4.2. Learning Objectives#

Here is what you should understand after this week

4.2.1. Conceptual#

After this week you should:

  • Understand what a null and alternative hypothesis are

  • Understand the terms Type I Error and Type II Error

  • Understand the procedure for null hypothesis testing (ie assume the null is true, work out the probability of our test statistic arising)

  • Understand the converse procedure for power analysis (ie assume an effect of a certain size, and work out the probability of obtaining a significant result)

  • Understand what an effect size is and that:

    • The effect size for a correlation is the correlation coefficient r

    • the effect size for a t-test is Cohen’s d

The conceptual material is covered in the lecture and recapped in the worked examples in Python

Additionally, in the lecture there was an overview of conceptual issues involved in selecting tests. You should understand this material, which will be assessed in the Hand In Assignment to be set in 3rd week.

4.2.2. Python skills#

The key skill practiced this week is running a power analysis for the t-test using statsmodels

We also create home made code to run a power analysis, which includes simulating data with a certain effect size.

This is rather involved and you would not be expected to reproduce it independently, but you should understand conceptually what was done and be able to copy the code blocks an modify them to run a power analysis for a different value of \(r\) or \(d\)

This material is covered in the Jupyter Notebooks in this section