Thank you, you are one of the very few channels that transcribes their videos by hand instead of letting youtube's horrible closed captioning do it for them. I am in the process of transcribing videos to help students in our statistics classes, this makes my job so much easier. Plus, you do an amazing job explaining everything here. Kudos!
It depends on whether the long-form formula or shortcut (smaller of the two sample sizes minus 1) is used. The calculator implements the longer formula, which is a pain to implement by hand.
Not pooling doesn't have notable downside, provided there is even just a very modest amount of data, while pooling relies on a strong assumption that may not be valid. Scenario #1 - If the variances are similar, then pooled or not pooled will give very similar results either way, so there isn't a meaningful downside to using the non-pooled variances approach. Scenario #2 - If the variances are not actually similar (which we might not realize), then it's better to use the unequal variance method. The only instance where not pooling becomes notably less good is when the sample sizes are extremely small in #1. In that scenario, checking the equal variance assumption is also not possible, so pooling would rely on a significant assumption and carries more risk.
thanks for the video! if im doing confidence intervals for a two sample t-interval for the difference between 2 means, does my direction affect my degrees of freedom?
Hi, is the question, does the subtraction order xbar_1 - xbar_2 vs using xbar_2 - xbar_1 affect degrees of freedom? If so, the answer is that it does not affect the degrees of freedom.
@@OpenIntroOrg yes, that was the question. I also just realized there is an option for testing if difference between the 2 means is 0, so that works too. Thanks!
For this example, the data aren't pooled, so that setting would be "No". A helpful video on paired data, if you're looking for more info on it: ua-cam.com/video/K0QZ9_4w0HU/v-deo.html
I was literally pulling my hair out because no one could explain this to me. Thank you so much for this video, it has been incredibly helpful.
Thank you, you are one of the very few channels that transcribes their videos by hand instead of letting youtube's horrible closed captioning do it for them. I am in the process of transcribing videos to help students in our statistics classes, this makes my job so much easier. Plus, you do an amazing job explaining everything here. Kudos!
This is by far the most straightforward example that I've seen so far, especially compared to what my professor is putting us through. Thank you!
what about one where they are asking for the difference being greater than a specific number. Like U1- U2 > 30? or U1 - U2 =/ 30?
The p value did not show on the calculations or the calculator making this confusing.
Thank you so much! By explaining everything as you went, I was able to figure out what I've been doing wrong! Thanks again!
What do you do when you only have the estimated standard deviation and not the S for each group?
I always get the incorrect P value. Everything else is spot on. What am I doing wrong
Thank you, God bless you!
Very nicely explained. Thanks Much appreciated
I am little confused how you got 46 for the df? is it just a random number between 32 and 45?
In this case, we use the value given by the calculator, and the spot it is shown is as the 3rd statistic (df) on the calculator at 4:45.
how do you get the digital version of the calculator?
Thank you so much! I spent all week looking for the right video that gets straight to the point. You saved me lots of time thank you new sub
if i get negative t- tesr what do i do?
what happened if I have two samples list ?? asap
How do you get the STS from this data?
Don’t you get a different df if you do it by hand though? Does it not matter as long as you state the df u sed?
It depends on whether the long-form formula or shortcut (smaller of the two sample sizes minus 1) is used. The calculator implements the longer formula, which is a pain to implement by hand.
Not clear why you didn't select pooled variance. Can you please explain this?
Not pooling doesn't have notable downside, provided there is even just a very modest amount of data, while pooling relies on a strong assumption that may not be valid.
Scenario #1 - If the variances are similar, then pooled or not pooled will give very similar results either way, so there isn't a meaningful downside to using the non-pooled variances approach.
Scenario #2 - If the variances are not actually similar (which we might not realize), then it's better to use the unequal variance method.
The only instance where not pooling becomes notably less good is when the sample sizes are extremely small in #1. In that scenario, checking the equal variance assumption is also not possible, so pooling would rely on a significant assumption and carries more risk.
you're a legend, tysm
thanks for the video! if im doing confidence intervals for a two sample t-interval for the difference between 2 means, does my direction affect my degrees of freedom?
Hi, is the question, does the subtraction order xbar_1 - xbar_2 vs using xbar_2 - xbar_1 affect degrees of freedom?
If so, the answer is that it does not affect the degrees of freedom.
@@OpenIntroOrg yes, that was the question. I also just realized there is an option for testing if difference between the 2 means is 0, so that works too. Thanks!
When I have to set "pooled" -> "Yes" ?
For this example, the data aren't pooled, so that setting would be "No".
A helpful video on paired data, if you're looking for more info on it:
ua-cam.com/video/K0QZ9_4w0HU/v-deo.html
how can I get that formula sheet?!?
You can find it in the last pages of the Advanced High School Statistics textbook. The PDF of this textbook is free at
www.openintro.org/book/ahss
thanks a lot this was very helpful
what program are u using for this video?
It's an emulator called Wabbitemu, available for windows, mac os x, and android
「コンテンツを調整する必要があります」、
where can I find a copy of the formula sheet referenced in the video?
The sheet is linked at
www.openintro.org/book/ahss
Very helpful. Thank you.
I keep getting a domain error can anyone help me?