Not sure if anyone gives a damn but if you are stoned like me atm then you can stream pretty much all the latest movies on InstaFlixxer. I've been binge watching with my gf for the last few months xD
I've been taught that not rejecting Ho is not the same as accepting it,there just isn't enough evidence to reject Ho.There certainly isn't enough evidence to suggest the coin is fair,we just can't be 95% sure it isn't.(we can reject Ho with a slightly larger a)
The P value looks at the probability of observing the results of the experiment *or a more extreme result* given that the null hypothesis is true. More extreme means "in the direction of the alternative hypothesis". I have asked myself this question as well, I do not have a really good principle / fundamental answer but here is how I live with this: 1) This how how hypothesis tests are designed. 2) It makes more sense when calculating p-values on continuous distributions since P(X = a number | Ho) = 0 when X is continuous. In this context it makes sense to ask for: P(X < a | Ho) (left-one-tailed, Ho: X=a, H1: X b | Ho) (right-one-tailed, Ho: X=a, H1: X>a)
Only 6 years late but it is intuitive, no? Rolling 1 six in 24 throws it pretty unlikely, but the chance of rolling 1 or less sixes (i.e. rolling 1 six or 0 sixes) is actually greater than the chance of rolling just 1 six because these are 2 of the 25 possible outcomes rather than just one. However, rolling 1 six or 0 sixes is *more* extreme than rolling just one six and it gives more evidence that the die is biased.
@@archjanasivasuthan6252 bit late to this but oh well. As you know, if the probability of getting a head on a supposedly fair coin is 1/2 and we have 24 trials, we would expect to get 12 heads. However, we could very well get much more than 12, like getting 22 heads or much less than 12, like 1, both of which are quite unlikely, causing you to say that it might be unfair. The question is, at what point do we say this is in fact not a fair coin? that's what the significance level is for. if we have a test statistic that has a probability of occurring which is less than the significance level given, we deem it "too" unlikely to support the null hypothesis. Bit of a lengthy explanation but hope that helps.
Why can't you just find the probability of X=1 and see if that is below the significance level, rather than doing the probability that X is less than or equal to 1? Thanks
12 years now and students are still learning from you. Thank you!
9 years old video and still helpful, thank you!
Yep. still going.
12 years now!!!
Thank you Max0r for introducing me to this absolute BANGER
I was always taught by my stats professors that you don't accept any hypothesis. You fail to reject. Great video though.
Thanks. Good to hear that it helped.
Not sure if anyone gives a damn but if you are stoned like me atm then you can stream pretty much all the latest movies on InstaFlixxer. I've been binge watching with my gf for the last few months xD
@Ellis Kase Definitely, I've been using InstaFlixxer for months myself =)
You do not "accept" H_0. You "do not reject".
"insufficent evidence to reject h_0"
I've been taught that not rejecting Ho is not the same as accepting it,there just isn't enough evidence to reject Ho.There certainly isn't enough evidence to suggest the coin is fair,we just can't be 95% sure it isn't.(we can reject Ho with a slightly larger a)
Thanks Professor! It was clear like always.
thanks but a doubt is its this situation a two tail test ?
Super helpful 👍
Great video. Super thank you
Why do you need to do less than or equal to 1? I can't find anywhere that explains this. Why don't you just find out P(X=1)?
I believe we should only count P(x=1), since the question is not no more than one time in 24 throws
The P value looks at the probability of observing the results of the experiment *or a more extreme result* given that the null hypothesis is true.
More extreme means "in the direction of the alternative hypothesis".
I have asked myself this question as well, I do not have a really good principle / fundamental answer but here is how I live with this:
1) This how how hypothesis tests are designed.
2) It makes more sense when calculating p-values on continuous distributions since P(X = a number | Ho) = 0 when X is continuous. In this context it makes sense to ask for:
P(X < a | Ho) (left-one-tailed, Ho: X=a, H1: X b | Ho) (right-one-tailed, Ho: X=a, H1: X>a)
Only 6 years late but it is intuitive, no? Rolling 1 six in 24 throws it pretty unlikely, but the chance of rolling 1 or less sixes (i.e. rolling 1 six or 0 sixes) is actually greater than the chance of rolling just 1 six because these are 2 of the 25 possible outcomes rather than just one. However, rolling 1 six or 0 sixes is *more* extreme than rolling just one six and it gives more evidence that the die is biased.
@@tamircohen1512 Thank you!!! Also why does the test statistic have to be less than the significance value?
@@archjanasivasuthan6252 bit late to this but oh well. As you know, if the probability of getting a head on a supposedly fair coin is 1/2 and we have 24 trials, we would expect to get 12 heads. However, we could very well get much more than 12, like getting 22 heads or much less than 12, like 1, both of which are quite unlikely, causing you to say that it might be unfair. The question is, at what point do we say this is in fact not a fair coin? that's what the significance level is for. if we have a test statistic that has a probability of occurring which is less than the significance level given, we deem it "too" unlikely to support the null hypothesis. Bit of a lengthy explanation but hope that helps.
thank you!
You're welcome.
Thanks
thanks again
@HanKooR Thanks
What book are you using?
Dunno
R u still alive
Thank you for the help :)
Why can't you just find the probability of X=1 and see if that is below the significance level, rather than doing the probability that X is less than or equal to 1? Thanks
Lol
Lol
Lol
LOL THANK YOU AGAIN
Please make a gaming channel
@hongabonga12345 Cool
:)
thanks