Is This Coin Fair? Hypothesis Testing Questions and Answers in Data Science Interviews

[@inna5649]

Thank you so much for your incredible videos!

[@emma_ding]

So glad you're enjoying them, Inna! Thanks for watching! 💛

[@sergeilogvinov5963]

This is way too superficial. Why you chose the z-test instead of the proportional one? What is the small number of samples? And my favorite: "Set number to 1000" - how about talking about mde and minimal sample size?

[@shilpaparmar3611]

Can you also address this problem using a binomial test?

[@normalhuman6260]

Nice fundamental explanation. You should do more problem breakdown.

[@emma_ding]

Thanks for the suggestion! 😊

[@normalhuman6260]

@@emma_ding Thank you. There is a serious lack of fundamentals specially hypothesis testing on education youtube.

[@mckinleypaul6943]

How do you know the variance of the null distribution here? i.e. how do you get var(H0) = np(1-p)?

[@jacob-zs9wc]

The weight is unbalaced. So does they heavier side land face down?

[@arojitdas8256]

Nice Explanation .😅 Can you make video on Law Of Large Number . Also a video comparing T test, z test, Annova , Chi Square test

[@emma_ding]

Thanks for the suggestion! I've added it to my list of content ideas. ✏️

[@yashsaxena7754]

The condition for binomial distribution to approximate like normal distribution is np>5 where n is the number of trials and p is the probability of success. Had the trials been less than 10 this condition would have failed, how would we do this test in that case?

[@bodwiser100]

quick question - for computing the test statistic, why did you not use the standard error in the denominator, i.e., sigma/sqrt(n)? Where sigma is sqrt(np(1 - p)). Basically, my question is that you used sigma in the denominator; instead, should it not be sigma/sqrt(n) ?

[@navishgoyal]

I also have this doubt

[@fenghu7452]

I have same question... can anyone explain?

[@emma_ding]

Good question Rajat! If you divide n in both the denominator and the numerator, you'll get the same result.

[@muthukamalan.m6316]

could you please suggest experiment design books for beginners

[@emma_ding]

Here you go! kzbin.info2Px-gRn8Zd0

[@CS_n00b]

what didnt you use the standard error? ie sqrt(p(1-p)/n)

[@markjones-qi5xs]

Could you not also use a one-sample proportion test for this as well? Where p_0 would be 0.5. Doing it this way I end up with k needing to be between 0.46 & 0.53 in order for the coin to be considered fair