Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Constrain mean of latent variable in GSEM

    Dear Stata users,

    I am using GSEM to run a probit model with a binary dependent variable (y), a set of observed covariates (x1, x2, x3) and a latent covariate (LV). The LV has one indicator (ind), measured on a 5-point ordinal scale. x1, x2 and x3 are also covariates in a linear equation of LV. See code below.

    Code:
    gsem (y <- x1 x2 x3 LV, family(bernoulli) link(probit)) ///
         (LV <- x1 x2 x3) ///
         (ind <- LV, family(ordinal) link(probit)) ///
         , latent(LV)  cov(e.LV@1)
    I manage to constrain the variance of LV to 1 and would like to constrain the mean of LV to 0. I have experimented with the means() option for LV by adding

    Code:
    , means(LV@1)
    but receive an error message: "invalid mean specification; LV does not identify a latent variable"

    How can I constrain the mean of LV?

    Many thanks,
    Tobias

    Last edited by Tobias Borger; 10 Oct 2018, 04:00.

  • #2
    Correction: To constrain the mean to 0 I have added:
    Code:
    , means(LV@0)

    Comment

    Working...
    X