I’ve fit the model and played with it a bit, and I believe the issue is that you fit the model and created the predictions based on the centered variable version weight_mc while in your plot you are sometimes calling for it to plot the original variable weight.
I’ve modified your plotting code to this:
ggplot() +
geom_abline(intercept=154.61, slope=0.9, color="red")+
geom_ribbon(
aes(ymin = .lower, ymax = .upper, x = weight_mc),
data = predicted_slopes_qi,
alpha = .25
) +
geom_line(
aes(x = weight_mc, y = .prediction),
data = predicted_slopes_qi
) +
geom_point(
aes(x = weight_mc, y = .prediction),color="skyblue2",pch="*", size=3,
data = pdraw%>%sample_n(100)
) +
geom_point(
aes(x = weight_mc, y = height),
data = data,
alpha = .25
) +
geom_ribbon(
aes(ymin = .lower, ymax = .upper, x = weight_mc),color="salmon",
data = linpred,
alpha = .5
)+
labs(x = "Weight (centered)", y = "Height")
Which outputs the following plot where the oringal geom_abline() matches up with the model predictions.
