Research design and statistics in life sciences

1-week intensive course with 20 lectures and 20 hours practical training primarily aimed at post-graduate students and researchers with at least basic statistical knowledge. The course is also open to students/researchers from other disciplines.

The course provides an overview of how to successfully design surveys and experiments in life sciences. The use of uni- and multivariate statistical techniques to explore patterns in large data sets. Provide consultation to individuals concerning designing their own research, analyse and present their data.

Main topics include conditions for statistical tests (Random sampling, independence of data, homogeneity of variances); Sampling (Number of samples, sampling strategies, pollution studies); Experimental and sampling design (Pseudoreplication, factorial experiments, block and nested design, power of statistical tests); Regression and correlation (Association between variables, multiple regression); Multivariate statistics (Measures of similarity, clustering, ordination techniques: PCA, MDS, relating multivariate analyses to environmental factors); Biodiversity (Measurements of diversity indices)

Practical training: Students work both in groups and individually. Calculation of sample sizes, power of tests, diversity indices etc. Students will read scientific papers and try to detect flaws in research designs. A number of statistical exercises will be available but students are encouraged to work on their own data sets.

Tools: Exercises will involve several computer packages: spreadsheets (Excel, Lotus) and statistical (Excel, Statistica, Primer).

Number of participants: Preferably no more than 3 students per computer. Maximum number of students is 20.