Announcement

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

  • Specifying both crossed and nested random effects in "mixed"

    I have a hierarchical design where politicians where each asked 9 randomly chosen questions from a pool of 18 questions. Therefore, the questions are crossed with politicians. Furthermore, the politicians are nested within districts. So I want to fit crossed random effects for politicians and questions and a nested random effect for politicians within districts.

    Would the following syntax fit such a model?

    Code:
    mixed ... || district: || _all: R.politician || _all: R.question
    Thanks for your help.
    Last edited by Denis Infanger; 23 May 2019, 02:24.

  • #2
    After reading section 9.7 in Rabe-Hesketh S & Skrondal A (2012), I think I want to fit the following random effects (the fixed effects are not shown):

    \begin{equation}
    y_{ijk} = \ldots + \theta_{1i} + \theta_{2k} + \theta_{3ij} + \epsilon_{ijk}
    \end{equation}

    Where i denotes the district, j the politicians and k the questions. I'm still unclear whether the code above will fit these effects. Any ideas?

    Comment

    Working...
    X