Dienstag, 28.04.2026, 12-13 Uhr in W9-109
Prof. Dr. Antonia Zapf
Institut für Medizinische Biometrie und Epidemiologie, Hamburg
Estimands and missing values in diagnostic accuracy studies
In 2020, the Estimand Framework was introduced in the appendix to the ICH E9 guideline for therapeutic trials. An estimand precisely defines the research question of interest, including the target treatment effect and strategies for handling intercurrent events that affect the interpretation or existence of measurements. Missing values represent a common challenge in clinical research, and multiple imputation is a frequently applied method to address this issue (Schaefer et al., 2026). However, when moving from therapeutic trials to diagnostic accuracy studies - where the objective is to assess the performance of a diagnostic test - statistical methodology remains less well developed. Recent work by Stahlmann et al. (2023) reviewed existing approaches for handling missing values in diagnostic studies and Stahlmann et al. (2025) and Juljugin et al. (2026) evaluated them in simulation studies. In addition, Fierenz et al. (2025) proposed an Estimand Framework tailored to diagnostic accuracy studies. Building on this work, the Estimand Framework has recently been evaluated in a simulation study in combination with different methods for addressing missing values. In this talk, I will introduce the concept of estimands in diagnostic accuracy studies and discuss approaches for handling missing values. Their practical relevance will be illustrated using an example study (Leutkens et al., 2025), and initial results from a simulation study will be presented (Stahlmann et al. 2026).
Dienstag, 12.05.2026, 12-13 Uhr in W9-109
Kurtuluş Kıdık
Universität Bielefeld
Rethinking Global VARs: Estimating Cross-Country Linkages with Matrix Autoregressions
This talk develops a unified perspective on two prominent econometric frameworks for analyzing matrix time series (MaTS): the global vector autoregressive (GVAR) model and the matrix autoregressive (MAR) model. We provide a systematic methodological comparison, highlighting their common foundations as restricted vector autoregressive (VAR) systems while clarifying key differences in structure, interpretation, and estimation. A key feature of GVAR models is the use of a pre-specified weight matrix to capture cross-country linkages, typically based on observed economic networks such as bilateral trade flows or financial linkages. While intuitive, this choice is crucial: different specifications can lead to very different empirical conclusions, and incorrect weights may undermine the model’s assumptions. We show that MAR models with multiple terms provide a natural and flexible framework for estimating such interaction structures directly from the data, while retaining a parsimonious representation relative to an unrestricted VAR. This framework also enables formal testing of competing assumptions about the underlying network of interdependencies. Finally, we characterize the restrictions under which GVAR and MAR models are observationally equivalent, and embed both within a common VAR representation with structured coefficient matrices. This unification not only clarifies the relationship between the two approaches but also yields practical diagnostic tools for assessing and improving GVAR specifications.
Dienstag, 26.05.2026, 12-13 Uhr in W9-109
Prof. Dr. Göran Kauermann
Institut für Statistik der Ludwig-Maximilians-Universität München
Deriving multiple Information from pure Count Data – The Skellam Approach
We consider the scenario where we have count data at observational units, but we are interested in the quantities leading to the counts. Examples include the filling of bikes in racks of a bike network, but we are interested in the trips between stations of the network. Another example relates to the COVID-19 pandemic, where we observed the number of occupied beds in ICUs, but we are really interested in the number of incoming and outgoing patients. The talk demonstrates how the problem can be solved by relying on the Skellam distribution, which allows us to infer the number of incoming and outgoing patients from the occupancy in the ICUs. The talks goes a step beyond and approaches the additional question of whether we can also estimate the average length of stay of ICU patients. Hence, the task is to derive not only the number of incoming and outgoing patients from total net counts, but also to gain information on the duration time of patients on ICUs. We make use of a stochastic Expectation-Maximisation algorithm and additionally include exogenous information which are assumed to explain the intensity of inflow.
Dienstag, 09.06.2026, 12-13 Uhr in W9-109
Maya Vienken
Universität Bielefeld
Inference on state occupancy in covariate-driven hidden Markov models
Hidden Markov models (HMMs) are natural and popular tools for analysing animal behaviour based on movement, acceleration and other sensor data. In particular, these models make it possible to infer how the animal's decision-making process interacts with internal and external drivers, by relating the probabilities of switching between distinct behavioural states to covariates. A key challenge arising in the statistical analysis of movement behavioural data using covariate-driven HMMs is the models' interpretation, especially when there are more than two states, as then many functional relationships between state-switching probabilities and covariates need to be jointly interpreted. The model-implied probabilities of occupying the different states as a function of the covariate of interest, also known as the stationary state probabilities, constitute a more concise and hence useful summary statistic for drawing ecological inferences. A pragmatic approximation of the state occupancy distribution, namely the hypothetical stationary distribution of the model's underlying Markov chain for fixed covariate values, has in fact routinely been reported in HMM-based analyses of ecological data. However, for stochastically varying covariate processes with relatively little persistence, we show that this approximation can be severely biased, hence potentially invalidating ecological inference based on the approximate version of this important summary statistic of interest. In this contribution, we develop three alternative approaches for obtaining the state occupancy distribution as a function of a covariate of interest - two based on resampling of the covariate process and the third obtained by regression analysis of the empirical state probabilities. The practical application of these approaches is demonstrated in simulations and a case study on Galápagos tortoise (Chelonoidis niger) movement data. Our methods enable practitioners to conduct unbiased inference on the relationship between animal behaviour and general types of covariates, thus allowing us to uncover the factors influencing behavioural decisions made by animals.
Dienstag, 23.06.2026, 12-13 Uhr in W9-109
Daniel Dzikowski
Technische Universität Dortmund
Titel folgt
Dienstag, 07.07.2026, 12-13 Uhr in W9-109
Dr. Alina Schenk
Institut für Medizinische Biometrie, Informatik und Epidemiologie der Universität Bonn
Titel folgt
Dienstag, 21.07.2026, 12-13 Uhr in W9-109
Dr. David Winkelmann
Universität Wien
Why bookmakers offer inefficient odds in football live-betting markets