Convenience samples are fine because interactions are mostly not real
Attacking a study for using a non-representative, typically online convenience sample is common practice. However, for most studies, this is wrong-headed and here I want to argue why that is…