Does the colon operator currently support only two-way interaction? I tried something like x1:x2:x3 and it does not work:
```
sigma <- matrix(.4, 4, 4)
diag(sigma) <- 1
set.seed(84324)
dat <- MASS::mvrnorm(100, rep(0, 4), sigma)
colnames(dat) <- paste0("x", 1:4)
mod <-
"
x4 ~ x1 + x2 + x3 + x1:x2 + x1:x3 + x2:x3 + x1:x2:x3
"
fit <- sem(mod, dat)
parameterEstimates(fit, remove.nonfree = TRUE)
```
The output:
```
> parameterEstimates(fit, remove.nonfree = TRUE)
lhs op rhs est se z pvalue ci.lower ci.upper
1 x4 ~ x1 0.041 0.095 0.436 0.663 -0.144 0.227
2 x4 ~ x2 0.232 0.115 2.014 0.044 0.006 0.458
3 x4 ~ x3 0.266 0.105 2.526 0.012 0.060 0.473
4 x4 ~ x1:x2 -0.054 0.084 -0.640 0.522 -0.219 0.111
5 x4 ~ x1:x3 -0.034 0.106 -0.326 0.745 -0.242 0.173
6 x4 ~ x2:x3 -0.005 0.111 -0.046 0.963 -0.222 0.212
7 x4 ~~ x4 0.737 0.104 7.071 0.000 0.533 0.941
```
-- Shu Fai