[R-sig-teaching] Rcmdr

[John Fox]

Dear Robert,

OK -- I see the problem. As a general matter, the Rcmdr is structured
around rectangular data sets. This format of the input for t.test()
only works with a rectangular data set if both groups have the same
number of observations.

Regards,
 John

On Fri, 9 Feb 2007 20:49:49 -0500 (EST)
 "Robert W. Hayden" <hayden at mv.mv.com> wrote:
> 
> Thanks.  I guess what was missing was an indication of how Rcmdr
> wanted the data laid out.  Mine was in two columns, and R is
> perfectly
> happy with this if I do "t.test(col1,col2)" so the test is certainly
> "applicable to the...data".  The catch in Rcmdr is that, because the
> menu item is greyed out, you can't get to the dialog box that asks
> for
> data in the format you mentioned, so you can't tell that Rcmdr is
> fussier than R.  This should not be a problem as long as I include it
> in the documentation for my students.
> 
> PS
> 
> I share your preference for how the data SHOULD be laid out, but that
> was not how the data I had was laid out, and I had no trouble with it
> in R and was curious that I could not reproduce my analysis in Rcmdr.
> Thanks again for your explanation.
> 
> Forwarded message:  From: "John Fox" <jfox at mcmaster.ca>
> > 
> > Dear Robert,
> > 
> > The Rcmdr only activates menu items that are applicable to the
> current
> > data set. In the case of a two-sample t-test, the data set must
> include
> > at least one two-level factor and one numeric variable. Does that
> > explain the behaviour you noted?
> > 
> > I hope this helps,
> >  John
> 
> 
> 
>  _______
> |   ^  | 		    
> |     /  		    Robert W. Hayden
> |    |   		    in the old library
> |    |   		    212 Main Street
> |   /    	            P. O. Box 450
> |  |     		  ^ North Troy, VT 05859
> L__L			    (802) 988-2587
>                             http://statland.org/      
> Map of VT                   bob at statland.org
> 
> Communications sent to Plymouth State will not reach me.
> 
> _______________________________________________
> R-sig-teaching at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-sig-teaching

--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario, Canada
http://socserv.mcmaster.ca/jfox/