On 6/23/15 4:07 PM, My List wrote:
All,
I am new to the vcd package and new to R too.
Welcome to R and glad you found the vcd package.
1) I have a lickert analysis based data set.
2) I am doing a hypothesis tests on the variables ( like, is there a
relationship between the choice of a Doctor based on the Patients Income ,
that's just one example)
3) I am building contingency tables ( R x C) and running chiqsqr.test on
these tables.
These are mostly 2 x 2 tables. In some cases I am getting very low counts
of the variables in these ( R x C ) tables. I wanted to know at what point
do I need to use
Your questions are reasonable concerns for someone doing this sort of
analysis with possibly small cell counts, but without any details,
you are really asking for statistical consulting help, rather than
R-help, and you would be better off posting your questions to
http://stats.stackexchange.com/
How do I show if there is any association between the variables in the R x
C setup?Should be I using oddratio() or or assocstats() or should I use
Cramers V test for 2 x 2 tables.
odds ratios, Cramer`s V etc. are different ways of quantifying the
degree of association, with different metrics and different
interpretations of the numbers.
Please, advice or lead me to a source.
Perhaps the notes from my old short course on categorical data might be
useful, particularly lecture 2
http://www.datavis.ca/courses/VCD/
More generally, since you`re using vcd, you may find more comfort with
the visual methods fourfold(), mosaic() and friends than with the
numerical summaries.
*Example of Case - Age of patients and If the Dr Justified a Diagnostic
Test on them.*
DrJustifyDgntTstNo
DrJustifyDgntTstYes
Age21-40
1
91
Age41-50
4
30
GtrThan51
3
50
Thanks in advance,
Harmeet
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
