STATISTICS FOR DATA ANALYTICS - 19
T-test - numerical - 2
T - test : Unpaired Two sample Mean Test :-
Mostly data is unpaired :-
Eg :- testing a new drug over the existing one.
Two different sample ( comparison )
It is divided into 2 parts :-
Equal variance
Unequal variance
Step 1 :-
Null hypothesis => mean of sample A = mean of sample B.
Alternate hypothesis => mean of sample A is not equal to mean of sample B.
Step 2 :-
T- statistics
Step 3 :-
t - tabular(critical) > t- calculated { reject the null hypothesis )
T - test : Two sample Proportion Test :-
It cannot be used directly in excel so, we have to use the extension XLSTATS.
It is used when sample output is in Categorical form.
SAMPLE A 0 1
SAMPLE B 0 1
Like :- A person SHOWS SYMPTOMS OF DISEASE AFTER MEDICATION OR NOT..
ITS BEST APPLICATION :-
A/B testing
It is a direct industry application of the two sample proportion tests.
Z vs T test
If standard deviation is given, go for the Z-test or T- test.
If standard deviation is not given, go for the T- test
If sample size is above 30 go for Z-test
If sample size is less than 30 go for T-test
Feel free to share this post to enhance awareness and understanding of these fundamental concepts in statistical analysis!
🙏 Thank you all for your time and support! 🙏
Don't forget to catch me daily at 6:30 Pm (Monday to Friday)for the latest updates on my programming journey! Let's continue to learn, grow, and inspire together! 💻✨
Top comments (0)