Published using Google Docs
SaS Codes for Experimental Designs.docx
Updated automatically every 5 minutes

SAS CODES  FOR EXPERIMENTAL DESIGNS

Completely Randomized Design (CRD)

Let us consider three treatments (A, B, C)

Step 1: Load your data 

   data crd;

   input Treatment $ Response;

   cards;

A 12

A 15

A 14

B 18

B 20

B 19

C 22

C 23

C 21

    ;

   run;

Step 2:  Perform the ANOVA for CRD

  proc glm data=crd;

    class Treatment;

    model Response = Treatment;

    means Treatment / tukey;

    title "Analysis of Completely Randomized Design";

 run;

Randomized Complete Block Design (RCBD)

Step 1: Load your data 

data rcbd;

input Block $ Treatment $ Response;

cards;

1 A 15

1 B 18

1 C 16

2 A 17

2 B 19

2 C 18

3 A 16

3 B 17

3 C 15

4 A 19

4 B 21

4 C 20

;

run;

Step 2:  Perform the ANOVA for RCBD

proc glm data=rcbd;

    class Block Treatment;

    model Response = Block Treatment;

    means Treatment / tukey;

    title "Analysis of Randomized Complete Block Design";

run;

Latin Square Design (LSD)

Step 1: Load your data 

 data lsd;

 input Row $ Column $ Treatment $ Response;

 cards;

1 1 A 10

1 2 B 15

1 3 C 14

2 1 B 12

2 2 C 18

2 3 A 13

3 1 C 11

3 2 A 16

3 3 B 19

 ;

 run;

Step 2:  Perform ANOVA using the following code

proc glm data=lsd;

    class Row Column Treatment;

    model Response = Row Column Treatment;

    means Treatment / tukey;

    title "Analysis of Latin Square Design";

run;

Split Plot Design (SPD)

/* Import the dataset */

data splitplot;

    input Block A B Y;

    datalines;

    1 1 1 50

    1 1 2 55

    1 2 1 60

    1 2 2 65

    2 1 1 52

    2 1 2 56

    2 2 1 61

    2 2 2 66

    /* Add more data as needed */

    ;

run;

/* PROC MIXED for split-plot design */

proc mixed data=splitplot;

    class Block A B;

    model Y = A B A*B;

    random Block A*Block;

    lsmeans A*B / pdiff adjust=tukey;

run;

Strip Plot Design

/* Import the dataset */

data stripplot;

    input Block A B Y;

    cards;

    1 1 1 50

    1 1 2 55

    1 2 1 60

    1 2 2 65

    2 1 1 52

    2 1 2 56

    2 2 1 61

    2 2 2 66

    /* Add more data as needed */

    ;

run;

/* PROC MIXED for strip-plot design */

proc mixed data=stripplot;

    class Block A B;

    model Y = A B A*B;

    random Block A*Block B*Block;

    lsmeans A*B / pdiff adjust=tukey;

run;

2 Factors Factorial Design

/* Import the dataset */

data factorial;

    input A B Y;

    cards;

    1 1 45

    1 2 50

    1 3 48

    2 1 55

    2 2 60

    2 3 58

    3 1 65

    3 2 70

    3 3 68

    /* Add more data as needed */

    ;

run;

/* PROC GLM for 2-factor factorial design */

proc glm data=factorial;

    class A B;

    model Y = A B A*B;

    lsmeans A*B / pdiff=all adjust=tukey;

run;

3 Factors Factorial Design

/* Import the dataset */

data factorial3;

    input A B C Y;

    datalines;

    1 1 1 45

    1 1 2 50

    1 2 1 55

    1 2 2 60

    2 1 1 48

    2 1 2 53

    2 2 1 58

    2 2 2 63

    3 1 1 52

    3 1 2 57

    3 2 1 62

    3 2 2 67

    /* Add more data as needed */

    ;

run;

/* PROC GLM for 3-factor factorial design */

proc glm data=factorial3;

    class A B C;

    model Y = A B C A*B A*C B*C A*B*C;

    lsmeans A*B*C / pdiff=all adjust=tukey;

run;

Pooled Analysis of RBD

/* Import the dataset */

data pooled_rbd;

    input Trial Block Trt Y;

    datalines;

    1 1 1 45

    1 1 2 50

    1 1 3 48

    1 2 1 55

    1 2 2 60

    1 2 3 58

    2 1 1 52

    2 1 2 57

    2 1 3 55

    2 2 1 60

    2 2 2 65

    2 2 3 63

    /* Add more data as needed */

    ;

run;

/* PROC MIXED for pooled analysis of RBD */

proc mixed data=pooled_rbd;

    class Trial Block Trt;

    model Y = Trt Trial Trt*Trial;

    random Block(Trial);

    lsmeans Trt / pdiff adjust=tukey;

run;