Plotting effects with library(effects) functions

[claravds]

Dear all,

When plotting effects determined by a logistic regression with the allEffects / Effect / effect functions, what are the levels of the other predictors for the plots of predicted probabilities for each variable?

According to the help function, the default defined by the argument typical  ("a function to be applied to the columns of the model matrix over which the effect is "averaged" ") is "mean" - so for interval predictors this will be the mean, but how to interpret "averaged effect" for categorical variables? Is this simply the reference level that is chosen for each variable?

Do I understand it correctly that this can be changed by the given.values argument?

Thank you all for any input!

CV

[Stefan Th. Gries]

"For factors, this choice of a typical value averages each of the
contrasts that represents the factor, which in a linear model is
equivalent to setting the proportional distribution of the fact to the
distribution observed in the data." (Fox & Weisberg 2011:173) Since
you were in the bootcamp: SFLWR2 now uses effect(...) to create what
you got to know as preds.hyp.

HTH,
STG
--
Stefan Th. Gries
-----------------------------------------------
University of California, Santa Barbara
http://www.linguistics.ucsb.edu/faculty/stgries
-----------------------------------------------