[R] when all I have is a contingency table....

[(Ted Harding)]

On 07-Jul-10 20:39:32, Andrei Zorine wrote:
> Hello,
> I just need a hint here:
> Suppose I have no raw data, but only a frequency table I have, and I
> want to run basic statistical procedures with it, like histogram,
> descriptive statistics, etc. How do I do this with R?
> For example, how do I plot a histogram for this table for a sample of
> size 60?
> 
> Value   Count
> 1       10
> 2       8
> 3       12
> 4       9
> 5       14
> 6       7
> 
> 
> Thanks,
> A.Z.

You could use the data with barplot(O) -- see '?barplot' -- to emulate
a histogram; or you could do it with your "bare hands" by using the
data to draw your own boxes.

Another approach (which is slightly "naughty") is along the following
lines. First a "dummy histogram" object H is created, then its main
components (for your purpose) 'breaks' and 'counts' are matched to
the data; and then you plot it:


  H <- hist(1,plot=FALSE)
  H$breaks <- 0.5+(0:6)
  H$counts <- c(10,8,12,9,14,7)
  plot(H,main="Example")

Note, however, that if you inspect what is in H by simply entering

  H

you will find that the components

  $intensities
  $density

have not been properly set.

Nevertheless, the above is a relatively painless way of getting
a standard histogram plot from such data.

Ted.


--------------------------------------------------------------------
E-Mail: (Ted Harding) <Ted.Harding at manchester.ac.uk>
Fax-to-email: +44 (0)870 094 0861
Date: 07-Jul-10                                       Time: 22:26:05
------------------------------ XFMail ------------------------------