Resilient, but from whose perspective?: like-for-like comparisons of objective and subjective measures of resilience

As resilience continues its rise to top of the international policy agenda, development funders and practitioners are under mounting pressure to ensure that investments in resilience-building are effective and targeted at those most in need. It is here that robust resilience measurement can make valuable contributions: identifying hotspots; understanding drivers; and inferring impact. To date, resilience measurement has been dominated by objectively-oriented approaches. These rely on external definitions of resilience (often informed by outside ‘experts’, literature reviews or resilience practitioners) and measured through observation or external verification. More recently, the potential for subjective approaches has been proposed. These take a contrasting approach, soliciting people’s judgements of what resilience means to them, and getting them to self-evaluate their own resilience. While both approaches have their strength and weaknesses, little is known about how objective and subjective modes of resilience measurement compare. To shed light on this relationship, we provide likefor- like comparisons of these two approaches using a regionally representative household survey of 2,308 households in Northern Uganda. In so doing, we introduce a new measurement approach named the Subjective self-Evaluated Resilience Score (SERS). Outcomes from SERS are directly compared with an objectively-evaluated approach, the Resilience Index Measurement Analysis (RIMA), widely used by resilience practitioners. Findings from the survey suggest a moderate correlation between objectively- and subjectively-evaluated resilience modules. More importantly, both approaches share similar associations with many key socioeconomic drivers of resilience. However, there are notable differences between the two. In some case, the approaches differ entirely regarding contributions of important traits, including coping strategies, levels of education and exposure to prior shocks. Our results highlight the need for resilience evaluators to consider a diversity of knowledge sources and seek greater use of evidence in indicator selection. We also investigate the properties of the SERS module itself. We find that characterisations of resilience that mimic various commonly-used frameworks produce similar resilience outcomes, suggesting that debates over the exact composition of resilience-characteristics may matter little. In addition, shorter SERS modules match the performance of the full set of SERS questions, allowing for quicker administration and reduced survey burden. Lastly, we call for evaluators to consider the strengths and weaknesses of subjective and objective measurement approaches, including options for combining both formats.