Date: Tue, 25 May 1999 11:00:05 -0400
Reply-To: "Fehd, Ronald J." <rjf2@CDC.GOV>
Sender: "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From: "Fehd, Ronald J." <rjf2@CDC.GOV>
Subject: Re: A few queries about FREQ
Content-Type: text/plain; charset="iso-8859-1"
From: Panos PAPANIKOLAOU [mailto:PapanikolaouP@CARDIFF.AC.UK]
/I would appreciate some comments on proc freq.
/ 1. suppose, the independnt var. is X whilst Y is the
/dependent var. then how i would write the table command:
/ a. tables Y * X ; OR
/ b. tables X * Y ;
Either way you will get the same information, that is, the counts will be
the same for every intersection of the levels of X and Y,
but the report will be sorted by your first variable.
be sure to check that missing are present and accounted for:
tables X * Y /list missing;
note: the parameter 'list' turns off the column and row sums;
if either X or Y are mising, then that summary row will be counted and noted
'Frequency missing = ??' line at the end of the summary.
Ron Fehd the FREQ maven CDC Atlanta GA