TY - JOUR

T1 - Representing human decision-making in agent-based simulation models

T2 - Agroforestry adoption in rural Rwanda

AU - Noeldeke, Beatrice

AU - Winter, Etti

AU - Ntawuhiganayo, Elisée Bahati

N1 - Funding Information: We thank the Rwandan farmers for their participation. Furthermore, we are thankful for Ronja Seegers' support and her work regarding the role-playing game. The research was conducted in the context of the project “Harnessing the potential of trees on farms for meeting national and global biodiversity targets” funded by the International Climate Initiative (IKI) (Grant number: BMUZ_1273 ).

PY - 2022/10

Y1 - 2022/10

N2 - Advancing the transition towards more sustainable agriculture requires policy interventions that support farmers' adoption of sustainable practices. Models can support policy-makers in developing and testing interventions. For these models to provide reliable support, their underlying assumptions need to reflect reality and hence adequately represent human decision-making. This study compares several approaches that represent human decision-making. The comparison is applied to farmers' decision to adopt agroforestry. An agent-based simulation model is calibrated to a case study in rural Rwanda, where socio-economic survey data was collected from 145 small-scale farmers. Of these farmers, 72 were randomly selected to participate in a role-playing game, during which the players decided about adopting agroforestry. The game was conducted to validate the tested decision-making approaches. The simulations show that the decision-making approaches predict significantly different agroforestry adoption rates. Compared with the role-playing game, the Theory of Planned Behaviour exhibits the highest validity. Rational choice theory and the econometric approach overestimate implementation. Bounded rationality approaches underestimate the share of adopters. The results highlight the importance of adequately representing farmers' adoption decisions in models for providing reliable forecasts and effective policy support.

AB - Advancing the transition towards more sustainable agriculture requires policy interventions that support farmers' adoption of sustainable practices. Models can support policy-makers in developing and testing interventions. For these models to provide reliable support, their underlying assumptions need to reflect reality and hence adequately represent human decision-making. This study compares several approaches that represent human decision-making. The comparison is applied to farmers' decision to adopt agroforestry. An agent-based simulation model is calibrated to a case study in rural Rwanda, where socio-economic survey data was collected from 145 small-scale farmers. Of these farmers, 72 were randomly selected to participate in a role-playing game, during which the players decided about adopting agroforestry. The game was conducted to validate the tested decision-making approaches. The simulations show that the decision-making approaches predict significantly different agroforestry adoption rates. Compared with the role-playing game, the Theory of Planned Behaviour exhibits the highest validity. Rational choice theory and the econometric approach overestimate implementation. Bounded rationality approaches underestimate the share of adopters. The results highlight the importance of adequately representing farmers' adoption decisions in models for providing reliable forecasts and effective policy support.

KW - Agent-based modelling

KW - Agroforestry adoption

KW - Bounded rationality

KW - Decision-making

KW - Rational choice theory

KW - Theory of planned behaviour

UR - http://www.scopus.com/inward/record.url?scp=85133164541&partnerID=8YFLogxK

U2 - 10.1016/j.ecolecon.2022.107529

DO - 10.1016/j.ecolecon.2022.107529

M3 - Article

AN - SCOPUS:85133164541

VL - 200

JO - Ecological economics

JF - Ecological economics

SN - 0921-8009

M1 - 107529

ER -