Question on Noncensored Variables

[jikaczmarski]

Hi there,

I have three goods where only two of them contain zero shares (censoring). One of the variables is never censored. As a result, lines 200 to 205 of quaidsce.ado return an exit,

foreach x of varlist `shares' {
    summ `x' if `touse', mean
    if r(min) > 0 {
        di as error "noncensoring for `x' found"
        exit 499
    }
...

Is this a limitation of the methodology or the program? I read through the 2021 publication and there is a mention in the methodology that,

First, a univariate probit equation 𝑑𝑑ℎ𝑖𝑖 = 𝑧𝑧ℎ′𝜃𝜃𝑖𝑖 ∀ 𝑖𝑖 is estimated for all categories... (page 3)

I'm trying to figure out what to do in the case where I have a partially censored model. Any advice would be much appreciated.