Reliability of Difference Between Uncorrelated Means (Nᡖ)
Performing the z test
- If the z calculated (using formulas from previous slide) is 1.96 or greater, we can reject the null hypothesis.
- This means that a difference as large or larger than the one we saw in our sample would occur by chance when there is no difference between the population mean (experimental group - control group) no more than 5 times out of 100.
- This is a 2 tailed test!