Announcement

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

  • How to make two samples have similar characteristics? weighting using mixed logit

    Hi all,

    I have two samples, cancer group and non-cancer group. I aim to compare whether people have different preferences for treatment between cancer and non-cancer groups using discrete choice experiment. Before performing analysis, I need to make sure social economic status (SES) between the two groups are similar (not statistically significant). However, I found income and education are significantly different. I want to ask how to make SES of two groups similar before doing analysis?

    I am thinking of using weight, such as propensity score matching. However, as I perform analysis on each group rather than pooling two samples together, generating weighting seems useless? Any suggestions?

    Many thanks,

    Ni



  • #2
    Coarsened Exact Matching will do the trick.
    Code:
    search cem
    cem produces weights (cem_weights). Then estimate model with [aw=cem_weights].

    Comment

    Working...
    X