Module 2, Lecture 1 • Research process: o Doing experimental biological research:  Clear process  Clear structure o Writing reports, theses, papers  Similar structure o Role of stats?  Summarise data, patterns Are patterns “real”? Important, but just a part • Key elements: o Types of variables o Unit of replication o Limits of inference (statistical vs biological) o Strength of evidence o Parametric vs non-parametric stats

Module 2, Lecture 2 Learning outcomes: • Dealing with class independent variables • Type of analysis driven by types of variables • ANOVA – one useful parametric test • Reliable analyses depend on assumptions met • Transforming can “improve” data • Today, there are alternatives to transforming

Module 2, Lecture 3 Learning outcomes: • Experimental design determines type of analysis • ANOVA vs Regression o Same principle, different detail • Simple vs Multiple Regression o Same principle, different detail • Stats: part of a bigger set of questions

Module 2, Lecture 4 Learning Outcomes: • Binary data: e.g. +/- • Binomial distribution (i.e. not normal) • Logistic regression suits Binary dependent data • Want probability of “success” • Model with logit link function (“transformed” probability) • Plot & interpret using probability of “success” equation • Today, great flexibility: o other statistical analyses for other data distributions

Module 2, Lecture 5 No Learning Outcomes, but Summary: • Bayesian and Information Theory approaches: alternatives to null-hypothesis testing • AIC determines the model of best fit with the number of parameters • AIC is NOT a measurement of support for a model • Model averaging recognises uncertainty in models by calculating average parameters weighted by Akaike weights • Note: CANNOT undo bad sampling or poor experimental design