I can't disclose details but I know for my office raw [short term] productivity was about 20% up for WFH (the situation differs in almost every aspect from that in the study your linked).
That's non statistically significant to make a general statement, but confirms to me that as a minimum there is high variability.
Data entry is low skill, high churn; in person I'd expect managers to put a lot of pressure on because they can drive people away in six months (it might even be preferable for employment law reasons) and still have plenty of people needing jobs to pick up the work.
I'm not going to research a paper here, only to note my direct experience makes the broad applicability of your single study highly questionable.
It's one study, I'm sure there are now more longitudinal indicative studies for most regions that anyone here can find with ease.
You (I assume) were probably right to suggest the OP might be cherry picking, but you didn't do much better IMO (nor did I, I suppose).
I'd expect pretty different results in different industries, demographics, countries.
Cherry picking indeed.