Lecturer(s)
|
-
Šmilauer Petr, doc. RNDr. Ph.D.
-
Lepš Jan, prof. RNDr. CSc.
|
Course content
|
Content of lectures: Observation and experiment, limitation of experiments in ecology, general rules of hypothesis testing, confirmatory and exploratory data analysis. Experimental designs and corresponding ANOVA models: Latin square design, nested designs, split plot, repeated measurements. General linear model and its implications as a basis of ordination methods. Multivariate methods in ecological research resemblance function, classification, gradient analysis. Ordination methods basics, techniques (CANOCO). Constrained, unconstrained, partial ordination in evaluation of designed experiments. Graphical display of results, ordination diagrams. Numerical classification. Content of practicals: Practical application of statistical methods introduced during the lectures, using Canoco 5 software, TWINSPAN, and cluster analysis and ANOVA in R software
|
Learning activities and teaching methods
|
Monologic (reading, lecture, briefing), Projection, Skills training, E-learning
- Preparation for credit
- 16 hours per semester
- Class attendance
- 56 hours per semester
- Preparation for exam
- 20 hours per semester
- Semestral paper
- 32 hours per semester
- Preparation for classes
- 45 hours per semester
|
Learning outcomes
|
The course provides basic principles of experimental and sampling design in ecology, and of the analysis of resulting data. The practical problems resulting from work in the field are stressed. A special attention is paid to understanding the difference between causal and statistical relationships and the role of experiments in revealing the causal relationships. Whereas in Biostatistics course the main emphasis was on the univariate statistical methods, in this course we focus on the more complex ANOVA models, and mainly on the multivariate statistical methods. The multivariate statistics will be demonstrated using the CANOCO and TWINSPAN software packages. In the presented framework, the multivariate methods are seen not only as the methods of exploratory data analysis, but mainly as a tool of hypothesis testing and analysis of designed ecological experiments.
Student will acquire the abilities to: (1) correctly plan and evaluate experiments in the fields of theoretical as well as applied ecology and taxonomy, particularly at the level of populations and communities, but also experiments producing multivariate data of molecular biology (2) distinguish research questions answerable using unconstrained and constrained ordination methods and correctly choose the corresponding methodology (3) interpret the results of permutation tests of multivariate hypotheses and also to correctly interpret graphical representation of ordination methods (ordination diagrams) (4) choose and apply appropriate methods of cluster analysis or TWINSPAN method
|
Prerequisites
|
Students need to have the expertise corresponding to the course of Biostatistics (KBE/012 or KBO/012), particularly in the area of hypothesis testing, analysis of variance, and linear models. Students have a sufficient command of English.
|
Assessment methods and criteria
|
Essay, Combined exam, Interim evaluation
Students solve homeworks during the term (5-7 times), which are then scored by the lecturer. At the end of term, students submit a written essay, representing analysis of own (or provided-by-lecturer) dataset, using the methods introduced during the course.
|
Recommended literature
|
-
Jongman R. H. et al. (1987): Data analysis in community and landscape ecology. - Pudoc, Wageningen.
-
P. Legendre, L. Legendre. Numerical Ecology. Third English Edition. Elsevier, Amsterdam, 2012. ISBN 978-0-444-53868-0.
-
Šmilauer P., Lepš J. Multivariate analysis of ecological data using Canoco 5. Cambridge University Press, Cambridge, UK, 2014. ISBN 978-1-107-69440-8.
|