bayestestR: how to plot a different null for a probability of direction?

Modeling
1

I haven’t figured out how to plot probability of direction with a changed “null” line.

NULL = 0

posterior <- distribution_normal(10000, 0.4, 0.2)

pd1 = p_direction(posterior, null = 0)

plot(pd1)

Probability of Direction: 0.98

image

NULL = 0.5 gives the same plot

pd2 = p_direction(posterior, null = 0.5)

plot(pd2)

Probability of Direction: 0.69

image

How to get a plot with a “null-line” on 0.5?

2

Interesting issue: looking at the R code for the p_direction() function, it seems that when method = "direct" (the default) then null = 0 (see lines 142-159 of bayestestR/p_direction.R at df0562c1919fbad241dc46c912ed9e18e2fb5695 · easystats/bayestestR · GitHub). I may be misreading that or missing a place where the null value gets overwritten somewhere else

Does it work if you use a different method for the posterior: e.g., p_direction(posterior, method = "kernel", null = 0.5)?

-----EDIT-----
Just realized that your question and issue is with the plot function, not the actual p_direction() function… that seems like an issue with the see package then, and I don’t see anywhere in that language for the null value (see/plot.p_direction.R at 6434c554799bc4cdc98874559fec970faab42bb0 · easystats/see · GitHub). You may need to create the desired plot in ggplot2 instead