> 500 employees were divided into two groups--a control group (who continued working at HQ) and volunteer work-from-homers (who had to have a private room at home, at least six-month tenure with Ctrip, and decent broadband access as conditions).
That's a huge selection bias - if the 'treatment' group are self-selecting volunteers from your pool of all employees then that causes a whole host of issues.
> After a lottery draw, those employees with even-numbered birthdays were selected to work from home, and those with odd-numbered birthdates stayed in
the office to act as the control group.
Not really, this increases its applicability as this is the usual case in which people are considering working from home - as an option at a non-fully-remote company, or as an option to join a fully remote company.
It is selection bias, but the study shows more work being done. If people who self-selected for working from home would do so to try to avoid working, then why there would be positive result (and more work done)?
That's a huge selection bias - if the 'treatment' group are self-selecting volunteers from your pool of all employees then that causes a whole host of issues.