Open coverton-usgs opened 3 years ago
Glad the package is (potentially) useful for your birds!
It's a bit tricky to tell what's going on with the axes without a minimal working example - and I have very little bandwidth at the moment to play with this too much. But if you just want to illustrate that the two fits are very similar (as the stats certainly show) you can just fit the separate models and use the add = TRUE
tag, e.g.
fit1 <- fit_cyclomort(T.morts.pop1, n.seasons = 1)
fit2 <- fit_cyclomort(T.morts.pop2, n.seasons = 1)
plot(fit1, hazcolor = "blue")
plot(fit2, add = TRUE, hazcolor = "red")
That's pretty simple and should be robust.
Also - if you want to tweak the plots in any other way, you can check out the guts of the plot.cmfit
and the - much fussier - plot.cmfactorfit
functions.
Maybe if you mess around, you can see where there might be an error?
Wonderful package! I am so glad to have just found it.
I am working on contrasting seasonal survival patterns of two populations of birds. When summarizing the cmfactorfit object there is no significant evidence of a difference between groups which I want to show with a graphic:
However, the resulting plot places the "Null" model centered on a y-axis scaled from 0-1 using fit = "both" or fit = "null" The "Alt" plots appear to be placed approrpiately for the 0-1 y axis when plotting fit = "both"
However when plotting fit = "alt" the y axis is only scaled for first factor.
The above graphic should have lines that almost completely overlap (with factor #2's confidence intervals covering the whole range of the y-axis
If I plot with a specified axis range it also doesn't match the figure below is with:
Sorry I know its bad form to post two issues at once but I didn't realize the null model wasn't scaled with appropriate ylims labels until writing this.