Modeling Relationship Between Variables: Regression
Problem
An automated system for marking large numbers of student computer
programs, called AUTOMARK, has been used successfully at
|
AUTOMARK GRADE x |
INSTRUCTOR GRADE y |
AUTOMARK GRADE x |
INSTRUCTOR GRADE y |
AUTOMARK GRADE x |
INSTRUCTOR GRADE y |
|
12.2 |
10 |
18.2 |
15 |
19.3 |
17 |
|
10.6 |
11 |
15.1 |
16 |
19.5 |
17 |
|
15.1 |
12 |
17.2 |
16 |
19.7 |
17 |
|
16.2 |
12 |
17.5 |
16 |
18.6 |
18 |
|
16.6 |
12 |
18.6 |
16 |
19 |
18 |
|
16.6 |
13 |
18.8 |
16 |
19.2 |
18 |
|
17.2 |
14 |
17.8 |
17 |
19.4 |
18 |
|
17.6 |
14 |
18 |
17 |
19.6 |
18 |
|
18.2 |
14 |
18.2 |
17 |
20.1 |
18 |
|
16.5 |
15 |
18.4 |
17 |
19.2 |
19 |
|
17.2 |
15 |
18.6 |
17 |
19.3 |
17 |
|
12.2 |
10 |
19 |
17 |
19.5 |
17 |
Question
Answer
SAS Program Used for the Analysis
*--- SAS program: Regression_Model_1.SAS
;
options nodate pageno=1;
*---Create SAS data set;
data automark;
input automark_grade instructor_grade @@;
cards;
12.2 10 18.2 15 19.3 17
10.6 11 15.1 16 19.5 17
15.1 12 17.2 16 19.7 17
16.2 12 17.5 16 18.6 18
16.6 12 18.6 16 19
18
16.6 13 18.8 16 19.2 18
17.2 14 17.8 17 19.4 18
17.6 14 18 17 19.6 18
18.2 14 18.2 17 20.1 18
16.5 15 18.4 17 19.2 19
17.2 15 18.6 17 19.3 17
12.2 10 19 17 19.5 17
;
run;
proc reg data=automark;
title "Regression
of instructor_grade on automark_grade";
model instructor_grade = automark_grade / p cli;
id automark_grade;
quit;
Note
SAS Output
Regression of instructor_grade on automark_grade
The REG
Procedure
Model:
MODEL1
Dependent
Variable: instructor_grade
Analysis
of Variance
Sum of Mean
Source DF Squares Square F Value
Pr > F
Model 1 153.09529 153.09529 97.01
<.0001
Error 34 53.65471 1.57808
Corrected Total 35 206.75000
Root MSE 1.25622 R-Square
0.7405
Dependent Mean 15.58333 Adj R-Sq 0.7329
Coeff
Var
8.06128
Parameter Estimates
Parameter Standard
Variable DF Estimate Error t Value
Pr > |t|
Intercept 1 -1.04264 1.70093 -0.61
0.5440
automark_grade 1
0.94406 0.09585 9.85
<.0001
Regression of instructor_grade on automark_grade
The
REG Procedure
Model: MODEL1
Dependent
Variable: instructor_grade
Output Statistics
automark_ Dep Var Predicted Std Error
Obs grade
instructor_grade Value
Mean Predict 95% CL
Predict Residual
1
12.2 10.0000 10.4749 0.5593 7.6804 13.2695 -0.4749
2
18.2 15.0000 16.1393 0.2168 13.5486 18.7300 -1.1393
3
19.3 17.0000 17.1777 0.2647 14.5688 19.7867 -0.1777
4
10.6 11.0000 8.9644 0.7039 6.0381 11.8908 2.0356
5
15.1 16.0000 13.2127 0.3190 10.5787 15.8467 2.7873
6
19.5 17.0000 17.3666 0.2768
14.7524 19.9807 -0.3666
7
15.1 12.0000 13.2127 0.3190 10.5787 15.8467 -1.2127
8
17.2 16.0000 15.1952 0.2130 12.6058 17.7846 0.8048
9
19.7 17.0000 17.5554 0.2897 14.9354 20.1753 -0.5554
10
16.2 12.0000 14.2512 0.2493 11.6484 16.8539 -2.2512
11
17.5 16.0000 15.4784 0.2096 12.8902 18.0667 0.5216
12
18.6 18.0000 16.5169 0.2298 13.9216 19.1122 1.4831
13
16.6 12.0000 14.6288 0.2307 12.0331 17.2244 -2.6288
14
18.6 16.0000 16.5169 0.2298 13.9216 19.1122 -0.5169
15
19 18.0000 16.8945 0.2481 14.2923 19.4968 1.1055
16
16.6 13.0000 14.6288 0.2307 12.0331 17.2244 -1.6288
17
18.8 16.0000 16.7057 0.2384 14.1072 19.3042 -0.7057
18
19.2 18.0000 17.0833
0.2589 14.4767 19.6899 0.9167
19
17.2 14.0000 15.1952 0.2130 12.6058 17.7846 -1.1952
20
17.8 17.0000 15.7617 0.2102 13.1732 18.3501
1.2383
21
19.4 18.0000 17.2722 0.2706 14.6606 19.8837 0.7278
22
17.6 14.0000 15.5728 0.2094 12.9847 18.1610 -1.5728
23
18 17.0000
15.9505 0.2127 13.3612 18.5397 1.0495
24
19.6 18.0000 17.4610 0.2832 14.8440 20.0780 0.5390
25
18.2 14.0000 16.1393 0.2168 13.5486 18.7300 -2.1393
26
18.2 17.0000 16.1393 0.2168 13.5486 18.7300 0.8607
27
20.1 18.0000 17.9330 0.3174 15.2998 20.5662 0.0670
28
16.5 15.0000 14.5344 0.2349 11.9372 17.1316 0.4656
29
18.4 17.0000 16.3281 0.2226 13.7354
18.9208 0.6719
30
19.2 19.0000 17.0833 0.2589 14.4767 19.6899 1.9167
31
17.2 15.0000 15.1952 0.2130 12.6058 17.7846 -0.1952
32
18.6 17.0000 16.5169 0.2298 13.9216 19.1122 0.4831
33
19.3 17.0000 17.1777 0.2647 14.5688 19.7867 -0.1777
34
12.2 10.0000 10.4749 0.5593 7.6804 13.2695 -0.4749
35
19 17.0000 16.8945 0.2481 14.2923 19.4968 0.1055
36
19.5 17.0000 17.3666 0.2768 14.7524 19.9807 -0.3666
Sum of
Residuals 0
Sum of Squared
Residuals 53.65471
Predicted Residual SS
(PRESS) 62.78104