Beskrivelse
Courses on statistical methods for Master students in the master program on agroecology.The course aims to provide the basic tools to design and perform analysis of experimental studies and use these tools to evaluate simple scenarios of observational studies.
More precisely, the course has two main goals:
1) To learn basic techniques for designing/planning simple experimental and observational studies including: the notions of randomization, replication, local control (blocking), confounding, complete random designs, complete and incomplete block designs, factorial designs, cohorts and simple case-control schemas for observational studies. Additionally, methods for power calculations (mainly based on simulations) and optimization of designs of experiments will be studied. These techniques will be used to discuss some scenarios of observational studies where the study is not completely determined beforehand. The student will learn to use simulation techniques to gain intuition and to compare different experimental designs and different scenarios of simple observational studies.
2) To learn some basic techniques to analyse data arising from the type of designs of experiments discussed in the previous item using suitable Gaussian linear mixed models for continuous responses (i.e. linear models with random components based on the normal distribution) and simple standard generalized linear mixed models for proportions and counts (binomial and Poisson mixed models). Basic methods of model control will be studied and discussed in real and simulated data.