I want to measure correlations between human ratings of second language speech and a number of quantitative measures in the speech, such as articulation rate. I have collected speech rating data for this: 12 listeners have rated 45 speech samples (using a 9-point scale) in a fully crossed design. In order to run a regression analysis of the listeners’ ratings and the quantitative measures, I want to collapse the 12 ratings for each speech sample. Rather than doing a simple average of the ratings for each sample, is there a straightforward way of correcting the averages for individual rater variables (e.g. severity vs leniency; tendency to use the whole 9-point scale vs using only the middle part, etc.)? What would be the best package to use in R?
Any ideas much appreciated, thanks!
- Badges
- Science topic
- Similar topics
- Linguistics
- Composition Studies
- Publication