Chi Square Test for Independence

Question 1

When Would you Use the Chi Square Test > t-Test?

Answer 1

When data is categorical.

Question 2

What are the Characteristics of the Chi-Squared Distribution?

Answer 2

Skewed to the right (long tail)
More symmetrical as DoF increase.

Question 3

Why is Chi Test Always 1-Tailed?

Answer 3

Looking for Differences, not whether expected values are less/greater than.

Question 4

Assumptions Made Regarding the Chi Square Distributions:

Answer 4

Expected Values calculated on basis that H0 is true and can be fractions/decimals.
Test only if expected values are atleast 5. Otherwise combine rows.

Question 5

What is the DoF When Testing for Independence?

Answer 5

(r - 1)*(c -1)
r = no. rows
c = no. columns

Question 6

What are Tips for Combining Categories?

Answer 6

Only combine similar categories.
Keep same sizes similar across rows/columns
Focus on variable of interest.