Post by andre on Mar 15, 2022 14:23:37 GMT
Dear All,
I really like the counterbalancing function, but we often do not know how many participants we might test (e.g. student projects with opportunistic sampling who collect until they run out of time). To avoid stopping the data collection earlier than necessary, I use rather large numbers, e.g. [50 50] for 50 per group, with 2 groups, so 100 in total.
But the assignment to groups seems random. I've done it now and got for 42 participants 17 in group 1 and 25 in group 2.
Wouldn't it make more sense (at least as an option) to simply fill both groups up in a balanced fashion? So participant 1 is group 1, participant 2 is group 2, participant 3 is group 1, participant 4 is group 2, participant 5 is group 1, and so forth... Then, no matter when data collection is stopped, the groups are as balanced as possible, with a max difference of 1 participant.
Best wishes,
Andre
I really like the counterbalancing function, but we often do not know how many participants we might test (e.g. student projects with opportunistic sampling who collect until they run out of time). To avoid stopping the data collection earlier than necessary, I use rather large numbers, e.g. [50 50] for 50 per group, with 2 groups, so 100 in total.
But the assignment to groups seems random. I've done it now and got for 42 participants 17 in group 1 and 25 in group 2.
Wouldn't it make more sense (at least as an option) to simply fill both groups up in a balanced fashion? So participant 1 is group 1, participant 2 is group 2, participant 3 is group 1, participant 4 is group 2, participant 5 is group 1, and so forth... Then, no matter when data collection is stopped, the groups are as balanced as possible, with a max difference of 1 participant.
Best wishes,
Andre