The article makes two good points re: bias and lack of research. To the first I think it's up to us as interviewees to mention this in all of our calls. "Do you have a non-take-home option? For some groups this could be an exceptional burden."
To the second there's no real answer other than a lot of companies need to try it out to see how they fare and then when they have a basic idea of logistics that work run a study on efficacy / predictive power.