Earlier small sample domain was taken for a sample size of 30 or less. But now some recent tables have increased the boundary to 500.
What is the reason and which one to follow?
Because there are 2 ends of the curve. You only need both sides if you don't know how to adjust from one side to the other. See an elementary statistics book. Best, D. Booth
David Eugene Booth, I get that but I was asking for two versions on the t table. where one ends at 29 and the other at 500 for a small sample domain. Why was there an upgrade in a small sample domain from 30 to 500?
Sorry someone noticed that df don't stop at 30 and that it takes longer than that for the t to converge to normal. Shown by simulation runs. BTW it takes longer than 500 as well but the approx. is better. Some stat computing books give this as an exercise. Best, D. Booth
So it's an ongoing process. Thanks, David Eugene Booth. Can you also share some related sources that you know about?
Hi Ashay Singh
It is well known the t converge to normal with the increasing of df; however, I cannot explain the usefulness of t-table with 500 df!
https://math.stackexchange.com/questions/3240536/convergence-of-students-t-distribution-to-a-standard-normal
- Badges
- Science topic