Announcement

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

  • Randomization - Balance check table

    Hi everybody,
    I have an easy question to ask, perhaps silly, but I'm really struggling to solve it!
    I am working with the Dehejia-La Donde paper about the NSW project, and I have to build a summary table checking for balance in covariates after randomization.
    In particular, I have to construct a table checking for balance across treatment and control, providing for each variable the mean for treated and controls with respective standard deviations, and the difference in means, with appropriate standard errors.

    I tried to build the table using tabstat or summarize, but still I cannot include the difference tn means.
    How can I do this?

    Thanks for the kind help!

  • #2
    your request is not completely clear to me; if you are asking for a descriptive table, with p-values,, see my response in the following thread:
    http://www.statalist.org/.../1371934...s-by-cross-tab

    if, however, balance is your main interest you need to decide what your criterion (criteria if more than one) for balance is; do not use p-values as they are affected by sample size as well as by differences; there are several routines in Stata, some user-written for check balance; if this is your real interest, I would start with (1) a search of this forum as this has been discussed before by others or (2) use
    Code:
    hsearch balance

    Comment


    • #3
      Please check out the module 'orth_out'. (ssc install orth_out)

      https://ideas.repec.org/c/boc/bocode/s457800.html

      Comment


      • #4
        Thanks for the replies!
        Originally posted by Rich Goldstein View Post
        your request is not completely clear to me; if you are asking for a descriptive table, with p-values,, see my response in the following thread:
        http://www.statalist.org/.../1371934...s-by-cross-tab
        The link is not working! By the way, I'm looking for a table with the means of some control variables with their respective standard deviations, both for treated and control group. Then, I also need the differences of those variables. The outcome should be similar to this table, diving the means for each variable for both treated and control (and difference)
        Click image for larger version

Name:	Cattura.JPG
Views:	1
Size:	43.7 KB
ID:	1380444

        Comment


        • #5
          I can't read your insert; please read the FAQ

          sorry about the bad url; go to google and enter:
          Code:
          goldstein sumtable site:statalist.org
          and go to #2 in the first thing found

          Comment

          Working...
          X