STA 402S 1998: Introduction to SAS (Lesson One)
tuzo.erin > ls tuzo.erin > ls cars.dat cars98a.sas class90.dat intro98a.sas tuzo.erin > less class90.dat 1 2 9 1 7 8 4 3 5 2 6 10 10 10 5 0 0 0 0 55 43 0 2 10 10 5 9 10 8 6 8 10 10 8 9 9 9 9 10 10 66 79 1 2 10 10 5 10 10 10 9 8 10 10 10 10 10 10 9 10 10 94 67 1 2 10 10 8 9 10 7 10 9 10 10 10 9 10 10 9 10 10 81 65 0 1 10 1 0 0 8 6 5 2 10 9 0 0 10 6 0 5 0 54 29
And so on. There are 21 columns and 62 rows of data; columns not aligned. This is a typical data file, in which the rows are cases, and the columns are variables. If there are lots of variables, you may need more than one line of data for each case. In this instance the cases are people, but they can be anything -- cities, families, farms, days (time series analysis).
Now the COMMAND FILE, a simple program that reads the data and specifies what statistics to compute.
tuzo.erin > cat intro98a.sas
options linesize=79 pagesize=100;
title 'Grades from STA 302: Fall, 1990';
proc format; /* Used to label values of the categorical variables */
value sexfmt 0 = 'Male' 1 = 'Female';
value ethfmt 1 = 'Chinese'
2 = 'European'
3 = 'Other' ;
data erindale;
infile 'class90.dat';
input sex ethnic quiz1-quiz8 comp1-comp9 midterm final;
/* Drop lowest score for quiz & computer */
quizave = ( sum(of quiz1-quiz8) - min(of quiz1-quiz8) ) / 7;
compave = ( sum(of comp1-comp9) - min(of comp1-comp9) ) / 8;
label ethnic = 'Apparent ethnic background (ancestry)'
quizave = 'Quiz Average (drop lowest)'
compave = 'Computer Average (drop lowest)';
mark = .3*quizave*10 + .1*compave*10 + .3*midterm + .3*final;
label mark = 'Final Mark';
format sex sexfmt.; /* Associates sex & ethnic */
format ethnic ethfmt.; /* with formats defined above */
proc freq;
tables sex ethnic;
proc means n mean std;
var quiz1 -- mark; /* single dash only works with numbered
lists, like quiz1-quiz8 */
proc freq;
tables sex*ethnic / chisq;
proc corr;
var final midterm quizave compave;
proc ttest;
class sex;
var mark;
/* Predict final exam score from midterm, quiz & computer */
proc reg simple;
model final = midterm quizave compave / ss1;
smalstuf: test quizave = 0, compave = 0;
That was the end of the command file. On tuzo we must use the command sasb instead of sas.
tuzo.erin > sasb class90 tuzo.erin > tuzo.erin > ls cars.dat class90.dat intro98a.lst cars98a.sas intro98a.log intro98a.sas tuzo.erin >
SAS produces two output files, the log and the list file. If you have errors, there may be no lst file, and your error messages will appear on the log file.
tuzo.erin > cat intro98a.log
The SAS System
13:58 Sunday, January 7, 1998
NOTE: Copyright(c) 1989 by SAS Institute Inc., Cary, NC USA.
NOTE: SAS (r) Proprietary Software Release 6.07 TS203
Licensed to UNIVERSITY OF TORONTO, Site 0008987003.
NOTE: AUTOEXEC processing beginning; file is /local/lib/sas/autoexec.sas.
NOTE: SAS initialization used:
real time 0.255 seconds
cpu time 0.120 seconds
NOTE: AUTOEXEC processing completed.
1 options linesize=79 pagesize=100;
2 title 'Grades from STA 302: Fall, 1990';
3
4 proc format;
4 /* Used to label values of the categorical variables
*/
5 value sexfmt 0 = 'Male' 1 = 'Female';
NOTE: Format SEXFMT has been output.
6 value ethfmt 1 = 'Chinese'
7 2 = 'European'
8 3 = 'Other' ;
NOTE: Format ETHFMT has been output.
9
NOTE: PROCEDURE FORMAT used:
real time 0.122 seconds
cpu time 0.040 seconds
10 data erindale;
11 infile 'class90.dat';
12 input sex ethnic quiz1-quiz8 comp1-comp9 midterm final;
13 /* Drop lowest score for quiz & computer */
14 quizave = ( sum(of quiz1-quiz8) - min(of quiz1-quiz8) ) / 7;
15 compave = ( sum(of comp1-comp9) - min(of comp1-comp9) ) / 8;
16 label ethnic = 'Apparent ethnic background (ancestry)'
17 quizave = 'Quiz Average (drop lowest)'
18 compave = 'Computer Average (drop lowest)';
19 mark = .3*quizave*10 + .1*compave*10 + .3*midterm + .3*final;
20 label mark = 'Final Mark';
21 format sex sexfmt.; /* Associates sex & ethnic */
22 format ethnic ethfmt.; /* with formats defined above */
23
NOTE: The infile 'class90.dat' is:
File Name=/res/jbrunner/class/sta402s98/class90.dat,
Owner Name=jbrunner,Group Name=research,
Access Permission=rw-------,
File Size (bytes)=4372
NOTE: 62 records were read from the infile 'class90.dat'.
The minimum record length was 65.
The maximum record length was 77.
NOTE: The data set WORK.ERINDALE has 62 observations and 24 variables.
NOTE: DATA statement used:
real time 0.160 seconds
cpu time 0.060 seconds
24 proc freq;
25 tables sex ethnic;
NOTE: The PROCEDURE FREQ printed page 1.
NOTE: PROCEDURE FREQ used:
real time 0.186 seconds
cpu time 0.060 seconds
26 proc means n mean std;
27 var quiz1 -- mark; /* single dash only works with numbered
28 lists, like quiz1-quiz8 */
NOTE: The PROCEDURE MEANS printed page 2.
NOTE: PROCEDURE MEANS used:
real time 0.025 seconds
cpu time 0.030 seconds
29 proc freq;
30 tables sex*ethnic / chisq;
NOTE: The PROCEDURE FREQ printed page 3.
NOTE: PROCEDURE FREQ used:
real time 0.291 seconds
cpu time 0.040 seconds
31 proc corr;
32 var final midterm quizave compave;
NOTE: The PROCEDURE CORR printed page 4.
2 The SAS System 13:58 Sunday, January 7,
1998
NOTE: PROCEDURE CORR used:
real time 0.022 seconds
cpu time 0.020 seconds
33 proc ttest;
34 class sex;
35 var mark;
36
37 /* Predict final exam score from midterm, quiz & computer */
38
NOTE: The PROCEDURE TTEST printed page 5.
NOTE: PROCEDURE TTEST used:
real time 0.015 seconds
cpu time 0.020 seconds
39 proc reg simple;
40 model final = midterm quizave compave / ss1;
41
42
43
NOTE: 62 observations read.
62 observations used in computations.
NOTE: The PROCEDURE REG printed pages 6-7.
NOTE: PROCEDURE REG used:
real time 0.382 seconds
cpu time 0.030 seconds
NOTE: The SAS System used:
real time 1.511 seconds
cpu time 0.460 seconds
NOTE: SAS Institute Inc., SAS Circle, PO Box 8000, Cary, NC 27512-8000
tuzo.erin >
tuzo.erin > cat intro98a.lst
Grades from STA 302: Fall, 1990
1
13:58 Sunday, January 7,
1998
Cumulative Cumulative
SEX Frequency Percent Frequency Percent
----------------------------------------------------
Male 39 62.9 39 62.9
Female 23 37.1 62 100.0
Apparent ethnic background (ancestry)
Cumulative Cumulative
ETHNIC Frequency Percent Frequency Percent
------------------------------------------------------
Chinese 41 66.1 41 66.1
European 15 24.2 56 90.3
Other 6 9.7 62 100.0
Variable Label N Mean Std Dev
------------------------------------------------------------------------
QUIZ1 62 9.0967742 2.2739413
QUIZ2 62 5.8870968 3.2294995
QUIZ3 62 6.0483871 2.3707744
QUIZ4 62 7.7258065 2.1590022
QUIZ5 62 9.0645161 1.4471109
QUIZ6 62 7.1612903 1.9264641
QUIZ7 62 5.7903226 2.1204477
QUIZ8 62 6.3064516 2.3787909
COMP1 62 9.1451613 1.1430011
COMP2 62 8.8225806 1.7604414
COMP3 62 8.3387097 2.5020880
COMP4 62 7.8548387 3.2180168
COMP5 62 9.4354839 1.7237109
COMP6 62 7.8548387 2.4350364
COMP7 62 6.6451613 2.7526248
COMP8 62 8.8225806 1.6745363
COMP9 62 8.2419355 3.7050497
MIDTERM 62 70.1935484 13.6235557
FINAL 62 49.4677419 17.5141327
QUIZAVE Quiz Average (drop lowest) 62 7.6751152 1.1266917
COMPAVE Computer Average (drop lowest) 62 8.8346774 1.1204997
MARK Final Mark 62 67.7584101 11.0235746
------------------------------------------------------------------------
TABLE OF SEX BY ETHNIC
SEX ETHNIC(Apparent ethnic background (ancestry))
Frequency|
Percent |
Row Pct |
Col Pct |Chinese |European|Other | Total
---------+--------+--------+--------+
Male | 27 | 7 | 5 | 39
| 43.55 | 11.29 | 8.06 | 62.90
| 69.23 | 17.95 | 12.82 |
| 65.85 | 46.67 | 83.33 |
---------+--------+--------+--------+
Female | 14 | 8 | 1 | 23
| 22.58 | 12.90 | 1.61 | 37.10
| 60.87 | 34.78 | 4.35 |
| 34.15 | 53.33 | 16.67 |
---------+--------+--------+--------+
Total 41 15 6 62
66.13 24.19 9.68 100.00
STATISTICS FOR TABLE OF SEX BY ETHNIC
Statistic DF Value Prob
------------------------------------------------------
Chi-Square 2 2.921 0.232
Likelihood Ratio Chi-Square 2 2.996 0.224
Mantel-Haenszel Chi-Square 1 0.000 0.995
Phi Coefficient 0.217
Contingency Coefficient 0.212
Cramer's V 0.217
Sample Size = 62
WARNING: 33% of the cells have expected counts less
than 5. Chi-Square may not be a valid test.
Grades from STA 302: Fall, 1990
4
13:58 Sunday, January 7,
1998
Correlation Analysis
4 'VAR' Variables: FINAL MIDTERM QUIZAVE COMPAVE
Simple Statistics
Variable N Mean Std Dev Sum
FINAL 62 49.467742 17.514133 3067.000000
MIDTERM 62 70.193548 13.623556 4352.000000
QUIZAVE 62 7.675115 1.126692 475.857143
COMPAVE 62 8.834677 1.120500 547.750000
Simple Statistics
Variable Minimum Maximum Label
FINAL 15.000000 89.000000
MIDTERM 44.000000 103.000000
QUIZAVE 4.571429 9.714286 Quiz Average (drop lowest)
COMPAVE 5.000000 10.000000 Computer Average (drop lowest)
Pearson Correlation Coefficients / Prob > |R| under Ho: Rho=0 / N = 62
FINAL MIDTERM QUIZAVE COMPAVE
FINAL 1.00000 0.51078 0.47127 0.14434
0.0 0.0001 0.0001 0.2630
MIDTERM 0.51078 1.00000 0.59294 0.41277
0.0001 0.0 0.0001 0.0009
QUIZAVE 0.47127 0.59294 1.00000 0.52649
Quiz Average (drop lowest) 0.0001 0.0001 0.0 0.0001
COMPAVE 0.14434 0.41277 0.52649 1.00000
Computer Average (drop lowest) 0.2630 0.0009 0.0001 0.0
Grades from STA 302: Fall, 1990
5
13:58 Sunday, January 7,
1998
TTEST PROCEDURE
Variable: MARK Final Mark
SEX N Mean Std Dev Std Error Minimum Maximum
-------------------------------------------------------------------------------
Male 39 67.62097070 10.11112521 1.61907581 43.61428571 89.93214286
Female 23 67.99145963 12.65945704 2.63967927 48.48214286 95.45714286
Variances T DF Prob>|T|
---------------------------------------
Unequal -0.1196 38.5 0.9054
Equal -0.1268 60.0 0.8995
For H0: Variances are equal, F' = 1.57 DF = (22,38) Prob>F' = 0.2190
Grades from STA 302: Fall, 1990
6
13:58 Sunday, January 7,
1998
Descriptive Statistics
Variables Sum Mean Uncorrected SS
INTERCEP 62 1 62
MIDTERM 4352 70.193548387 316804
QUIZAVE 475.85714286 7.6751152074 3729.6938776
COMPAVE 547.75 8.8346774194 4915.78125
FINAL 3067 49.467741935 170429
Variables Variance Std Deviation
INTERCEP 0 0
MIDTERM 185.60126917 13.623555673
QUIZAVE 1.2694341618 1.1266916889
COMPAVE 1.2555195664 1.1204996949
FINAL 306.744844 17.514132693
Model: MODEL1
Dependent Variable: FINAL
Analysis of Variance
Sum of Mean
Source DF Squares Square F Value Prob>F
Model 3 6216.33032 2072.11011 9.618 0.0001
Error 58 12495.10516 215.43285
C Total 61 18711.43548
Root MSE 14.67763 R-square 0.3322
Dep Mean 49.46774 Adj R-sq 0.2977
C.V. 29.67112
Parameter Estimates
Parameter Standard T for H0:
Variable DF Estimate Error Parameter=0 Prob > |T|
INTERCEP 1 0.554234 16.09660613 0.034 0.9727
MIDTERM 1 0.498057 0.17318547 2.876 0.0056
QUIZAVE 1 5.371214 2.24349814 2.394 0.0199
COMPAVE 1 -3.086879 1.99437533 -1.548 0.1271
Variable
Variable DF Type I SS Label
INTERCEP 1 151718 Intercept
MIDTERM 1 4881.795290
QUIZAVE 1 818.430992 Quiz Average (drop lowest)
COMPAVE 1 516.104040 Computer Average (drop lowest)
Dependent Variable: FINAL
Test: SMALSTUF Numerator: 667.2675 DF: 2 F value: 3.0973
Denominator: 215.4328 DF: 58 Prob>F: 0.0527