While working on the FetchMaker project in the Data Analyst career path, I found one of the task descriptions to be odd. For task 10 it says:

Run a hypothesis test for the following null and alternative hypotheses:

Null: There is an association between breed (poodle vs. shihtzu) and color.

Alternative: There is not an association between breed (poodle vs. shihtzu) and color.

Save the p-value as pval and print it out. Do poodles and shihtzus come in significantly different color combinations? Use a significance threshold of 0.05.

The solution uses a Chi2 test. However, I thought that H0 of a Chi squared test is always “there is no association between the categorical values, any variation is pure chance” and that if the p value is significant (here <0.05), we reject H0 and assume the alternative H1 “there is an association between the categorical variables”.

Am I correct here? Or can I perform the Chi2 test in way where H0 and H1 are as they are given in the description?

the null is that there is no association between the variables (or, status quo) and the alternative is that there is a statistically significant association between the two variables.

Task 10 on my end is the Tukey’s range test for dog breed weights.

Most likely not. It was probably just overlooked and not updated in my version.
I completed this project and these lessons as part of an Intro to DA back in Dec. 2017. The whole course was in Python 2. (This was before they had Pro content.) They used to offer these class intensives that were 6 weeks or so & $199. This content was in one of them. When I look at it now my code is still mostly in Python 2. (ie: print pval rather than print(pval)).