topic Re: statistical models in New SAS User

statistical models

always-good — Fri, 04 Aug 2023 15:04:28 GMT

Hello everyone, Please I need an urgent answer!

I have a table from an experiment in which, we tested the effect of the season ( winter: conducted in November and December) and summer: conducted in July and August) on the milk production of two groups of animals reared in the same conditions. Each group is randomly assigned according to season winter (30 animals) and summer(30 animals). Both groups have the same characteristics of animals(age, weight..), so can only test the effect of season. The results are presented in the table below :

My question is which one of the two suggested models is better describing the result;

1/ Mixed model for repeated measurements where y (milk yield)= µ+Cow+Season+samling time+Season*sampling time+error;

2/One-way Anova where y= µ+season+error

Please, I'm counting on the SAS community!

	winter (Nov-Dec)	summer (July-Augst)	SEM	P
Milk yield
Fat
protein

Re: statistical models

PaigeMiller — Fri, 04 Aug 2023 15:54:55 GMT

Your problem description never mentions repeated measures, so its hard to see how model 1 would fit here. Now, if you accidentally left the repeated measures out of the description, well then my response would likely change.

Re: statistical models

always-good — Fri, 04 Aug 2023 16:15:55 GMT

Thank you for your answer, the measurements were repeated for each cow, for each season. The data before analysis per animal. Can we consider the cow as a random effect, (two groups) and the sampling time as a fixed effect but the problem there is no result for the effect of the time of sampling? what do you think, please? Looking forward to your answer!

Re: statistical models

PaigeMiller — Fri, 04 Aug 2023 16:32:04 GMT

I suppose the answer in my mind depends on what you want to learn from this experiment. If you just want averages by time period, as your original table implies, the repeated measures aspect of the study doesn't seem to be relevant. If you want to understand the relationship of the first measurement of a cow to the second measurement of a cow and so on, and see if this time relationship between the measurements on the same cow differs from summer to winter, then you would definitely need a repeated measures design of some sort.

Re: statistical models

always-good — Fri, 04 Aug 2023 16:50:58 GMT

Please, how to see if this time relationship between the measurements on the same cow differs from summer to winter? and can we present the table this way for the mixed analysis?

Re: statistical models

Ksharp — Sat, 05 Aug 2023 09:47:57 GMT

It would be better to post it at Stat Forum:

https://communities.sas.com/t5/Statistical-Procedures/bd-p/statistical_procedures

since you are asking Stat Model.

Maybe @StatDave would give you a good advice .