>It is important to be cognizant of potential sampling bias. The only way to collect this data is for users to self-select, i.e., opt-in, to the project. Self-s

>It is important to be cognizant of potential sampling bias. The only way to collect this data is for users to self-select, i.e., opt-in, to the project. Self-selection bias, especially in online polling, for example, will often skew responses in favor of the most motivated group. Credit to these guys for stepping out in front of the most glaringly obvious shortcoming of the study. I was hoping the Reddit admins would have given them access to data that they would then retroactively ask the users’ permission to use.