There is a great temptation to reduce complexity by combining a range of indices into a single measurement. The Human Influence Index (Sanderson et al. 2002) is an example. Strictly speaking, doing this by summation or averaging makes mathematical sense only if the individual indices have the same units and measurement scales. Many social and economic indicators (such as the Human Development Index, UNDP 2003) are of such a multifactorial nature. Sometimes the problem of incommensurability is addressed by normalization of all the numbers to some benchmark or weighting of the various components (in which case the weightings are implicitly unit conversion factors). Axis scores in principal component analysis (or similar techniques) fall into this category. Multiplicative indices may make more mathematical sense but should be based on a sound conceptual model, or else they will make no ecological sense.
Given the complexity and variety of biodiversity in and between ecosystems, composite indicators probably are one of the few ways in which information can be made digestible and understandable for politicians and the public (CBD 2003a, 2003c). However, such indicators must be transparent about the way the various factors are transformed and combined. It must be possible to delve down into the individual components that make up the index and disaggregate them spatially.
Was this article helpful?