Evaluation of keyness metrics: performance and reliability

Abstract The methodological debates surrounding keyword analysis have given rise to a wide range of keyness metrics. The present paper delineates four dimensions of keyness, which distinguish between frequency- and dispersion-related perspectives. Existing measures are then organized according to these dimensions and evaluated with regard to their performance on a specific keyword analysis task: The identification of key verbs in academic writing. To this end, the rankings produced by 32 different metrics are evaluated against an established academic word list. Further, the reliability of measures is assessed, to determine whether they produce stable rankings across repeated studies on the same pair of text varieties. We observe notable differences among metrics with regard to these criteria. Our findings provide further support for the superiority of the Wilcoxon rank sum test and text-dispersion–based measures, and allow us to identify, within each dimension of keyness, metrics that may be given preference in applied work.