~/teaching
teaching
MA-level courses for students of Linguistics, Speech & Language Processing and Multilingualism at Universität Konstanz.
summer semester 2027
Advanced Statistics
A follow-up to Statistics. It starts with a refresher on linear and mixed models, continues with honest, publication-ready figures in ggplot2, and covers generalised additive (mixed) models for curves such as formant trajectories and pupil dilation, with simulated data and a hands-on session. Two sessions on random forests, one theoretical and one hands-on, follow, then a session on power analysis by simulation, with simr, and on why post hoc power is uninformative. Later sessions will cover factor analysis and cluster analysis. Each deck is a self-contained HTML file: click a lecture to open the slides.
- 01–05 Models, figures and curves
- 06–10 Random forests, power and multivariate methods
- 01 Statistics refresher Linear models, interactions and mixed models
- 02 Figures for publication Honest, accessible and ready for print
- 03 Generalised additive models Modelling curves with smooth terms
- 04 Generalised additive mixed models Random smooths, autocorrelation and non-linear interactions
- 05 GAMMs in practice A guided analysis of pupil curves
- 06 Random forests Trees, forests and variable importance
- 07 Random forests in practice A guided analysis of accent features
- 08 Power analysis Planning sample sizes by simulation
-
09 Factor analysis Latent variables behind many measures in preparation
-
10 Cluster analysis Finding groups in data in preparation
winter semester 2027/28
Introduction to Forensic Speech Science
An introduction to forensic speech science: speaker profiling and speaker comparison, earwitness evidence, the authentication and enhancement of recordings, transcription, language analysis in the asylum procedure, language in police interviews and courtrooms, authorship analysis, and the work of expert witnesses. The sessions follow the chapters of Leemann, Perkins, Sullivan Buker & Foulkes (2024), An Introduction to Forensic Phonetics and Forensic Linguistics, with cases from the book and simulated data. The last two sessions are Praat practicals with the students’ own recordings. Click a session to open its slides.
- 01 Course introduction Forensic phonetics, forensic linguistics, and this course
- 02 Phonetic analysis Segments, prosody and acoustic measurement
- 03 Speaker profiling What a voice can and cannot tell us
- 04 Speaker comparison Methods, evidence and conclusions
- 05 Earwitness evidence Voice recognition and voice parades
- 06 Authentication, enhancement and speech content Is the recording genuine, and what was said?
- 07 Linguistic analysis in the asylum procedure Where did this speaker grow up?
- 08 Language in the legal process From the police caution to the courtroom
- 09 Authorship analysis Profiling and comparing the writers of texts
- 10 Expert witnesses Reports, courts and legal systems
- 11 Praat practical 1 Measuring your own voice
- 12 Praat practical 2 Casework with the class recordings
winter semester 2026/27
Statistics
An introduction to data analysis in R, from the tidyverse and descriptive statistics to linear regression, interactions, significance testing, mixed models and mixed-effects logistic regression. The lectures follow the structure of Winter (2019), Statistics for Linguists: An Introduction Using R, with new examples and data sets. Each deck is a self-contained HTML file: click a lecture to open the slides.
- 01–03 R, data and descriptive statistics
- 04–08 The linear model
- 09–13 Inference and mixed models
- 01 Course introduction Organisation, R and RStudio, and first steps in R
- 02 Data wrangling Reading and wrangling data with the {tidyverse}
- 03 Descriptive statistics Describing distributions: centre and spread
- 04 Linear models Intercepts, slopes and residuals
- 05 Transformations Centering, standardising, correlation and logarithms
- 06 Multiple regression More than one continuous predictor
- 07 Categorical predictors Treatment coding, sum coding and factors
- 08 Interactions When the effect of one predictor depends on another
- 09 Significance testing Effect sizes, standard errors, confidence intervals and p-values
- 10 Mixed models 1 A conceptual introduction
- 11 Mixed models 2 Fitting, testing and troubleshooting
-
12
Mixed logistic regression
Binary outcomes with
glmer() - 13 Power and model building p-values, power, uncertainty and building models
Taught courses
- Statistics (2021+, every semester): R, linear regression, linear mixed models, logistic regression, analysis of experimental data.
- Advanced Statistics (2021+): non-linear modelling (GAMMs), cluster analysis, factor analysis, random forests.
- Pupillometry (2026, with Yining Yan): hands-on pupillometry for cognitive and language-related tasks.
- Introduction to Forensic Speech Science (2023, 2025): phonetic analysis for speaker comparison and speaker profiling.
- Eye-tracking and Pupillometry (2024, with Dr Ilaria Venagli): both methods applied to the study of language, with a practical focus.
Course evaluations
The chart shows the Lehr-Lern-Index of my Statistics course; the window shows a selection of positive comments from the evaluations of my courses and workshops, quoted verbatim, one at a time.
Means from the evasys reports, 1 = trifft zu (best) to 5 = trifft nicht zu. Error bars are ±1 SE; n is the number of questionnaires. ★ Mittelbau Teaching Award 2024.
show as table
| Semester | n | LLI | SD |
|---|---|---|---|
| WS 2023/24 | 6 | 1.83 | 1.28 |
| SoSe 2024 ★ | 19 | 1.22 | 0.38 |
| WS 2024/25 | 16 | 1.44 | 0.66 |
| SoSe 2025 | 13 | 2.06 | 1.06 |
| WS 2025/26 | 12 | 1.35 | 0.55 |
| SoSe 2026 | 9 | 1.40 | 0.58 |
Translation:
Statistics, 2022