Get the same results with bonferroni and LSD

[msherman]

Dear List:  I just performed a factorial  (2 x 2)ANOVA (Using GLM) and requested some comparisons of the interaction cell means using LSD.

   /EMMEANS = TABLES(Spirituality_Level*Group) compare(spirituality_level) adj(lsd)

I then repeated the above but changed LSD to Bonferroni

   /EMMEANS = TABLES(Spirituality_Level*Group) compare(spirituality_level) adj(bonferroni)

and got the exact same results. In particular the comparison of two specific means provided in the outcome probability of p = .047. for both LSD and Bonferroni.

Next I decided to alter the factorial and make it into one way with four groups. I now used the ONEWAY procedure and requested post hoc tests of LSD and Bonferroni.  Results for the LSD come out the same as the Factorial with p= .047, while the results of the Bonferroni come out with a p value of .18.  If I ran the Bonferroni test and never ran the LSD within the GLM procedure I would have concluded that I had obtained significance.  Has anyone had this problem before? Is it a bug or is it possible that I am doing something wrong.?

martin sherman

[Leitzel, Jeffrey]

Martin,
Unless I am missing something here...
For a factor with two levels there is no adjustment to make for multiple
comparisons, because there are only two groups to compare, which the
main effect of the ANOVA does. A four group ONEWAY anova essentially
loses the ability to assess the possible interaction of the variables.
Best, Jeff
Jeffrey D. Leitzel, Ph.D.
Assistant Professor, Department of Psychology
Office: McCormick 2123
Bloomsburg University
400 East Second Street
Bloomsburg, PA  17815
Office Phone:[phone],fax:[phone]
Off Hrs:MWF 8:20 am - 10 am
Alt. Office (Tues): 570 348-6100 ext:3216

-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Martin Sherman
Sent: Friday, April 20, 2007 11:37 AM
To: [hidden email]
Subject: Get the same results with bonferroni and LSD

Dear List:  I just performed a factorial  (2 x 2)ANOVA (Using GLM) and
requested some comparisons of the interaction cell means using LSD.

   /EMMEANS = TABLES(Spirituality_Level*Group)
compare(spirituality_level) adj(lsd)

I then repeated the above but changed LSD to Bonferroni

   /EMMEANS = TABLES(Spirituality_Level*Group)
compare(spirituality_level) adj(bonferroni)

and got the exact same results. In particular the comparison of two
specific means provided in the outcome probability of p = .047. for both
LSD and Bonferroni.

Next I decided to alter the factorial and make it into one way with four
groups. I now used the ONEWAY procedure and requested post hoc tests of
LSD and Bonferroni.  Results for the LSD come out the same as the
Factorial with p= .047, while the results of the Bonferroni come out
with a p value of .18.  If I ran the Bonferroni test and never ran the
LSD within the GLM procedure I would have concluded that I had obtained
significance.  Has anyone had this problem before? Is it a bug or is it
possible that I am doing something wrong.?

martin sherman