summary.estimate_slopes() no longer working.

[strengejacke]

m <- lm(Sepal.Width ~ Sepal.Length, data = iris)
modelbased::estimate_slopes(m, trend = "Sepal.Length", backend = "marginaleffects") |> summary()
#> Error in if ((data$Confidence[i] != sig) || ((centrality_signs[i] != centrality_sign) && : missing value where TRUE/FALSE needed

^{Created on 2025-01-16 with reprex v2.1.1}

[DominiqueMakowski]

Ah yes, I should have seen that coming

This is blocked by easystats/parameters#1063

Basically, summary.estimate_slopes() will create "segments" of similar slope (negative, positive, flat) based on coef sign and significance. Since we don't have significance with marginaleffects yet...

Either we fix the above, OR another option would be that in the absence of indices of significance, we only use the sign of the coefficient to make the groups

[strengejacke]

estimate_contrasts() returns p values also for slopes... Or we use the CI

[DominiqueMakowski]

The people have spoken on the most desired easystats feature 🫡

[strengejacke]

Which comment?

[DominiqueMakowski]

easystats/parameters#1063 (comment)