Timeline for Analysis of interaction with multiple levels in each factor (emmeans in mixed model)
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 11, 2019 at 12:58 | comment | added | Russ Lenth |
Yes, you could do that. BTW, del.eff contrasts are proportional to eff ones, so in s way it doesn’t matter which you use.
|
|
| Sep 11, 2019 at 11:55 | comment | added | Cuenco |
Thanks! This is very helpful. I think your first suggestion compares the treatment effects in between each two environments, but I wanted to say something general about the different effect in 4 vs 1+2+3. Your second suggestion is much more in this direction, but by manually constructing only the comparison of 4 to 1+2+3, I seem to assume that this is the only difference possible, no? Wouldn't I want to compare the effect in each environment to that of all others and see which one is different than the others? Perhaps something like contrast(emm, interaction = c("del.eff","del.eff"))?
|
|
| Sep 11, 2019 at 11:42 | vote | accept | Cuenco | ||
| Sep 11, 2019 at 2:41 | history | edited | Russ Lenth | CC BY-SA 4.0 |
Addendum
|
| Sep 11, 2019 at 1:06 | history | answered | Russ Lenth | CC BY-SA 4.0 |