Mastering pandas - Second Edition

A t-test example

In simplified terms, to do Null Hypothesis Significance Testing (NHST), we need to do the following:

1. Formulate our null hypothesis. The null hypothesis is our model of the system, assuming that the effect we wish to verify was actually due to chance.
2. Calculate our p-value.
3. Compare the calculated p-value with that of our alpha, or threshold value, and decide whether to reject or accept the null hypothesis. If the p-value is low enough (lower than the alpha), we will draw the conclusion that the null hypothesis is likely to be false.

For our real-world illustration, we want to investigate whether manual transmission vehicles are more fuel efficient than automatic transmission vehicles. In order to do this, we will make use ...