TY - JOUR

T1 - Extrapolating from experiments, confidently

AU - Khosrowi, Donal

N1 - Funding Information: Open Access funding enabled and organized by Projekt DEAL. My work on this paper was financially supported by an AHRC Northern Bridge Doctoral Studentship (grant number: AH/L503927/1), a Durham Doctoral Studentship, and a Royal Institute of Philosophy Jacobsen Studentship. Funding Information: I wish to thank Wendy Parker, Julian Reiss, Nancy Cartwright, Katherine Puddifoot, Robert Northcott, Mathias Frisch, William Peden, Joe Roussos and members of the CHESS and K4U research groups at Durham University for valuable discussions that helped shape the ideas developed in this paper. Moreover, I would like to thank the audiences at BSPS 2021 and EPSA 2021 for raising several important points that helped improve my arguments. Finally, I would like to thank two anonymous reviewers for their many very helpful comments, criticisms and suggestions, which helped significantly improve the ideas developed in this paper.

PY - 2023/3/22

Y1 - 2023/3/22

N2 - Extrapolating causal effects from experiments to novel populations is a common practice in evidence-based-policy, development economics and other social science areas. Drawing on experimental evidence of policy effectiveness, analysts aim to predict the effects of policies in new populations, which might differ importantly from experimental populations. Existing approaches made progress in articulating the sorts of similarities one needs to assume to enable such inferences. It is also recognized, however, that many of these assumptions will remain surrounded by significant uncertainty in practice. Unfortunately, the existing literature says little on how analysts may articulate and manage these uncertainties. This paper aims to make progress on these issues. First, it considers several existing ideas that bear on issues of uncertainty, elaborates the challenges they face, and extracts some useful rationales. Second, it outlines a novel approach, called the support graph approach, that builds on these rationales and allows analysts to articulate and manage uncertainty in extrapolation in a systematic and unified way.

AB - Extrapolating causal effects from experiments to novel populations is a common practice in evidence-based-policy, development economics and other social science areas. Drawing on experimental evidence of policy effectiveness, analysts aim to predict the effects of policies in new populations, which might differ importantly from experimental populations. Existing approaches made progress in articulating the sorts of similarities one needs to assume to enable such inferences. It is also recognized, however, that many of these assumptions will remain surrounded by significant uncertainty in practice. Unfortunately, the existing literature says little on how analysts may articulate and manage these uncertainties. This paper aims to make progress on these issues. First, it considers several existing ideas that bear on issues of uncertainty, elaborates the challenges they face, and extracts some useful rationales. Second, it outlines a novel approach, called the support graph approach, that builds on these rationales and allows analysts to articulate and manage uncertainty in extrapolation in a systematic and unified way.

KW - Bayesian evidence amalgamation

KW - Causal inference

KW - Confidence

KW - Economics

KW - Evidence-based policy

KW - External validity

KW - Extrapolation

KW - Uncertainty

UR - http://www.scopus.com/inward/record.url?scp=85150908531&partnerID=8YFLogxK

U2 - 10.1007/s13194-023-00520-1

DO - 10.1007/s13194-023-00520-1

M3 - Article

AN - SCOPUS:85150908531

VL - 13

JO - European Journal for Philosophy of Science

JF - European Journal for Philosophy of Science

SN - 1879-4912

IS - 2

M1 - 18

ER -