User Experience
The user experience was measured through both formal and informal user testing.  Users were explained the system and given free range of the compound to test the tricycle under various conditions.  For the formalized testing, users were asked to compare the figure-eight drive to the standard tricycle.  The formalized testing provided documented positive results, with all users preferring the figure-eight drive to the single-speed tricycle.   Nine users and four staff members were formally surveyed and their responses are on record at APDK and are also available in the laboratory notebook, which can be downloaded here.  Copies of the survey forms can be found on the Documents Page.
In addition to short term testing, we began the distribution of tricycles equipped with the figure-eight drive to users for long-term testing.  After the first week, we checked in with our first long-term recipient (he was given Prototype C2).  He reported that the system was still performing well, though the chain tension did need adjustment.  He was able to adjust the chain tension himself, with the spanner provided with the tricycle.
The durability of the figure-eight drive was tested extensively over a six-week period.  Of primary concern was chain wear due to the guidance tube.  Secondarily, the effects of chain elongation were monitored.  The all-around integrity of the frame and fork were was tested as the tricycle was run over grueling terrain, purposely driven into potholes and curbs to test their effect.
In order to extensively test the effects of the chain guidance tube and chain elongation, a testing fixture was constructed to mechanically manipulate the drivetrain.  This fixture consisted of a stand to lift the front wheel off of the ground and a coupling to attach the pedals to a power drill.  The drivetrain could be rotated by applying power to the drill.  A video of this testing stand can be seen here (4.5 Mb).
The 8-side of the system was manipulated for the equivalent of approximately 100 km (this was done on Prototype C1).  This was well below the goals of the testing regime, but was still enough to show us several important trends.  The first was that the chains seemed to continuously elongate.  Both the o-side and 8-sides were able to absorb reasonable amounts of chain elongation, though periodically the chains would need to be retightened (which can be accomplished with the fork’s vertical dropouts) to prevent derailment.  This is similar to the maintenance required for the standard single speed tricycle.
Chain wear was not found to be significant.  Though the drivetrain was not tested against resistance during the manipulation on the testing stand, chain wear could still be accurately tested because the chain guidance tube, the primary source of chain wear, was on the chain’s untensioned return side, the tension of which is not dependent on resistance applied to the system.
To test adverse real-world conditions, sand was repeatedly added to the chain guidance tube.  The sand did not noticeably cause chain wear, as it was driven out of the tube by the chain.
There was only one mechanical failure in the system.  The bottom freewheel on the o-side of Prototype C1 failed after five weeks of riding and approximately 80 km of manipulation using the testing stand.  The reason for the failure is undetermined, though it resulted in the freewheel no longer catching consistently.  No similar failures occurred on Prototypes A or B.  It cannot be determined without long term testing whether or not the failure was due to a manufacturing flaw on the part of the freewheel manufacturer, or was induced by a heavy amount of freewheeling on the bottom o-side freewheel (which effectively rotates three times for every revolution of the crank arm when utilizing the 8-side).  Changing the freewheel was simple, as it is mounted in the standard fashion.  
Power Output
Some tests were made of the power output of the figure-eight drive (when utilizing the 8-side) compared to the standard tricycle drivetrain.  These tests came back inconclusive because of the lack of incline variety within the APDK compound and surrounding area.  Of the inclines available, they were either shallow enough to allow one to be able to climb easily using the o-side, or were too steep for the standard drive to move up the hill at all.  
Desire for Further Testing
Tests should be done to compare a single speed tricycle, the figure-eight drive, and a 21-speed derailleur system on a moderate incline that can allow all systems to be taken advantage of.  Also of interest would be to compare the power output of 1:1 gear ratio in forwards pedaling to that of a 1:1 gear ratio in backwards pedaling (which is found on the 8-side of the figure-eight drive).  
Biomechanically it seems as if pedaling backwards is advantageous both in terms of increased muscle strength in the pulling direction (as opposed to pushing) as well as decreased pressure on the wrists (since they are being used in tension rather than compression, the possibility of accidentally bending your wrist inappropriately when pushing is eliminated).  It would be beneficial to this project and tricycle users if further research can be done in this area.