Navigation auf uzh.ch

Suche

LiRI - Linguistic Research Infrastructure

Workshops & Guest Lectures

Coming Up:

Grid containing content elements

Meet the Families, Part 3: Count Data

In this workshop, we learn how to analyse count data in a generalised linear modelling framework (GLM). We  discuss the basics of count GLMs, how to fit the model, how to visualise and interpret your model output, as well as how to check model assumptions. We further talk about common issues with count data and how to deal with them, such as exposure variables, overdispersion, and zero-inflation.

Date: Wednesday, April 17, 2024, 14:15-16:00
Location: UZH, Andreastrasse 15, 8050 Zurich, room will be communicated later
Speaker: Andri Manser
Please register by April 12, 2024 toSandra Schwab

 

Setting appropriate contrasts in linear models (linear regression, mixed-effects models)

Every time a linear model is used to study the impact of a categorical predictor (or factor) on a dependent variable, it is necessary to transform this categorical predictor into numbers. R (or any other statistical software) cannot do math with words. In other words, we need to set the contrasts.  Setting the contrasts right will ensure that the model answers the right questions. In this workshop, I will introduce and illustrate different ways to set contrasts and show how contrasts are directly related to research hypotheses.  

Date: Wednesday, April 24, 2024, 14:15-16:00
Location: Zoom
Speaker: Audrey Bürki
Please register by April 19, 2024 to Sandra Schwab.

Meet the Families, Part 4: Ordinal Data

Ordinal data are common in linguistic research (for example in questionnaires). In this workshop, we l learn how to analyse ordinal data using ordinal regression. We discuss the basics of ordinal regression method, how to fit an ordinal model, how to visualise and interpret your model output. We further discuss an approach to check model assumptions (using surrogate residuals) as well as how to deal with assumptions violations (e.g. proportional odds).

Date: Wednesday, May 15, 2024, 14:15-16:00
Location: UZH, Andreastrasse 15, 8050 Zurich, room will be communicated later
Speaker: Andri Manser
Please register by May 10, 2024 to Sandra Schwab.

Convergence issues with mixed-effects models in lmer

Mixed-effects models are often used to analyze linguistic data. These models are complex, with many parameters and therefore require quite a lot of data. It may happen (and happens quite often) that the, the statistical model which would be appropriate given the design is too complex for the data at hand. As a consequence, some parameters cannot be estimated (properly).  In such cases, and when these models are run in lmer, R throws a convergence warning. In this workshop we will discuss these warnings, their meaning and possible consequences. We will discuss possible strategies. 

Date: Wednesday, May 29, 2024, 14:15-16:00
Location: Zoom
Speaker: Audrey Bürki
Please register by May 24, 2024 to Sandra Schwab.

Meet the Families, Part 5: Proportion Data

In this workshop, we  learn how to analyse proportion data using logistic regression. We discuss the basics of binomial GLMs, how to fit the model, how to visualise and interpret your model output, as well as how to check model assumptions. We further talk about common issues with count data such as overdispersion, zero-inflation, and complete separation.

Date: Wednesday, June 19, 2024, 14:15-16:00
Location: UZH, Andreastrasse 15, 8050 Zurich, 4th floor, room AND 4.55/57
Speaker: Andri Manser
Please register by June 14, 2024 to Sandra Schwab.

Past Workshops and Guest Lectures

Further information regarding workshops, guest lectures, and trainings we have offered in the past, including the most recent ones listed below, can be found in our archive.

Weiterführende Informationen

Meet the families: Beta distribution, March 20, 2023, 14:15-16:00

Proportion or percentage data is ubiquitous in linguistic research. In this workshop, we will learn about the different types of proportion data, and then introduce beta regression models as a way of analysing (continuous) proportion data. We will discuss how to interpret your model output, check model assumptions, and visualise model predictions. The workshop will have a practical part at the end where you will learn to implement a beta model first hand.

Date: Wednesday, March 20, 2023, 14:15-16:00

Location: UZH, Andreastrasse 15, 8050 Zurich, 4th floor, room AND 4.55

Speaker: Andri Manser

Please register by March 15, 2023 to Sandra.Schwab@uzh.ch

flyer bild

March 13, 2024, 14:15-16:00 on Zoom

Online Workshop "Rethinking statistics: Bayesian approach"

The aim of this course is to provide participants with a basic understanding of Bayes Rule and of the reasoning behind the Bayesian approach to statistics. For this, we will need to review a few important concepts, including variables, distributions, probability, and parameters of statistical models. I will present examples of analyses performed in the Bayesian context. This workshop is an introduction to the Bayesian approach. Participants will not learn to perform Bayesian analyses on their own.

Date: Wednesday, March 13, 2024, 14:15-16:00

Location: Zoom (link will be provided after registration)

Speaker: Prof. Audrey Bürki

Please register by March 8, 2024 to Sandra.Schwab@uzh.ch

flyer bild

Meet the families: finding the distribution to model your data, 13 December, 14:15-16:00

Not everything in life is normal(ly distributed). Finding a distribution that adequately represents your outcome variable is a (the?) key challenge in our quest for an appropriate statistical model. In this little workshop, we will discuss some of the data types typically encountered in linguistic research, introduce statistical distributions that are commonly used for them, as well as briefly discuss how to implement, check and address typical challenges with distributional choices.

Date: Wednesday, December 13, 2023, 14:15-16:00
Location: Andreastrasse 15, 8050 Zurich, 3rd floor, room AND 3-46
Speaker: Andri Manser

Please register by December 8, 2023 to Sandra.Schwab@uzh.ch.