choice-metrics.com

Hi there,

Question about the sample size estimate we get using Ngene. I frequently conduct choice experiments in which each subject sees only one task. That is, in essence, the number of blocks is the same as the number of runs. I want to use the sample size calculation in Ngene, along with my priors, to estimate the required sample size. But I'm a bit stuck. Is that the sample size overall? Per task, so I should multiply by the number of runs to get the sample size needed?

I'd appreciate any advice.

Thanks,
Paul

The sample size reported is the sample size if each respondent would face all choice tasks in the design. Suppose that there are 10 choice tasks in the design and you only give 1 choice tasks to a single respondent, then you need to multiply the sample size with 10.

Got it. Makes sense. Thanks.