Comparing objective and subjective measures of household resilience

Evidence from Ethiopia

Bibliographic Data

ID22027935
AuthorsTseday Jemaneh Mekasha (0000-0001-5329-0898, Addis Ababa University), Tseday Mekasha, Alemu Mekonnen (0009-0002-2861-9828), Alemu Abibi Mekonnen (Addis Ababa University), Peter Fisker (0000-0003-4061-400X, Bispebjerg Hospital), Finn Tarp (0000-0002-6247-4370, University of Copenhagen, corresponding author), Mintewab Bezabih Ayele (North Carolina State University), Mintewab Ayele, Tagel Gebrehiwot (Ethiopian Public Health Institute)
Year2026
Volume136
Pages106071
Publication date2026-04-01
Peer ReviewedYes
Open AccessYes
TypeARTICLE
VenueInternational Journal of Disaster Risk Reduction (JOURNAL)
Journal identifiersISSN: 2212-4209
PublisherElsevier BV (PUBLISHER)
DOI10.1016/j.ijdrr.2026.106071
OpenAlexW7131656124
LanguageEN
Citations received1
References cited25

Resilience measurement has become central to targeting and evaluating development interventions, yet different measurement approaches may identify different households as vulnerable. This study compares objective and subjective measures of household resilience using data from 2000 households in rural Ethiopia. We analyze two objective measures (FAO’s RIMA and the TANGO method) alongside subjective indicators based on households’ self-assessed capacity to cope with shocks, examining both aggregate indices and disaggregated capacities (absorptive, adaptive, and transformative). Using correlation analysis, kernel density estimation, and tests of distributional equality, we find that while objective and subjective measures are positively correlated (0.305–0.606), their statistical distributions differ significantly. More than 50% of households classified in the bottom 20% by one measure are not similarly classified by others. Alignment is strongest for absorptive capacity and weakest for transformative capacity. Analysis of resilience determinants shows that while some factors (such as social networks and savings) consistently enhance resilience across measures, others exhibit varying or opposite effects. These findings highlight the need for caution when applying resilience measures for policy targeting and impact assessment, as the choice of measure can significantly affect analytical outcomes and subsequent decision making in resilience-enhancement programs. • We compare resilience measures using survey data from 2000 rural households. • Measures correlate but differ; > 50% discrepancy in identifying the least resilient. • Resilience drivers show overlaps but key differences between measurement methods. • Findings warn against using resilience measures interchangeably in policymaking. • Larger and diverse sample improves reliability and address prior study limitations

Psychological resilience · Survey data collection · Agricultural risk and resilience · Climate change impacts on agriculture · Income, Poverty, and Inequality