Datasets:    Cases by Variables
 
 
 
 

In SAS (or SPSS) the basic data unit is the case.
( This is a leftover from the era of Computer Cards)
Data is stored in
 

• SAS data files.

• Text files.

• Locations on the network. Suppose you work with data from Hong Kong Telecom, the census bureau, Network applications, Market research.  Data is stored in large databases and you access the data through the network

• Relational databases. They can be used in SAS by combining SAS Datasets.

Next we will review basic SAS procedures by example.
 
 
 

• Creating Datasets in SAS.

1. From a file

EXAMPLE:  CRIME DATA
VARIABLES :
MURDER RAPE ROBBERY ASSAULT BURGLARY LARCENY AUTOTHFT REGION
DATA  FILE : crime.dat
 1.500 7 12.600 62 562 1055 146 1
 2     6 12.100 36 566  929 172 1
 1.300 10.300 7.600 55 731 969 124 1
         ...............
 
 
======= SAS CODE =========

OPTIONS PS=55 LS=80;
 

DATA CRIME;
filename myf 'crime.dat';
Infile myf;

INFILE 'crime.dat';
INPUT MURDER RAPE ROBBERY ASSAULT BURGLARY LARCENY AUTOTHFT REGION $;
CARDS;
 
 
 

 

2. Dataset inside the program file:
 
 

DATA CRIME;
INPUT MURDER RAPE ROBBERY ASSAULT BURGLARY LARCENY AUTOTHFT REGION $;
CARDS;
 1.500 7 12.600 62 562 1055 146 1
 2     6 12.100 36 566  929 172 1
 1.300 10.300 7.600 55 731 969 124 1
         ...............
 ;
 3. Open a file from DBASE 4
options linesize=80 pagesize=55;

filename product 'ULTIMVN4.DBF';
proc dbf db4=product out=new;
run;

data temp;
set new;
if _N_ NE 6 then output;
if _N_ ne 6;
if _N_ ne 7;
if _N_ ne 77;
if _N_ ne 78;
if _N_ ne 101;
if _N_ ne 143;
lage = log(age18t24);

run;
 
 


4. Reading a file that is compressed:

This is from a SAS program that reads the 5% sample from the US Census.
 

filename in pipe 'gunzip -c
      /rci/archive/MRD/ICPSR/81xx/8170/NY.DATA.gz';

infile in lrecl=193;

input RECTYPE $ 1 @;

if RECTYPE='H' then
  do;
   input SAMPLE $ 2 DIVISION $ 3 STATE $ 4-5
         COGRP $ 6-8 AREATYPE $ 9   SMSA $ 10-13
         .............................
         ARENT1 $ 162 FILLER $ 163-193;
  end;
else if RECTYPE ='P' then
  do;
   input RELAT1 $ 2-3 RELAT2 $ 4 SUBFAM1 $ 5
        SUBFAM2 $ 6 SEX $ 7   AGE 8-9 QTRBIR $ 10
        ...............................
        AINCOME6 $ 192 AINCOME7 $ 193
;
run;
 
 

• Print the data

proc print;
var murder robbery;
• Univariate Statistics

proc univariate plot;
var murder robbery;
run;

Output file from "univ.sas"

 
• How To Generate A Random Sample From A List

Title Generate a random sample of size 10;
options ls=80 ps=65;

data records;

infile 'records.txt';
input name $ ;
id = _N_;
x = uniform(0);
cards;

proc print;
run;

proc sort; by x;
data sample;
set records;
if _N_ <= 10;
proc print;
run;

Output file from "sample.sas"

 

 
 
 

• Multiple Linear Regression Example

 

 
 
 

This is the standard model in many statistical problems. The data  contains several predictors and one or more responses.
 
 

 
•  In practice we have a group of predictors that are candidates for the regression equation but we need to determine which ones are appropriate.
• There maybe collinearities among the predictors
• Regression diagnostics must be used to find out if there are any outliers in the data.

Example:

libname mylib spss 'Reg.por';
options linesize=70 pagesize=55;

data a;
format  IN80BO IN80C IN90C INBBID INBP90 INCENP INCODE INCR80 INCR90
INEMP INMAIN INMAIN2 INMAIN3 INNV INO INREST INT79 INT80 INT90 INV90
INVAL SEV79 SEV80 SEVR80 SFAMV VAPT79 VAPT80 VCOM79 VCOM80 VFRM79
VFRM80 VIND79 VIND80 VRES79 VRES80 VTOT79 VTOT80 VVAL79 VVAL80 comma8.
NAMEV1 $CHAR200.;
set mylib._first_;
if _N_ < 100;
run;

proc plot;
plot ravlchm*(roadacc roadcap transacc transcap);
run;
 

proc reg data=a;
model  ravlchm = hhinc black hslds setr vac indexr pstu
           roadacc roadcap transacc transcap/P R COLLIN;
 
 

Output file from "reg.sas"


 

We will look at regression analysis in the next chapter.
 
 
 

• General Linear Models Example

 

 

PROC GLM;
 CLASS REGION ;
 MODEL
   MURDER RAPE ROBBERY ASSAULT BURGLARY = REGION;
   MEANS REGION/ BON DUNCAN SCHEFFE;
   TITLE ' CRIME INCIDENCE IN THE UNITED STATES BY REGION' ;
run;
 

Output file from "crime.sas"