A meta-analysis of vowel formants in Turkish: Methodological differences and their impacts
Synopsis
Although research on vowel formants in Turkish constitutes a substantial body of work, there is a lack of methodological consensus among these studies. This study presents a meta-analysis examining methodological differences in research on vowel formants in Standard Turkish and the effects of these differences on formant values. Within the scope of this study, 16 studies on Standard Turkish were adopted, and the methodological criteria of these studies were coded in a data table to create the dataset. In the data analysis, separate mixed-effects models were fitted for each vowel in the R environment. The results indicated that random effects (e.g., study and study-vowel interaction) in the models were statistically more determinant in explaining variability compared to fixed effects (e.g., gender, material type, number of speakers, recording environment, recording equipment, sampling rate, and normalisation method). This finding indicates that variability in formant values largely stems from implicit methodological choices that are neither clearly reported nor standardised across studies. The study emphasises the necessity of methodological transparency, open data sharing, and standardised measurement protocols to ensure the reliability and comparability of phonetic descriptions in Turkish.
Keywords: Vowel formants, Phonetic description, Meta-analysis, Methodology
Pages
Published
License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.