These were other, earlier experiments, which the current experiment is attempting to duplicate. Importantly, while the new experiment has superficially "confirmed" the previous ones, it shows a much smaller effect.
When the magnitude of the observed effect scales roughly with the margin of error of the experiment, one has very good reason to be skeptical ...
The previous tests were on an EmDrive, while the current one was a Cannae drive, which is a related but different design. The EmDrive guy claims the lower thrust is to be expected due to the design difference.
<Cough> You shocked me so much that I read the article! :-)
(I follow the subject by browsing news now and then at http://talk-polywell.org/ which really is about the small fusion projects. That place should have more details then at Wired.)
When the magnitude of the observed effect scales roughly with the margin of error of the experiment, one has very good reason to be skeptical ...