[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
st: RE: Random Group Resampling
Suppose you have a variable -group-. Then
egen rank = rank(uniform()), unique
gen group2 = group[rank]
creates such pseudo-groups. (The -unique-
is paranoia, based on a worry that there
might be ties in the random numbers.)
However, you still need machinery
on top of this.
In addition to what you mention,
-permute- might help.
> -----Original Message-----
> From: email@example.com
> [mailto:firstname.lastname@example.org]On Behalf Of
> Gabriel Kaplan
> Sent: 12 May 2004 20:09
> To: email@example.com
> Subject: st: Random Group Resampling
> Does anyone know of a user created ado file to implement Bliese's RGR
> (Random Group Resampling) procedure?
> Essentially, this involves randomly creating pseudo groups of
> equal size
> to the groups in your data by reassigning the observations to groups
> randomly. One then calculates test statistics for the pseudo
> groups such
> as mean and variance and contrasts these with the means and
> variances of
> the actual groups.
> I am interested in running RGR in STATA. Bliese wrote some
> language for S-Plus, but I wanted to try this in STATA.
> Before I go off
> and try to implement this, has anyone done this? More info is
> Paul D. Bliese and Ronald R. Halverson. 2002. "Using Random Group
> Resampling in multilevel research*1: An example of the
> buffering effects
> of leadership climate" in Leadership Quarterly, (13(1):53-68.
> My sense is I will need to run a combination of SIMUL and BSTRAP 1000
> times to get this done. I know that the Boston College
> Archive maintains some
> ado files for random group variance estimators labeled RGROUP
> and RANVAR,
> but I am not sure this is the same thing since it is form the
> literature and RGR is from the organizational literature. Any ideas?
* For searches and help try: