The purpose of a biological assay, or bioassay, is to measure and compare the response or activities of organisms as a function of physical, chemical, biological or temporal stimuli. Often, summaries of biological assays require linear or non-linear regression models. For example, fertilizer effect on crop yield is often analysed using ANOVA whereas multiple linear regression models are useful for describing weight increase adjusting for the initial weight and other baseline variables. Concentration/dose-response experiments in ecotoxicology and human toxicology, weed competition in crops and Michaelis-Menten kinetics are just few of the biological responses to be modelled with non-linear regression models. Seed germination in response to time or chemical stimuli has to be analysed using models for categorical data such as generalised linear models including logistic regression. Extensions towards random effects modelling and multivariate analysis will be discussed.

The way experimental response data are summarised, however, varies much among scientists and is often restricted to traditional approaches and available statistical software rather than theoretical considerations of the best way to exploit the information contained in the data. During the past decade the development of statistical software has been swift, and today reliable free statistical software is available; any scientist in any place of the world with an internet access can download the programmes. The advances in statistical software allow both standard statistical methods and more advanced methods to be applied, moving the focus in the application of statistical analysis from pure computational aspects to more relevant aspects concerning interpretation of results (biological implications). In this course our approach is to base all statistical analyses on a single statistical software programme, namely the open source environment R (http://www.r-project.org).

The course will discuss different types of endpoints (binary, count or continuous), transformation of the response/independent variables, model specification and interpretation of the model parameters, checking the model assumptions, remedies for model violations, estimation of relevant effects and hypothesis testing.

We will discuss modelling biological phenomena like baseline measurements, hormetical effects, natural mortality etc. The topics will be illustrated through small case studies. The course is intended for PhD students, researchers and scientists in agricultural, biological and environmental sciences.