Studies that quantify speech tempo on acoustic grounds typically use one of various rate measures. Theavailability of multiple measurement techniques yields ‘researcher degrees of freedom’ which call therobustness of generalisations across studies into question. However, explicit assessments of thepossible impact of researchers’ choices among the available measures are rare. In this study we attemptsuch an assessment by comparing the distributions of five common rate measures―canonical andsurface syllable and phone rates, and CV segment rate―calculated over fluent stretches of unscriptedspeech produced by 100 English speakers. We assess the measures’ inter-correlations across the corpusas a whole as well as in relevant data samples to simulate multiple analysis scenarios. We also reporton deletion rates in our corpus, as they determine the relationship between canonical and surface rates;we assess the impact on rate figures of variable assumptions as to what constitutes deletion; and wecompare the measures’ discriminating powers in a forensic analysis context using Bayesian likelihoodratios. Our results suggest that in a sizeable English corpus with normal deletion rates, the five ratesare closely inter-correlated and have similar discriminating powers; decisions as to the segmentalmake-up of canonical forms also have limited impact on distributions. Therefore, for commonanalytical purposes and forensic applications the choice between these measures is unlikely tosubstantially affect outcomes.
|Publication status||Accepted/In press - 6 Jan 2021|