Reproduction Analysis: Reflections on Additive vs. Multiplicative Vulnerability Scores
A reanalysis of Malcomb et al. (2014) compares multi-criteria analysis results when using additive versus multiplicative climate vulnerability models. The original analysis uses an additive model while the reanalysis uses a multiplicative model. Results differed between the analyses both in the range of scores and in the number of households classified as highly vulnerable. For example, visually inspecting the spatial distribution of both scores indicates that the additive vulnerability score documents more regions with moderate to high vulnerability. Meanwhile, the multiplicative vulnerability score is more selective in classifying regions as highly vulnerable due to the larger range of possible scores. These findings indicate potential sources of uncertainty in the results of the study due to researcher subjectivity in model conceptualization, parameterization, and measurement of the complexity of the real world.
Previous studies have found that generating an additive vulnerability score is effective when estimating the impact of adaptive capacity interventions, however the multiplicative index is more robust when exposure and sensitivity metrics are priorities as is the case in the Malcomb (2014) study. Furthermore, the methodologies used to assess risk and vulnerability, particularly how a model is conceptualized, critically impact the outcome of results. Findings from the reanalysis have significant implications when considering the validity of results and for the prioritization of development projects. In conclusion, comparing results from additive and multiplicative vulnerability methods can assist in identifying the most vulnerable regions to climate change where climate adaptation projects should be prioritized.