Calcite | Level 5

## Split plot design with whole-plot factor in a CRD - allowing for repeated measures

How can I extend the following code for a split-plot design with whole-plot factor (A) in a Completely Randomized Design (CRD) to allow for repeated measures over time at the subplot level (B)?

``` proc mixed data=yourdata;
class subject A B;
model y = A B A*B;
random subject(A);
lsmeans A B A*B;
run;```

Thanks!

5 REPLIES 5
Diamond | Level 26

## Re: Split plot design with whole-plot factor in a CRD - allowing for repeated measures

You really need to explain your design better.

For example, what is the split-plot factor called? Is the repeated measures factor within A, or within A*(split plot factor). What does a completely randomized design have to do with split plots (I thought these were separate designs that can't be used with one another)?

But really, show us (a portion of) your experimental design, with explanation.

--
Paige Miller
Calcite | Level 5

## Re: Split plot design with whole-plot factor in a CRD - allowing for repeated measures

The plots are patients and the sub-plots are the left and right eye within each patient. The whole-plot factor A is a dosing regimen (once vs twice daily) and the sub-plot factor B is dose concentration (high vs low). Individual patients are randomized to dosing regimen and the left eye is randomized to dose concentration level with the right eye receiving the other dose concentration level. Level of myopia is measured in each eye over time, hence repeated measures.

Diamond | Level 26

## Re: Split plot design with whole-plot factor in a CRD - allowing for repeated measures

Sometimes these things require a lot of thought for me (and maybe for you), and discussion.

Patients do not seem to be plots in my understanding. They are crossed with A. Eyes are sub-plot within patient, and the repeated level of myopia measurements are repeated measures.

What do you think?

--
Paige Miller
Calcite | Level 5

## Re: Split plot design with whole-plot factor in a CRD - allowing for repeated measures

I think the patients are plots because each patient is randomized to receive either once-daily or twice-daily dosing of across both eyes.The only possible combinations are (Eye 1; Eye 2):

(Low concentration once daily; High concentration once daily)

(High concentration once daily; Low concentration once daily)

(Low concentration twice daily; High concentration twice daily)

(High concentration twice daily; Low concentration twice daily)

It is not possible to receive, for example:

(Low concentration once daily; High concentration twice daily)

What do you think?

Diamond | Level 26

## Re: Split plot design with whole-plot factor in a CRD - allowing for repeated measures

Patients are nested within treatment regimen (once-daily or twice-daily). Perhaps this is mathematically the same as patients are plots, but I'm not sure.

--
Paige Miller
Discussion stats
• 5 replies
• 956 views
• 0 likes
• 2 in conversation