- Franco Arda

# Old school vs. New School Statistics

Do you remember when you learned statistics? For a hypothesis test, we had to pull tables, define degrees of freedom, set the p-value ... this is old school statistics. The last 5-10 years have changed dramatically. I know it well because I learned statistics the old way as well.

Today, we favor simulations.
Let's take an example of a __Chi-Squared-Test__ where we compare the distribution of two categorical variables. In this case, we compare college vs. family income (college = yes or no, and family income = low, medium, high).

On the left, I implemented the Chi-Squared-Test the old school way and on the right, the new school way:

You see that we needed quite a few more lines of code for simulating the Chi-Squared-Test the modern approach. I used the awesome R package __infer__ for creating the permutation test and visualization.

If you're interested in learning more, Jake Vanderplas published a fantastic video on __YouTube__. For those who do Data Science in Python, they might know his popular book "Python Data Science Handbook." In other words, Jake is seriously good at Data Science.

Franco