[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: Re: st: create a contingency table
> Thank you Nick, but my question is :
> How to obtain a data table and not to display the results ?
> How, with a simple syntax, create the variables of the second table
> Jean-Benoit Hardouin
> On Tue, 4 Nov 2003 23:05:33 -0000, Nick Cox <email@example.com>
> > Jean-Benoit Hardouin
> >> I search a method to transform a data as :
> >> Item1 Item2 Item3 Item4 0 0 1 2 1 3 0 1
> >> 3 0 2 1
> >> (table of the responses of individuals to four items) in
> >> Item1 Item2 Item3 Item4 Count
> >> 0 0 0 0 125
> >> 0 0 0 1 88
> >> 0 0 0 2 72
> >> 0 0 0 3 52
> >> 0 0 1 0 28
> >> 0 0 1 1 65
> >> 0 0 1 2 42
> >> ...
> >> where 125 is the number of individuals who have respond 0 to each
> > item, 88
> >> is the number of individuals who have respond 0 to the first
> > item and
> >> 1 to the last item, etc.
> >> I don't find how to obtain this kind of table with "reshape" or
> > "egen" (for
> >> example).
> >> I have idea to program that with four "forvalues..." but perhaps
> > exist a
> >> specific command.
> > For a first-principles solution:
> > bysort Item? : gen Count = _N
> > by Item? : gen tag = _n == 1
> > list Item? Count if tag
> > For a canned solution, with extra bells and whistles:
> > ssc inst groups
> > groups Item? , show(f) showhead(Count)
> > whelp groups
_EITHER_ the same answer
bysort Item? : gen Count = _N
by Item? : gen tag = _n == 1
list Item? Count if tag
If you want just to keep the
table as new dataset:
keep if tag
keep Item? Count
See [R] contract.
I guess I would have suggested that earlier
if I had sensed interest in dataset as result,
rather than table as result.
* For searches and help try: