Intermediate statistics for life scientists


This two day course will be suitable for life-scientists, in particular PhD students, who wish to extend their understanding of statistics. Some basic concepts such as tests of significance and confidence intervals will be revisited but the emphasis will be on using good graphical displays and intelligent models to gain insight in to data. Time will be devoted to planning studies also. The examples are all genuine and mainly based on in-vivo animal and in vitro experiments. A few examples will also be included from clinical trials. The analyses for the course have all been implemented in GenStat?, although use of this is not an essential aspect of the course. The emphasis will be on using statistics to help understand data rather than as a black box. Mathematical development will be kept to a strict minimum.


Stephen Senn


Day 1

1. Introduction: types of data, summarising data, standard deviations and standard errors,
Normal distribution, transformations
2. Simple examples: presentation. One way layouts, dotplots, histograms, scatterplots, trellis plots
3. Simple examples: analysis. Analysis of variance, residuals, transformations, pairwise comparisons
multiplicity, using covariates
4. Some complications. Repeated measures. Multivariate analysis of variance, split plot,
summary measures, dose response, Emax models

Day 2

5. Inference. Significance tests, P-values, confidence intervals, multiplicity (again), invalid inversion
6. Planning. Power, sample size, blocks, randomisation, regression, outliers
7. Modelling. Dose finding, non-linear models, Emax, extended Emax
8. To be determined. This will either cover suggestions from delegates or, if none are received,
extend the modelling theme of module 7.

Return to Statwise courses