A building’s effect on gravity experiments

Amidst the descriptions by Clive Speake and Terry Quinn (“The search for Newton’s constant,” Physics Today, July 2014, page 27 ) of eliminating effects of the planet’s pull and handling a plethora of electronic and mechanical problems in order to measure gravity, one issue seems missing: the effects of the mass of the building in which the experiments were performed. Buildings are massive compared with the test and source masses, and the distances from walls and ceilings are small enough to make their effects noticeable in, say, the fourth decimal place of the measurement. Wouldn’t you need to use a massive spherical chamber devoid of other mass so that the effects would cancel each other?

Perhaps I’ve misunderstood or failed to notice any discussion of this in the article. Could the experiments be done outdoors above a flat plain with no mountains or buildings within miles? The problem seems way too simple a thing to have been overlooked by so many diligent people; perhaps it was considered and compensation made. I would be interested in how, if the authors could take time to explain.