They only had 3 participants. You can’t make any claims or generalizations from these results since they will likely be caused by individual differences of the participants. There may also be ordering effects.
Not to mention that we don't know if people naturally adjust to their environment over time so those "efficiency" gains disappear as users adapt to the environment (eg virtual desktops). Also no indication of operating system or task involved.
Studies in general assume a certain level of scientific rigour, reproducibility and verifiability.
To further your point, the discussion here is as valuable as the original article: people sharing their individual anecdotal experiences.