Methodological issues in the study of word stress correlates

The investigation of acoustic correlates of word stress is a prominent area of research. The literature is rife with studies of the acoustic exponents of what is often referred to as stress but the methodological diversity of this research has created an unclear picture of the properties robustly associated with it. The present paper explores the methodological issues involved in examining word stress correlates with the goal of proposing a set of recommendations for future research. Based on a survey of 110 (sub-) studies on 75 languages, desiderata for research on the acoustics of stress are identified: descriptions of employed methods should be as detailed as possible, speech material should be designed to allow for teasing apart word level stress from phrase level prominence, and sample sizes should be chosen according to statistical considerations