CODEWITHSUNDEEP
rspsscontrast
Mina
Mina

Reputation: 111

Simple contrasts output from SPSS in R

I estimated a mixed model with the GENLINMIXED command in SPSS. In the model options I specified that I want to estimate simple contrasts for my interaction variable (category*group) by the variable "group". The output of the "simple contrasts" looks like this: enter image description here

I would like to get the same output in R where I estimate the model with glmer from the lme4 package. Does anyone know how to do that?

Upvotes: 0

Views: 66

Answers (1)

Ben Bolker
Ben Bolker

Reputation: 226742

In general you would use the emmeans package for this. Here's an example using lm() and a built-in data set: it should work equally well for a glmer model, substituting ~group | category in the emmeans() statement and ref="0" (or something like that) in the contrast() statement.

library(emmeans)
warp.lm <- lm(breaks ~ wool*tension, data = warpbreaks)
warp.emm <- emmeans(warp.lm, ~ tension | wool)
cc <- contrast(warp.emm, "trt.vs.ctrl", ref = "L")
summary(cc, infer = TRUE)  ## infer = TRUE to get CIs

You can choose the method of multiple comparisons correction using the adjust argument to summary(); I think you might use adjust = "tukey" to match your SPSS results, but I'm not sure.

wool = A:
 contrast estimate   SE df lower.CL upper.CL t.ratio p.value
 M - L     -20.556 5.16 48    -32.4    -8.74  -3.986  0.0005
 H - L     -20.000 5.16 48    -31.8    -8.19  -3.878  0.0006

wool = B:
 contrast estimate   SE df lower.CL upper.CL t.ratio p.value
 M - L       0.556 5.16 48    -11.3    12.37   0.108  0.9863
 H - L      -9.444 5.16 48    -21.3     2.37  -1.831  0.1338

Confidence level used: 0.95 
Conf-level adjustment: dunnettx method for 2 estimates 
P value adjustment: dunnettx method for 2 tests

Upvotes: 3

Related Questions