Bonferroni/Tukey's-how to do it in Excel

P

Per Madsen

I'm afraid I need a little help on this!

I've done a series of measurements on tissues from the Green Shorecrab.

The test specimens were divided into four groups (A->D), with 12
individuals in each. The crabs in each group were destroyed and 8
different tissue sample's were taken from each individual. An average
were calculated for each group. I then performed an ANOVA test to see if
there were differences between the groups in regards to metal content in
the tissues. The test showed that there indeed were differences between
the groups (p < 0.05). The task now, is to determind which groups shows
a significant difference from one another (is it A and B, A and C,
or...ect), for every type of tissue. This can be done with a Bonferroni
test or a Tukey's test. Unfortunately these tools are not included in
the Data Analysis Toolpack. So, my question is: how can I perform a
Bonferroni or Tukey's test in Excel?

E.g.

Tissue: Gills

A B C D


12 13 12 12
.. . . .
.. . . .
..
..
..
..
..
..
..
..
..
13 12 14 15
--------------------------------------------------

Average:
12 13 14 15

P.S. Sorry 'bout my broken english :)

Thx in advance!!!

Per Madsen, Denmark

www.madsen.blogdrive.com
 
D

David J. Braden

Per,
John McGimpsey alerted me to your post.

Sorry about my broken Danish <bg>.
My stat books are still in storage; I've never heard of the Bonferroni
test, forgot the Tukey test (I'm a resampling sort of guy when it comes
to "frequentist" , as opposed to Bayesian, stats).

Frankly, you should be glad those tools aren't in the ATP, because if
they were, they would probably be as screwed up as the other stat tools
in the ATP.

I suggest you repost this in microsoft.public.excel.misc, where you are
much more likely to catch the eye of some folks who either know this off
the top of their head, or can simply reach over to their bookshelf for
the solution, as I would normally be able to.

On the other hand, you can research the underlying formulas yourself,
realizing, that for your data set, Excel's matrix functions will more
than suffice. That strikes me as more worthwhile than the suggestion
above, in the long run.

Also, if you plan on doing this sort of thing a lot, I suggest you look
into R, which is a very, very powerful, and free(!) open-source version
of S. It has a number of versions specific to the Mac's different
operating systems, most of which are pre-compiled. You can run it, for
example, off of Classic, and there are several variants that take
advantage of different things available under OS-X (pure Unix/gnome,
Quartz interface, etc.) Look for it under SourceForge, or Google search
for "R stat" or something like that. The Mac side of the project is
extremely active. It will provide not only the analyses, but give you
*far* better graphical output: it is a real statistics "environment".

Sorry to not be of more help.
Tschuss!
Dave Braden
MVP - Excel
 
P

Per Madsen

David J. Braden wrote:

Hi David
I suggest you repost this in microsoft.public.excel.misc...
Click to expand...
On the other hand, you can research the underlying formulas yourself,
realizing, that for your data set, Excel's matrix functions will more
than suffice. That strikes me as more worthwhile than the suggestion
above, in the long run.
Click to expand...

I think I'll do both. But you're absolutely right. In the long run it
indeed would be extremely valuable
to be able to take a given formula and implement it into Excel myself. I
do have the underlying formulas for
Bonferroni and Tukey,- I'll see if I can figure out to how to use them
in Excel (I'm not an Excel "power-user").
Also, if you plan on doing this sort of thing a lot, I suggest you look
into R, which is a very, very powerful, and free(!) open-source version
of S. It has a number of versions specific to the Mac's different
operating systems, most of which are pre-compiled. You can run it, for
example, off of Classic, and there are several variants that take
advantage of different things available under OS-X (pure Unix/gnome,
Quartz interface, etc.) Look for it under SourceForge, or Google search
for "R stat" or something like that. The Mac side of the project is
extremely active. It will provide not only the analyses, but give you
*far* better graphical output: it is a real statistics "environment".
Click to expand...

Great,- I've been looking for a good stats program for the mac for quite
some time now. As a
student I don't have the finansial means to buy expensive programs, so
an OpenSource app.
would certainly be great. I'll look into it right away....
Sorry to not be of more help.
Tschuss!
Dave Braden
MVP - Excel
Click to expand...

Thank you for taking you the time to reply to my post!!

kindly

Per Madsen
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Similar Threads

Bonferroni/Tukey's-How to do it in Excel 3
Problem doing t-tests in Excel 6
How do I set up my data extraction 1
Data grouping and averaging in excel 2003 1
Data grouping and averaging questions in excel 2003 0
How to lock records until specific conditions are met? 1
How do I graph a pivot table of sparse values? 0
Creating a Macro 9

Top