Split plot plus control

[SAIUPPALA]

Greetings,

I was testing 9 different media for fungus each at 4 different concentrations (1, 2, 3, 4) for their effect on radial growth of fungus. 

                           9 x 4 = 36 combinations each with 3 Replications

 I want to compare how these 36 combinations effect the growth in comparison with control_1 and control_2.

Control_1 is = Media '10'

Control_2 is = Media '11'

Media	Conc	Rep	Radial growth
1	25	1	1.05
1	25	2	0.85
1	25	3	0.8
1	50	1	0.95
1	50	2	0.85
1	50	3	0.9
1	100	1	1.05
1	100	2	0.85
1	100	3	0.85
1	200	1	1
1	200	2	1
1	200	3	1.05
2	25	1	0.8
2	25	2	0.7
2	25	3	0.65
2	50	1	0.7
2	50	2	0.65
2	50	3	0.8
2	100	1	0.55
2	100	2	0.7
2	100	3	0.65
2	200	1	0.75
2	200	2	0.65
2	200	3	0.6
3	25	1	1.15
3	25	2	0.95
3	25	3	1
3	50	1	1
3	50	2	1.1
3	50	3	1.15
3	100	1	1.25
3	100	2	0.85
3	100	3	0.9
3	200	1	0.95
3	200	2	0.95
3	200	3	0.95
5	25	1	0.95
5	25	2	0.9
5	25	3	0.95
5	50	1	0.9
5	50	2	1
5	50	3	0.85
5	100	1	0.85
5	100	2	0.9
5	100	3	0.65
5	200	1	0.95
5	200	2	0.6
5	200	3	0.9
6	25	1	0.75
6	25	2	0.55
6	25	3	0.6
6	50	1	0.5
6	50	2	0.5
6	50	3	0.5
6	100	1	0.4
6	100	2	0.45
6	100	3	0.45
6	200	1	0.5
6	200	2	0.5
6	200	3	0.5
7	25	1	0.65
7	25	2	0.6
7	25	3	0.65
7	50	1	0.6
7	50	2	0.5
7	50	3	0.6
7	100	1	0.55
7	100	2	0.6
7	100	3	0.6
7	200	1	0.55
7	200	2	0.6
7	200	3	0.6
8	25	1	0.4
8	25	2	0.6
8	25	3	0.45
8	50	1	0.4
8	50	2	0.4
8	50	3	0.35
8	100	1	0.4
8	100	2	0.4
8	100	3	0.45
8	200	1	0.4
8	200	2	0.5
8	200	3	0.5
9	25	1	0.4
9	25	2	0.4
9	25	3	0.45
9	50	1	0.4
9	50	2	0.4
9	50	3	0.35
9	100	1	0.45
9	100	2	0.45
9	100	3	0.5
9	200	1	0.4
9	200	2	0.45
9	200	3	0.5
10	0	1	0.65
10	0	2	0.65
10	0	3	0.6
11	0	1	0.75
11	0	2	0.6
11	0	3	0.65

I never used contrasts before. Could someone help me with code?

Thank you,

Sai

[sld]

This paper is a good resource for the technical details

https://support.sas.com/resources/papers/proceedings11/351-2011.pdf

The challenge will be deciding what you want to compare to what, which is context-specific and thus falls to you, the researcher.

If your intent is to compare each of 36 treatment means to each of two control means, then you could consider mean comparisons controlled for Type I error using the Dunnett method, possibly with additional adjustment due to having two controls rather than one.

Good luck!