face criticisms; detractors & proponents often speak past each other. Symposium showcases these themes; also the substantial promise and the road ahead for RCTs to become “normal” in dev. studies 5/n
Reply Retweet Like
Arun Agrawal Dec 16
To realize promise of RCTs, critical to collaborate across methods & disciplines; overcome persistent hurdles between scholarship & practice; impact on most pressing challenges of sustainability and development needs enormous additional development of RCTs; 6/n
Reply Retweet Like
Arun Agrawal Dec 16
The best users of RCTs are well aware of common criticisms around challenges of scale, short term assessments, questions about internal and external validity, costs, ethics, and at times technocratic orientation; Broadly, RCT research trying & moving beyond these limitations; 7/n
Reply Retweet Like
Arun Agrawal Dec 16
Many contributors focus on future; but also ask what made this latest episode of RCTs so influential when similar earlier efforts failed at least twice to gain traction-WHY? see Leao and Eyal – but also Levitt and List 2009 (authors without year ref are in symposium); 8/n
Reply Retweet Like
Arun Agrawal Dec 16
Kapur reflects on changing scholarly emphases in development over past decades; asks if current attention to RCTs has similar half life of decade to two; Cox considers whether advocates and critics of RCTs are missing productive opportunities because of in-group commitments; 9/n
Reply Retweet Like
Arun Agrawal Dec 16
Drèze discusses importance of understanding, value judgements, & deliberation for sound policy. These go beyond straightforward formulation of evidence-based policy making; Avdeenko & Frölich examine development RCTs against experimentation in medicine; also (Favereau 2016); 10/n
Reply Retweet Like
Arun Agrawal Dec 16
Ravallion has balanced+informative account of main virtues/ limitations of RCTs; Barrett & Carter update 2010 review on importance of RCTs for development but caution readers re ethical risks, high heterogeneity in intervention impacts, and non-classical measurements errors; 11/n
Reply Retweet Like
Arun Agrawal Dec 16
McKenzie’s is an elegant response to the question of why care about experiments when it is not RCTs that led to the most prominent development success today or historically – think China, UK, US, Europe; 12/n
Reply Retweet Like
Arun Agrawal Dec 16
Glewwe similarly discusses the relationship between theory and experiments with a focus on education - and how they may inform each other. Much ink on this also spilt by other Econ Nobelists such as Heckman and Deaton – alas not in the SI; 13/n
Reply Retweet Like
Arun Agrawal Dec 16
Muller acknowledges positive contributions of RCTs, while questioning extrapolation of RCT research results-highights role of machine learning; but Cole and coauthors more skeptical about ML; instead putting forward digital delivery and closed loop evaluations; 14/n
Reply Retweet Like
Arun Agrawal Dec 16
Baird+co-authors focus on RCT policy relevance with reference to school-based deworming program intervention by Miguel and Kremer (2004); Banerji and Chavan examine policy impacts of education RCTs in India; these examples demonstrate long-term engagement of BDK; 15/n
Reply Retweet Like
Arun Agrawal Dec 16
Rao discussed the paradoxical risk to development policies when specific technical demands drive research; critique by Kvangraven hone in on assumptions about individual utility-maximizing behavior, point to need to account for context & human agency; See also Zaveri; 16/n
Reply Retweet Like
Arun Agrawal Dec 16
Pieters+Klasen, de Haan & co-authors focus on women’s economic empowerment; highlight lessons from RCTs for achieving gender equality, especially thru microcredit, business services, and education; focus on how to marry strengths of RCTs with alternative designs and analyses 17/n
Reply Retweet Like
Arun Agrawal Dec 16
Akram-Lodhi shows how influence of rural property relations and asset inequalities hard to capture in RCT designs; Achi suggests failure to examine these factors often makes for incremental rather than more profound changes where RCTs are concerned; 18/n
Reply Retweet Like
Arun Agrawal Dec 16
Bulte+co-authors highlight concerns about INTERNAL validity in RCTs when outcomes hinge on changes in behavioral responses - using the case of seed distribution in Africa; Other authors also focus on implementation; 19/n
Reply Retweet Like
Arun Agrawal Dec 16
Eg. Dillon+co-authors talk about data quality+need for investments in higher quality data from RCTs; similarly, Parker+co-authors’ example of large-scale RCT with multistakeholder fora highlights adjustments in experiment around data, impact measurement & collaboration; 20/n
Reply Retweet Like
Arun Agrawal Dec 16
Pieterse cautions as well around implementation, including design failures even before data collection occurs; she and Kabeer both highlight role of collaborative qualitative work; Kabeer points to “RCT misbehavior” 21/n
Reply Retweet Like
Arun Agrawal Dec 16
Surendran & Kumar counter the view that RCTs have brought empirical work back into economics; 22/n
Reply Retweet Like
Arun Agrawal Dec 16
Hoffmann’s essay raises insistently the ethical issues flagged by Barrett and Carter; she uses dataset on large number of published RCTs & points to their lack of discussion of informed consent and challenges in obtaining informed consent more broadly; 23/n
Reply Retweet Like
Arun Agrawal Dec 16
The long history of power asymmetries between researchers and research subjects highlights the critical need for much greater attention by RCT researchers to informed consent; Kaplan+co-authors also highlight ethical issues from PoV of research staff; 24/n
Reply Retweet Like
Arun Agrawal Dec 16
Stevano argues that small questions are worth asking if the answers are big! Humphreys & Scacco suggest how evidence from experimental findings can lead to inferences relevant to larger theoretical questions; 25/n
Reply Retweet Like
Arun Agrawal Dec 16
Other contributors also call for explicit cross-fertilization and collaboration across the RCT-observational divide. Gisselquist believes single experiments are “single case studies” – and highlights linking of experiments to theory as well; 26/n
Reply Retweet Like
Arun Agrawal Dec 16
Williams suggests mechanism mapping as a practical tool for policy makers to integrate experimental evidence from other locations with information about their own local contexts (see also Muralidharan and Niehaus 2017; Acemoglu 2010, Card et al. 2011, Rodrik 2008); 27/n
Reply Retweet Like
Arun Agrawal Dec 16
Das reflects on the political process of policymaking; Davis & Mobarak examine complexities in moving from RCT results to anti-poverty policy. Krause & Licona narrate an example of how evidence-based approaches influenced dev. policies in Mexico; 28/n
Reply Retweet Like
Arun Agrawal Dec 16
The Transfer Project contributions similarly highlight need for close engagement with decn makers for RCTs (and indeed other research designs) to be effective in policy change; the same concern motivates Escobal & Ponce for Peru; 29/n
Reply Retweet Like
Arun Agrawal Dec 16
Naritomi+co-authors discuss how RCTs create opportunities for collaboration that do not happen with other research methods; Samii highlights benefits of experiments around more deeply engaged learning about policy formulation and implementation; 30/n
Reply Retweet Like
Arun Agrawal Dec 16
In contrast, both Kabeer and Akram-Lodhi suggest RCT-teams often do not collaborate with qualitative researchers, suggesting that there is still work to be done to foster more cooperative cultures of research and evaluation; 31/n
Reply Retweet Like