Biostatistical models applied to animal experimentation

Abstract

The continual advances seen in agricultural sciences are due to uninterupted experimentation worldwide. For these studies to produce intelligible answers, they need to be planned, conducted and interpreted according to the principles of Experimental Statistics. Frequently these analyses are carried out by professionals trained in statistics with little experience in biology, leading to a gap in understanding the peculiarities of this area of study and so researchers and students do not make correct use of available methodologies. Also, with the increase availability and use of computers, there is a tendency to implement new data analysis methods such as REML (Restricted Maximum Likelihood) and multivariate analyses are increasingly demanded by the publications of greater impact. Thus, the use of statistical methods in experimentation is key in planning studies, statistical analysis and interpretation of the results. This also involves the correct presentation of the data in the final report. Thus, the present work plan is intended to assist in design, analysis and presentation of results of experiments with animals, offer lectures on scientific methodology, experimental design and data analysis for students / researchers from CENA / USP and other interested parties. (AU)