Open eimichae85 opened 2 weeks ago
Hi, Even after a further search I could not figure this out. Is anybody able to help me solve my issue above (i.e. what command should I use for "transform=...." to backgransform the estimates after a sqrt transformation "^2" does not work)?
Thanks a lot Michael
Can you try
power_of_2 <- function(x) x^2
plot_model(modRob,
vline.color = "red", colors = c("black", "black"),
sort.est = F, show.values = TRUE,
value.offset = .45, value.size = 6, title = "",
dot.size = 5.3, line.size = 0.8,
transform = "power_of_2",
order.terms = c(2, 3, 1, 4, 6, 7, 5), auto.label = FALSE
)
Hi thanks a ton for your reply.
I already tried this approach. Unfortunately, if I use your code the estmiate-values and the corresponding SEs are not aligned (I think the power-transformation is only applied to the SE bars). See image below
Hello, I would like to use the plot_model() command to plot the estimates & SE of the following model
modRob<-lm(sqrt(C50)~Treatment*Site,data=tab)
I would like to get the untransformed/backtransformed estimates (not the sqrt-transformed estimates) in my final plot_model output. How do I do that? Under
help(plot_model)
it says: "Estimates of linear models remain untransformed. Use transform= NULL if you want the raw, non-transformed estimates". However, if I use "transform=NULL" I still get the sqrt-transformed values. Also "transform="^2" does not work.What command do I have to use to back-transform my sqrt-transformed estimates?
Thanks a lot for your help Michael
Below is the code and a dummy data set I am working with