Recovering information from synthetic air quality indices

Synthetic. indices are often used to condense complex situations into a single figure. However, this condensing process risks losing potentially useful information, especially when the index is to be milised by public decision-making bodies. The present study proposes a general strategy, combining a number of different methods, designed to recover information from air-quality indices: graphical methods to reconstruct the composition of pollution, multinornial logit analysis to study the influence of meteorological covariates on air-quality indices, and finally, a probability distribution for the index itself as a basic tool with which to interpret the index's crucial values. Copyright (c) 2007 John Wiley & Sons, Ltd.