Re: Type III Sum of Squares
- From: Hong Ooi <hong@xxxxxxxxxxxxxxx>
- Date: Sun, 21 Aug 2005 20:12:14 +1000
On 18 Aug 2005 08:22:03 -0700, ggroups.anoop@xxxxxxxxxxxxxxx wrote:
>How do I calculate type III Sum of Squares by hand? I am using this in
>ANOVA analysis.
>
>Any help will be appreciated. Thanks!
To calculate the type III SS:
1) Fit the model without the term you want the type III SS for. Find the
error SS.
2) Fit the model with the term you want the type III SS for. Find the error
SS.
3) Subtract one error SS from the other. The result is the type III SS.
--
Hong Ooi | Research and Development
hong@xxxxxxxxxxxxxxx | Insurance Australia Group Ltd (IAG)
Ph: (02) 9292 1566 | 388 George St, Sydney NSW 2000
.
- Follow-Ups:
- Re: Type III Sum of Squares
- From: Frank E Harrell Jr
- Re: Type III Sum of Squares
- References:
- Type III Sum of Squares
- From: ggroups . anoop
- Type III Sum of Squares
- Prev by Date: Ordinal probit analysis or PLUM with probit link
- Next by Date: Which kernel to use for data in [0;1]?
- Previous by thread: Type III Sum of Squares
- Next by thread: Re: Type III Sum of Squares
- Index(es):
Relevant Pages
|
