The overlooked effect of amplitude on within-speaker vowel variation

Code and analysis repository
NZILBB Vowel Clustering Project

Overview

This repository contains analysis and code for the paper "The overlooked effect of amplitude on within-speaker vowel variation". The bulk of this repository is made up of four R markdown files corresponding to each stage of the analysis in the paper and found in the supplementary_material folder along with knitted html versions which can be read without rerunning the code yourself.

Directory structure

• supplementary material: contains the primary R markdown files running through the four stages of our analysis:
  1. preprocessing (hosted here)
  2. interval representation (hosted here),
  3. corpus level PCA analysis (hosted here), and
  4. GAMM models (hosted here).
• labbcat_data: Folder for anonymised raw data from LaBB-CAT. Empty directory. To get data go to https://osf.io/m8nkh/ and download the file merged_data.rds and place it in this directory.
• scripts: Contains R scripts for running permutation tests referred to in the supplementary materials.
• processed_data: Data output at various stages of the analysis by the scripts of supplementary materials.
• plots: Plots output by scripts or supplementaries.
• models: Folder for models reported in paper. Empty directory. To get data go to https://osf.io/m8nkh/ and download the files in the models directory.

How to use this repository

The above files are sufficient to repeat the analysis we have carried out. Some models take a long time to fit. You can download our models here. Place these in the project directory in a directory named models.

It is best to open this repository as an Rproject. One way to do this in RStudio is to go to 'File -> New project', select 'Existing directory', and navigate to the downloaded repository. wherever