Enjoy learning Statistics Online! Please be sure to share and subscribe to our YouTube channel.

Hi Professor Im struggling with a questions that seems pretty simple!

Ch. 10 #3 in the homework. Im doing everything right except for the SST, I'm adding 1368squared/4 + 1279squared/4 + 1471squared/4 + 1454squared/4 = 1946305.5

Then i go on to do the SSE and I subtract 31121.333-1946305 and get -1915183.667

I don't know what I'm doing wrong

Posted to STATS 2 on Monday, November 10, 2014   Replies: 6 Professor Mcguckian
11/10/2014
10:06 PM EST

Hi Sandra,

You haven't subtracted your correction factor from the 1,946,305.5. Don't forget the correction factor is (Σy)² /N.

Hope that helps,

Professor McGuckian .
11/11/2014
9:59 PM EST
Thank you!!!!

I have another quick question, for number 12 in chapter 11 homework, why is it that you subtracted the times 150-146 to find the prediction error? Professor Mcguckian
11/11/2014
10:47 PM EST

The prediction error is y - ^y . The variable y represents the actual y value from the original data that corresponds to the x value they are using. The ^value is the value the regression equation you created gives you when you plug the x into the equation. .
11/12/2014
10:20 AM EST
Thank you again!

Hopefully this will be my last question!

for number 62 b in the chapter 11 mixed review
62. b What is the largest deviation you might expect between any one of the 6 points and the least
squares line (find a deviation that 95% of the observed values will fall within from our line)?

you posted the answer to be 2S which stands for 2 standard deviations? Sorry if it seems obvious but just want to make sure! .
11/12/2014
10:45 AM EST
Just in case, can you show me how you got the answer for 62b. I keep getting the wrong one. Professor Mcguckian
11/12/2014
11:08 AM EST

Hi Sandra,

If you understood part a, then part b is easy. Part a gives us the variance, s², so we just take the square root of the answer in part a and then multiply by two. The reason we multiply by two is that we assume the error terms have a normal distribution with a mean of zero. The empirical rule from STATS I tells us that approximately 2 standard deviations capture 95% of the data under a bell curve.

Professor McGuckian