>2. Bring in remaining candidates over the next 2-3 days and have them go through the job that they will be performing if they were to be hired.

>...

>4. Bring in final candidates (select few) based on multiple days of feedback, have them perform the same job again for one day and make the final decision.

So if I'm doing arithmetic correctly, you're asking candidates to participate in a trial run for a total of 3 to 4 full working days so you can evaluate them.

I wonder if the writer of the article considered that it will skew the candidate pool towards recent college grads trying to land their first job, and/or unemployed people who can't quickly find a replacement job (because they are undesirable.)

The ones who won't apply are the valuable stars already employed. It's very likely they only have 10 PTO days. The writer needs a reality check if he thinks desirable candidates would be willing to burn up 4 vacation days for a trial work period.

No job recruiting strategy is perfect and most have unintentional side effects (e.g. distorting the candidate pool) that are not explicitly discussed by articles extolling whatever method they're proposing.