# Regression in Statistics Question

• Feb 28th 2011, 02:15 PM
jpd750
Regression in Statistics Question
For a multiple regression model, SST = 189, and SSE = 30, then the proportion of the total variation in the observed y’s that is not explained by the model is?

I've tried different formulas such as sst - sse and found a site saying ss/sst

Thanks
• Feb 28th 2011, 03:58 PM
theodds
You should say what SST and SSE are. Better notation is something like SS(Total) and SS(Error), which is what I'm assuming those two stand for - SST is also sometimes used as SS(Treatments).

$\displaystyle
\underbrace{SS(\mbox{Total})}_{\mbox{Total Variability}}
=
\underbrace{SS(\mbox{Error})}_{\mbox{Unexplained Variability}}
+
\underbrace{SS(\mbox{Regression})}_{\mbox{Explaine d Variability}}
$
• Mar 1st 2011, 03:12 PM
chaotic
I looked at the question and even I'm confused. Wouldn't the unexplained part of the model be SSE? The explained variation of the model is SSR.