Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Total # Respondents plus subgroup quotas

Sorry if this is a basic question, but I'm not sure how to approach.  

I am working on a survey that has a maximum sample size of 600 and two broad quota categories (A and B), each with a limit of 300 each.  That part is straightforward enough.  However, within the two quota categories, A and B, are 3 additional quotas that were quoted to me as ranges (min and max) instead of absolute limits.  

Say, for the sake of example, that the 3 subgroups can have between 90-110 respondents.  The intent is to allow some groups to have more than others, as long as the total sample and the Group A and Group B totals are as defined.  However, if these subgroups all hit the maximum, the total sample (and category A and B quotas) will be exceeded.

Is it possible to impose a fixed total respondent and broad category quotas, while allowing some flexibility in the finer subgroups?  If so, how would I set this up in quota control?
asked Dec 20, 2019 by dpatek (220 points)

1 Answer

+1 vote
You can setup a quota question that defines the following ...

1/ A sub-group 1
2/ A sub-group 2
3/ A sub-group 3
4/ B sub-group 1
5/ B sub-group 2
6/ B sub-group 3

Set all targets to 100.

You can monitor the quotas during surveying and make quota target changes during field via the admin module. As the quotas start to fill, you may notice one particular quota filling faster than others, so you bump that to a target of 110. Once that target is reached, you will have the other targets to monitor. And you can lower the targets on the other sub-groups.

The important thing is the A target is always set to 300 and the same goes for the B target.

It's a little bit of manual work but should be straight forward to manage.

I've had to contend with this on many projects in the past and it works fine.

So the fixed quotas on A+B and fine, with a soft target on the sub-group quotas allowing for target adjustment via the admin module.
answered Dec 20, 2019 by Paul Moon Platinum (78,075 points)
Thank you, once again!  I currently have the quotas set up as you describe but temporarily set to their max values (until I could find an answer).  Your suggestion makes a lot of sense.
Adjust the quotas as necessary during field work, but prevent going over the target of 300 for each of A and B. The individual quotas will prevent that anyway.
Could this be done with three quota questions, a main A/B one and two follow-up ones?  The first quota question would force 300 respondents into both main groups, then the follow-ups could have three cells, 110 max per cell, and maybe least fill / randomize order.
Zachary, I'm not sure how the allocation of A+B and the sub-groups are meant to be handled? Are they based on survey questions or a random approach (where least fill would come in handy)?

If the count for a sub-group goes over 100, then that's one less for one of the other sub-groups.
Good point.  How respondents are meant to be categorized is definitely a key consideration here.
...