topic Re: How to create BIB and making some of the Treatments blank or missing values in SAS Programming

How to create BIB and making some of the Treatments blank or missing values

Shomo_olongus — Sat, 15 Sep 2018 01:33:19 GMT

How can I create a BIBD with some of the treatments replace with missing values. So let say I have 4 treatments but I am selecting 3 at a time. With each of the 3 selected I can have 6 different arrangement but with one of the treatments missing. Let me illustrate it

trt1 trt2 trt3 trt4

1 1 1 block 1

1 1 1 block 2

1 1 1 block 3

1 1 1 block 4

for block 1 I can have using ranks (trt4 is missing)

trt1 trt2 trt3 trt4

1 2 3

1 3 2

2 1 3

2 3 1

3 1 2

3 2 1

for block 1 I can have using ranks (trt3 is missing)

trt1 trt2 trt3 trt4

1 2 3

1 3 2

2 1 3

2 3 1

3 1 2

3 2 1

this will go own up to block 4.

Re: How to create BIB and making some of the Treatments blank or missing values

VDD — Sat, 15 Sep 2018 01:37:26 GMT

have you looked at call missing, or setting options missing = 0;

Re: How to create BIB and making some of the Treatments blank or missing values

Shomo_olongus — Sat, 15 Sep 2018 01:59:09 GMT

No please, I don't know that function. The problem is, I will calculating modified jonckheere terpstra (mJT), for each block.
That is I will go through each row and if a rank is less than another rank in a trt i will assign 1, for example
trt1 trt2 trt3 trt4

1 2 3

MJT = (1) + 2(1) + 3(0)+1+2(0)+(0)
for trt1 and trt2 1
for trt1 and trt3 is (3-1)*1

the factor is the difference between the treatments . and because the trt4 is missing any value with it is 0.

I will then sum the MJT for each block and find the mean and variance.

Re: How to create BIB and making some of the Treatments blank or missing values

ballardw — Mon, 17 Sep 2018 18:32:27 GMT

@Shomo_olongus wrote:

How can I create a BIBD with some of the treatments replace with missing values. So let say I have 4 treatments but I am selecting 3 at a time. With each of the 3 selected I can have 6 different arrangement but with one of the treatments missing. Let me illustrate it

trt1 trt2 trt3 trt4

1     1    1               block 1

1     1            1       block 2

1           1      1       block 3

        1    1      1       block 4

for block 1 I can have using ranks (trt4 is missing)

trt1 trt2 trt3 trt4

1      2    3

1      3    2

2      1    3

2      3    1

3      1    2

3      2    1

for block 1 I can have using ranks (trt3 is missing)

trt1 trt2 trt3 trt4

1      2          3

1      3          2

2      1          3

2      3          1

3      1          2

3      2          1

this will go own up to block 4.

I may be dense today but from you post I 1) cannot tell what your starting data actually looks like or 2) what you want the output to look like.

I think you are looking for PROC PLAN but you aren't clear as to which type of design you may be looking for.