How to do statistical procedures & test assumptions

The summaries (listed left)These summaries  describe the 'nuts and bolts' of how to do statistical procedures, ranging from simple summary statistics, statistical tests & intervals, to the use of appropriate study designs, and generalized linear modelling.

We detail the procedure used, and formula(e) for calculating the appropriate test statistic(s). For the more sophisticated analyses we outline the mathematical model underlying each analysis. When considering study designs, we describe the main characteristics of each design, and summarize its pros and cons.

These 'how to' sections provide sample data sets, from medical, veterinary and ecological studies, appropriate for each of the statistical procedures. We detail how assumptions of each procedure may be checked, including tests for normality and homogeneity of variances. We provide 'worked examples' using both the 'calculator' formulae and using R. We also provide the R code for each analyses.

• Although these summary pages provide R code, the book  from which they are extracted is intended to enable biologists to understand statistical use and misuse - it is not a course on learning R.

• We use R to demonstrate the analyses, and perform simulations.

• Our code should prove useful to beginners - but it cannot substitute for gaining a good working knowledge of the language and procedures.

• There are many freely available on-line resources which should assist you in learning R - just Google 'learning R' and you will have more than enough.

For the more sophisticated analyses we perform diagnostics such as residual plots to examine the model's adequacy. When considering study designs, we describe the main characteristics of each design - and its pros and cons.

These on-line extracts from Brightwell & Dransfield (2012)  are in essence the bare bones of these statistical procedures - which you may feel is quite sufficient for you to design your research, and analyse your data.

If you prefer some understanding what you are doing (and why) you you need to understand what lies behind the formulae, what the underlying model assume, how those assumptions might be relaxed.

References  Brightwell, R. & Dransfield, R.D. Avoiding and Detecting Statistical Malpractice (or How to Get On Top of Statistics): Design & Analysis for Biologists, with R. [full text].