Prev: How to read a large odd looking file
Next: Importing Excel into SAS and the MIXED= option
From: "Eamonn O'Brien" on 27 Jan 2010 13:08
Hi all

I am attempting to design an unbalanced staggered nested study to estimate
variance components. There are 6 factors A B C D E F with a maximum of
levels of 6 3 2 2 3 2 respectively.

Staggered Nested designs seem to be an alternative to fully nested designs
in that the degrees of freedom are more evenly balanced in the study.
Heckler and Rao (1985) have extended the concept of staggered nested designs
to allow for more than two levels for any factor. There is a particular
design by Heckler-Rao that I want to investigate/simulate. It is
pictured on page 358 fig 16.6 Heckler-Rao five-stage extended staggered
(ESN) design in the book 'Analysis of Variance for Random Models Volume II:
Unbalanced Data' by Hardeo Sahai and Mario Miguel Ojeda. The number of
levels and degrees of freedom for different stages in the given example
are given by [a, 2, 3, 2, 3] and [a - 1, a(1), 2a(2), 2a(1), 2a(2)],
respectively. In general, a five-stage ESN design where the first two stages
have balanced levels consists of a levels at the first stage, b levels at
the second stage, c levels at the third stage, d levels at the fourth stage,
and e levels at the fifth stage. As you can see I am actually trying to
extend it to 6 levels. The corresponding degrees of freedom are a-1, a(b-1),
ab(c-1), ab(d -1), and ab(e-1), respectively

Has anyone experience of designing these types of studies? Is Proc Plan
appropriate and can do it?
Thanks
Eamonn
 | 
Pages: 1
Prev: How to read a large odd looking file
Next: Importing Excel into SAS and the MIXED= option