lme4 formula to test levels under an independent variable

Question

I am running a linear mixed effects model in R using (lme4). I have two independent variables: word types (five levels) and lexicality (two levels), and one dependent variable; reaction time. I used the following formula to look at the overall interaction between the dependent variables.

mydata.mod1=lmer(RT~lexicality*wordType*(1|Item)+(1|Subject), mydata)
summary(mydata.mod1)

My question now is about the possibility of looking at each level under word types separately and compare it to lexicality based on reaction time. For example, I want to take Type Two (under word types) and compare it to lexicality. What terms should I include to run this type of analysis. I have tried the following formula, but it did not show what I needed:

mydata.mod2=lmer(RT~wordType*(1+lexicality|Subject)+(1|Item),mydata)

Is there something I missed?

Answer 1

I cannot add this as a comment, because it is too long. It should be slightly helpful to get started at least.

You have missed quite a lot =) Terms in brackets are the random effects. (1+...|...) means that you make a random slope model together with random intercept as opposed to just a random intercept model (1|...). Use the ranef() command to see the difference.

Here are couple of useful references you should read before doing the analysis (unless you have done so already):

• Winter, lme4 tutorial
• Baayen et al. 2008
• Barr et al. 2013

All of them are written in a language understandable for non-statisticians. See especially the difference of ANOVA and mixed model equations from Barr et al. 2013.