[R] Smartest way to evaluate question forms

[John Kane]

It really in not an R question.  It's much more complicated.

You need to consult with a subject matter specialist and a statistical consultant for this. If you do not have access to a statistical specialist you might want to ask for advice on the news group  sci.stats.consult.  

I'd suggest having the equivalent of an introduction and methods section for the study written up and available (say on a handy posting site like media fire http://www.mediafire.com/ or mytempdir.com http://www.mytempdir.com ) so that anyone reading your posting has some idea of what is the purpose of the study and the general details of what was done.





--- On Fri, 8/1/08, vlasto <vlasto.pis at gmx.net> wrote:

> From: vlasto <vlasto.pis at gmx.net>
> Subject: [R]  Smartest way to evaluate question forms
> To: r-help at r-project.org
> Received: Friday, August 1, 2008, 10:56 AM
> Hi,
> I'm trying to help a friend who is doing a thesis in a
> nurse college, to
> evaluate medical question forms.
> There are about 30 questions giving more than 110
> parameters to describe
> each responding person's (gender, health etc.) and
> there are about 120
> question forms to evaluate.
> I have basically 2 questions.
> 1. What to search for.
> 2. How to evaluate it statisticaly.
> 
> As for No. 1. I have these ideas. 
> To search for significant groups. Meaning, that i would
> like to find all
> "significant" groups that have a certain criteria
> in common. F.e. All men,
> that have a good doctor patient relationship. The idea is
> to fix 1,2,3,4 or
> five parameters f.e. white divorced men in their 60s and to
> look for any
> other significant parameters (meaning one or multiple) they
> have in common
> (where I can set some significance boundary)
> 
> Later on, i would like to look up question forms with the
> highest number of
> common parameters and find the parameters with the highest
> and lowest rate
> of divergence.
> Eventually it might be interesting to look for some
> correlations between 2
> and more parameters.
> 
> As for No. 2 I would like to know if there is a R module
> having performing
> this kind of tasks.
> I think the problem could be analyzed by treating all the
> params as a
> binomial tree and then measure length and repetition of
> certain path
> segments
> 
> I have written a simple prog in VBasic to do the first part
> of the analysis,
> but i would be thankful for any hint or advice regarding
> this problem,
> especially any info about existing solutions with R.
> -- 
> View this message in context:
> http://www.nabble.com/Smartest-way-to-evaluate-question-forms-tp18776233p18776233.html
> Sent from the R help mailing list archive at Nabble.com.
> 
> ______________________________________________
> R-help at r-project.org mailing list
> 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.