Statistics 21 - Lecture 6

[couchtalk]

He is a wonderful teacher who tries his best to clarify for his students what is abundantly crystal clear to him. He also seems to love what he does and that is contagious.

[Koni804]

camera guy; I like your work however I find it hard to follow the prof. if you dont follow him, since this video is on youtube I can pause and rewind if I need to

thanks

[Jabartanha]

Great lecture, :)

[edgeArchitect]

Note to self: 54:42 - Regression Method

[shamelesstill25]

looking back on my college experience i really zoned out lots during power point presentations..

my stat teacher used blackboard tho <3

[edgeArchitect]

@shamelesstill25 yea, for me it was not a PowerPoint Presentation, but a PowerNap time, ;(

[ZR666]

Why is the division by n and not

(n-1) for the sample correlation coefficient?

[ZR666]

I see what you did...why use a nonstandard definition of the sd?

[kleinbogen]

You are correct! for sample correlation coefficient, it should be divided by n-1 not n. I think he is wrong here. Interestingly, if I give you a list of order pairs without telling you if they are the whole population or they are just a sample. You go ahead and calculate sample correlation coefficient (divide by n-1) and then calculate the population correlation coefficient (divide by n). You will see in either case you will get the same result.