# SST Calculation

• Oct 21st 2010, 09:56 AM
SST Calculation
Ok, so I am thoroughly confused atm. Right now I am trying to do three things.

Calculate SS Total, and the explained sum of squared variation for SSR.

Ok so I had the outputs given to me of

Sum x = 30
Sum x squared = 104
Sum of y = 40
Sum of y squared - 178
Sum of x times y = 134

I used that data to solve and get the following
SSxy = 14
SSxx = 14
b1 = 1
ybar = 4
xbar = 3
b0 = 1
yhat = 1 + 1x
Ok so based on this the sqares regression line is 1+x
so y intercept is 1 and slope is 1

I calculated sse to be 9...

When I try and calculate sst which is xbar1-xbar
I get 10(30-3)^2 = 7290

Im pretty sure this number is wrong. Can anyone tell me what I am doing wrong?
• Oct 21st 2010, 11:10 AM
Guy
We need to know n, but I guess it is supposed to be 10.

$SS_{total} = \sum Y_i ^ 2 - n \bar{Y}^2 = 178 - 160 = 18$.

Assuming you calculated SSE correctly, the regression SS is 9.
• Oct 21st 2010, 11:41 AM
Quote:

Originally Posted by Guy
We need to know n, but I guess it is supposed to be 10.

$SS_{total} = \sum Y_i ^ 2 - n \bar{Y}^2 = 178 - 160 = 18$.

Assuming you calculated SSE correctly, the regression SS is 9.

Ah ok. Thanks, I didn't have that formula. And yes, N was 10.
• Oct 21st 2010, 12:06 PM
Guy
In nice situations (regression being one) the relevant sums of squares all have easy computational forms that are worth knowing so that you can do problems like this quickly. They all look basically the same as this one does.
• Oct 21st 2010, 12:58 PM