A bit tricky - Input needed 
Given a tibble my_dat created like so:
# Simple function for generating data
mk_dat = function(m, d){
d %>% rnorm %>% list %>% rep(m) %>% return
}
# Dimensions of example data
n_row = 5
n_col = 3
depth = 10
# Create tibble example
my_dat = tibble(1:n_row) %>% select
for( var_name in letters[1:n_col] ){
my_dat = my_dat %>% mutate(!!var_name := mk_dat(m = n_row, d = depth))
}
I.e. basically a tensor disguised as a tibble:
> my_dat
# A tibble: 5 x 3
a b c
<list> <list> <list>
1 <dbl [10]> <dbl [10]> <dbl [10]>
2 <dbl [10]> <dbl [10]> <dbl [10]>
3 <dbl [10]> <dbl [10]> <dbl [10]>
4 <dbl [10]> <dbl [10]> <dbl [10]>
5 <dbl [10]> <dbl [10]> <dbl [10]>
I want to create a new variable, which is the rowwise and elementwise mean, such that the first element of the first list in the new variable will be something like:
mean(c(my_dat$a[[1]][1], my_dat$b[[1]][1], my_dat$c[[1]][1]))
and the second element in the first list will be:
mean(c(my_dat$a[[1]][2], my_dat$b[[1]][2], my_dat$c[[1]][2]))
etc...
I could code something complicated, but am interested in input to a more clever solution?