Interaction Terms controlling Influence case in lmer

Since your DV is a proportion, it may not really be normally distributed in
it's original form. Perhaps you should transform the dependent variable
before running the analysis. Using an arcsine square root transformation may
help (you should be able to find details via Google). 


Steven J. Pierce
Associate Director
Center for Statistical Training & Consulting (CSTAT)
Michigan State University
178 Giltner Hall
East Lansing, MI 48824
Web: http://www.cstat.msu.edu


-----Original Message-----
From: jungck [mailto:jungck at gmail.com] 
Sent: Sunday, March 28, 2010 9:36 PM
To: r-sig-mixed-models at r-project.org
Subject: [R-sig-ME] Interaction Terms controlling Influence case in lmer

Dear all,

I am working on interaction terms between two levels.
But when I plot them with the package languageR and following function, I
could have "impossible" predicted values, which is going beyond my dependent
variable.
My DV ranges 0 to 1 in continuous variable. But lines appear above 1.

lmerPlotInt.fnc(model.after.control.influence, "x", "y", "y:x",
which="matplot", ylabel="DV", ylim=c(0,1.1))

The issue here seems only to appear when I control influence country by
using the package influence.ME. (Actually there is another issue involved
this package like I couldn't plot after using "exclude.influence" function.
In R, it only makes coordinates without any lines)

When I include influence case(s), I can see those lines within the range of
DV.
Can anyone familiar with lmerPlotInt.fnc let me know what's going on?

And, I wonder if there is another alternative package/function for plotting
interaction terms in mixed effect model using LMER or others.
Thanks much in advance.

-Chang

An embedded and charset-unspecified text was scrubbed...
Name: not available
URL: <https://stat.ethz.ch/pipermail/r-sig-mixed-models/attachments/20100329/13f4422f/attachment.pl>