I am planning on launching 15 surveys with 100 respondents each all to be scheduled at the same time. How can I set it up so I don’t have any repeat respondents across the surveys (without launching them sequentially, and excluding participants from past completed surveys?). I’d like to have 1,500 unique respondents.
I’d probably combine them into one survey with random branching to 15 different forms. Are these surveys very different?
I think that allocate monster not only distributes randomly but also equally.
But the problem is that it only links to two or three surveys. I can be nested but I don’t think that there is a way of reaching 15 using multiples of two and three.
If you were to use
Level 1: 2 linking to level 2
Level 2: 3 2 linking to level 3
Level 3: 333 33 linking to the surveys
that would link to 15 surveys but the first allocate monster would allocate equally to the allocate monsters on level 2 So I think you would have 750 applicants going to 9 surveys, and 750 going to 6 surveys.
If you have random branching in the survey software you are using then that may be the way to go.
Thank you Tim. In fact, the random branching was something I explored, but it became complex as I had to make variations to each question stem/answer choice - when I was making a variation. As I work my way through a variation, I had to take an answer choice out and replace it with something else.
I ended up launching the surveys sequentially. Since Prolific is quite quick in collecting data, I thought I would ensure I exclude participants from previous surveys and just launch them that way. Hopefully there isn’t an issue in doing this methodologically.
Whether it’s methodologically fine is a function of the study itself and whether you think the time of survey administration biases the sample, like whether Friday morning responders would operate in the context of your survey from Thursday afternoon responders.
Hmm…One issue I can imagine is that due to time zone differences there may be a difference in global residence of the participants (particularly between the US and the UK the two main subject pools) depending upon the time of day. So it may be an idea to check the nationalities of the participants and equalise the number of studies at any times of day where the demographics (nationality or rather time zone of domicile) is skewed, or start with 75 places and run another 75 places later, if nationality/domicile is an issue.