Unit 8 - Data Analysis and Visualisation

This unit refined my conceptual and practical understanding of statistical inference, particularly hypothesis testing. Building on earlier statistical foundations, I explored how to formulate null and alternative hypotheses, interpret p-values, and apply decision criteria using significance levels.

I found the analogy to a courtroom trial (starting with the presumption of innocence) particularly helpful in conceptualising the logic of starting from the null hypothesis. Setting a significance level (e.g. 5%) as a threshold for rejecting the null enabled a clearer understanding of error control, especially Type I and Type II errors. The decision matrix on page 5 provided a structured view of outcomes, reinforcing the probabilistic and uncertain nature of inference.

We also covered the test statistic and its role in measuring the distance between a sample mean and a hypothesised population mean, adjusted by standard error. This gave me a stronger appreciation of how distributions and sample size affect inference power. Visual examples, such as the normal distribution curves on pages 2 and 3, helped consolidate these concepts.

Importantly, the worksheet also stressed the importance of setting hypotheses in advance and the ethical pitfalls of post hoc test selection. This was a valuable reminder of professional integrity in statistical research. Similarly, understanding why related samples tests can offer greater power (while acknowledging dropout risks) added a useful dimension for study design in my own project work.

Overall, this unit sharpened my ability to interpret significance with caution, avoid overstatement, and apply formal inferential logic. These are directly relevant to evaluating experimental outcomes and forecasting model performance in my MSc project, where statistical justification is essential.