The primary problem with the NASA test is that their control engine produced the same amount of thrust as their actual engine.
Which strongly suggests that no thrust at all was produced, and there is some kind of zero calibration error, or the whole thing was off balance in some way, giving a false reading.
Actually, this isn't correct. NASA's Control Group was a regular source of Radio-Frequency energy without one of the prototypes attached, which didn't register as providing force. The news article Alway linked seems to have confused their Control Group with one of the prototypes they called a "Null" test; it was designed in a way that they didn't expect to produce thrust, but it produced thrust anyway. This could be due to a lack of understanding behind what's going on in the device, faulty construction in the "Null" prototype that made it actually work, or some other property of the system we don't know yet. But the conclusion seems to be that 30-50 micronewtons of force was registered when the devices were there, and nothing registered when it was a plain RF load. Something happened, and it doesn't seem to have been a measurement error.
EDIT: We don't have access to the whole publication, the data, or the images yet, but this is according to the the abstract we can see so far. And note that NASA concluded that thrust was registered by the system. They have a good track record about accuracy, and I'd assume they wouldn't make a claim based on a completely rookie and unscientific error, which ignored their experimental results. That makes no sense.
Several different test configurations were used, including two different test articles as well as a reversal of the test article orientation. In addition, the test article was replaced by an RF load to verify that the force was not being generated by effects not associated with the test article.
...
Thrust was observed on both test articles, even though one of the test articles was designed with the expectation that it would not produce thrust. Specifically, one test article contained internal physical modifications that were designed to produce thrust, while the other did not (with the latter being referred to as the "null" test article).