Classification of species sensitivity for biomonitoring has been approached under two different frameworks, using either empirical data or expert opinion. Two tools for fine sediment (i.e. clogging and colmation) biomonitoring in the UK tend towards these contrasting approaches. The Proportion of Sediment‐sensitive Invertebrates (PSI) index was developed using expert judgement. Empirical weightings were subsequently added at genus or species (EPSI) and mixed (EPSImixed) taxonomic levels, but scores remain constrained by the original categories. In contrast, the Combined Fine Sediment Index, composed of separate taxon scores along organic matter and total fine sediment gradients, was developed using a purely empirical approach. We tested the mechanistic bases for these indices by relating taxon scores to species traits. We compared the results with those for the well‐established Whalley Hawkes Paisley Trigg index of organic pollution. After controlling for varying sample sizes, Whalley Hawkes Paisley Trigg could be better predicted by a linear combination of all available traits (mean R2 = 0.92) than any of the fine sediment indices (0.68 < mean R2 < 0.76). When only traits expected to respond to fine sediment were offered as independent variables, the goodness of fit was substantially reduced for all fine sediment indices (0.27 < mean R2 < 0.46). Our findings demonstrate the lack of integration between the literature on macroinvertebrate responses to fine sediment, the available trait data and taxon scores. Refinement of the trait database is recommended to build on the valuable work performed to date. As the UK has taken the lead in embedding fine sediment into routine biomonitoring programmes, these findings have important international implications. Copyright © 2017 John Wiley & Sons, Ltd.