STAT1201 Semester 2 2023 Summary Module 1: Variability: - Natural variability - Measurement variability - Error variability - Group variability (difference between two distinct groups) - Sampling variability (difference due to sampling method) To minimize the effect of variability we must replicate our experiments. Variables: characteristics we can record about subjects or objects in a study. - Discrete or continuous - Categorical Nominal variables: arbitrary groups with no order between them e.g., gender Ordinal variables: categories that do have an order like age group Variable Categorical Quantitative Nominal Ordinal Discrete Continuous Designing studies: - Observational study (association) - Experiment (causation) Randomisation: splitting a sample into experimental and control groups randomly Evidence from data: Null hypothesis: IV has no significant effect on the DV and any variability is due to chance (p>0.05) Experimental hypothesis: IV has significant effect on DV and variability is not due to chance (p<0.05). - If the p-value is small, we have evidence against the null hypothesis.
Strong Moderate Weak 0 0.01 0.05 Inconclusive 0.1 1 Hypthetico-deductive method: Karl Popper, making hypothesis and testing predictions. Randomisation tests: to test if data was a result of chance, shuffle all data points between the two groups and see if the pattern appears - if it does it may be chance if it doesn't it is likely due to differences in the groups.