ABSTRACT

So far in this book we have tried to make clear the underlying reasoning behind the analysis of simple experimental and observational data. The basic notion is that most hypothesis tests concern the distribution of test statistics, which are quantities we can estimate from our data and which have a known distribution if the Null hypothesis is true. By careful choice of the test statistic, we can find out whether sets of measurements conform to expectations (Chapter 2), or whether groups of organisms produce different results when they are subjected to different treatments (we have seen how to compare two groups in Chapter 3). In addition to the factors we control in our experiments, there are many other influences that may affect the results (as discussed in Chapter 4). Some extraneous influences can be counteracted by using appropriate sampling methods, but others must be taken into account in the statistical analysis. This chapter deals with how we analyse the experimental designs described in Chapter 4.