AllExperts > Probability & Statistics 
Search      
Probability & Statistics
Volunteer
Answers to thousands of questions
 Home · More Probability & Statistics Questions · Answer Library  · Encyclopedia ·
More Probability & Statistics Answers
Question Library

Ask a question about Probability & Statistics
Volunteer
Experts of the Month
Expert Login

Awards

About Us
Tell friends
Link to Us
Disclaimer

 
 
 
 
About ronny fisher
Expertise
general questions on probability and statistics. please do not send intro prob/stat homework questions.

Experience
have taught probability and stats for 25 years

Education/Credentials
ba in math, phd in stats

Past/Present Clients
federal government, state AG, start-up pharma companies, engineering consulting firms, academic researchers (the list goes on).

 
   

You are here:  Experts > Science > Mathematics > Probability & Statistics > New Studentized Range table?

Probability & Statistics - New Studentized Range table?


Expert: ronny fisher - 11/2/2009

Question
Hi!
Good day!

I just want to ask if the table used for the Tukey's test which is the studentized range table is the same as with the table used for the Duncan's Multiple Range test? It just confuses me because somebody have mentioned that Duncan's Multiple Range test does not use the Studentized Range table instead it makes use of the New Studentized Range tables.

I just want to confirm if this is true. Thanks for your time.

If it will not take much time. Can you please give some links to articles about the new studentized range table? I have been searching the internet for hours with no positive outcome. Thank you.

Answer
mamoy -

i haven't heard of 'new studentized range' tables per se. but there is an article that
appeared in the journal biometrics [pp 671-685] in 1960 by h.l.harter entitled 'critical
values for duncan's new multiple range test'.

tables are a bit old-fashioned these days - one usually uses software to get things like
critical values. usually one needs specialized statistical software for this. for example,
sas can be used to get duncan critical values [see proc glm], as can R - among other
software packages.

however, you should be aware that duncan's procedure does not control error rates properly
and is currently out of favor among statisticians. in this regard, you may want to look at
an article by john tukey in the 1991 journal statistical science [pp100-116], entitled
'the philosophy of multiple comparisons'. [articles in statistical science tend to be somewhat
less technical than in other statistics journals.] there is also a 1992 follow-up article
by tukey called 'where should multiple comparisons go next?', which appeared in a book
called 'multiple comparisons, selection and applications in biometry: a festschrift in
honor of charles w. dunnett' [pp187-208], edited by f.m. hoppe and published by dekker.

ronny  

Add to this Answer   Ask a Question


 
User Agreement | Privacy Policy | Kids' Privacy Policy | Help
Copyright  © 2008 About, Inc. AllExperts, AllExperts.com, and About.com are registered trademarks of About, Inc. All rights reserved.