Error bars using marginal_effects()

[Edward Patzelt]

Hi brms users,

When a categorical interaction is plotted using marginal_effects(). What is the definition of the error bars? Are these credible intervals for the joint posterior distribution marginalizing for other variables in the model?

Best,

Edward

[Paul Buerkner]

95% credible intervals. It is not really marginalization. See ?marginal_effects for some details in that regard.

--
You received this message because you are subscribed to the Google Groups "brms-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to brms-users+unsubscribe@googlegroups.com.
To post to this group, send email to brms-...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/brms-users/b70a07d2-418d-4dda-acfc-643298637224%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Edward Patzelt]

Thanks Paul!

On Tuesday, April 17, 2018 at 3:08:45 AM UTC-4, Paul Buerkner wrote:

95% credible intervals. It is not really marginalization. See ?marginal_effects for some details in that regard.

2018-04-16 19:02 GMT+02:00 Edward Patzelt <pat...@g.harvard.edu>:

Hi brms users,

When a categorical interaction is plotted using marginal_effects(). What is the definition of the error bars? Are these credible intervals for the joint posterior distribution marginalizing for other variables in the model?

Best,

Edward

--
You received this message because you are subscribed to the Google Groups "brms-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to brms-users+...@googlegroups.com.