Allow me to be more specific: there are no published studies showing that Q rings or Rotorcranks improve performance.
This may be the case, but that does not mean there is no evidence for a performance enhancing effect, the repeatability of the reported gains across several tests would give the indication that there is reason for further study to quantify and explore the exact reasons for the reported improvements.
As for this unpublished preliminary report, I find it interesting that power was measured using a Computrainer, which like the SRM ergometer assumes a constant angular velocity when calculating power. Such an assumption obviously doesn't hold when using non-round chainrings, which is why the SRM tends to overestimate power when used with such products - the Computrainer could very well be the same.