Sas code and output

The following code uses the same variable names specified at the end of Section 7.4. The code is given below initially for the compound symmetry model, with inclusion of carry-over effects.

PROC MIXED; CLASS treat period patient carry; TITLE 'COMPOUND SYMMETRY';

MODEL lvet=treat period carry/ DDFM=KENWARDROGER; REPEATED period/ SUBJECT=patient TYPE=CS RCORR; LSMEANS treat carry/DIFF PDIFF;

The code for the other covariance pattern models is identical except for the REPEATED statement.

TITLE 'STRUCTURED BY PERIOD';

REPEATED period/ SUBJECT=patient TYPE=UN RCORR;

TITLE 'STRUCTURED BY TREATMENT';

REPEATED treat/ SUBJECT=patient TYPE=UN RCORR;

The changes to remove carry-over effects are obvious. The following output is for the above code, where structuring by treatment is employed.

STRUCTURED BY TREATMENT

The Mixed Procedure

Model Information

Data Set

Dependent Variable Covariance Structure Subject Effect Estimation Method Residual Variance Method Fixed Effects SE Method

WORK.NEW lvet

Unstructured patient

REML

None

Prasad-Rao-Jeske-

Kackar-Harville

Kenward-Roger

Degrees of Freedom Method

Class Level Information

Class

Levels

Values treat period patient

1234 1234

1234

carry

Dimensions

Covariance Parameters Columns in X Columns in Z Subjects

Max Obs Per Subject

13 0

14 4

Number of Observations

Number of Observations Read Number of Observations Used Number of Observations Not Used

56 56 0

Iteration History

Iteration

Iteration History

on

Evaluations

-2 Res

Log Like

0

1

531.

19328518

1

2

508.

00614950

2

1

502.

46576070

3

1

499.

65259977

4

1

498.

42652859

5

1

498.

03992412

6

1

497.

97162927

7

1

497.

96709658

8

1

497.

02010791 01043261 00468615 00154375 00028985 00002064 00000023 00000000

Convergence criteria met. Estimated R Correlation Matrix for patient 1

Row Coll Col2 Col3 Col4

1 1.0000 0.6367 0.8333 0.09861

2 0.6367 1.0000 0.8184 -0.09949

3 0.8333 0.8184 1.0000 0.1628

4 0.09861 -0.09949 0.1628 1.0000

Covariance Parameter Estimates

Cov

Parm

Subject

Estimate

UNi

1

,1)

patient

1681.

02

UNi

2

,1)

patient

322.

45

UNi

2

,2)

patient

6360.

25

UNi

3

,1)

patient

390.

80

UNi

3

,2)

patient

3891.

90

UNi

3

,3)

patient

3429.

52

UNi

4

,1)

patient

-205.

26

UNi

4

,2)

patient

2555.

15

UNi

4

,3)

patient

2411.

65

UNi

4

,4)

patient

2531.

-2 Res Log Likelihood 498.0

AIC (smaller is better) 518.0

AICC (smaller is better) 524.3

BIC (smaller is better) 524.4

Null Model Likelihood Ratio Test DF Chi-Square Pr > ChiSq

9 33.23 0.0001

Type 3 Tests of Fixed Effects

Num Den

Effect

DF

DF

F Value

Pr >

F

treat

3

10

.3

7.

75

0

.0054

period

3

11

8.

91

0

.0028

carry

3

14

.1

1.

82

0

.1890

Least

Squares

Means

Standard

Effect

treat carry

Estimate

Error

DF

t

Value

Pr

> |t|

treat

1

388.

46

11.

5910

13

.4

33.

51

0001

treat

2

344.

64

21.

7665

12

.8

15.

83

0001

treat

3

380.

15

17.

9227

16

.2

21.

21

0001

treat

4

322.

10

13.

9013

10

.8

23.

17

0001

carry

1

338.

38

16.

3402

21

.4

20.

71

0001

carry

2

361.

92

18.

4008

23

.7

19.

67

0001

carry

3

392.

39

21.

0386

33

.6

18.

65

0001

carry

4

342.

66

15.

5753

19

.2

22.

00

0001

Differences of Least Squares Means

Standard

Effect

treat

carry

-treat

-carry

Estimate

Error

DF t

Value

treat

1

2

43.

8265

23

.5349

12

.4

1.86

treat

1

3

8.

3143

19

.2337

12

.1

0.43

treat

1

4

66.

3686

18

.8818

9

.8

3.51

treat

2

3

-35.

5123

14

.4029

13

.3

-2.47

treat

2

4

22.

5421

18

.1408

14

.1

1.24

treat

3

4

58.

0543

13

.2261

15

.3

4.39

carry

1

2

-23.

5450

18

.7010

17

-1.26

carry

1

3

-54.

0162

22

.6449

17

.6

-2.39

carry

1

4

-4.

2811

16

.4810

7.

98

-0.26

carry

2

3

-30.

4712

21

.5069

17

-1.42

carry

2

4

19.

2639

21

.4383

11

.6

0.90

carry

3

4

49.

7351

24

.6643

30

.2

2.02

Differences of Least Squares Means

Effect

treat

carry

-treat

_carry

Pr

> |t|

treat

1

2

0.

. 0864

treat

1

3

0.

. 6731

treat

1

4

0.

. 0058

treat

2

3

0.

0280

treat

2

4

0.

2343

treat

3

4

0.

0005

carry

1

2

0.

2250

carry

1

3

0.

0286

carry

1

4

0.

8016

carry

2

3

0.

1747

carry

2

4

0.

3871

carry

3

4

0.

0527

Was this article helpful?

0 0

Post a comment