Hi,

I am so happy to have found this package, thank you very much!

My question is about the generation of "non-stimulus splits" using a (relatively) short list of materials.

generate() is not converging even with very relaxed constraints. I wonder whether I am doing something wrong. The details are pasted below.

Thank you in advance for your help

Xavier

I have a file dxE that may have between 30 and 70 Item (say N = 32 in the example below).

I want to split it into two groups matched for their frequency. The range of frequency values is fairly limited 0.07 - 6.96.

I most often get the message:

n is too large; requested n of 16, but the condition with the fewest members has 13 entries. Ensure n < 13. Will generate as many stimuli as possible.

This is the case even when I use very broad constraints in control_for(), that I expect to not constrain the matching. Why is it that "n is too large"?

Sample code

set.seed(13331)
d = tibble(Item = 1:32, frequency = runif(32, min = .07, max = 6.96) )
           
d %>%
    set_options(id_col = "Item") %>%
    split_random(2) %>%
    control_for(frequency, -100:100) %>%
    LexOPS::generate(n=32/2) ->
    foo