A collection of five short tutorials designed for post-graduate students (who were taking an online course) with limited background in a) the statistical analysis of data and b) evidence from trial data. Students tell us they typically can read and critique papers, but reported that they don't have a concrete understanding of how to take a set of real data, analyse it, while linking the concepts they use for critical appraisal with the actual techniques of basic modeling and analysis. Note, these aim to show how it's done, but don't interpret the methods discussed as being advocated for, similarly, although instructions for SPSS are given, I can't claim to have much expertise with it. If you want to analyse your own real trial data, consult a statistician.
- Part One : Introduction, assumptions and basics of the example trial data (Kane's classic 1988 chlorpromazine and clozapine RCT) used to illustrate the concepts and methods that follows.
- Part Two : A primer on linear models, regression with continuous and categorical independent variables. Shows how data is structured so that a linear model can be constructed and interpreted.
- Part Three : Explains repeated measures ANOVA, showing how terms in a linear model can be interpreted as mean effects and how this relates to structured data and the output from statistics software.
- Part Four : Explores how to frame hypothesis testing with linear models and checking randomisation assumptions.
- Part Five : Shows how to formulate the preceding analyses as ANOVA of change (in outcome) as well as ANCOVA (analysis of covariance). As before, the relationship between the linear model and data structure is emphasised, and some sample data is provided