Anyone familiar with Dickey-Fuller test?

Does it follow exactly the same t-test steps? i.e. calculated t values compared with critical t values? reject if greater than critical t. OR am I missing something here? I don’t see any actual example on the materials.


I don’t think we actually need to calculate the DF test for the exam. The LOS says to “Describe the steps of the unit root test for nonstationarity.” we can do this by running an AR model and examining the autocorrelations, or performing the DF.

What is important to know about DF is this:

  1. we have a unit root if b1 - 1 is 0. (in the denominator of our test for a unit root)

  2. DF null hypothesis is that the denominator, g, is equal to 0.

  3. Thus, if they tell us that we cannot reject the null (g=0) we will assume that there is a unit root.

Many thanks!

The exam is probably not going to test the actual calculations, they can still ask if the null is rejected or not based on some stats. I am not clear on how to determine that.