Disallowed combinations

[#1]

Hello, I have a factor with three levels and I am trying to create a design in which this factor levels are not equal in any given block. So for instance, Factor A should not have the same level repeated in any block. How can I do this please? Thanks. 

[#2]

Hi @aus22,

There is not much information to help you precisely.
Do you want to have all 3 levels individually in each block ? If yes, setting the number of runs per block to 3 could enforce the creation of block with each of the three levels of this factor, but it will also depend on the other factors (and their number of levels + difficulty to change/randomize).
What is the reason behind this question ? What is the purpose of the block ?

With more information about the other factors, objectives of the design and any practical constraints, it will be easier to help and guide you.

[#3]

Hi Victor, thank you for your response much appreciate it. I will try to restate as I made an error: I need to create choice sets of two choices. The factor of interest has 3 levels. And I dont want the levels repeated in any of the choice sets. My purpose is to compare and contrast each level with the other. Again thank you! Regards. Amit

[#4]

If you set the number of profiles to 2 per choice set, is the constraint respected for the 3-level factor ?
If not, I guess there are other factors involved ? How many are there ? You can maybe influence the change of levels for your factors by setting the "number of attributes that can change within a choice set" to the number of factors you have.

[#5]

Yes there are 3 other factors with 2, 2, and 6 levels respectively. So in total 4 factors as follows: Factor 1 (3 levels); Factor 2 (2 levels), Factor 3 (2 levels) and factor 4 (6 levels). In a choice set of 2 I want the first factor 1 levels to not be repeated in the choice set with 2 choices. So for instance, in a choice set of 2 options, Factor 1, level 1 cannot appear in both the choices offered. Etc. Here is an example of what is disallowed - the red marked repeated levels in Factor 1 should not happen. Other factors can repeat the levels and thats fine, just not factor 1. Thank you. 

[#6]

Hi @aus22,

Sorry for the delay of my response.
I wonder if it wouldn't be easier to build the design with the Custom Design platform, and then add analysis scripts from Choice models to analyze the results ?

From your experimental setup using Custom Design platform, it's possible to create a design respecting your constraint by defining Factor 1 with 3-levels as "Hard to Change" and the 3 other factors as "Easy to Change". Then, if you specify a number of runs that is the double of the number of whole plot, you have a design structure that meets your needs :

Whole Plot would then become your "Choice set" parameter.
You could add other factors to take into account the structure of Choice designs, like "Respondent" and "Survey" to make sure you have a similar design structure like the one you want.
Please find attached the design proposed, with only one survey and one respondent, and using column "Whole Plot" as "Choice Set ID".

Hope this answer may help you,

[#7]

Thanks a lot Victor. Let me try this out. Much appreciate it.