Permutation Hypothesis Test in R with Examples | R Tutorial 4.6 | MarinStatsLectures

[@marinstatlectures]

👋🏼 Hello there! In this R statistics video we learn how to use R (and RStudio) to do a permutation hypothesis test following an example where we compare a numeric variable for two categorical/qualitative variable. In another video we explain the concept of permutation test in statistics (bit.ly/2E0BOZ9); There is also an awesome R Script available for those of you who like to learn by doing with some extra material for keen learners! (bit.ly/2E0BOZ9) If Like to support us you can Donate (bit.ly/2CWxnP2), Share our Videos, Leave us a Comment and Give us a Like 👍🏼 ! Either way We Thank You!

[@janethompson6612]

honestly the best stats videos ive come across on youtube, cannot say thank you enough!!

[@gabrielrobert7110]

These videos and R script/tutorial are really great and I watched a lot of them, Marin is very clear and seem to follow people mind when he explains complicated concepts, many thanks !

[@SealionPrime]

These videos are fantastic! You explain the concepts so clearly and concisely and then back that up with some excellent well commented code. You guys rock!

[@marinstatlectures]

thanks, we appreciate that!

[@daesoolee1083]

I love how you explain things.

[@llorensbf]

you the man!

[@팬더-n4w]

thank you for the great video! reviewing the video, I was wondering why did you take the absolute value of the mean difference? When calculating the difference of two-sided hypothetical tests, absolute values would not fit into the hypothetical test... Am I right??

[@larissacury7714]

Hi! What is the difference between a permutation test and a Kruskal test for the analysis of an ordinal dependent variable?

[@ameliawake4361]

can you explain more about setting the seed. As in i understand why we set the seed and what the function does, but I dont understand the number in the brackets

[@haneulkim4902]

Thanks for an amazing video! I still have confusion about this topic. My understanding is that if you reject the null hypothesis(say when alpha=0.05) you could conclude with alternative hypothesis being true since in statistical sense having p-value of 4% says that observed result has 4%chance of happening by random chance therefore there must be a difference between two groups(alternative). However not fully understanding why "fail to reject Null" does not allow you to conclude that Null is True. Is there a way to conclude with Null is true? My goal is to prove that two different groups actually comes from same distribution(have same mean).

[@conrardyryan]

This video is really good, and does a great job of explaining permutation tests at a base level. I have a question though, for larger datasets with multiple variables it can often be beneficial to perform a t-test on each column (such as with the colttests() function), and in cases where you would want a permutation test you may want to do so for all columns. Do you know of a good method to do that, cause right now I generally have to manually run a permutation t-test one column/variable at a time.

[@marinstatlectures]

Thanks. To do that you can use the apply() function to apply a permutation test to columns. You can either write a permutation test function yourself or you can find a package with a perm test function that you can then apply to the columns. The apply function is essentially a more efficient version of a loop

[@conrardyryan]

@@marinstatlectures thanks! I'll try doing that.

[@shahidsayyed2824]

Can we use COIN package here?

[@muhammadkamil3558]

How can we do the Welch test in this example ?

[@rashawnhoward564]

infer package makes this easier

[@marinstatlectures]

That is true. But there is value in learning how to code certain things yourself, so that you are able to build or adapt what you need to

[@rashawnhoward564]

@@marinstatlectures You're right! This comment was for people that learned the hard way and now want to make it easier, so to speak.