[R] analyzing qualitative data sets

Bert Gunter gunter.berton at gene.com
Tue Jul 29 16:27:46 CEST 2014


1. If you are asking about statistics, this is the wrong list. Post
here instead: stats.stackexchange.com.

2. If you you are asking about what sorts of statistical analyses are
available in R, check the CRAN task views here:
http://cran.r-project.org/web/views/

3. If you are asking about how to program in R and have not already
done so, please read "An Introduction to R" or R web tutorial of your
choice before posting here further.

Bert

Bert Gunter
Genentech Nonclinical Biostatistics
(650) 467-7374

"Data is not information. Information is not knowledge. And knowledge
is certainly not wisdom."
Clifford Stoll




On Tue, Jul 29, 2014 at 6:01 AM, Sun Shine <phaedrusv at gmail.com> wrote:
> Hello list
>
> I'm just beginning my PhD and am likely to be using lots of surveys in my
> data collection, and am wanting to get my head around the ideas about how
> best to approach the tasks in R.
>
> The data sets I have collected so far for some preliminary practise with are
> made up of the following survey data:
>
> (1) 25 observations x 15 variables of dichotomous nominal (categorical) data
> [basically, yes/ no responses with a couple of missing values]
>
> (2) 25 obs x 14 var of ordinal rank data [5 item Likert-scale, with some
> missing values], and
>
> (3) 23 observations of free text, typically in the form of one sentence or
> statement, and I will be using RQDA for that part.
>
> So far, I have been able to piece together that I can use the Spearman
> method of the wilcox.text for #2 (ordinal data), but have yet to find
> anything that I can do for the nominal data. I was thinking of using
> frequency tables, but I don't seem to be able to find out too much info on
> it/ how to do that.
> Anyway, I have three questions that I'd appreciate members of this list
> taking a swing at for ideas please.
>
> (a) what types of analyses are available to apply to the data types above? I
> have been thinking about MCA using FactoMineR as well as MDS using MASS to
> visualise the data in high dimensional space, but I think that I haven't
> (yet!) figured out how to properly prepare my data sets for these, and most
> texts and tutorials seem to focus mostly on quantitative data analysis.
>
> (b) is there anyway that I can automate the Spearman process so that it
> iterates across the set, otherwise it looks like I may have to manually take
> the two columns and keep comparing pairs until I have correlated all of the
> columns with all of the other columns - so is there anyway that I can
> automate this and get the test statistics and p values dumped in a table for
> summarising?
>
> (c) after using RQDA to code the statements, is it feasible to reintroduce
> those codes back into the data set to explore correlations among the other
> columns and the units of coded text to see what variables co-occur?
>
> Well, thanks for taking the time to read this - and I look forward to any
> thoughts/ suggestions that might help.
>
> Cheers
>
> ______________________________________________
> 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.



More information about the R-help mailing list