tab by order of frequency across multiple variables

[tim bradshaw]

Hi there
If I want to tab values by order of frequency for one variable, I simply do this:
tab var1, sort

But what if I want to tab across multiple variables var1 to var7?
One way might be to restructure my data from wide to long, so that I have var1 to var7 listed as observations.
Can anyone suggest a nice alternative that doesn't involve having to restructure my data?

Any thoughts much appreciated!

With thanks
Tim

[tim bradshaw]

To clarify, I simply want a list of values by order of frequency, across var1-var7.

[Clyde Schechter]

Well, I don't understand what "a list of values by order of frequency, across var1-var7" means. Can you show an example of what the results you want might look like?

[tim bradshaw]

Apologies.

OK, so my dataset looks like this:

var1	var2	var3
1	5	3
4	4	3
3	7	4
5	8	3
6	2	3

I want to know which values come up most frequently across all these variables, so I get a list of results like this:

3 comes up 4 times
4 comes up 3 times
5 comes up 2 times
etc

I hope that clarifies!

Thanks
Tim

[Nick Cox]

Look at tabm from

ssc inst tab_chi

[tim bradshaw]

...except that, in my example, the value "3" actually comes up 5 times (evidently I can't count!)

[Carole J. Wilson]

A couple of ways to get this:


1. Series of count commands

clear
input float(var1 var2 var3)
1 5 3
4 4 3
3 7 4
5 8 3
6 2 3
end

count if var1==1 | var2==1 | var3==1
count if var1==2 | var2==2 | var3==2

2. Temporarily stack your variables into one long variable and tabulate

clear
input float(var1 var2 var3)
1 5 3
4 4 3
3 7 4
5 8 3
6 2 3
end

preserve
stack var1 var2 var3, into(_onevar)
list
tab _onevar
restore

[Carole J. Wilson]

Nick Cox -- tabm is a neat solution!

[tim bradshaw]

Thank you all very much.

I went for a slightly modified version of Carole's second suggestion, as follows:

preserve
stack var1-var7, into(tempvar)
tab tempvar, sort
restore


I'd not used the "stack" command before, but it was a neat alternative to "reshape" on this occasion.

I experimented with tabm, but this command didn't seem to like the large number of different values in my data, nor was I able to get it to sort the values by frequency.

Thanks again
Tim

[Nick Cox]

As tabm is an implementation of precisely the strategy you adopted, there should be no real problem.

tabm by default will put the distinct values as columns; in your case you really need that to be rows. That's why there's a transpose option.

 
tabm var1-var7, transpose