Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
Philip Ender <[email protected]> |

To |
[email protected] |

Subject |
Re: Re: Re: st: RE: SEM with bootstrapping for analysis of mediation |

Date |
Wed, 11 Jul 2012 22:34:36 -0700 |

Oops, my posting with subject "st: Re: question about graphs" should have been "Re: Re: Re: st: RE: SEM with bootstrapping for analysis of mediation". I am reposting with the correct Subject lis "Iyer, Neeraj N" <[email protected]> wrote: Thank you. I did not want to assume that the direct and indirect effects would be adjusted for the covariates; your note is helpful in that regard. However, in the command line sem (MV <- IV CV1 CV2 CV3 CV4 ) (DV <- MV IV CV1 CV2 CV3 CV4), I was unable to fathom how STATA would know which one of IV/CV1/CV2/CV3/CV4 is the true Indep. Var. In other words, since " IV CV1 CV2 CV3 CV4" are common to both paths how can STATA differentiate between the variables, which will answer the question, "Whose indirect effect are we observing"? Hence, I was looking for bootstrap CI for the indirect effect of each individual variable, or a partitioned bootstrap CI. Is this partitioning plausible? ------ Dear Neeraj, Okay, now I can see what you're asking and I also see that the -indireff.ado- program needs to be modified due to the four covariates. The program get a vector of direct, indirect, and total effects for each bootstrapped sample, so the number of covariates affects the position of the coefficient of interest. With four covariates you will need to change the 3 to 7 in following lines return scalar indir = el(bi,1,7) return scalar direct = el(bd,1,7) return scalar total = el(bt,1,7) These coefficients will give you the indirect, direct and total effect for the IV on the DV. The easiest way to figure this out is to rum the -sem- and -estat effects- the normal way without bootstrapping and look at the vectors r(indirect), r(direct) and r(total) I hope may explanation is clear. -- Phil Ender UCLA Statistical Consulting Group * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/statalist/faq * http://www.ats.ucla.edu/stat/stata/

- Prev by Date:
**Re: st: estadd with margins** - Next by Date:
**RE: st: OLS pooled regression** - Previous by thread:
**Re: Re: st: RE: SEM with bootstrapping for analysis of mediation** - Next by thread:
**st: OLS pooled regression** - Index(es):