## Learn everything about Analytics

Home » Your Guide to Master Hypothesis Testing in Statistics

# Your Guide to Master Hypothesis Testing in Statistics

This article is quite old and you might not get a prompt response from the author. We request you to post this comment on Analytics Vidhya's Discussion portal to get your queries resolved

### 23 Comments

• rkohli0 says:

Thanks for posting this article – An excellent read

• meenakshi says:

Nice article.

• Gaurav says:

Hi Sunil,

Very nice article. Keep up the good work!

Regards
Gaurav

• Pamba says:

excellent analysis

• Hossein says:

Hey Sunil,
I think this article was the best because you start with a challenge in work that every person may face

• Kishore says:

Great Article Sunil. This was helpful.

Keep it coming.

Thanks,
Kishore

• Tarunoday says:

It was really helpful.Its great to see how you give real life examples from your work.Want to see more statistics in your blog.

Cheers!!

• Rajeev says:

Hi Sunil,

Thanks for the excellent article. I wanted to understand this from long time however thickness of stat books kept me away. You did great job in keeping a story along with fundamental of stats.

Cheers

• chandrashekhar says:

good job!!!!

• Sunil Ray says:

Hi All,

Thanks for the comment …

Regards,
Sunil

• Shilpi says:

Very insightful!! I have one question
So does it means that if random chance probability is less than 5 % then there is difference in the behavior of two population and it is due to randomness?

• Avinash says:

If the p-value (random chance probability) is less than 5% then yes very likely you are correct in your conclusion that there is systematic difference between the two groups and the difference is NOT due to random chance. Or you can also say that you are 95% confident in your conclusion.

• Rahul Sharma says:

Hi Shilpi,
Suppose I have observed data with me..lets say I have 48 files comprising of 24 male gender files and 24 female gender files which will be represented to a board of members of the firm who will evaluate these files based on their performance for promotion. Lets say out of 24 males, 18 were promoted and out of 24 females only 11 were promoted to higher ranks. so there is a 30% observed difference in promotion of males and females. based on our observed finding I know wants to test the hypothesis if there is any relationship between promotion and gender? Is their any bias towards males than females when it comes to promotion of employees? I start by saying that lets assume there is no relationship between gender and promotion i.e. there is no effect in the population. So I initially assume my null hypothesis to be true. The other hypothesis which is my alternative hypothesis says that there is an effect in the population i.e. there is a relationship between gender and promotion for which i want to conduct hypothesis testing. By simulating this process for 100 times, lets say that it happened on less than 5% chances(Only on less than 5 instances in 100) that such a large difference of 30% occurred in gender promotion which points out that this observed difference of 30% in our original data didn’t happened by chance but is a case of bias towards the gender when it comes to their promotion..so p-value is the probability of occurrence of an effect in the population assuming null hypothesis to be true which in this case is less than 5%(very rare in my simulations that such a high value will occur)..Hope it gets clear!

• Avinash says:

Easy to understand and concise. I had this learning from my course I took from Jigsaw Academy but had lost some touch. Thanks for putting across this very important article.

• Amanpreet Singh says:

That’s by far the best content I’ve ever read regarding hypothesis testing.
Thankyou Sunil and team!!
Looking forward to reading more of your articles.

• Sathya says:

Nice job!
Where do we find the next article?

• AVTK says:

Hi,
Nice post! I am interested of how you answered your boss? What was the solution to the “signal or noise” original problem?

• Pooja says:

Thank you for this article !

• Danny says:

Thanks for such great Article, you explained this concept very easily

Regard

• Delon says:

Great read, although some images are missing.
Please will you make these images available again.

• Oren says:

Hi Sunil,

Nice article, you explain it nicely with simple worlds, and nice charts, I like it 🙂

I want to mention that the examples are based on the assumption that the population distribution is normal, this assumption also should be checked.

Also, I want to mention that in the example we know the population’s standard deviation,
usually, we don’t know the population’s standard deviation, and we need to estimate the value based on the sample. in this case, we should use the t distribution instead of the normal distribution (z)

• Samson Mangin says:

today is my exam on bio-statistics, and i am not ready….. but suddenly i viewed this article. it helps me a lot….. thank you so much

-Samson-
fr: Philippines

• […] mean is right around zero, with a standard deviation of 7.3. You’ll see the formula broken out here, as we snag the z-score for this value and produce -0.328: our tanking rookies barely differ from […]