topic Re: Randomized Complete Block Design in SAS Programming

Randomized Complete Block Design

JUMMY — Mon, 06 May 2019 10:05:49 GMT

I have a randomized complete design problem where Six litters of hamsters with 2 hamsters from each litter (considered blocks) were available for an experiment examining differences between length of day light on the NI enzyme level (ENZYME). The treatments (DAYLENGTH) of short (1) and long (2) are randomly assigned within litter (LITTERS). How do I test for differences in treatments taking litter into account

I am not sure if my code is set up the code properly to answer the question.

data have;
input litter daylength enzyme;
datalines;

1 short 2.1 
2 short 1.8 
3 short 1.4 
4 short 1.2 
5 short 1.9 
6 short 2.4 
1 long 2.6 
2 long 2.2 
3 long 2.4 
4 long 1.7 
5 long 2.9 
6 long 2.3 
;
run;


proc glm data=have plots=(meanplot(connect));
  class daylength litter;
  model enzyme=daylength litter;
  means daylength / tukey;
  lsmeans daylength / stderr pdiff adjust=tukey;
title 'RCB ANOVA';
run;

Does the above code answer the question on testing the differences in treatments while taking BLOCKS into account?

Re: Randomized Complete Block Design

PaigeMiller — Mon, 06 May 2019 10:39:28 GMT

@JUMMY wrote:

I have a randomized complete design problem where Six litters of hamsters with 2 hamsters from each litter (considered blocks) were available for an experiment examining differences between length of day light on the NI enzyme level (ENZYME). The treatments (DAYLENGTH) of short (1) and long (2) are randomly assigned within litter (LITTERS). How do I test for differences in treatments taking litter into account

I am not sure if my code is set up the code properly to answer the question.
data have;
input litter daylength enzyme;
datalines;

1 short 2.1 
2 short 1.8 
3 short 1.4 
4 short 1.2 
5 short 1.9 
6 short 2.4 
1 long 2.6 
2 long 2.2 
3 long 2.4 
4 long 1.7 
5 long 2.9 
6 long 2.3 
;
run;


proc glm data=have plots=(meanplot(connect));
  class daylength litter;
  model enzyme=daylength litter;
  means daylength / tukey;
  lsmeans daylength / stderr pdiff adjust=tukey;
title 'RCB ANOVA';
run;
Does the above code answer the question on testing the differences in treatments while taking BLOCKS into account?

Yes.

Re: Randomized Complete Block Design

JUMMY — Mon, 06 May 2019 10:56:02 GMT

The overall F-Test shows a p-value of 0.0648 which is insignificant. However, the p-value for the treatment shows a significant difference while block is insignificant. Can I still go ahead and find which treatments are different even though the F-Test suggest otherwise?