1
$\begingroup$

This is the R output I was given for a split plot analysis. I think I'm on the right track with my interpretation, but I want to make sure.

enter image description here

Q1: What should you look at first?

If the treatment*year interaction is significant. It is marginally significant at the an alpha-level of 0.05 with a p-value of 0.0454.

Now things start to get a little muddy for me...

Q2: How much evidence is there that the year effects change from treatment to treatment?

There is slight evidence at the alpha = 0.05 level that the year effects change from treatment to treatment with a p-value of 0.044. I am using this number because each whole plot (treatment) requires one observation for the RCB analysis, and this number is the average soil moisture (in this particular case) of the subplots (years) in each whole plot.

Q3: How much evidence is there that the treatment effects change from year to year?

Going along the line of thinking of the previous answer, I want to say there is a significant amount of evidence because I'm getting this p-value from the Year line and the p-value is very small (3.68e-11). Is the right line of thinking for both of these questions?

Q4: If you decide that there is NO Treatment by Year interaction, is there evidence that the soil moisture changes from year to year?

I'm not quite sure what to do with this question. I do know that if there is no interaction that main effects comparisons can be made on whole plot and subplot levels; however, I'm not sure what that means with regard to this output...

$\endgroup$
0
$\begingroup$

A note on p-values: a statistician once told me that there's no such thing as "marginal" significance. A result is either significant or it isn't. Our data either allow to reject the null hypothesis (knowing that there's an alpha*100 % chance of us being wrong) or they don't. So for Q1 I'd say that interaction is significant; for Q2 I'd say that there either is or isn't evidence (there is). Same for Q3. Q4 sounds as a "what if" scenario to me. If there were no significant interactive effect but there were a main effect for "Year", then I'd say that there is evidence that the soil moisture changes from year to year. My two cents.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.