Example SAS routines

EXAMPLE SAS ROUTINES

Analysis of Variance (PROC ANOVA)

This is an example of a simple one-way ANOVA. The example data are taken from the Agresti data set.

DATA INCOME;
INPUT INCOME EDUC RACE Z1 Z2;
CARDS;
8 10 b 1 0
9 7 b 1 0
13 9 b 1 0
.
.
.
9 10 w 0 0
12 12 w 0 0
28 20 w 0 0
;
PROC ANOVA;
  CLASSES RACE;
  MODEL INCOME=RACE;
  MEANS TYPE/ BON LINES;
RUN;

Descriptive Statistics (PROC UNIVARIATE)

Use for generating descriptive statistics on quantitative data. The example data are taken from the Agresti data set.

DATA HOMES;
INPUT Price Size Bedrms Bathrms New;
CARDS;
48.5 1.10 3 1 0
55.0 1.01 3 2 0
68.0 1.45 3 2 0
137.0 2.40 3 3 0
.
.
.
280.0 3.85 4 3 0
PROC UNIVARIATE;
VAR Price Size;
RUN;

Histograms (PROC CHART)

Use for generating histograms for both quantitative and qualitative data. The example data are taken from the Agresti data set.

DATA HOMES;
INPUT Price Size Bedrms Bathrms New;
CARDS;
48.5 1.10 3 1 0
55.0 1.01 3 2 0
68.0 1.45 3 2 0
137.0 2.40 3 3 0
.
.
.
280.0 3.85 4 3 0
PROC CHART;
VBAR New; Simple Bar Chart
PROC CHART;
VBAR Price/TYPE=PERCENT; Relative Frequency Histogram
RUN;

Random numbers

DATA RANNUMS;
DO N = 1 TO 10;
NUMBER = RANUNI(0);
OUTPUT;
END;
PROC PRINT; VAR NUMBER;
RUN;

To obtain more or less random numbers, change the line "DO N = 1 TO 10;" to "DO N = 1 TO X;" where X is how many random numbers you need.

Multiple Regression (PROC REG)

Use for doing multiple regression analysis of quantitative data. For more elaborate procedures, such as ANCOVA, use PROC GLM.

DATA HOMES;
INPUT Price Size Bedrms Bathrms New;
S-Bed = Size * Bedrms;  Creates interaction between Size and no. of bedrooms. CARDS; 48.5 1.10 3 1 0
55.0 1.01 3 2 0
68.0 1.45 3 2 0
137.0 2.40 3 3 0
.
.
.
280.0 3.85 4 3 0
PROC REG;
  MODEL Price = Size Bedrms Bathrms New S-Bed;
  OUTPUT OUT=RESIDS P=YHAT R=RESID;  Saves residuals.
PROC PLOT;
  PLOT RESID*(YHAT SIZE);  Plots residuals by predicted values and Size.
PROC CHART;
  VBAR RESID;  Bar chart of residuals.
RUN;

Scatterplots (PROC PLOT)

Use for generating crude scattergrams for quantitative data. For high quality graphics, use PROC GRAPH.

DATA HOMES;
INPUT Price Size Bedrms Bathrms New;
CARDS;
48.5 1.10 3 1 0
55.0 1.01 3 2 0
68.0 1.45 3 2 0
137.0 2.40 3 3 0
.
.
.
280.0 3.85 4 3 0
PROC PLOT;
PLOT Price * Size;
RUN;

t-test (Independent Samples)

Use for testing hypotheses of the variety: µ₂ - µ₁= 0.

DATA TWOSAMPL;
INPUT GROUP $ OUTCOME;
CARDS;
EXP 52 CONTR 60
EXP 77 CONTR 80
.
.
.
EXP 66 CONTR 60
PROC TTEST;
CLASS GROUP; VAR OUTCOME;
RUN;

t-test (Matched Pairs)

Use for testing hypotheses of the variety: µ_d= 0.

DATA MATCHED;
INPUT SUBJECT $ TIME1 TIME2;
DIFF = TIME2 - TIME1; CARDS;
Sue 52 60
Bob 77 80
.
.
.
Sam 66 60
PROC MEANS T PRT;
VAR DIFF;
RUN;

Last modified: August 16, 1999