RANK VARIABLES=PROEVENT (A) /NTILES (4) /PRINT=YES /TIES=MEAN .

Greetings

Jan

-----Original Message----- From: SPSSX(r) Discussion [mailto:SPSSX-L@LISTSERV.UGA.EDU]On Behalf Of Arnaldo Ferrari Nasi Sent: Wednesday, July 28, 2004 11:49 AM To: SPSSX-L@LISTSERV.UGA.EDU Subject: automatic recode using quartiles

Hi List,

I have have a discrete numeric variable that ranges from -5 to 8.

PROEVENT Valid Cum Value Label Value Frequency Percent Percent Percent

-5 37 ,3 ,3 ,3 -4 83 ,6 ,6 ,9 -3 372 2,8 2,8 3,7 -2 349 2,6 2,6 6,3 0 316 2,4 2,4 8,6 1 888 6,6 6,6 15,3 2 3121 23,3 23,3 38,5 3 2590 19,3 19,3 57,8 5 483 3,6 3,6 61,4 6 2464 18,4 18,4 79,8 7 2245 16,7 16,7 96,6 8 461 3,4 3,4 100,0 ------- ------- ------- Total 13409 100,0 100,0

Percentile Value Percentile Value Percentile Value

25,00 2,000 50,00 3,000 75,00 6,000

Valid cases 13409 Missing cases 0

I want a new variable in 4 classes (0, 1, 2, 3) obtained recoding the first one with quartiles. As new cases are daily added in the dataset the quartiles could change, so I would do this job automatically. Spss has to find himself the quartile values and create a new variable using these recoding points.

Anyone could help? I use Spss 6.0.

TIA

Arnaldo.

Arnaldo Ferrari Nasi Facolta' di Scienze Politiche Universita' degli Studi Genova - Italia, EEC www.unige.it