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

  • Clustering standard errors, panel data, when households move to different clusters over time

    I have a household panel dataset, with a two stage sample design, first at village and then household. If no households moved out of villages, I would cluster at the village level. However, households move to different villages over time, and I am unsure of whether to still cluster at the (original) village cluster. Any advice?

  • #2
    Welcome to Statalist. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stat output, and sample data using dataex. Knowing whether you have more than one observation per household is essential to answering your question.

    There are some debates about the level of clustering. Some would argue that clustering at a lower level (household) may make sense but is not robust to differences at the higher level (village). It might also depend on how many households change villages - if it is very few, you might just go with village.