Modern statistics in natural sciences, 5 credits
Modern statistik i naturvetenskaperna
Course information
Language of instruction: English
Course period: January-March 2025.
Course structure: On campus. Lectures, independent practical statistical problem solving, group discussions, reading and discussion of course book and selected papers, tutoring in the software environment R.
Recommended prerequisites
Basic statistical knowledge.
Learning outcomes
To give an introduction to the most commonly applied modern statistical techniques and tools used in a wide range of natural sciences, to analyze empirical data and test hypotheses. In addition to providing an overview of the statistical “tool-box”, the course generates an understanding of the philosophy and reasoning behind statistical design, modelling and inference. Practical elements (exercises) and group discussions gives the students substantial hands-on experience, deeper insights, confidence and training in discussing statistical analyses. This is a general course in applied statistics with an emphasis on experimental data that primarily attracts PhD students from biology/life sciences, geosciences, chemistry, information technology, experimental physics and related fields.
Learning outcomes for doctoral degree
The course develops and discusses a number of fundamental and general goals in postgraduate education, such as scientific inferences, limitations of dualism, hypothesis testing, scientific transparency, good statistical practice, experimental design and numerical analysis.
More specific goals include insights into choice, and execution, of statistical models and interpretation of statistical analysis, which is key in all fields leaning on empirical data. In summary, the course significantly improves the participants’ knowledge and understanding of scientific methodology, statistical analysis and critical evaluation of scientific inferences. These are all central goals in post-graduate training.
Course contents
The course is focused on analyses of experimental data, but an overview of analyses of observational data is also given. The course focusses on linear models and includes: experimental designs leading to ANOVA or ANCOVA, mixed models, blocked experiments, repeated measurement designs, nested and factorial designs, multiple regression including strategies for selecting variables and evaluating models, generalized linear models (GLIM) including logistic and Poisson regression, contingency table tests, power analysis, multivariate analysis and ordination techniques, resampling and permutation statistics, Bayesian model fitting, MCMC techniques, geometric morphometrics and a few other topics. The philosophical basis of hypothesis-testing and statistical inferences is covered at the start of the course and the course closes with considering good scientific practice in terms of statistical analyses.
Instruction
The core of the course is built around a series of 13 half-day and interactive lectures. In addition, the students then work off-schedule with a series of common practical elements/problems that are then discussed during a series of tutored group discussions. The participants deliver written individual practical reports on these exercises. Hands-on advice and individual tutoring of scripting in the statistical environment R is also offered at several occasions during the course. The course closes with a group discussion of a series of statistical problems built around the course book.
Assessment
Attendance at all lectures (or substitute assignments) and approved individual practical reports that students hand in.
Course examiner
Prof. Göran Arnqvist, Goran.Arnqvist@ebc.uu.se
Department with main responsibility
Department of Ecology and Genetics
Contact person
Göran Arnqvist, Goran.Arnqvist@ebc.uu.se
Application
Submit the application for admission to:
Register in google sheet sent out to all faculty PhD students by email during the fall or by email to course responsible teacher.
Submit the application not later than: December 31, 2024