1
$\begingroup$

For a Two Way ANOVA with two factors A and B, the Sum of Square equation is:-

SSTO = SSA + SSB + SSAB + SSE

Where,
SSTO = Total Sum of Squares
SSA = Sum of Squares due to Factor A
SSB = Sum of Squares due to Factor B SSAB = Sum of Squares due to Interaction of Factor A and Factor B SSE = Amount of Variability left unexplained taken as error

Now my question is, why does One Way ANOVA not have a SSE Term?

FYI,
Im reading this book, Intermediate Statistics for Dummies. I was reading about ANOVA. For one way ANOVA the book says that the Total Sum of Squares is the Sum of Squares within and Sum of Squares Between. There is no Sum of Square Error Term in one way ANOVA. That's why I am asking this question.

$\endgroup$
  • $\begingroup$ What leads you to believe it doesn't? $\endgroup$ – gung - Reinstate Monica Jan 3 at 17:32
  • $\begingroup$ Im sorry gung, but I dont know. Im just starting to learn advance statistics. Im reading from this book, Intermediate Statistics for dummies. The book says that for One way ANOVA, SSTO = SSW + SSB. So i just assumed that SSE is not there. If you could suggest me a better book to read, that will be helpful :) $\endgroup$ – Rishi Sharma Jan 3 at 18:05
  • $\begingroup$ I'm not familiar with that book. The quote sounds bizarre. FWIW, we have some statistics textbook recommendations here: Free statistical textbooks. $\endgroup$ – gung - Reinstate Monica Jan 3 at 18:05
  • $\begingroup$ Please do not delete & repaste identical comments. $\endgroup$ – gung - Reinstate Monica Jan 3 at 18:06
  • $\begingroup$ @gung It's not really that bizarre. See my answer. This notation is not uncommon. $\endgroup$ – Peter Flom - Reinstate Monica Jan 4 at 9:49
2
$\begingroup$

One way ANOVA does have an SSE term, see e.g. ANOVA notes.

EDIT in response to comments:

The labeling can get confusing; it's important to remember what is actually going on in ANOVA. I wrote about this (in very simple terms) here. For your particular question:

When we divide a sample into groups, there will be variation within each group and variation between the groups. The latter is what we want to model, the former is noise. So, the variation within each group is sometimes abbreviated SSW (W for Within) and sometimes SSE (E for error).

Even worse, when there are two or more factors, the second one is sometimes called SSB (B for factor B).

And SST! Sometimes the T is total and sometimes it is treatment.

$\endgroup$
  • $\begingroup$ Im sorry Sir, but I dont know. Im just starting to learn advance statistics. Im reading from this book, Intermediate Statistics for dummies. The book says that for One way ANOVA, SSTO = SSW + SSB. So i just assumed that SSE is not there. If you could suggest me a better book to read, that will be helpful :) $\endgroup$ – Rishi Sharma Jan 3 at 18:05
  • $\begingroup$ SSB is SSE. The labeling does get confusing. I will edit my answer. $\endgroup$ – Peter Flom - Reinstate Monica Jan 3 at 22:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.