[Corpora-List] Help in Applying Appropriate Statistical Test and Its Interpretation

[polyglot at bu.edu]

Hi True Friend,

I would highly recommend that you take a look at:

Joan Bresnan, Anna Cueni, Tatiana Nikitina, and Harald Baayen. 2007.  
"Predicting the Dative Alternation."   In Cognitive Foundations of 
Interpretation, ed. by G. Boume, I. Kraemer, and J. Zwarts.  Amsterdam: 
Royal Netherlands Academy of Science, pp. 69--94.

or at some of the other papers by Bresnan that analyze the dative alternation.

http://www.stanford.edu/~bresnan/publications/index.html

Good luck!

Linnea Micciulla
lmicciulla at comcast.net

----- Original Message -----
From: "Angus B. Grieve-Smith" <grvsmth at panix.com>
To: "Corpora list" <corpora at uib.no>
Sent: Monday, June 28, 2010 9:15:45 AM
Subject: Re: [Corpora-List] Help in Applying Appropriate Statistical 
Test and        Its Interpretation

True Friend wrote:

Good Day to All Copora Members
I am a masters in applied linguistics student, currently working on my 
thesis. The topic of research is the use of ditransitive constructions. 
To authenticate the results I want to apply statistical techniques on 
the research. For example I am trying to see whether there is a 
significant difference in the usage of two alternative ditransitive 
patterns in PWE (Pakistani Written English, the corpus I am working on 
for the research). The alternative ditransitive patterns here mean 
Double Object (He gave me a pen) and To Dative (He gave a pen to me). I 
am pasting the table here, which contains genre names and frequencies 
of all verbs (used ditransitively) in that genre.
Genre D. Object To Dative
ALT 0 4
ART 210 344

     First of all, let me applaud your question.  I think too many 
linguists are reluctant to ask about their statistics.  It's important 
for us to know what these things mean and how they work.  At UNM we 
were required to take at least a semester of statistics, and it helped 
tremendously, but I can tell that we just scratched the surface.  I try 
to check all my tests with a statistician to make sure they're 
appropriate.  If your university has a statistics clinic, I strongly 
recommend a visit.

     I agree with what Adam and Thomas wrote, but I'm going to focus on 
a different aspect, relating to the envelope of variation.  Here's a 
paper I wrote about it!


The Envelope of Variation in Multidimensional Register and Genre Analyses
Author: Grieve-Smith, Angus B.
Source: Language and Computers, Corpus Linguistics Beyond the Word: 
Corpus Research from Phrase to Discourse. Edited by Eileen Fitzpatrick 
, pp. 21-42(22)
Publisher: Rodopi
http://www.ingentaconnect.com/content/rodopi/lang/2006/00000060/00000001/art00003?crawler=true

http://www.grieve-smith.com/Academic/AAACL-grvsmth.060225.pdf

     In this case, correlation tests are not appropriate, because you 
would expect the number of tokens to vary with the total number of 
words in each genre.  Running a correlation test on per-word frequency 
counts is also not appropriate, because these are two different 
strategies for doing the same thing, and you would expect them to vary 
inversely with one another.  The writers are describing events where a 
thing is being given to a person (or similar).  The two constructions 
have the same envelope of variation.

     I think you need a better hypothesis.  It is unlikely that any two 
constructions will occur with comparable frequencies, especially 
constructions that have the same conceptual meaning, so if you find 
that to be true, it doesn't tell you much.  If you are breaking it out 
by genre, does that mean that you expect the percentage of 
ditransitives to vary with genre?  In that case, I think you need to 
figure out which genres you would expect to do what, and why.  Then you 
will have a good hypothesis, and you can find a statistical test based 
on that.

     I hope this helps.

-- 
				-Angus B. Grieve-Smith
				grvsmth at panix.com


_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora



_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora